Actin is a surface component of calf pulmonary artery ... - Europe PMC

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 3815-3819, May 1993 Cell Biology

Actin is a surface component of calf pulmonary artery endothelial cells in culture (angiogenin/binding protein/angiogenesis)

JUNONA MOROIANU, JAMES W. FETT, JAMES F. RIORDAN, AND BERT L. VALLEE Center for Biochemical and Biophysical Science and Medicine, Harvard Medical School, 250 Longwood Avenue, Boston, MA 02115

Contributed by Bert L. Vallee, January 29, 1993

ABSTRACT An angiogenin binding protein isolated previously from endothelial cells has been shown to be a member of the actin family. Calf pulmonary artery endothelial (CPAE) cells were investigated for the presence of surface actin by inmmunoblotting of isolated surface proteins and by immunofluorescence. CPAE cell surface proteins were isolated by selective apical biotinylation and recovery of biotinylated proteins by avidin affinity chromatography. Immunoblotting with a specific smooth muscle a-actin antibody detected the presence of this type of actin among the isolated cell surface proteins. Immunofluorescence confirmed that smooth muscle a-actin is localized at the surface of nonpermeabilized CPAE cells. Exposure of CPAE cells to angiogenin prior to cell surface immunostaining diminished the signal. When CPAE and rat aortic smooth muscle cells were made permeable before staining, stress fibers could be recognized by the antibody in smooth muscle cells but not CPAE cells. The results indicate that a smooth muscle type of a-actin is localized specifically on the surface of cultured CPAE cells where it might interact with angiogenin and other actin binding proteins present in the extracellular environment.

function as an AngBP involved in the process of angiogenesis.

MATERIALS AND METHODS Materials. Bovine muscle actin and monoclonal antibodies to amoeba actin and to smooth muscle a-actin were from Sigma; fluorescein isothiocyanate (FITC)-labeled goat F(ab')2 anti-mouse IgG was from Caltag (South San Francisco, CA); 125I-labeled species-specific anti-mouse F(ab')2 was from Amersham; bodipy-phallacidin was from Molecular Probes; and 125I-labeled streptavidin, sulfosuccinimidyl 2-(biotinamido)ethyl-1,3-dithiopropionate (NHS-SS-biotin), and immobilized avidin were from Pierce. Cell Culture. CPAE cells (CCL 209, American Type Culture Collection) and rat aortic smooth muscle (RASM) cells were cultured in 75-cm2 culture flasks (Nunc) in minimal essential medium (MEM; GIBCO) with 20% heat-inactivated fetal bovine serum (GIBCO), 50 units of penicillin per ml, and 50 jig of streptomycin per ml (CPAE cells) or in Dulbecco's modified Eagle's medium supplemented with 10% heatinactivated fetal bovine serum (GIBCO) and the same antibiotics (RASM cells) in a humidified atmosphere of 5% C02/95% air at 37°C as described (13, 14). CPAE cells were used between passages 20 and 25 and RASM cells were used up to passage 10. Isolation of Surface Proteins. Selective surface biotinylation. Confluent CPAE cells were washed four times with ice-cold MEM and incubated with NHS-SS-biotin (0.5 mg/ ml) in phosphate-buffered saline (PBS). After two 15-min treatments at 4°C (15), the monolayers were washed four times with PBS [or in some experiments with 0.1% bovine serum albumin (BSA) in PBS (16)] and extracted with 5 ml of ice-cold lysis buffer (150 mM NaCl/1% Nonidet P-40/0.5% deoxycholate/0.1% SDS/10 mM Tris, pH 7.5/1 mM phenylmethylsulfonyl fluoride/10 ,ug of aprotinin per ml) for 1 hr on ice. Cell extracts were clarified by centrifugation (14,000 x g, 10 min) at 4°C. Detection of biotinylated proteins. Cellular extract containing 100 ,ug of protein (BCA protein assay; Pierce) was subjected to SDS/PAGE in 10% gels. After electrophoresis, proteins were transferred to nitrocellulose membranes in 0.041 M Tris/0.040 M boric acid buffer, pH 8.3 (17), and incubated with 125I-labeled streptavidin essentially as described (18). Briefly, following transfer the blots were blocked with 3% BSA in PBS containing 0.5% (vol/vol) Tween 20 and 10% (vol/vol) glycerol (buffer G). Subsequently, the blots were incubated with 1251-labeled streptavidin (1 ,uCi/ml in 0.3% BSA in buffer G; 1 Ci = 37 GBq) for 2 hr at room temperature, washed with PBS containing 0.5%

Actin is an abundant, highly conserved protein that polymerizes into filaments that are essential for cellular motility and cell division as well as for the structure and mechanical properties ofthe cytoplasmic matrix. Whereas the majority of actin is found in the intracellular compartment, the protein has also been reported to be present on the cell surface of brain endothelial cells (1) and lymphocytes (2, 3), in the extracellular matrix of bovine aorta (4), cultured chicken embryo fibroblasts (5), or cultured rat smooth muscle cells (6), and in the basement membrane of endothelial cells, pericytes, and smooth muscle cells of arteries and arterioles

(7).

Angiogenin, a potent angiogenic molecule present in tumor cell conditioned medium (8), plasma (9), and milk (10), binds a 42-kDa dissociable cell surface component of calf pulmonary artery endothelial (CPAE) cells and GM7373 cells, a transformed bovine endothelial cell line (11). The angiogenin binding protein (AngBP) can be released from endothelial cells by exposure to heparin, heparan sulfate, or angiogenin (11, 12). Tryptic peptide mapping and amino acid sequence analyses have indicated that AngBP is a member of the muscle-type actin family (12). The present study was undertaken to establish that actin is associated with the surface of endothelial cells in culture. Two different approaches were used: (i) immunoblotting of proteins isolated specifically from the cell surface and (ii) immunofluorescence with nonpermeabilized cells. The results indicate that a smooth muscle type of a-actin is present on the surface of endothelial cells in culture where it may

Abbreviations: CPAE, calf pulmonary artery endothelial; RASM, rat aortic smooth muscle; AngBP, angiogenin binding protein; NHSSS-biotin, sulfosuccinimidyl 2-(biotinamido)ethyl-1,3-dithiopropionate; BSA, bovine serum albumin; FITC, fluorescein isothiocyanate.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 3815

3816

Cell Biology: Moroianu et al.

Tween 20 (three times, 10 min each), dried, and autoradiographed on Kodak XAR-5 film. Avidin recovery of biotinylated proteins. Cell extracts containing biotinylated surface proteins were incubated with immobilized avidin for 1 hr at 4°C, with end-over-end rotation. Following the binding step, the immobilized avidin complex was recovered by centrifugation and washed six times with lysis buffer. The biotinylated surface proteins were eluted by cleaving the disulfide bond of NHS-SS-biotin with 100 mM dithiothreitol in lysis buffer, over 1 hr at 4°C (16, 19). Immunoblotth4g of Surface Proteins. Proteins recovered from the immobilized avidin column were subjected to SDS/ PAGE in 10% gels and transferred to nitrocellulose membranes. The membranes were blocked with PBS containing 2% BSA and 0.05% Tween 20 (quenching buffer) and incubated with antibody to smooth muscle a-actin or to amoeba actin (a 1 mg/ml solution diluted 1:100 with quenching buffer) for 2 hr at room temperature. After rinsing, 125I-labeled F(ab')2 anti-mouse IgG (diluted 1:100 in quenching buffer) was added. Following a-1-hr incubation at room temperature the membranes were washed with PBS, dried, and developed by autoradiography on Kodak XAR-5 film. Immunofluorescence. For surface staining, CPAE and RASM cells grown on glass microscope coverslips were washed three times with MEM and fixed for 5 min with 3.7% paraformaldehyde in PBS. After fixation, the cells were washed and equilibrated in PBS containing 1.5% BSA. For intracellular staining, the cells were fixed for 5 min with 3.7% formaldehyde and permeabilized for 30 s with acetone cooled on dry ice (20). For surface and intracellular staining, cells were incubated sequentially with antibody to smooth muscle a-actin (1:400-1:1000 dilution) followed by FITC-labeled goat F(ab')2 anti-mouse IgG (1:25 dilution). All incubations were performed on ice for surface staining and at room temperature for intracellular staining. Cells were incubated with the first antibody for 1 hr, washed five times in PBS, and incubated with FITC-labeled goat anti-mouse antibody. After 1 hr, the cells were washed five times with PBS and mounted in glycerol/PBS, 1:1 (vol/vol). Controls were (i) omission of the first antibody, (ii) substitution of the first antibody by nonimmune mouse IgG at the same concentration, and (iii) preadsorption of the monoclonal antibody to RASM cell smooth muscle a-actin that had been immobilized on a nitrocellulose membrane by blotting. Labeled cells were observed with a Nikon Labophot microscope equipped for fluorescein fluorescence and photomicrographs were taken with Kodak EKTAR-1000 film at an exposure index of 800 or with TMY-400 film at an exposure index of 400. Bodipyphallacidin staining of F-actin was performed according to the manufacturer's instructions based on the procedure of Barak et al. (21).

Proc. Natl. Acad. Sci. USA 90 (1993) kDa

66 45

29

12.5

1235 1

2

3

FiG. 1. Surface biotinylation of CPAE cells. Confluent CPAE cel monolayers were labeled with NHS-SS-biotin (0.5 mg/ml) and extracted with lysis buffer. Surface biotinylated proteins were resolved by SDS/PAGE and analyzed by 125I-labeled streptavidin blotting. Lane 1, biotinylated molecular mass markers; lane 2, biotinylated bovine muscle actin; lane 3, biotinylated CPAE celi surface proteins.

streptavidin blotting of the unretained fraction (data not shown). To ensure that unbiotinylated actin was not retained nonspecifically on immobilized avidin, 50 jAg of bovine muscle actin was subjected to affinity chromatography as above and the bound and eluted fraction was analyzed by electrophoresis. Actin could not be detected in this material. Immunoblotting of affinity-isolated surface proteins with the monoclonal antibody to amoeba actin, which recognizes an epitope that occurs in muscle and nonmuscle actins, indicated the presence of a 42-kDa band corresponding to actin (Fig. 2, lane 1). Moreover, when a monoclonal antibody to smooth muscle a-actin specific for this isoform (22) was used, the same 42-kDa protein was again recognized (Fig. 2, lane 2). These results indicate that a type of smooth muscle a-actin is present on the cell surface of CPAE cells. Though endothelial cells most commonly express nonmuscle type actins (20, 23-25), they have been shown to express smooth muscle a-actin when grown under certain culture conditions (26, 27). LocIzation of Actin to the CPAE Cell Surface by Immuoftluorescence. To determine whether smooth muscle a-actin kDa

RESULTS Detection of Actin by Immunoblotting of Isolated Cell Surface Proteins. Treatment of CPAE cells with NHS-SS-biotin under conditions designed to be selective for cell surface

protein biotinylation (15, 16) followed by washing to remove excess reagent led to the isolation of biotinylated proteins by avidin affinity chromatography. Essentially identical results were obtained when washing was carried out in the presence or absence of BSA. Analysis of biotinylated surface proteins by 1251-labeled streptavidin blotting revealed the presence, among others, of a 42-kDa protein band (Fig. 1). The biotinylated proteins were bound to immobilized avidin and specifically eluted with dithiothreitol, which reduces the disulfide bond of the NHS-SS-biotin moiety and releases the associated protein. All biotinylated proteins bound to the avidin gel since no bands were visualized by 125I-labeled

42

1

2

FiG. 2. Immunoblotting of isolated CPAE celi surface proteins. Surface proteins were subjected to SDS/PAGE and blotted onto

nitroceliulose membranes. The blots were incubated with antibody to amoeba actin (lane 1) or antibody to smooth muscle a-actin (lane 2). Rainbow protein molecular mass markers (Amersham) were used for molecular mass determination.

Cell

Biology: Moroianu et al.

Proc. Natl. Acad. Sci. USA 90 (1993)

3817

is specifically localized to the cell surface, immunofluorescence was performed on nonpermeabilized and permeabilized CPAE cells. Since smooth muscle a-actin is a constituent of the cytoskeleton of vascular smooth muscle cells (20, 23, 24, 26, 29), RASM cells were also included in this study as a control. Surface fluorescence was observed with fixed, nonpermeabilized CPAE cells stained with antibody to smooth muscle a-actin (Fig. 3 A and B). Both controlsabsence of the first antibody and substitution with nonimmune mouse IgG for the first antibody-were negative. Moreover, surface immunostaining was almost completely abolished by preadsorption of the antibody to RASM cell actin (data not shown). These results indicate that a smooth

FIG. 4. Immunofluorescence with antibody to smooth muscle a-actin on permeabilized RASM cells. (X300.)

muscle type of a-actin is present on the CPAE cell surface. Preexposure of the CPAE cells to bovine angiogenin (100 ng/ml) for 30 min at 37°C greatly diminished the surface staining with antibody to smooth muscle a-actin (Fig. 3C). The presence of smooth muscle a-actin was also detected on the surface of RASM cells (data not shown). Since actin is present in the fetal bovine serum that is included in the medium used for cell culture, control experiments were carried out to exclude this as the source of cell surface-bound material. Thus, CPAE cells were maintained in medium minus serum for 24 hr prior to antibody staining and essentially the same immunofluorescence was observed as for cells maintained in 20% fetal bovine serum. In addition, actin was isolated from a sample of fetal bovine serum by affinity chromatography on immobilized DNase I (28). Sequence analysis of the protein obtained after hydroxylamine cleavage to remove the N-terminal dodecapeptide identified it as a cytosolic actin (D. J. Strydom, personal communication). Immunofluorescence of permeabilized RASM cells with antibody to smooth muscle a-actin showed strong staining of actin stress fibers typical for cultured smooth muscle cells (Fig. 4). In contrast, virtually no staining of actin stress fibers was detected when permeabilized CPAE cells were used (Fig. SA). Increasing the concentration of the muscle actin antibody from a 1:1000 to a 1:400 dilution stained cytoplasmic components but still failed to stain CPAE stress fibers. However, when bodipy-phallacidin, which specifically stains F-actin (21), was used the CPAE cytoskeleton was clearly visible (Fig. SB). The findings indicate that in CPAE cells the stress fibers do not contain a-smooth muscle actin and that this actin type is specifically localized to the cell surface.

DISCUSSION Two methods have been used to investigate the presence of actin on the surface of CPAE cells in culture. The first involves selective surface biotinylation followed by isolation

FIG. 3. Surface immunofluorescence of CPAE cells with antibody to smooth muscle a-actin at 1:1000 dilution. (A, x50; B, x500.) (C) CPAE cells were preincubated with bovine angiogenin (100 ng/ml) for 30 min at 37°C prior to surface immunostaining. (x50.)

of biotinylated proteins by avidin affinity chromatography and immunoblotting. This approach has previously been used with success to study apical or basolateral surface protein compositions of a variety of cell types (15, 16, 18, 30). The labeling is sensitive (biotinylation of lysines) and is confined strictly to cell surface proteins. Moreover, the biotinylated proteins can be recovered by adsorption on immobilized avidin/streptavidin. A cleavable biotin reagent (NHS-SSbiotin) allows specific elution of the biotinylated proteins from the avidin gel under nondenaturing conditions, thereby avoiding possible contamination by proteins that bind to the avidin affinity matrix nonspecifically (16). Immunoblotting of

3818

Cell Biology: Moroianu et al.

Proc. Natl. Acad Sci. USA 90 (1993)

FIG. 5. Staining of permeabilized CPAE cells with antibody to smooth muscle a-actin. (A) No stress fibers were visible when the cells were stained with the antibody at a 1:1000 dilution. (B) Staining of CPAE actin fibers with bodipy-phallacidin. (x300.)

isolated CPAE surface proteins with monoclonal antibodies to amoeba actin and to smooth muscle a-actin indicated that actin is a component of these surface proteins. Immunofluorescent techniques with an antibody to smooth muscle a-actin detected the presence of this muscle actintype molecule on the surface of fixed, nonpermeabilized CPAE cells. With permeabilized CPAE cells only diffuse cytoplasmic staining with the antibody was observed, but staining of intracellular stress fibers could not be detected. The present studies identify a smooth muscle a-actin type as a surface component of CPAE cells. This supports the recent work of Hu et al. (12), who found that the AngBP isolated from cultured endothelial cells (11) is a type of muscle actin (12). Although actin is mainly an intracellular constituent, a surface actin has also been identified in bovine brain microvessels (1). Interestingly, the amino acid composition reported for that protein is closely similar to that found for the AngBP isolated from transformed fetal bovine aortic endothelial cells, GM7373 (12). Fetal bovine serum is a component of the culture medium used for growing CPAE and RASM cells, but it is not the source of actin found on the cell surface: virtually the same immunofluorescence was observed when cells were cultured for 24 hr in MEM with or without serum prior to antibody staining. Moreover, although actin is present in fetal bovine serum, its amino acid sequence identified it to be of the cytosolic type (D. J. Strydom, personal communication). Thus, the surface actin most likely derives from the cells themselves. Preliminary metabolic labeling studies with CPAE cells (J.M., unpublished observations) are consistent with this view. Smooth muscle a-actin is the major isoform of actin found in smooth muscle cells (20, 26, 29). Although endothelial cells contain cytoplasmic 3- and y-actin (20, 23, 24), microvascular endothelial cells can also express a-actin when grown under specific culture conditions (26, 27). It is therefore possible that CPAE cells express either a smooth muscle a-actin type or a closely related form of actin. Recently, closely related new actin isoforms were reported to be present in vertebrates (31, 32). Northern blot analysis with specific probes will clarify if the actin identified here is smooth muscle a-actin or a new related actin. The presence of this actin on the CPAE cell surface suggests that it might be specifically targeted to this site following biosynthesis. Indeed, sorting of muscle and nonmuscle actins was reported for cells that express both proteins simultaneously (33, 34). For example, in pericytes and smooth muscle cells, smooth muscle a-actin was localized inside the cell on the stress fibers, whereas nonmuscle actins were found also in other regions of the cytoplasm (20,

33, 34).

In vivo, actin has been specifically detected in the extracellular matrix and in the limiting basement membrane of smooth muscle cells, in arteries, and in large-sized arterioles (7). In small arterioles, the presence of actin was restricted to the endothelial basement membrane and to the intercellular matrix of smooth muscle cells. In capillaries, specific staining for actin localized it to the basement membrane surrounding the endothelial cells and pericytes (7). Moreover, actin has been reported to bind to fibronectin (35, 36). Thus, it was suggested that the association between actin, fibronectin, and other extracellular proteins could be functional-i.e., actin might participate in the control of cell adhesion and cell movement (7). Incubation of CPAE cells with angiogenin strongly diminished the surface immunostaining with smooth muscle a-actin antibody, indicating that angiogenin binds to surface actin. Earlier studies had shown that AngBP is a cell surface protein that can be detached from the cell by treating with heparan sulfate or probably with angiogenin itself (11). Hence, the diminished staining could reflect loss of actin/AngBP from the cell surface. Indeed, cross-linking experiments have demonstrated that 125I-labeled angiogenin releases at least some of the AngBP from CPAE and GM7373 cells. Moreover, an anti-actin polyclonal antibody greatly reduces the amount of cross-linked complex that is formed (J.M., unpublished observations). Together, these data support the view that the AngBP from endothelial cells is a dissociable cell surface muscle actin (12). Of course, diminished staining might also mean that angiogenin blocks the antibody recognition site of bound AngBP/actin or the complex might be internalized within the cell, or both. Further experiments should help to clarify this issue. Angiogenin -appears to dissociate some AngBP from the endothelial cell surface but the mechanism by which this occurs is not clear. It may be a consequence of the actin polymerization that is induced by angiogenin (12). The inhibitory effects of actin and anti-actin antibodies on angiogenin-induced angiogenesis in the chicken chorioallantoic membrane assay strongly suggest that the binding of angiogenin to surface actin is physiologically significant and may be an important step in angiogenesis (12). Moreover, previous studies have shown that the interaction of actin with angiogenin is inhibited by protamine and platelet factor 4 (12). Both of these proteins are inhibitors of in vivo angiogenesis (37, 38), which further supports the view that the angiogeninactin interaction is a critical component of this process. Another feature of angiogenin that is indispensable for angiogenesis is its ribonucleolytic activity. Angiogenin is a member of the ribonuclease superfamily with 33% sequence identity to bovine pancreatic ribonuclease A and it contains

Cell Biology: Moroianu et al. all of the principal amino acids that constitute the enzymatically active site (39). Chemical modifications and sitedirected mutagenesis of these residues have established that abolition of catalytic activity is accompanied by loss of angiogenic activity (40). Although the relationship between actin binding and ribonucleolytic activity remains to be established, it has been shown that angiogenin contains a putative cell surface receptor binding site that is distinct from the catalytic site and is crucial for angiogenic but not for ribonucleolytic activity (41). Indeed, it has been suggested to be the site responsible for binding AngBP/actin (12). In this regard, exogenous actin appears to prevent the interaction of angiogenin with endothelial cells. A 100-fold molar excess of bovine actin completely abolished the ability of angiogenin to induce new blood vessel formation as assessed on the chicken chorioallantoic membrane (12). This inhibitory activity of actin toward angiogenin-induced angiogenesis may presage a new class of agents with therapeutic potential for pathological conditions associated with aberrant blood vessel growth. We thank Dr. Daniel J. Strydom for information on fetal bovine serum actin. This work was supported by funds from Hoechst, A.G., under agreement with Harvard University. 1. Pardridge, W. M., Nowlin, D. M., Choi, T. B., Yang, J., Calaycay, J. & Shively, J. E. (1989) J. Cereb. Blood Flow Metab. 9, 675-680. 2. Sanders, S. K. & Craig, S. W. (1983) J. Immunol. 131, 370377. 3. Owen, M. J., Auger, J., Barber, B. H., Edwards, A. J., Walsh, F. S. & Crumpton, M. J. (1978) Proc. Natl. Acad. Sci. USA 75, 4484-4488. 4. Bach, P. R. & Bentley, J. P. (1980) Connect. Tissue Res. 7, 185-190. 5. Chen, L. B., Murray, A., Segal, R. A., Buschnell, A. & Walsh, M. L. (1978) Cell 14, 377-391. 6. Jones, P. A., Scott-Burden, T. & Gevers, W. (1979) Proc. Natl. Acad. Sci. USA 76, 353-357. 7. Accinni, L., Natali, P. G., Silvestrini, M. & De Martino, C. (1983) Connect. Tissue Res. 11, 69-78. 8. Fett, J. W., Strydom, D. J., Lobb, R. R., Alderman, E. M., Bethune, J. L., Riordan, J. F. & Vallee, B. L. (1985) Biochemistry 24, 5480-5486. 9. Shapiro, R., Strydom, D. J., Olson, K. A. & Vallee, B. L. (1987) Biochemistry 26, 5141-5146. 10. Maes, P., Damont, D., Rommens, C., Montreuil, J., Spik, G. & Tartar, A. (1988) FEBS Lett. 241, 41-45. 11. Hu, G.-F., Chang, S.-I., Riordan, J. F. & Vallee, B. L. (1991) Proc. Natl. Acad. Sci. USA 88, 2227-2231. 12. Hu, G.-F., Strydom, D. J., Fett, J. W., Riordan, J. F. & Vallee, B. L. (1993) Proc. Natl. Acad. Sci. USA 90, 1217-1221. 13. Soncin, F. (1992) Proc. Natl. Acad. Sci. USA 89, 2232-2236.

Proc. Natl. Acad. Sci. USA 90 (1993)

3819

14. Moore, F. & Riordan, J. F. (1990) Biochemistry 26, 228-233. 15. Le Bivic, A., Real, F. X. & Rodriguez-Boulan, E. (1989) Proc. Natl. Acad. Sci. USA 86, 9313-9317. 16. Busch, G., Hoder, D., Reutter, W. & Tauber, R. (1989) Eur. J. Cell. Biol. 50, 257-262. 17. Gershoni, J. M. & Palade, G. E. (1982) Anal. Biochem. 124, 396-405. 18. Sargiacomo, M., Lisanti, M., Graeve, L., Le Bivic, A. & Rodriguez-Boulan, E. (1989) J. Membr. Biol. 107, 277-286. 19. Savage, D. M., Mattson, G., Desai, S., Nielander, G. W., Morgensen, S. & Conklin, E. J. (1992) Avidin-Biotin Chemistry: A Handbook (Pierce, Rockford, IL), p. 41. 20. Herman, I. M. & D'Amore, P. A. (1985) J. Cell Biol. 101, 43-52. 21. Barak, L. S., Yocum, R. R., Nothangel, E. A. & Webb, W. W. (1980) Proc. Natl. Acad. Sci. USA 77, 980-984. 22. Skalli, O., Ropraz, P., Trzeciak, A., Benzonana, G., Gillessen, D. & Gabbiani, G. (1986) J. Cell Biol. 103, 2787-2796. 23. Chamley, J. H., Groschel-Stewart, U., Campbell, G. R. & Burnstock, G. (1977) Cell Tissue Res. 177, 445-457. 24. Gown, A. M., Vogel, A. M., Gordon, D. & Lu, P. L. (1985) J. Cell Biol. 100, 807-813. 25. Skalli, O., Vanderkerckhove, J. & Gabbiani, G. (1987) Differentiation 33, 232-238. 26. Kocher, 0. & Madri, J. A. (1989) In Vitro Cell. Dev. Biol. 25, 424-434. 27. Amberger, A., Bauer, H., Tontsch, U., Gabbiani, G., Kocher, 0. & Bauer, H. C. (1991) FEBS Lett. 287, 223-225. 28. Lindberg, U. & Eriksson, S. (1971) Eur. J. Biochem. 86, 474-479. 29. Kocher, 0. & Gabbiani, G. (1987) Differentiation 34, 201-209. 30. Lisanti, M. P., Caras, I. W., Davitz, M. A. & RodriguezBoulan, E. (1989) J. Cell Biol. 109, 2145-2156. 31. Lees-Miller, J. P., Helfman, D. M. & Schroer, T. A. (1992) Nature (London) 359, 244-246. 32. Clark, S. W. & Meyer, D. I. (1992) Nature (London) 359, 246-250. 33. DeNofrio, D., Hoock, T. C. & Herman, I. M. (1989) J. Cell Biol. 109, 191-202. 34. Hoock, T. C., Newcomb, P. M. & Herman, I. M. (1991) J. Cell Biol. 112, 652-664. 35. Keski-Oja, J., Sen, A. & Todaro, G. J. (1980) J. Cell Biol. 85, 527-533. 36. Koteliansky, V. E., Glukhova, M. A., Morozkin, A. D., Musatov, A. P., Shirinsky, V. P., Tskhovrebova, L. A. & Smirnov, V. N. (1981) FEBS Lett. 133, 31-35. 37. Taylor, S. & Folkman, J. (1982) Nature (London) 297, 307-312. 38. Maione, T. E., Gray, G. S., Petro, J., Hunt, A. J., Donyer, A. L., Bauer, S. I., Carson, H. F. & Sharpe, R. J. (1990) Science 247, 77-79. 39. Strydom, D. J., Fett, J. W., Lobb, R. R., Alderman, E. M., Bethune, J. L., Riordan, J. F. & Vallee, B. L. (1985) Biochemistry 24, 5486-5490. 40. Shapiro, R. & Vallee, B. L. (1989) Biochemistry 28, 7401-7408. 41. Hallahan, T. W., Shapiro, R. & Vallee, B. L. (1991) Proc. Natl. Acad. Sci. USA 88, 2222-2226.