Fibronectin-Like Protein in Porifera

Proc. NatL Acad. Sci. USA Vol. 78, No. 10, pp. 6261-6265, October 1981 Cell Biology

Fibronectin-like protein in Porifera: Its role in cell aggregation (phylogenesis/celi membrane/intercellular matrix/differentiation/morphogenesis) JACQUELINE LABAT-ROBERT*, LADISLAS ROBERT*, CAROLE AUGERt, CLAIRE LETHIASt, AND ROBERT GARRONEt *Laboratoire Biochimie du Tissu Conjonctif (GR Centre National de la Recherche Scientifique, No. 40), Facult6 de MWdecine, UniversitM Paris-Val de Marne, 8 rue du Ceneral Sarrail, 94010 Crkteil Cedex, France; and tLaboratoire Histologie et Biologie Tissulaire (LA Centre National de la Recherche Scientifique, No. 244), UniversitM Claude Bernard, 43 Bd. du 11 Novembre, 69621 Villeurbanne, France Communicated by D. H. R. Barton, June 15, 1981 ABSTRACT Experiments were carried out on a freshwater sponge cells-were used to show whether fibronectin plays a sponge (Ephydatia muderi) in order to demonstrate the presence role in this process. of fibronectin in Porifera. By using antibodies to highly purified Previous experiments carried out in our laboratory had dem- human plasma fibronectin, the presence of a similar or identical onstrated the presence of structural glycoproteins in various protein could be demonstrated in the membranes of E. mullen marine sponges (12-14). These glycoproteins were obtained in cells such as epithelial cells, fibroblast-like cells, and choanocytes. a highly purified form and were shown to have an amino acid The reaction was specific, could be abolished by the addition of composition quite similar to that found in structural glycopro- excess fibronectin, and was not observed with nonimmune rabbit from vertebrate tissues These serum. The immune fluorescent reaction became stronger when teins isolated (8-10, 15). sponge the sponge cells were pretreated with acetone and could also be glycoproteins, however, were of a relatively low molecular observed, although with a less intense staining, on the intercellular weight. matrix. This shows the predominant presence of a sponge fibronectin-like protein in the cell membranes and also its presence to MATERIALS AND METHODS a lesser extent in the intercellular matrix. When dissociated Materials. Human plasma CIg was purified by the method sponge cells were led to reassociate under the microscope, reas- of Vuento and Vaheri (16). The purity of the preparation was sociation could be completely inhibited by anti-human fibronectin checked by acrylamide gel electrophoresis and by antiserum up to a dilution of 1:120 and partially inhibited up to immunodiffusion. a dilution of 1:240. The reassociation of dissociated sponge cells The antibody to highly purified human plasma CIg was pre- could also be inhibited by the addition of purified gelatin but not pared in rabbits. Other immune serawere obtained through the with serum albumin or with a normal, nonimmune rabbit serum. courtesy ofA. Vaheri (Helsinki, Finland) and ofMosesson (New These results clearly indicate that a sponge cell fibronectin-like York). All three sera gave a single line with total human plasma protein may play an important role as the (or one of the) recog- which showed an identity with the line given by highly purified nition site(s) of the aggregation factor(s) and can therefore be di- plasma fibronectin. rectly involved in cell association, morphogenesis, and Gelatin from pig skin collagen type I and bovine serum al- differentiation. bumin were obtained from Sigma. Immunofluorescence. These experiments were carried out Fibronectin, cold-insoluble globulin (CIg), and large external according to the indirect Coon's technique using sheep antisera transformation sensitive protein (LETS protein) are more or less to rabbit IgG labeled with fluorescein (Pasteur Institute, Paris). synonymous expressions for a family of closely related glyco- Sponge Culture. Freshwater sponges (Ephydatia mulleri) proteins (1-4). These glycoproteins were found in blood plasma were grown from gemmules collected in local brooks near Lyon. and on the cell membranes of several different kind of mes- The sponges were cultivated in the laboratory at 20°C in a 1:1 enchymal cells such as fibroblasts as well as in the intercellular (vol/vol) mixture of soft mineral water ("Evian") and distilled matrix (5-7). water, between two glass cover slips, according to the "sand- On the basis of their amino acid and carbohydrate compo- wich" method of Ankel and Eigenbrodt (17). Some gemmules sition, their solubility behavior, and their presence in the in- were allowed to develop simply on cover slips in the same tercellular matrix, they appear to belong to the class ofstructural medium. glycoproteins (8-10). They play a role in the interaction be- Preparation of Cells for Fluorescence Microscopy. The tween cell membrane and intercellular matrix as far as they as- sponges were cultivated for 8 days. Just before use, the two sociate more or less specifically with collagen fibers. cover slips were separated so that each carried a thin layer of Collagen, in a microscopically and chemically well-recogniz- sponge tissue. Cells were exposed in sequence to dilute (100 able form, appears first during phylogenesis in Porifera or mosM) phosphate-buffered saline (PjNaCl), 3.5 % paraformal- sponges (11). It was therefore important to investigate whether dehyde in PjNaCl (for 30 min), and three 10-min rinses in P/ fibronectin is also present at this early stage of phylogenesis. NaCl. For surface staining, cells were not made permeable; for In order to demonstrate the presence of fibronectin-like pro- internal staining, cells were made permeable with acetone (18): teins in Porifera, we used antibodies prepared against human 2 min in 50% acetone at -20°C, 5 min in pure acetone at 40C, plasma CIg (1) as well as highly purified fibronectin obtained and 2 min in 50% acetone at room temperature. Antifibronectin from human plasma. Two different approaches-immune flu- antiserum was applied at 1:60 and 1:120 dilutions for 45 min. orescence and the inhibition of reassociation of dissociated Staining was indirect, with fluorescein isothiocyanate-conjugated sheep anti-rabbit IgG. Controls were performed by re- The publication costs ofthis article were defrayed in part bypage charge payment. This article must therefore be hereby marked "advertise- Abbreviations: CIg, cold-insoluble globulin, P/NaCl, phosphate-buf- ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact. fered saline. 6261 Downloaded by guest on September 29, 2021 6262 Cell Biology: Labat-Robert et aPProc. Nad Acad. Sci. USA 78 (1981) placing the antiserum with one of the following: normal rabbit EDTA/34 mM NaCV1.34 mM KCV1.38 mM glucose/1.07 serum, a mixture offibronectin and antifibronectin antiserum, mM NaCOJ0.7 mM K2HPO4, buffered to pH 8 with 6 mM and PINaCl. Each treatment was followed by three 10-min Tris. rinses in PJNaCl. Cell suspensions were prepared by replacing the culture Fluorescence Microscopy. Slides were examined with a Leitz medium by the dispersion medium during 10 min. Dispersion photomicroscope equipped with epifluorescence illumination was ensured mechanically by pipetting. The suspension was (Ploemopak system). Fluorescence was excited with the output then centrifuged at 200 x g for 5 min. The supernatant was ofa high-pressure Hg lamp (50 W) filtered through a Leitz FITC discarded and the pellet was resuspended in the culture me- interference filter. Cells were observed through X25 and X40 dium. This procedure was repeated twice and the final pellet lens. was resuspended in 0.5 ml ofculture medium. A high-density Reaggregation Experiments. The dispersion medium, pre- population ofabout 2 x 107 cells per ml was obtained. In order pared according to Curtis and Van de Vyver (19), was 0.25 mM to study the interference of antifibronectin antiserum, of gel-

FIG. 1. Cell cultures grown from gemmules examined with epifluorescence illumination after treatment with rabbit antifibronectin antiserum and fluorescein isothiocyanate-conjugated sheep anti-rabbit IgG. (A) The fluorescence is located largely at the periphery (arrows) of the large epithelial cells. Aperinuclear area (presumably the Golgiregion) is also stained (G). This patternoffluorescence isobviouslydifferentfromthestaining of the small motile cells which are heavily labeled (Inset). (x360; Inset, x200.) (B) Fluorescent staining of the cells ensuringthe waterflow through the sponge. (x500.) (C) Dense staining of a fibroblast-like cell (lophocyte). (X800.) Downloaded by guest on September 29, 2021 Cell Biology: Labat-Robert et al. Proc. Natd Acad. Sci. USA 78 (1981) 6263

_|| ,\-,,fflti,:~~~~~~~~*' at

_ _ 4~~~~~VI Lma" 'Ei FRifioltiaeb t ) i l 3 inndt..)- DQ c nmut1i

(E)hadbeade.Ian,mscelaermndsnl.(akfedeaia

All.

FIG. 2. Reassociation of dissociated sponge cells and its inhibition by anti-Clgantiserum and by gelatin. (A) Dissociated cells after 30 min culture in normal medium. Numerous cells have adhered into small, clumps. (Dark-field examination; x200.) (B-E) Dissociated cells after 30 min culture in medium to which 1:40 antifibronectin antiserum (B), 1% gelatin (C), 1:40 normal nonimmune rabbit serum (D), or 1% bovine serum albumin (E) had been added. In B and C, most cells have remained single. (Dark-field examination; x200.) Downloaded by guest on September 29, 2021 6264 Cell Biology: Labat-Robert et aL Proc. Nad Acad. Sci. USA 78 (1981) atin, or of bovine serum albumin with the reassociation of dis- hibited reaggregation in a manner similar to the inhibition pro- sociated sponge cells, 0.05 ml ofcell suspension was mixed with duced by antifibronectin antiserum at 1:60. No such inhibition the same volume ofeither the normal culture medium (control), was obtained with normal (nonimmune) rabbit serum or a 1% diluted normal rabbit serum, a mixture of fibronectin and an- solution of bovine serum albumin (Fig. 2D and E). tifibronectin antiserum, several dilutions (up to 1:480) of the antifibronectin antiserum, 1% gelatin, and 1% bovine serum DISCUSSION albumin (final concentration). The reaggregation was checked The above experiments clearly demonstrate the presence of a at 30 min through a light microscope with phase-contrast or fibronectin-like protein in cell membranes of freshwater dark-field optics. sponges. Similar results have been obtained with a marine sponge, Tethya. The most striking fact of the localization offlu- RESULTS orescence is the difference between motile and nomotile cells. Immunofluorescence Studies. The indirect immunofluores- The epithelial cells react at their edges. Unlike published data ence method using rabbit antibodies to human plasma CIg led on fibronectin location in vertebrate cells (18), in sponges fi- to clear fluorescence ofmost cells in freshwater sponge cultures. bronectin does not seem to be arrayed in fibrillar structures. However, the nature of the pattern observed depended on the However, this is consistent with the fact that internal contractile cell type. A major difference was observed between epithelial filaments of sponges cells usually show a more diffuse pattern cells and all other recognizable cell types such as choanocytes than do filaments of vertebrate cells. and lophocytes. The large epithelial cells stained mainly at the These experiments also show that this protein appears only region where they make contact with the other epithelial cells during differentiation in various specialized sponge cells such (Fig. 1A). The nonepithelial cells showed a more regular stain- as the large epithelial cells, the choanocytes, and the fibroblast- ing of the surface, apparently without any local concentration like cells. Freshly dividing undifferentiated gemmular archeo- (Fig. 1B and C). This staining was increasingly brighter in skel- cytes do not appear to contain this protein, at least not in a form etal cells (sclerocytes) secreting the siliceous spicules, in flag- recognizable by the antibodies applied in these experiments. ellate cells (choanocytes) ensuring the water flow through the The strong fluorescence observed after acetone treatment in the sponge, and, above all, in the highly motile cells which are the interior of the cells and on the cell membranes shows the pre- fibroblast-like cells (lophocytes) and multipotent cells (archeo- dominantly membrane localization of this protein. The weaker cytes). Only a weak fluorescence was detected on the fibrous fluorescence observed throughout the matrix, however, shows matrix. The intensity of the fluorescence was increased when that the sponge fibronectin-like protein is also associated with cells were made permeable with acetone. Only slight nonspe- the intercellular matrix and probably with the collagen fibrils cific fluorescence was detected in the shell ofthe gemmules and of the sponges. The fact that the antibodies used were raised in some cell nuclei. against human plasma fibronectin shows that the sponge fibro- In control experiments carried out with normal, nonimmune nectin-like protein possesses some of the immunodominant an- rabbit serum, no fluorescence could be detected with the flu- tigenic sites present in the human plasma glycoprotein. Evi- orescent sheep anti-rabbit IgG preparation. dence was presented recently by Vuento et aL (21) that the Fluorescence also could be completely abolished when pu- antigenic sites on human plasma fibronectin (CIg) are confor- rified fibronectin was added to the sponge cell cultures before mationally determined. This suggests strong conservation of at addition of the specific anti-CIg antiserum. These experiments least some of the structural features of fibronectin all through suggest that the fluorescence detected is specific and can be phylogenesis. attributed to the presence of a protein similar to human CIg in The experiments with dissociated sponge cells show that sponge cell membranes. sponge fibronectin-like protein probably plays an important Appearance of Fibronectin During Differentiation of Sponge role during the reassociation of dissociated sponge cells and Cell. When the immune fluorescent experiments were carried during morphogenesis and differentiation of sponges. This was out with recent sponge cell cultures (age, 3 days or less), no demonstrated by the inhibition of the reassociation of sponge specific fluorescence could be detected in the freshly divided cells by highly dilute specific anti-human plasma antifibronectin cells that had emigrated from the original explant. It appears antiserum and also by the inhibition of the reassociation by ex- therefore that the gemmular archeocytes do not yet contain the cess pure fibronectin or pure gelatin. It therefore is possible that fibronectin-like protein. This protein appeared as soon as the sponge cell membrane fibronectin-like protein is one ofthe rec- cells started to differentiate and took on an endothelium-like ognition sites that react specifically with the aggregation factors or fibroblast-'like appearance. It appears therefore that synthesis during the reassociation of the sponge cells. and membrane localization of fibronectin-like protein accom- pany the differentiation ofthe sponge cells during morphogenesis. The antifibronectin antisera were kindly provided by Prof. Vaheri Role ofFibronectin in the Association ofDissociated Sponge (Helsinki, Finland) and by Prof. Mosesson (New York). We gratefully Cells. Dissociation of the cells of freshwater sponges usually acknowledge the microscope facilities provided by Wild and Leitz and leads to the reaggregation and the subsequent formation of one the skillful technical assistance of Miss Dominique Brechemier in the or several new sponges (20). This process takes several days but purification of human plasma fibronectin. This work-was supported by the initial formation of the aggregates is accomplished in less Centre National de la Recherche Scientifique (Groupe de Recherche than an hour. When dissociated cells ofE. muller reaggregated, 40), D616gation Generale a la Recherche Scientifique et Technique, well-recognizable clusters of cells were formed in half an hour Institut National de la Sante et de la Recherche medicale, and the Sci- (Fig. 2A). The addition of antifibronectin antiserum to the me- entific Council of the Universite Paris-Val de Marne. dium of the reaggregating cells considerably modified the process of aggregation. After half an hour, at a dilution of 1:40 it 1. Morrisson, P. R., Edsall, J. T. & Miller, S. G. (1948) J. Am. largely inhibited the aggregation, and only some small aggre- Chem. Soc. 70, 3103-3108. 2. Hynes, R. 0. & Bye, J. M. (1974) Cell 3, 113-130. gates were visible (Fig. 2B). Most cells remained dissociated. 3. Vaheri, A. & Mosher, D. F. (1978) Biochim. Biophys. Acta 516, At increasing dilutions, this inhibition gradually decreased, and 1-25., at a dilution of 1:480 it practically disappeared. Addition ofsmall 4. Yamada, K. M. & Olden, K. (1978) Nature (London) 275, amounts of gelatin (Fig.. 2C) to the dissociated sponge cells in- 179-184. Downloaded by guest on September 29, 2021 Cell Biology: Labat-Robert et aL Proc. Nati Acad. Sci. USA 78 (1981) 6265

5. Ruoslahti, F., Vaheri, A., Kuusela, P. & Linder, E. (1973) 12. Junqua, S., Fayolle, J. & Robert, L. (1975) in Protides ofthe Bi- Biochim. Biophys. Acta 322, 352-358. ological Fluids, ed. Peeters, H. (Pergamon, Oxford), Vol 22, pp. 6. Engvall, E. & Ruoslahti, E. (1977) nt. J. Cancer 20, 1-5. 337-341. 7. Linder, E., Vaheri, A., Ruoslahti, E. & Wartiovaara, J. (1975)J. 13. Junqua, S., Fayolle, J. & Robert, L. (1975) Comp. Biochem. Phys- Exp. Med. 142, 41-49. iol 50B, 305-309. 8. Robert, L., Darrell, R. W. & Robert B. (1970) in Chemistry and 14. Junqua, S. (1979) Dissertation (Univ. des Sciences et Techniques Molecular Biology of the Intercellular Matrix, ed. Balazs, E. A. de Lille, Lille, France). (Academic, london), Vol. 3, pp. 1591-1614. 15. Robert, L. & Comte, P. (1968) Life Sci. 7, 493-497. 9. Robert, L. & Robert. B. (1973) in Protides of the Biological 16. Vuento, M. & Vaheri, A. (1979) Biochem. J. 183, 331-337. Fluids, ed. Peeters, H. (Pergamon, Oxford), pp. 125-132. 17. Ankel, W. E. & Eigenbrodt, H. (1950) Zool. Anz. 145, 195-204. 10. Robert, L., Junqua, S. & Moczar, M. (1976) in Frontiers ofMa- 18. Mautner, V. M. & Hynes, R. 0. (1977)J. Cell Biol 75, 743-758. trix Biology Burkitt Lymphoma, Haemostasis and Intercellular 19. Curtis, A. S. G. & Van de Vyver, G. (1971)J. Embryol Exp. Mor- Matrix, eds. Robert, A. M. & Robert, L. (Karger, Basel), Vol. 3, phol. 26, 295-312. pp. 113-142. 20. Van de Vyver, G. (1971) Ann. Embryol Morphogen. 4, 373-381. 11. Garrone, R. (1978) in Frontiers of Matrix Biology, Phylogenesis 21. Vuento, M., Salonen, E., Salminen, K., Pasanen, M. & Sten- of Connective Tissue, ed. Robert, L. (Karger, Basel), Vol. 3. man, U. H. (1980) Biochem. J. 191, 719-727. Downloaded by guest on September 29, 2021