INTRODUCTION

Intercellular adhesion and cellular adhesion to extracellular matrices are biological processes that are now known to play critical roles in a number of normal physiological events, which include immune recognition (1, 2, 3), embryonic development (4, 5), and wound healing (6), as well as in a number disease states such as inflammatory disease (1, 3), tumorigenesis (7, 8), angiogenesis (9), and metastasis (10). The molecules responsible for mediating cell-to-cell or cell-to-substrate adhesions, including the immunoglobulin family (11, 12), the cadherin family (13), the selectin family (14), and the integrin family (15, 16) have been intensively studied during the last 10 years.

The mechanisms of intercellular adhesion in primitive species have also been under investigation for many years. The first cell adhesion study was performed at the turn of the century using a simple animal model, the sponge. Wilson (17) demonstrated that when disrupted into single cells, suspensions of sponge cells rapidly readhered to one another and reformed new sponges. By mixing single cell suspensions of different sponge species that were identifiable by their different intrinsic colors, Wilson (18) showed that sponge cells sorted out by color; this study represented the first demonstration of the specificity of cell adhesion. Since those initial studies, investigators utilized invertebrate or nonmammalian vertebrate animal models of embryogenesis and development to uncover the molecular mechanisms governing cell adhesion. Nonmammalian animal models of cell adhesion include slug formation in the slime mold, Dictyostelium discoideum (19, 20, 21), and development of echinoderm (22, 23, 24), amphibian (25, 26, 27, 28), and avian (29, 30) embryos.

Although advancements in cell and molecular biology have recently enabled a better understanding of the nature of cell adhesion in mammalian species, continued studies into the mechanisms of adhesion in more primitive species is important to gain a basic understanding of the rigorously conserved features of cell adhesion molecules that have been selected by evolution. Studies on the mechanism(s) of cell adhesion in the sponge have indicated that the reaggregation of sponge cells is mediated by a large proteoglycan-like molecule termed ``aggregation factor'' (31, 32). Aggregation factor has glycan- and calcium-dependent self-interaction sites (33, 34, 35) as well as glycan-dependent (36) and calcium-independent cell-binding sites (37). This molecule has been characterized as a large (2 × 10⁷ daltons) stable proteoglycan on the basis of chromatographic and ultracentrifugation studies (32, 33).

This aggregation factor binds to sponge cells with a high affinity (33). In this report, a 210-kDa protein was isolated and was shown to be a high affinity cell surface and extracellular matrix glycoprotein which plays a significant role in cell adhesion. Characterization of this protein suggests that it may represent one of the earliest extracellular matrix adhesion proteins to have arisen in metazoan evolution.

EXPERIMENTAL PROCEDURES

Sponges were obtained from the supply department of the Marine Biological Laboratory in Woods Hole, MA, and maintained in artificial seawater at 4 °C. Nitrocellulose was obtained from Schleicher & Schuell. IODO-BEADS were from Pierce, and peanut agglutinin-Sepharose was from Vector Labs, Inc. (Burlingame, CA). Ampholines, lentil lectin, concanavalin A, and CNBr-Sepharoses were obtained from Pharmacia Biotech Inc. X-AR film was obtained from Kodak. Na¹²⁵I was from Amersham Corp. Prestained molecular mass markers were from Life Technologies, Inc. Peptide N-glycosidase was from Boehringer Mannheim. The Phast-gel system was from Pharmacia. Isoelectric focusing standard proteins were from Sigma. Polyvinylidene difluoride membrane (Problott) was from Promega (Madison, WI).

SDS-Polyacrylamide Gel Electrophoresis and Western Blotting Assay

Protein samples were solubilized in SDS sample buffer (2% SDS, 0.5 mM EDTA, 100 mM Tris, pH 8.8, 5% glycerol, with 5% 2-mercaptoethanol only where specifically described), boiled 3 min, and electrophoresed on 7.5% polyacrylamide gels using the buffers of Laemmli (38). Aggregation factor binding proteins were assayed by Western blotting using minor modifications of the procedures described by Towbin et al. (39). Transfers to nitrocellulose were for 4 h at 400 mA in 192 mM glycine, 25 mM Tris base at 4 °C. The blots were blocked at 4 °C overnight in 3% (v/v) bovine serum albumin, 1% casein, 500 mM NaCl, 50 mM Tris, pH 8.0, incubated in iodinated aggregation factor in blocking buffer, washed for 10 min per each of five changes of 500 mM NaCl, 50 mM Tris, pH 8.0, and autoradiographed for 2 days.

Dot Blotting

Aliquots of extracts or column fractions were spotted onto a sheet of nitrocellulose. The sheet was allowed to dry, was wetted by capillary action, and was blocked for 24 h at 4 °C in block buffer. Blots were then incubated in iodinated aggregation factor in block buffer for 24 h at 4 °C, washed five times in 0.5 M NaCl, 0.05 M Tris, pH 8.0, and then autoradiographed for 2 days.

Purification of the Aggregation Factor Binding Proteins

Cell suspensions from five large sponges containing 10¹² cells were three times frozen at 80 °C and thawed and homogenized in a final volume of 320 ml of 50 mM Tris, pH 8.0, 1 mM phenylmethylsulfonyl fluoride with 10 strokes of a hand-held Potter-Elvejhem type homogenizer. Membranes were recovered from the homogenate by centrifugation at 45,000 rpm for 1 h in a Ti60 rotor at 4 °C and extracted in 320 ml of 0.25% octylpolyoxyethylene, 0.5 M NaCl, 50 mM Tris, pH 8, 1 mM phenylmethylsulfonyl fluoride, aprotinin, and leupeptin for 12 h at 4 °C. Insoluble material was removed by centrifugation at 45,000 rpm for 1 h in a Ti60 rotor at 4 °C. Extracts were concentrated by ultrafiltration with an Amicon device at 4 atm of N₂ using YM30 membranes at 4 °C until the final volume equalled 20 ml. The extract was dialyzed versus 150 mM NaCl, 50 mM Tris, pH 8.0, for 12 h at 4 °C with three changes of buffer, and centrifuged at 45,000 rpm in a Ti60 rotor for 1 h at 4 °C to remove insoluble proteins.

A 15-ml QAE-Sephadex column was prepared in a plastic syringe plugged with glass wool and was washed with a gradient of 70 ml of 150 mM NaCl, 50 mM Tris, pH 8.0, to 1 M NaCl, 50 mM Tris, pH 8.0. After re-equilibration to starting conditions, the 20-ml concentrated membrane extract was applied to the column. After washing the column to remove unbound proteins, adsorbed proteins were eluted with a linear 70-ml salt gradient (150 mM to 1 M NaCl). Two-ml fractions were collected, and 100 µl of every fraction were assayed by Coomassie Blue staining of 7.5% SDS-polyacrylamide gels and by Western blotting and 50 µl of every fraction by dot blotting. Fractions containing 210-kDa aggregation factor-binding activity were pooled, dialyzed versus 0.5 M NaCl, 50 mM Tris pH 8.0, and concentrated by Amicon ultrafiltration using YM30 membranes.

Concentrated QAE-Sephadex fractions containing 210-kDa protein were chromatographed on a 160-ml (1.5 × 90 cm) Sepharose 6B column in 0.5 M NaCl, 50 mM Tris, pH 8.0, at a flow rate of 0.5 ml/min, and 2.3-ml fractions were collected. One hundred-microliter aliquots of each fraction were assayed by Western blotting, and 50-µl aliquots of each fraction were assayed by dot blotting. The optical density profile at 280 nm for the column elution was also determined. The 210-kDa protein-containing fractions were concentrated by Amicon ultrafiltration on YM30 membranes and were chromatographed on a 110-ml (1.5 × 70 cm) Sepharose 4B column in 0.5 M NaCl, 50 mM Tris, pH 8, at a flow rate of 0.5 ml/min. The optical density of each 2-ml fraction was determined. Protein peaks were pooled and concentrated and samples assayed after electrophoresis on 7.5% SDS-polyacrylamide gels by Western blotting.

Inhibition of Aggregation Assays

Inhibition of aggregation assays were performed by incubating 3 µg of aggregation factor (3 aggregation units, AU) with 200-µl aliquots of column fractions in wells of a 24-well culture plate for 30 min at room temperature. 4 × 10⁷ cells of a single cell suspension in 200 µl were added, and the plates were shaken on a rotary shaker for 30 min at room temperature. The degree of aggregation was determined by comparison with that induced by dilutions of aggregation factor in a standard aggregation assay. Percent inhibition was plotted versus fraction number.

Iodination of Purified 210-kDa Protein and Isoelectric Focusing

Fifty µl (0.5 µg) of purified 210-kDa protein was combined in a 1.5-ml microcentrifuge tube with 0.5 mCi of Na¹²⁵I and one IODO-BEAD for 30 min at room temperature. Free iodine was removed by chromatography on a 2-ml Sephadex G-50 column in 0.5 M NaCl, 50 mM Tris, pH 8. Twenty-microliter samples of the iodinated 210-kDa protein were electrophoresed on a nonreducing 7.5% SDS-polyacrylamide gel and on a 1-mm isoelectic focusing tube gel in a 9 M urea, 1% 2-mercaptoethanol sample buffer for 13.5 h. The isoelectic focusing tube gel was then electrophoresed on a reducing second dimension 7.5% SDS-polyacrylamide gel. After fixation and drying, the gels were analyzed by autoradiography.

Peptide-N-Glycosidase Treatment of Purified 210-kDa Protein

Duplicate aliquots of 0.1 µg of purified 210-kDa protein were placed in 40 µl of 0.5% SDS, 100 mM Tris, pH 8.8, and 10 mM EDTA and then denatured by boiling for 3 min. Samples were adjusted to 1% Nonidet P-40 by addition of 5 µl of 10% Nonidet P-40. Five µl (0.25 units) of peptide N-glycosidase was added to one test, and 5 µl of glycerol was added to the other (control) samples. Both samples were incubated for 12 h at 37 °C, and then 25 µl of 50% glycerol, 10% SDS, 0.1% bromphenol blue was added. Samples were boiled for 3 min, loaded onto 7.5% SDS-polyacrylamide gels, and analyzed by silver staining.

Lectin Affinity Chromatography

One hundred µl of packed lentil lectin, concanavalin A, peanut agglutinin, and unconjugated Sepharose resins were washed in 1 ml of 1% bovine serum albumin, 0.5 M NaCl, 0.05 M Tris, pH 8.0, at 4 °C overnight and then rinsed three times in 1% Nonidet P-40, 0.5 M NaCl, 0.05 M Tris, pH 8.0 (lysis buffer) by centrifugation for 5 min at 1000 rpm. 5 × 10⁷ cells were solubilized in 100 µl of lysis buffer. One hundred microliters of lysates were incubated with each resin with rotation for 2 h at 4 °C. Lectins were always maintained in the presence of 1 mM CaCl₂ and 1 mM MnCl₂. Unbound lysates were removed and resins washed four times with lysis buffer. Bound protein was solubilized in SDS sample buffer, as was a sample of the lysate. Samples were evaluated by the aggregation factor Western blotting assay.

RESULTS

Sponge cell adhesion depends on an extracellular matrix proteoglycan that is thought to promote adhesion based on its ability to bind to a putative cell surface ligand. A potential cell surface ligand for this proteoglycan, a sponge proteoglycan-binding protein of 210 kDa, was previously identified by a ligand binding Western blotting assay (40). This protein was characterized as a nonintegral membrane, cell surface-bound extracellular matrix protein ligand for the major sponge proteoglycan (40). The protein was characterized as a nonintegral membrane protein on the basis of its solubility in the absence of detergent, failure to bind lipid, and presence in preparations of sponge extracellular matrix as well as in detergent and nondetergent extracts of sponge membranes. A 68-kDa peripheral membrane/extracellular matrix protein ligand for the sponge proteoglycan was also identified using this assay (40) and was recently purified and characterized as a cell adhesion protein (41). Therefore, experiments were designed to purify the 210-kDa sponge proteoglycan ligand and to establish its role as a sponge adhesion protein.

A 210-kDa aggregation factor binding protein was purified by a combination of anion exchange and gel filtration chromatography using an extraction procedure that was selected in order to maximize the ratio of 210-kDa protein to contaminating proteins. An octylpolyoxyethylene extract of sponge membranes was dialyzed to remove detergent, and the resulting soluble extract was initially applied to a 15-ml QAE-Sephadex column. The QAE-Sephadex column was eluted with a linear salt gradient (0.15-1.0 M NaCl). While the majority of the extracted protein eluted in fractions 1-21 of an anion exchange column eluate (Fig. 1, upper panel), the 210-kDa protein eluted in fractions 36-47 (0.8-1.0 M NaCl), as detected by the binding of radiolabeled aggregation factor to Western and dot blots (Fig. 1, middle and lower panels). The 210-kDa protein was clearly separated from the 68-kDa aggregation factor ligand that eluted in fractions 29-35 and that was previously purified and characterized (40, 41).

To purify the 210-kDa protein further, the anion exchange column fractions 36-47 were applied to a Sepharose 6B sizing column. The 210-kDa aggregation factor binding activity eluted in the column void volume, in fractions 20-30 (Fig. 2, middle and lower panels) but apart from the majority of the total protein (Fig. 2, upper panel).

The 210-kDa Sepharose 6B column void volume fractions 20-30 were then chromatographed on a Sepharose 4B column. Two major protein peaks were detected by optical density measurements at 280 nm (Fig. 3A). The 210-kDa protein peak with aggregation factor binding activity was detected exclusively in the first included peak, in fractions 18-24, using the aggregation factor Western blotting assay (Fig. 3B). A yield of 500 µg of 210-kDa aggregation factor ligand was obtained from 5 g of total cell protein using this procedure, with an enrichment of 1600-fold, based on the aggregation factor Western blot binding assay (Table I).

Table I. Purification of 210-kDa MAF ligand Aliquots of cells, membranes, dialyzed opoe extract, and the 210-kDa protein containing peaks from QAE-Sephadex, Sepharose 6B, and Sepharose 4B column chromatography were assayed for protein concentration with a commercially available dye binding assay and for 210-kDa aggregation factor binding protein by the previously described Western blotting assay using iodinated aggregation factor at 3 × 106 cpm/g (or 6 × 1019 cpm/mol). Iodinated aggregation factor bound to the 210-kDa protein was determined by excising the 210-kDa protein of washed blots and counting cpm of 125I. Total activity is expressed as the total cpm of aggregation factor each fraction can bind, and specific activity is expressed as the cpm of aggregation factor bound per µg of protein in each fraction. Enrichment is the fold increase in specific activity of each step in the purification over the specific activity of starting material (cells). Fractionation step Yield Total activity Specific activity Enrichment µg protein cpm MAF bound cpm MAF/µg protein Cells (1012) 5  × 106 7  × 106 2 Membranes 2.5  × 105 6.4  × 106 256 12.8 Extract 5  × 104 3.3  × 106 66 33 QAE peak 1.8  × 104 3  × 106 170 85 Sepharose 6B peak 1.8  × 103 2  × 106 1111 555 Sepharose 4B peak 500 1.6  × 106 3200 1600

Purification of the 210-kDa protein ligand permitted further analysis of its biological and biochemical characteristics. A sample of the Sepharose 4B protein peak was radiolabeled by iodination and evaluated for purity by gel electrophoresis. Electrophoretic analysis of the pooled and concentrated 210-kDa protein on 7.5% SDS-polyacrylamide gels followed by autoradiography (Fig. 4, ¹²⁵I) and silver staining (Fig. 4, Silver) revealed a single protein band of 210-kDa in both cases, indicating that the 210-kDa protein had been purified to homogeneity. Analysis of an unlabeled sample by the aggregation factor Western blot binding assay (Fig. 4, MAF Overlay) verified that this purified protein retained the aggregation factor binding activity and that no other aggregation factor binding protein was present in these protein preparations.

The 210-kDa protein-aggregation factor interaction is sensitive to reduction; the 210-kDa protein binds aggregation factor when it has been electrophoresed on nonreducing but not on reducing SDS-polyacrylamide gels prior to Western blotting (40). In contrast to the effect of reduction on its binding properties, no change in the migration of the 210-kDa protein under reducing conditions was observed (Fig. 5A). Because the migration of the protein, in contrast to its interaction with aggregation factor, does not change upon reduction, these results suggest that one or more intrachain disulfide bonds that do not significantly affect SDS gel migration play a role in the secondary structure of the active site.

An analysis of the 210-kDa protein by one- and two-dimensional isoelectric focusing revealed an isoelectric point of 4.3 (Fig. 5B), with a slight heterogeneity of focused bands at that pI. Minor heterogeneity of isoelectric points in a purified protein is often attributed to slight differences in glycosylation or other post-translational modifications (42). An analysis of the glycosylation state of the 210-kDa protein undertaken by digestion of the purified 210-kDa protein with glycosidases suggests that it is not highly glycosylated (data not shown; Ref. 40). However, an analysis of the affinity of the 210-kDa protein for lectins by lectin affinity chromatography demonstrates that the 210-kDa protein contains some complex glycan. The 210-kDa protein bound lentil lectin Sepharose (Fig. 6, lane A) and peanut agglutinin Sepharose (Fig. 6, lane B), but not concanavalin A-Sepharose (Fig. 6, lane C), which selectively binds high mannose-type glycans. These results indicate that the 210-kDa protein is a glycoprotein with complex-type glycans, but also that it is not highly glycosylated.

To demonstrate that the 210-kDa protein purified from the cell surface is indeed the same as that found in the extracellular matrix of the sponge, aggregation factor was allowed to bind preparations of the extracellular matrix and to purified 210-kDa protein by incubation of Western blots with iodinated aggregation factor (not shown). On the basis of molecular weight and affinity for aggregation factor in Western blotting assays, the purified cellular 210-kDa protein appears to be identical to the 210-kDa protein that is present in preparations of extracellular matrix. These results verify the previously published observation (40) that the 210-kDa protein is in fact an extracellular matrix protein with binding sites for the proteoglycan aggregation factor and the cell surface.

Purified 210-kDa protein was examined for its ability to inhibit aggregation factor-mediated cell adhesion. The 210-kDa protein was able to inhibit completely sponge cell aggregation (Fig. 7A). A concentration of 8 µg/ml was required to inhibit 50% of the aggregation induced by 1 µg of aggregation factor. In contrast, other protein fractions such as those from the second included peak of the Sepharose 4B chromatography column, a purified 220-kDa sponge protein which does not bind aggregation factor in the Western blotting assay and which is distinct from the 210-kDa protein in that it elutes from the anion exchange column in the flow-through and contains a covalently attached chromophore,¹ as well as purified mammalian extracellular matrix proteins such as fibronectin, laminin, and gelatin, were not able to inhibit the aggregation factor-mediated aggregation of sponge cells (Fig. 7A). These findings suggest that the 210-kDa protein is specifically associated with sponge cell adhesion.

The 210-kDa protein is a high affinity adhesion ligand for aggregation factor. A, the 210-kDa protein inhibits cell adhesion. Sepharose 4B peak one (pooled fractions 18-24) containing 210-kDa protein and peak two (pooled fractions 30-55) as well as samples of a purified 220-kDa sponge protein and mammalian fibronectin, laminin, and gelatin were assayed for inhibition of aggregation factor-mediated cell aggregation. Results are shown as percent inhibition of aggregation as a function of protein concentration. B, Scatchard analysis of the 210-kDa protein-aggregation factor interaction. Dilutions of iodinated aggregation factor were incubated for 8 h at 4 °C with 210-kDa protein from sponge cell extracts that had been electrophoresed and blotted onto nitrocellulose and then blocked with 3% bovine serum albumin in 0.5 M NaCl, 50 mM Tris, pH 8, 2 mM CaCl₂. After washing in 0.5 M NaCl, 50 mM Tris, pH 8, 2 mM CaCl₂ four times to remove unbound aggregation factor, bound cpm were determined. After subtracting nonspecific cpm, bound nmol/liter were calculated and plotted versus total nmol/liter (inset) and bound/free ratio was plotted versus bound nmol/liter. A dissociation constant (K_d) of 7 × 10⁹ M was determined from the slope, according to Scatchard (44). C, inhibition of aggregation factor cell binding by antisera to cell surface proteins. Aliquots of 10⁶ cpm of aggregation factor were incubated for 2 h at 4 °C with live dissociated sponge cells in the presence of dilutions of preimmune or immune antisera reactive with 210-kDa protein. Cell-bound aggregation factor was determined by gamma radiation counting after washing four times with cold calcium- and magnesium-free sea water. Bound cpm were plotted versus dilution factor of antisera.

Scatchard analysis (43) of the affinity of aggregation factor for immobilized 210-kDa protein indicates that the interaction is of high affinity (K_d = 7 × 10⁹ M; Fig. 7B). In further support of the role of the 210-kDa protein in sponge cell adhesion are experiments that demonstrate that immune sera containing antibodies that immunoprecipitate the 210-kDa protein (40), but not preimmune sera, potently inhibit the binding of aggregation factor to sponge cell surfaces (Fig. 7C). These antisera also inhibit sponge cell adhesion.¹ Taken together, these results indicate that the 210-kDa protein plays a major role in sponge cell adhesion.

DISCUSSION

Cell-to-cell and cell-to-substrate adhesion molecules are key elements of both normal and pathological events that include the immune response, development, tumorigenesis, and metastasis (4, 8, 10, 15, 16). Several families of integral membrane cell adhesion proteins including the integrin (15, 16), selectin (13), immunoglobulin (12), and cadherin (13) superfamilies have been extensively characterized during the past decade.

Although much is now known about mammalian adhesion molecules, invertebrate and nonmammalian animal models of cell adhesion have been the focus of cell adhesion research since the turn of the century. An understanding of the events and the evolutionary forces that drive change among adhesion molecules can lead to a better understanding of mammalian adhesion molecules and their mechanisms of action. For example, purified extracellular matrix proteins have been recently studied from a number of nonmammalian or invertebrate species to acquire insight into the functions and evolution of these kinds of molecules (44, 45, 46).

The first cell adhesion study, however, was performed as early as 1907 using the sponge, a primitive metazoan, as a simple animal model of cell aggregation (17). Research on the mechanisms of cell adhesion in the sponge later demonstrated that sponge cell adhesion is mediated by a large, calcium-dependent proteoglycan-like molecule termed aggregation factor (31, 32) which has high affinity cell-binding sites (33). A 210-kDa cell surface ligand for the aggregation factor was first identified and characterized as a cell surface-associated extracellular matrix protein using iodinated aggregation factor as a probe in a Western blotting assay (40). The protein was characterized as a nonintegral membrane, extracellular matrix protein on the basis of its solubility in the absence of detergent, failure to bind lipid, and presence in preparations of sponge extracellular matrix as well as in detergent and nondetergent extracts of sponge membranes.

In the studies described in this report, this 210-kDa invertebrate extracellular matrix ligand for the adhesion-mediating proteoglycan aggregation factor was purified to homogeneity. Several lines of evidence suggest that the 210-kDa protein plays a major role in cell adhesion in sponges. First, the 210-kDa protein has a high affinity for aggregation factor (K_d = 7 nM), and it binds aggregation factor in affinity chromatography experiments (40) and on Western and dot blots. Second, the 210-kDa protein can completely inhibit the aggregation of cells mediated by aggregation factor. Third, antibodies that immunoprecipitate the 210-kDa protein also block aggregation factor binding to cells and aggregation factor-mediated cell adhesion. Together, these observations suggest that the 210-kDa ligand functions to anchor aggregation factor to the cell surface and to mediate cell adhesion.

The 210-kDa ligand for the sponge proteoglycan has a subunit molecular mass of 210-kDa, but a larger native molecular mass, suggesting that it forms macromolecular complexes, as do extracellular matrix proteins such as tenascin (46). The function of the protein is dependent on intramolecular, but not intermolecular, disulfide bonds because reduction decreases its ability to bind its ligand, aggregation factor, but not its migration on SDS-polyacrylamide gels. The aggregation factor-interaction site is also dependent on a structure that is sensitive to proteolysis (40) and cyanogen bromide cleavage.¹ A slight charge and size heterodispersity exhibited on isoelectric focusing gels may be due to a slight amount of glycosylation. Although its migration on gels is not altered by glycosidase digestion, the protein binds lentil lectin and peanut agglutinin Sepharoses and hence appears to be a complex-type glycoprotein. Although early experiments suggested that the 210-kDa aggregation factor ligand did not bind lectins (40), a more detailed analysis of lectin binding indicated that the 210-kDa protein does weakly bind lentil lectin and peanut agglutinin.

The 210-kDa protein is found on the cell surface as well as in the extracellular matrix, suggesting its role as an extracellular matrix protein that serves as an intermolecular bridge between a cell surface receptor and the proteoglycan-like aggregation factor. Although sponges are among the most primitive of metazoans, their extracellular matrix resembles that of higher species. The sponge extracellular matrix is composed of collagen fibrils (47), proteoglycans, and perhaps structural proteins. It is possible that homologues of the mammalian structural extracellular matrix proteins such as laminin, fibronectin, vitronectin, or fibrinogen that bind to both cell surface receptors, collagen, and extracellular matrix proteoglycans (48, 49) exist in sponges. In fact, a 230-kDa tenascin-like hexamer with immunoreactivity to anti-tenascin antibodies has recently been described in one sponge species (50). The 210-kDa protein described in this article may play a role similar to that of these structural matrix molecules in that it associates with the cell surface and with a matrix proteoglycan; it may be a homologue of the sponge tenascin molecule recently described (50).

The evidence for the role of this 210-kDa protein in cell adhesion and for its presence in the extracellular matrix as well as on the cell surface suggests that the 210-kDa protein may be an ancestral extracellular matrix molecule. This novel 210-kDa extracellular matrix protein ligand for the sponge proteoglycan aggregation factor may be one of the most evolutionarily primitive extracellular matrix proteins yet described. Sequence analysis (via protein and cDNA sequencing) should help establish the ancestry of this matrix adhesion protein and to define the evolutionary lineages of some of the more well characterized vertebrate adhesion proteins.