Journal of Cell Science 107, 2323-2333 (1994) 2323 Printed in Great Britain © The Company of Biologists Limited 1994

Tenascin-R (J1 160/180) inhibits fibronectin-mediated cell adhesion - functional relatedness to -C

Penka Pesheva1,*,†, Rainer Probstmeier1, Amy P. N. Skubitz2, James B. McCarthy2, Leo T. Furcht2 and Melitta Schachner1 1Department of Neurobiology, Swiss Federal Institute of Technology, Hönggerberg, 8093 Zürich, Switzerland 2Department of Laboratory Medicine and Pathology, and Biomedical Engineering Center, University of Minnesota, Medical School, 420 Delaware Street SE, Minneapolis, Minnesota 55455, USA *Present address: Laboratory of Biochemistry, Swiss Federal Institute of Technology, 8092 Zürich, Switzerland †Author for correspondence

SUMMARY

Cell adhesion and neurite outgrowth on fibronectin is a neural cells. This effect is partially due to interactions at multistep process modulated by different extra- and intra- the substrate level that result in a steric hindrance and/or cellular signals. -mediated cell attachment and conformational change of the cell binding sites of the spreading can be affected in a negative way by tenascin-C, fibronectin molecule. In addition, tenascin-R 180 and an expressed in a tempo- tenascin-C interact with cells by an RGD- and β1 - rally and spacially restricted manner during early mor- independent mechanism, leading to cell rounding and phogenesis. Tenascin-R (J1-160/180), consisting of two detachment from such substrata. The expression of major isoforms of 160 kDa (tenascin-R 160) and 180 kDa tenascin-R and tenascin-C in the nervous system at times (tenascin-R 180) in mammals, is an extracellular matrix and locations where fibronectin-mediated cellular glycoprotein of the central nervous system that shares high processes take place may be related to the role of inhibitory structural homologies with tenascin-C. Here we show that signals in the extracellular matrix in the regulation of cell in relation to fibronectin-mediated adhesion, the two extra- migration and differentiation in general. cellular matrix molecules are also functionally closely related. When offered as mixed substrata with other extra- cellular matrix molecules, the two tenascin-R isoforms and tenascin-C derived from mouse brain selectively inhibit Key words: cell adhesion, extracellular matrix, fibroblast, fibronectin-dependent cell adhesion and neurite outgrowth, fibronectin, β1 integrin, inhibitory substrate, neurite outgrowth, and affect cell morphology of different mesenchymal and tenascin

INTRODUCTION First, during pattern formation in the peripheral (PNS) and central (CNS) nervous systems, FN supports cell adhesion and Morphogenetic events during normal development are criti- neurite outgrowth mediated by interaction with cell surface cally dependent on cellular interactions with the surrounding (an RGD-dependent process) and proteoglycans (an extracellular matrix (ECM), which result in RGD-independent process), two major groups of FN receptors via transmembrane , regulation and altered (FNRs) (reviewed by Reichardt and Tomaselli, 1991). Second, cell behavior. These interactions could thus navigate cells TN interferes with FN-mediated cell adhesion in vitro and may through the different stages of proliferation, migration and final thus modulate cellular morphology and differentiation state differentiation (for review, see Hynes and Lander, 1992). (Chiquet-Ehrismann et al., 1988; Chiquet-Ehrismann, 1991; Among the most intensely studied ECM molecules, fibronectin Lightner and Erickson, 1990; Probstmeier et al., 1990b). Third, (FN) and tenascin (TN) are two structurally related multido- J1 160/180, an ECM molecule in the CNS of mammals that main glycoproteins that affect cellular adhesiveness and differ- consists of two major isoforms of 160 and 180 kDa (Pesheva entiation in a positive and negative way (reviewed by Chiquet- et al., 1989), and its species homolog restrictin in the chicken Ehrismann, 1990; Adams and Watt, 1993). During (Rathjen et al., 1991), are structurally related to TN in that they development of the nervous system, their action may influence consist of EGF- and FN type III-like repeats, and a C-terminal neuronal cell migration and neurite extension (Reichardt and region homologous to fibrinogen (Nörenberg et al., 1992; Fuss Tomaselli, 1991; Wehrle and Chiquet, 1990; Lochter et al., et al., 1993). Together with the newly identified TN-like ECM 1991; Husmann et al., 1992; Prieto et al., 1992). The two encoded by a gene in the human major histocompati- molecules deserved our interest for the following reasons. bility locus, termed TN-X (Bristow et al., 1993), they have 2324 P. Pesheva and others recently been defined as members of the TN multigene family of Minnesota using a synthesizer (System 990; Beckman with so far identified three family members: TN-R (restrictin Instruments Co., Fullerton, CA; McCarthy et al., 1990) and conju- and J1 160/180), TN-C (tenascin/cytotactin) and TN-X (the gated to ovalbumin (Drake et al., 1992). These included the heparin large TN-like protein in human) (for review, see Erickson, binding FN-C/H-I and FN-C/H-II, containing the sequences 1993). In contrast to TN-C, TN-R has so far been found YEKPGSPPREVVPRPRPGV and KNNQKSEPLIGRKKT, respec- expressed only in the CNS by oligodendrocytes and some tively (McCarthy et al., 1988, 1990), and peptide CS1 containing the sequence DELPQLVTLPHPNLHGPEILDVPST (Humphries et al., neuronal subpopulations (Fuss et al., 1993; Pesheva et al., 1987), which does not bind heparin and promotes adhesion of 1989; Rathjen et al., 1991). In the mouse CNS, TN-R and TN- melanoma cells (McCarthy et al., 1990) and neurite extension of C are expressed in different temporal windows and by different dorsal root ganglion (DRG) neurons (Humphries et al., 1988), all cell types during development (Pesheva et al., 1989; Prieto et located within the 33/66 kDa FN fragment. The molar ratio of each al., 1990). As well as their structural relatedness, the two ECM peptide bound to ovalbumin was 1:12 for CS1, 1:4 for FN-C/H-I and glycoproteins display functional similarities. When applied as 1:5 for FN-C/H-I (Drake et al., 1992). The RGD-containing peptide substrates for neural cells in vitro, they cause repulsion of CNS (GRGDS) was purchased from Boehringer Mannheim. neurons and inhibit neurite outgrowth into substrate areas Antibodies enriched in these molecules (Faissner and Kruse, 1990; Pesheva et al., 1989, 1993). It was therefore an intriguing Monoclonal antibodies 596 and 597 from mouse, each recognizing different epitopes on TN-R 160 and TN-R 180, have been character- question as to whether TN-R would display similar functional ized (Pesheva et al., 1989). Polyclonal antibodies to TN-R 160, which properties to TN-C in relation to FN-mediated adhesion of cells react with both TN-R 160 and TN-R 180, but not with TN-C, were and, if so, what are the mechanisms behind it. produced in rabbits (Pesheva et al., 1991). Rat monoclonal antibody The aim of the present study was to investigate the func- J1/tn2 recognizes all isoforms of TN-C (Faissner and Kruse, 1990). tional consequences of the association between the TN-R Polyclonal antibodies specific for TN-C were raised in rabbits by three isoforms or TN-C with FN in FN-mediated cell adhesion and subcutaneous injections (the first in complete and the following two neurite outgrowth, and compare the molecular mechanisms in incomplete Freund’s adjuvant) at two-week intervals with 40 µg/ml involved. of brain-derived TN-C per animal. Polyclonal antibodies to LN from Here we show that the two TN-R isoforms, TN-R 160 and EHS sarcoma and to human plasma FN (both from Bethesda Research TN-R 180, and TN-C derived from mouse brain, inhibit FN- Laboratories) were produced in rabbits (Pesheva et al., 1989). Poly- clonal rabbit antibodies directed against the FNR derived from CHO dependent cell adhesion and neurite outgrowth, and affect cell cells (a kind gift from Dr R. L. Juliano; Brown and Juliano, 1986) and morphology. For TN-R 160, this effect is due to interaction recognizing the 140 kDa heterodimeric (α5β1) integrin complex in the with FN, resulting in a covering and/or conformational change mouse (Pesheva et al., 1988) were also used. in the cell binding sites on the FN molecule. TN-R 180 and TN-C bind to putative cellular receptor(s) different from β1 Analytical procedures integrin FNRs, thereby leading to cell rounding and detach- Protein determinations were carried out according to Bradford (1976). ment, which imply signal transduction mechanisms that follow Protein samples were separated by SDS-PAGE using 7% polyacry- the initial recognition event. We propose that TN-R 180 and lamide slab gels, and protein bands were visualized by the reducing TN-C are functionally closely related and that their deposition silver staining method (Oakley et al., 1980). in the ECM can modulate cell migration and differentiation. Cell cultures Mouse L 929 fibroblast cells (L cells; Pantazis and Jensen, 1988) were maintained in Eagle’s basal medium (BME) containing 10% horse MATERIALS AND METHODS serum. CV-1 (monkey kidney epithelial cell line; Schneider-Schaulies et al., 1990), WISH (human amniotic epithelial cell line, a kind gift Extracellular matrix from Dr G. Keilhauer, Knoll, Ludwigshafen) and mouse N2A neu- The two TN-R isoforms, TN-R 160 and TN-R 180, were purified from roblastoma cells (Rathjen and Schachner, 1984) were maintained in adult mouse brain by immunoaffinity chromatography using 597 and RPMI containing 10% fetal calf serum (FCS, Gibco). TSC, trans- 596 monoclonal antibody columns (Pesheva et al., 1989). TN-C was formed rat Schwann cells (pMT4 SVneo H1, a kind gift from Dr G. obtained from early postnatal (P3) mouse brain by immunoaffinity I. Tennekoon; Peden et al., 1990), were maintained in RPMI con- chromatography on a monoclonal TN-C antibody (J1/tn2) column taining 5% FCS and 100 µM ZnCl2. Primary cultures of Schwann (Faissner and Kruse, 1990; Lochter et al., 1991). Human plasma FN cells and DRG neurons were prepared from neonatal ICR mice and was purified by sequential ion-exchange and affinity chro- maintained in Dulbecco’s modified Eagle’s medium containing 10% matography (McCarthy et al., 1986). (LN) from EHS FCS and 20 ng/ml (Boehringer Mannheim; Seil- sarcoma was purchased from Sigma. type I was obtained heimer and Schachner, 1987). from Serva as a protein solution of 2 mg/ml in 0.1% acetic acid. Collagen type IV from mouse EHS sarcoma was isolated after mild Treatment of L cells with glutaraldehyde pepsin treatment whereby the C-terminal NC1 domain was destroyed L cells from monolayer cultures were gently trypsinized with 10 (Fahrig et al., 1987). µg/ml of trypsin in Ca2+- and Mg2+-free Hanks’ balanced salt solution (CMF-HBSS) for 5 minutes at room temperature. To eliminate the FN-derived fragments and synthetic peptides intracellular response mechanisms of live cells on one hand, and to The 75 kDa tryptic fragment containing the central cell-binding preserve the activity of cell surface integrins on the other, single cell domain of FN was generated and purified as described previously suspensions were washed with a divalent cation-containing buffer, i.e. (Hayashi and Yamada, 1983; Humphries et al., 1988). The 33/66 kDa HEPES buffer (10 mM HEPES, 100 mM NaCl, 3 mM MgCl2, 2 mM heparin binding fragment was purified from tryptic/catheptic digests CaCl2, pH 7.4), and treated with 1% glutaraldehyde (Serva) in HEPES of FN (McCarthy et al., 1988, 1990). The 45 kDa collagen binding buffer for 1 hour on ice. Cells were then washed three times in 0.1 M 2+ 2+ fragment was purchased from Chemicon (La Roche). Peptides from glycine in Ca /Mg -containing Tris-buffered saline (3 mM MgCl2, FN were synthesized at the Microchemical Facility of the University 2 mM CaCl2, 20 mM Tris-HCl, 150 mM NaCl, pH 7.2) followed by

Tenascin-R and tenascin-C block fibronectin-mediated adhesion 2325 two washing steps with HEPES buffer. Cells were stored at 4¡C as a with 1% BSA in PBS for 1 hour at room temperature. TN-R and TN- 5% suspension in HEPES buffer containing 0.02% sodium azide. For C (5 µg/ml in PBS containing 1 mg/ml BSA), with or without binding studies, single cell suspensions were used at a final density additives, were added to each well (100 µl/well) and incubated for 1 of 1×106 cells/ml as described in the following section. hour at 37¡C. Plates were washed three times with PBS and specific binding was determined by using monoclonal antibodies 596 and Cell-to-substrate adhesion assays J1/tn2, or polyclonal rabbit antibodies to TN-R and TN-C, respec- Cell-to-substrate adhesion assays were performed essentially as tively, and secondary antibodies coupled to horseradish peroxidase described (Pesheva et al., 1989, 1991). Briefly, TN-R and TN-C, alone (HRP, Promega). Each value was determined in at least three inde- (40 µg/ml in PBS) or mixed with other ECM molecules or fragments pendent experiments carried out in triplicate. (1:1 (w/w) on a protein basis at a final protein concentration of 20 The coating efficiencies of the 45 kDa collagen binding, 75 kDa µg/ml for 1 hour at room temperature), were coated as 2-3 µl droplets cell binding, and 33/66 kDa heparin binding fragments of FN on onto tissue culture Petri dishes (30 mm in diameter) or into the wells microtiter plates were determined by using 125I-labeled FN fragments of 24-well plates (Nunclon) for 60 minutes at 37¡C. The oligomer as described above. At equimolar concentrations, the coating effi- status of the purified molecules is as follows: dimers, for TN-R 160 ciency of the 75 kDa fragment was the highest observed (about 20% and FN; trimers, for TN-R 180; and hexamers, for TN-C (mainly rep- of the input). Compared to this value, the coating efficiencies of the resented by the 200 and 220 kDa isoforms in an approximate ratio of 45 and 33/66 kDa fragments comprised about 95% and 65%, respec- 1:1). At a final protein concentration of 20 µg/ml in mixed substrates, tively. the calculated molarities of the individual components were: 62 µM, for TN-R 160; 37 µM, for TN-R 180; 16 µM, for TN-C; and 45 µM, for FN. When TN-R 160 (31 µM) and TN-R 180 (18.5 µM) were coated in a mixture with FN (45 µM) at a final concentration of 10 RESULTS µg/ml, the same adhesion patterns were observed. Overlayed sub- strates were prepared in the following way: BSA (lipid-free bovine Selective inhibition of FN-mediated adhesion and serum albumin, from Sigma), TN-R and TN-C (10 or 20 µg/ml in neurite outgrowth by TN-R and TN-C PBS) were coated onto culture dishes for 30 minutes at 37¡C. FN and To study the effect of TN-R on FN-mediated adhesion, we LN (10 or 20 µg/ml) were then coated on top of the first substrates examined the behavior of different cells on substrates contain- for 2 hours at 37¡C. Unsaturated protein binding sites were blocked ing TN-R 160 or TN-R 180 compared to that on substrates con- with 1% BSA in PBS for 60 minutes at 37¡C and the dishes were sub- taining TN-C derived from early postnatal mouse brain. First, sequently washed three times with CMF-HBSS. Probe cells were obtained from monolayer cultures by gentle trypsinization (10 µg TN-R 160, TN-R 180 and TN-C were coated onto tissue trypsin/ml CMF-HBSS) for 5 minutes at room temperature. culture dishes either alone or in a mixture with other ECM Trypsinization was terminated by addition of ice-cold BME contain- molecules and the adhesion of L cells to these substrates was ing 10% horse serum. Cells were centrifuged at 600 g for 10 minutes examined 2 hours after plating (Fig. 1A). In a mixture with FN, at 4¡C and resuspended in the same medium (incubation medium). both TN-R 160 (a) and TN-R 180 (b), as well as TN-C (c), Alternatively, BME containing 0.1% BSA was used as an incubation drastically inhibited cell adhesion and spreading. The number medium with identical results. Single cell suspensions were plated of cells on these substrates comprised about one third of those into the substrate-coated dishes at densities of 1×106 cells/ml and adhering to FN/BSA substrates (d, Fig. 1B). The inhibitory incubated for one or two to three hours at 37¡C. Nonadherent cells effect of TN-R- and TN-C-containing substrates correlated were removed by gently washing the dishes with CMF-HBSS and specifically with FN-mediated cell adhesion, since mixed sub- cell-to-substrate adhesion was evaluated microscopically. For estima- tion of the number of cells adhering to the substrate, cells from micro- strates of TN-R 160/180 and TN-C with LN (Fig. 1A, e-g) or graphs were counted in microscope fields corresponding to 800 µm2. collagen type I (Fig. 1A, i-k) did not produce different Mean values ± s.d. of adherent cells in 6 to 10 different microscopic adhesion patterns from those on LN (Fig. 1A, h) and collagen fields from at least three independent experiments are presented. type I (Fig. 1A, l) mixed with BSA (Fig. 1B). Like TN-C, TN- R 160 and TN-R 180 substrates did not support the adhesion Determination of the coating efficiency of TN-R and FN in of L cells after 2 hours of incubation (not shown). mixed substrates A possible explanation for these observations could be that TN-R and FN were radioactively labeled with 125I-Bolton-Hunter TN-R and TN-C bring about their inhibitory effect by an reagent (from Amersham) (Bolton and Hunter, 1973) as described 125 interaction with: (i) FN, which leads to a steric hindrance or (Probstmeier et al., 1990a). To determine the coating efficiency, I- conformational change of the cell binding sites of the FN labeled TN-R or FN was used as a tracer and mixed with unlabeled protein at a ratio of 1:20. These proteins alone or in a mixture with molecule; and/or (ii) cell surface receptor(s), which thus BSA, LN or TN-C were coated to tissue culture plates as described prevents cell attachment and spreading on FN. Supporting the under ‘Cell-to-substrate adhesion assays’. The coating efficiency of second possibility are recent studies on the kinetics of L cell 125I-labeled TN-R in a mixture with FN was the same as in a mixture adhesion to TN-R and TN-C demonstrating that cells initially with BSA or LN. The coating efficiency of 125I-labeled FN in a attach to such substrates and are rapidly repelled already after mixture with TN-R or TN-C was slightly reduced (less than 20%) 15 minutes of incubation, suggesting an interference with when compared with its mixture with BSA. second signaling pathways (Pesheva and Schachner, unpub- lished data). For TN-C, it has been shown that the adhesion Interaction of TN-R and TN-C with other ECM molecules and spreading of cells of different origin on FN are negatively Binding of TN-R 160, TN-R 180 and TN-C to FN and FN-derived affected by the molecule, a process in which interference with fragments, LN, collagen types I and IV and BSA was examined by a solid-phase binding assay (Douillard and Hoffman, 1983). Wells of both FN substrate and cellular receptor(s) has been suggested microtiter plates (Falcon 3912) were coated with different ECM (reviewed by Chiquet-Ehrismann, 1991). To investigate molecules (5 or 10 µg/ml in PBS, 100 µl/well) or FN-derived whether TN-R displays the same inhibitory effect as TN-C, fragments (at equimolar concentrations in PBS, 70 µl/well) and the we studied the attachment of different cell lines (L 929, CV- unsaturated binding sites on the plastic were blocked by incubation 1, WISH, TSC, N2A) and cells from primary cultures

2326 P. Pesheva and others

B

120 + BSA Fig. 1. Selective substrate inhibition of FN-mediated adhesion + TN-R 160 of L cells by TN-R 160, TN-R 180 100 and TN-C. (A) L cell adhesion to + TN-R 180 substrate mixtures of TN-R or TN- C with FN, LN or 80 + TN-C (coll). Substrate mixtures were coated onto plastic and L cell 60 adhesion and spreading were examined after 2 hours in vitro. (a) FN/TN-R 160; (b) FN/TN-R 180; (c) FN/TN-C; (d) FN/BSA; 40 (e) LN/TN-R 160; (f) LN/TN-R 180; (g) LN/TN-C; (h) LN/BSA; (i) coll/TN-R 160; (j) coll/TN-R 180; (k) coll/TN-C; (l) coll/BSA. Bar, 50 µm. (B) Quantitative

ADHERENT CELLS ( % ) CELLS ( ADHERENT 20 data on the adhesion of L cells to the different substrate mixtures shown in (A). The number of cells adherent on 0 the substrate mixtures with BSA was set as 100%. Mean FN LN coll. I values of three independent experiments ± s.d. are shown.

(Schwann cells and DRG neurons) using the same cell- neuron attachment was reduced only on mixed FN substrates substrate adhesion assay as described above. FN-mediated (a-c), but not on mixed LN (e-h) or type I collagen substrates adhesion of all cell types was strongly reduced (more than (i-l). In a substrate mixture with FN, TN-R and TN-C almost 50%) by the presence of substrate-bound TN-R 160, TN-R completely inhibited neurite outgrowth, but had no effect on 180 or TN-C, while the molecules did not interfere with the the process formation mediated by LN or type I collagen. LN-mediated adhesion of these cells (not shown). The same Thus, TN-R and TN-C display inhibitory substrate proper- was true for the adhesion of DRG neurons to mixed substrates ties towards FN-mediated adhesion independently of the cell containing TN-R or TN-C (Fig. 2). As for the cell lines, type capable of binding and spreading on FN. Tenascin-R and tenascin-C block fibronectin-mediated adhesion 2327

Fig. 2. Effect of TN-R and TN-C on the adhesion and neurite outgrowth from DRG neurons on mixed substrates. Substrate mixtures of TN-R or TN-C with FN, LN or type I collagen (coll) were prepared as described in the legend to Fig. 1, and cell adhesion and neurite outgrowth were examined after 20 hours in vitro. Bar, 100 µm.

Binding of TN-R and TN-C to FN and FN-derived Further, we asked whether TN-R and TN-C interact with fragments correlates with inhibition of adhesion domains on the FN molecule involved in the interaction of FN The results described above suggest that in mixed substrates with its cellular receptors. Cells bind to FN mainly by two TN-R and TN-C inhibit FN-mediated cell adhesion by an inter- different mechanisms in which interactions of cell surface action with and/or steric blocking of the cell binding domain(s) integrins with the RGDS sequence (RGD-dependent) or of gly- on the FN molecule. We therefore examined the interaction of cosaminoglycans (membrane-bound proteoglycans) with the TN-R and TN-C with FN and other ECM molecules by a solid heparin binding sites (RGD-independent) on the FN molecule phase binding assay (Fig. 3). TN-C, TN-R 160 and TN-R 180 are involved (for review see Reichardt and Tomaselli, 1991). bound to FN, but not to LN or BSA. TN-R 160, TN-R 180 and Therefore, we examined the binding of TN-R and TN-C to FN- TN-C also bound to type I collagen, but not to type IV derived proteolytic fragments: (1) the 33/66 kDa heparin collagen, which is in agreement with previous data (Faissner binding fragment, which contains two heparin binding sites et al., 1990; Probstmeier et al., 1990a). (FN-C/H-I and FN-C/H-I; McCarthy et al., 1988, 1990), and

2.5 FN

LN 2 coll. I

1.5 coll. IV BSA Fig. 3. Interaction of TN-R and TN-C with 1 ECM molecules examined by solid-phase binding assay. Substrate-coated FN, LN, type I collagen, type IV collagen and BSA were incubated with TN-R 160/180 and TN-C for 1 hour at 37¡C. Maximal 0.5 absorbance at 405 nm for the binding of each component to FN is presented as 1.0. Values represent mean values of five independent experiments with four different preparations of RELATIVE ABSORBANCE (405 nm) (405 ABSORBANCE RELATIVE 0 TN-R 160 and TN-R 180, and three different preparations of TN-R 160 TN-R 180 TN-C TN-C. Standard deviations were less than 1%. 2328 P. Pesheva and others

2 FN

33/66 kD fr. 1.5 75 kD fr.

45 kD fr. 1 BSA

0.5 Fig. 4. Binding of TN-R 160, TN-R 180 and TN-C to FN- derived fragments (fr.) as determined by solid-phase binding assay. The absorbance at 405 nm for the binding to FN is presented as 1.0. Values represent mean values of three RELATIVE ABSORBANCE (405 nm) (405 ABSORBANCE RELATIVE 0 independent experiments. Standard deviations were less than TN-R 160 TN-R 180 TN-C 1%. kD (kilodaltons). connecting segment CS1, which is involved in the interaction the cell binding sites and/or a conformational change in the FN with RGD-independent integrins (α4β1) (Drake et al., 1992, molecule. However, our experiments do not exclude the pos- and references therein); (2) the 75 kDa fragment containing the sibility that TN-R and TN-C may interfere with the activity of RGDS cell binding sequence (Hayashi and Yamada, 1983; the β1 integrin FN receptor by either binding to it or to a Humphries et al., 1988); and (3) the 45 kDa collagen binding cellular receptor that in turn regulates the activity of this fragment (Pierschbacher et al., 1981). The two TN-R isoforms integrin. For hexabrachion, integrin (αβ1 and αvβ3) and non- and TN-C bound to both the 33/66 and the 75 kDa fragments integrin (heparan sulfate proteoglycans) receptors have already of FN, with greater binding to the 33/66 kDa fragment been postulated (Aukhil et al., 1993; Bourdon and Ruoslahti, compared to the intact FN molecule (Fig. 4). No significant 1989; Mendler et al., 1991; Prieto et al., 1992; for review, see binding was observed to the 45 kDa fragment when compared Erickson, 1993). To distinguish between these two modes of with BSA. inhibitory action, we performed experiments that address this To investigate whether the binding of TN-R or TN-C to FN issue. occurred at the sites responsible for binding of heparin or First, if the interference with FN-mediated adhesion by RGD-dependent integrins, binding of TN-R and TN-C to FN substrate-bound TN-R and TN-C is due to steric inhibition of and its fragments was assayed in the presence of competing the substrate, then less cells should initially bind to mixed concentrations (10 µg/ml to 1 mg/ml) of the peptides FN-C/H- FN/TN-R and FN/TN-C substrates. By eliminating intracellu- I, FN-C/H-I, CS1 and RGD-containing peptide. None of these lar response mechanisms, one should be able to prevent a peptides interfered with the interaction between TN-R or TN- possible repulsion of cells and thus register only the initial C and FN (not shown). recognition at the cell surface. Therefore, we studied the Binding of TN-R and TN-C to the FN fragments further cor- binding of L cells that had been gently treated with trypsin related with reduced adhesion and spreading of L cells on these (allowing the preservation of an intact cell surface membrane) fragments when offered as a mixed substrate with TN-R 160, and subsequently with glutaraldehyde in a way that preserves TN-R 180 or TN-C (Fig. 5A). L cells adhere to the 33/66 kDa the activity of the integrin receptors to different ECM FN fragment by a FN-C/H-I- and FN-C/H-I-dependent molecules (Fig. 6A). Glutaraldehyde-treated L (GA-L) cells mechanism, possibly involving cell surface heparan sulfate bound to mixed FN/BSA, FN/TN-R 180 and FN/TN-C sub- proteoglycans (Fig. 5B). Adhesion to this fragment was not strates equally well, while the number of cells binding to affected by RGD and CS1 synthetic peptides. Adhesion to the FN/TN-R 160 was strongly reduced by approximately 60%. 75 kDa fragment was almost abolished by the RGD-contain- Binding to LN, by contrast, was not affected by either TN-R ing peptide, but not by the other peptides, a process that form or TN-C. GA-L cells did not bind to TN-R 160, but bound involves an integrin interaction with the cell binding site of FN better to TN-R 180 and TN-C than to FN alone or to the mixed (Fig. 5B). FN/BSA substrate (Fig. 6A). These observations suggest that the three molecules block GA-L cells attached to FN via a β1 integrin, since in the cell attachment by a steric hindrance and/or conformational presence of the RGD-containing peptide or polyclonal anti- change in two major cell binding sites on the FN molecule, the bodies to the α5β1 integrin (anti-FNR) cell binding was heparin- and RGD-dependent integrin binding sites, rather than strongly inhibited (Fig. 6B). These antibodies did not interfere by a direct interaction with these sites. with L cell attachment to LN, TN-R 180 and TN-C, nor did the presence of the RGD-containing peptide reduce cell Two modes of inhibition of FN-mediated adhesion binding to these substrates (Fig. 6B). Similar results were by TN-R and TN-C obtained by using live cells for the adhesion assays, suggest- The results described above suggest that TN-R and TN-C ing that the GA-L cells expressed intact cell surface receptors inhibit FN-mediated adhesion of cells by steric hindrance of for FN (not shown). Tenascin-R and tenascin-C block fibronectin-mediated adhesion 2329 A 120 + BSA 100 + TN-R 160

+ TN-R 180 80 + TN-C

60

40

ADHERENT CELLS ( % ) ADHERENT 20

0 33/66 kD fr. 75 kD fr. Fig. 5. Effect of TN-R and TN-C on the adhesion of L cells to heparin- (33/66 kD) and cell-binding (75 kD) B fragments of FN. (A) L cell adhesion to substrate mixtures of 33/66 and 120 75 kDa FN fragments with BSA, + OVA TN-R 160, TN-R 180 and TN-C. The number of adherent cells 100 + RGD examined 2 hours after plating for + FN-C/H-I the substrate mixtures with BSA 80 (maximal adhesion) was set as 100%. Mean values of three Fig. 5A. - Pesheva et al. + FN-C/H-II independent experiments ± s.d. are 60 + CS 1 shown. (B) Mechanism of L cell adhesion to 33/66 and 75 kDa FN fragments. L cells were incubated on 40 33/66 and 75 kDa substrates in the presence of either ovalbumin (+ OVA) or the peptides RGD, FN-C/H-I, FN-C/H-I and CS1 µ ADHERENT CELLS ( % ) CELLS ( ADHERENT 20 conjugated to ovalbumin (all at 50 g/ml). The number of adherent cells was examined after 30 minutes of culture. The number of cells adhering to the respective substrate in 0 the presence of OVA was set as 100%. Mean values ± s.d. 33/66 kD fr. 75 kD fr. of two experiments are shown.

These results suggest that TN-R 160 interferes with FN- inhibits FN-mediated adhesion by a steric hindrance of and/or mediated adhesion by a steric hindrance of and/or a confor- a conformational change in the cell binding sites of FN when mational change in the cell binding sites of FN. For TN-R 180 the two molecules are present as mixed substrates. TN-R 180 and TN-C, an interaction with a cellular receptor different from and TN-C bring about their inhibitory effect by an interaction β1 integrins may influence cell adhesion to the FN molecule. with putative cellular receptor(s) different from the FNR (β1 To investigate this possibility, we examined L cell adhesion to integrin) but probably associated with it. overlayed substrates produced by coating FN or LN on top of TN-R 160, TN-R 180 and TN-C substrates (Fig. 7). This coating procedure permits FN to be present as a co-substrate DISCUSSION with TN-R or TN-C, but reduces the possibility of interaction between the molecules. After a 2 hour incubation, L cells Our present study demonstrates that TN-R 160 and TN-R 180 adhered and spread on TN-R 160 + FN as well as on BSA + exhibit inhibitory effects on FN-mediated adhesion. We have FN substrates, i.e. the inhibitory effect of TN-R 160 was com- shown that: (1) the two TN-R isoforms, TN-R 160 and TN-R pletely abolished. By contrast, overlayed substrates containing 180, presented as substrates in a mixture with the adhesive TN-R 180 and TN-C were still inhibitory. As already observed ECM glycoproteins FN, LN and collagen type I selectively for the mixed substrates, TN-R and TN-C did not interfere with inhibit attachment of different cell types and neurite outgrowth L cell adhesion to LN-overlayed substrates (Fig. 7). on FN; (2) the inhibition of cell attachment and spreading by From the data obtained, it becomes clear that TN-R 160 TN-R 160 results from its interaction with FN, leading to a 2330 P. Pesheva and others A 160 BSA 140 TN-R 160 120 TN-R 180 100 TN-C

80

60

40 ADHERENT CELLS ( % ) CELLS ( ADHERENT 20

0 FN LN - Fig. 6. Glutaraldehyde-treated L cells interact with TN-R 180 and TN-C by a RGD- and β1 integrin-independent B mechanism. (A) Binding of GA-L cells to substrate mixtures of FN or LN with 140 BSA, TN-R 160, TN-R 180 and TN-C - additive and to BSA, TN-R 160, TN-R 180 and 120 TN-C alone (-). Cell binding to + RGD FN/BSA and LN/BSA, respectively, + anti-FNR after 1 hour of incubation at 37¡C was 100 set as 100% and all other binding Fig. 6A. - Pesheva et al. events were related to these values. 80 Cell binding to BSA, TN-R and TN-C alone was related to the values obtained for FN/BSA. Mean values ± s.d. of three 60 independent experiments are shown. (B) Binding of GA-L cells to FN, LN, TN-R 180 and TN-C in the absence (− 40 additive) and presence of the RGD-containing peptide (+ RGD) or polyclonal FNR (α5β1) antibodies (+ anti-FNR).

ADHERENT CELLS ( % ) CELLS ( ADHERENT µ α β 20 Cells were incubated with GRGDS (10 g/ml) or anti- 5 1 antibodies (0.5 mg/ml) for 30 minutes at room temperature prior to plating. Binding of cells after 1 hour of incubation at 0 37¡C in the absence of additives was set as 100%. Mean FN LN TN-R 180 TN-C values ± s.d. of two independent experiments are shown. steric masking of and/or a conformational change in the cell terminal sequences involved in the formation of disulfide binding sites on FN; (3) TN-R 180, in addition, can interact bonds, might contribute to the different structural organization with putative cell surface receptor(s) in a RGD- and FNR(β1 and functional properties in relation to FN-mediated adhesion. integrin)-independent way, which may regulate the activity of The finding that the two TN-R isoforms and TN-C selec- the FNR. We have compared these effects of TN-R with those tively inhibit FN-dependent adhesion of different cells to elicited by TN-C from early postnatal mouse brain and mixed substrates and can interact with at least two sites on the confirmed the functional relatedness between the two ECM FN molecule suggested to us that the anti-adhesive effect glycoproteins, whereby TN-R 180 is more related to TN-C resulted either from a substrate-mediated inhibition of cell than TN-R 160. attachment through an interaction with FN or from interfering TN-R 160 does not display the same functional properties with the activity of one or several FN receptors, or both. To towards FN-mediated cell adhesion as TN-R 180, although the examine the first possibility, we made use of the knowledge of two isoforms are structurally and functionally very similar the functional domains of FN. As a multifunctional protein, FN (Pesheva et al., 1989, 1991, 1993; Fuss et al., 1993). The actual has been found to contain an increasing number of interactive structural difference between the two TN-R components is not sites for cells, ECM components or assembly of fibrillar matrix known. The existence of one known alternatively spliced FN (Schwarzbauer, 1991). For binding studies, we used synthetic type III-like domain (Fuss et al., 1993) and the appearance of peptides derived from FN and proteolytic fragments of 45, 75 TN-R 160 in electron micrographs as a di- but never as a tri- and 33/66 kDa (from N to C terminus) containing the collagen- brachion (Pesheva et al., 1989), suggesting a lack of N- binding site (Pierschbacher et al., 1981), the cell-binding site Tenascin-R and tenascin-C block fibronectin-mediated adhesion 2331

140 BSA

120 TN-R 160

100 TN-R 180 TN-C 80

60

40 Fig. 7. L cell adhesion to FN- and LN-overlayed substrates. FN and LN

ADHERENT CELLS ( % ) CELLS ( ADHERENT were coated onto culture dishes precoated with BSA, TN-R 160, TN-R 180 20 or TN-C. L cells were plated onto these overlayed substrates and cell adhesion was examined after 2 hours of incubation at 37¡C. The number of 0 cells adherent to BSA + FN and BSA + LN, respectively, was set as 100%. FN LN Mean values ± s.d. of two independent experiments are shown. for integrins (Humphries et al., 1988) and binding sites for the repellent effect of TN-R 160, TN-R 180 (Pesheva et al., heparin, proteoglycans and RGD-independent integrins (Drake 1993) and probably the 190 kDa TN-C isoform (Zisch et al., et al., 1992), respectively. The two TN-R isoforms and TN-C 1992), is not expressed by fibroblasts, Schwann cells and DRG bound to FN and to the 75 and 33/66 kDa fragments, but not neurons derived from early postnatal mice (Gennarini et al., to the 45 kDa fragment. Interestingly, this binding profile cor- 1989). It is thus very likely that the selective inhibition of FN- related with inhibition of cell adhesion to FN, when the dependent adhesion is due to an interference with the activity fragments were offered as mixed substrates with TN-R or TN- of the β1 integrin, since the cell types used in our study all C. This substrate-induced inhibition could be interpreted in two adhere to FN in an RGD-dependent manner (shown only for L ways: as a specific interaction with the cell-binding sites of FN cells, Fig. 6B). that antagonizes binding of the cellular receptors; or, as TN-R Our results raise several important questions as to the role and TN-C binding to parts of the FN molecule different from of TN-R, TN-C or other, unknown, related molecules in the the cell-binding sites, leading to a steric block and/or confor- modulation of such processes as cell migration and differen- mational change affecting the correct exposure of the cell- tiation, tumor growth and invasive behavior of tumor cells. The binding sites of FN. It appears that TN-R and TN-C do not act possible involvement of TN-C in pattern formation during as antagonists of the cell-binding sites of FN, since binding of embryogenesis and early postnatal development, as well as in the two molecules to the FN fragments could not be inhibited regeneration and oncogenesis, has been stated in a number of by RGD or other synthetic peptides (FN-C/H-I, FN-C/H-II and studies (Chiquet, 1989; Erickson and Bourdon, 1989; Chiquet- CS1) representing the sequences within the 33/66 kDa Ehrismann, 1990, 1991). In the PNS, TN-C expression is fragment that are active in cell binding (see also Chiquet- upregulated during regeneration (Martini et al., 1990), a Ehrismann et al., 1988). Rather, TN-R 160, when binding to process that correlates with increased expression of FN and the fragments containing distinct cell-binding sites stericly FNR (Lefcort et al., 1992), thus providing the molecular basis blocks and/or conformationally changes these sites on the FN for the control of axonal regrowth and Schwann cell migration molecule. In contrast to TN-R 160, TN-R 180 and TN-C and differentiation in the sciatic nerve. TN-C can potentiate induce cell repulsion by binding to an RGD-independent neurite outgrowth from both PNS and CNS neurons (Wehrle cellular receptor and thereby modifying the RGD-dependent β1 and Chiquet, 1990; Lochter et al., 1991) and appears to be integrin-mediated attachment to FN. Thus, TN-R 180 and TN- involved in the process of epithelial cell shedding in the small C influence cellular behavior by an active cellular response intestine by inducing cell detachment from the basal lamina specifically to the FN substrate. It is likely, therefore, that the (Probstmeier et al., 1990b). At the neuromuscular junction, the cell surface receptors for TN-R 180 and TN-C are coupled to highly localized appearance of TN-C at the denervated subsy- the β1 integrin receptor for FN by signal transduction, thereby naptic site after peripheral nerve lesion may be responsible for modifying it by ‘inside-out’ mechanisms (Ginsberg et al., attracting regrowing axons to this site and, at the same time, 1992). We have recently shown that disialogangliosides most prevent axonal growth beyond this site by the combined likely represent these TN-R 180 and TN-C receptors (Pesheva influence of TN-C and FN on neurite outgrowth (Sanes et al., and Schachner, unpublished data). 1986; Gatchalian et al., 1989). It is also conceivable that the The repellent effect of TN-R and TN-C on FN-dependent fibroblasts around the denervated neuromuscular junction that adhesion is different from that on CNS neurons. TN-C and both actively secrete TN-C may use their FN substrate-modifying TN-R 160 and TN-R 180 inhibit neuronal cell adhesion and capacity to support proliferation of perisynaptic fibroblasts neurite outgrowth independently of the adhesive substrate (Gatchalian et al., 1989). In the CNS, the coexpression of TN- present (Pesheva et al., 1989, 1991, 1993; Faissner and Kruse, R and TN-C with FN has not been systematically studied 1990). Moreover, the F3/11 cell adhesion molecule (Brüm- during ontogenesis. At developmental stages, when FN mendorf et al., 1989; Gennarini et al., 1989), which mediates expression is restricted to the basal lamina, a concerted action

2332 P. Pesheva and others of TN-R and TN-C with FN may be possible. Such coexpres- Brown, P. J. and Juliano, R. L. (1986). Expression and function of a putative sion is found for TN-C in the early postnatal rodent cerebel- cell surface receptor for fibronectin in hamster and human cell lines. J. Cell lum, when granule cell precursor cells adjoining the basal Biol. 103, 1595-1603. Brümmendorf, T., Wolff, J. M., Frank, R. and Rathjen, F. R. (1989). Neural lamina in the external granular layer may allow detachment of cell recognition molecule F11: homology with fibronectin type III and these actively proliferating cells from the basal lamina and imuunoglobulin type C domains. Neuron 2, 1351-1361. thereby favour their migration along the radial glial processes Chiquet, M. (1989). Tenascin/71/cytotactin: The potential function of (Bartsch et al., 1992; see also Hausmann and Sievers, 1985). hexabrachion proteins in neural development. Dev. Neurosci. 11, 266-275. TN-R 180, the first TN-R component expressed during devel- Chiquet-Ehrismann, R., Kalla, P., Pearson, C. A., Beck, K. and Chiquet, M. (1988). Tenascin interferes with fibronectin action. Cell 53, 383-390. opment (Pesheva et al., 1989; Bartsch et al., 1993), is Chiquet-Ehrismann, R. (1990). What distinguishes tenascin from detectable in the ventricular zone of neonatal mouse brain (P. fibronectin? FASEB J. 4, 2598-2604. Pesheva, unpublished observations) and its co-expression with Chiquet-Ehrismann, R. (1991). Anti-adhesive molecules of the extracellular FN and FNR (Stallcup et al., 1989) may fulfil the same function matrix. Curr. Opin. Cell Biol. 3, 800-804. Douillard, J. Y. and Hoffman, T. (1983). Enzyme-linked immunosorbent as already proposed for TN-C. assay for screening monoclonal antibody production using enzyme-labeled The finding of our study that, in relation to FN-mediated cell second antibody. Meth. Enzymol. 92, 168-174. adhesion, TN-R is functionally closely related to TN-C raises Drake, S. L., Klein, D. J., Mickelson, D. J., Oegema, T. R., Furcht, L. T. and the possibility of a functional substitution or redundancy. This McCarthy, J. B. (1992). Cell surface phosphatidylinositol-anchored question is even more pertinent in the light of recently heparan sulfate proteoglycan initiates mouse melanoma cell adhesion to a fibronectin-derived, heparin-binding synthetic peptide. J. Cell Biol. 117, published work on the lack of abnormal development in TN- 1331-1341. C-deficient mouse mutants (Saga et al., 1992). TN-R, or Erickson, H. P. and Bourdon, M. A. (1989). Tenascin: an extracellular matrix unknown molecules structurally or functionally related to TN- protein prominent in specialized embryonic tissues and tumors. Annu. Rev. C, could thus be instrumental in molecular and functional sub- Cell Biol. 5, 71-92. Erickson, H. P. (1993). Tenascin-C, tenascin-R and tenascin-X: a family of stitution. Identification of the cellular receptors mediating the talented proteins in search of functions. Curr. Opin. Cell Biol. 5, 869-876. repellent action of TN-R 180 and TN-C, and the mechanism Fahrig, T., Landa, C., Pesheva, P., Kühn, K. and Schachner, M. (1987). underlying a particular cellular response, are topics for further Characterization of binding properties of the myelin-associated glycoprotein investigation. In fact, we recently showed that the interference to extracellular matrix constituents. EMBO J. 6, 2875-2883. of the two molecules with FN-mediated adhesion involves a Faissner, A. and Kruse, J. (1990). J1/Tenascin is a repulsive substratum for central nervous system neurons. Neuron 5, 627-637. disialoganglioside-dependent mechanism, as the interaction of Faissner, A., Kruse, J., Kühn, K. and Schachner, M. (1990). Binding of the both TN-R 180 and TN-C with cell surface disialogangliosides J1 adhesion molecules to extracellular matrix constituents. J. Neurochem. 54, is coupled to the inhibition of RGD-dependent cell attachment 1004-1015. and neurite outgrowth on FN substrates (Pesheva and Fuss, B., Wintergerst, E. S., Bartsch, U. and Schachner, M. (1993). Schachner, unpublished data). Molecular characterization and in situ messenger RNA localization of the neural recognition molecule J1-160/180 - a modular structure similar to tenascin. J. Cell Biol. 120, 1237-1249. The authors are grateful to Iris Bahnmüller for expert technical Gatchalian, C. L., Schachner, M. and Sanes, J. R. (1989). Fibroblasts that help, R. L. Juliano for the kind gift of FNR (α5β1) antibodies, Lloyd proliferate near denervated synaptic sites in skeletal muscle synthesize the Vaughan for critically reading the manuscript, Hermann and Lilly adhesive molecules N-CAM, J1, fibronectin, and a heparan sulfate Schilling Stiftung for a fellowship (P.P.), and Kommission zur proteoglycan. J. Cell Biol. 108, 1873-1890. Förderung der wissenschaftlichen Forschung for support. This work Gennarini, G., Cibelli, G., Rougon, G., Mattei, M.-G. and Goridis, C. was also supported by grants CA29995 and CA21463 from the (1989). The mouse neuronal cell surface glycoprotein F3: a National Institutes of Health (to L.T.F.). phosphatidylinositol-anchored member of the immunoglobulin superfamily related to chicken contactin. J. Cell Biol. 109, 775-788. Ginsberg, M. H., Du, X. and Plow, E. F. (1992). Inside-out integrin signalling. Curr. Opin. Cell Biol. 4, 766-771. REFERENCES Hausmann, B. and Sievers, J. (1985). Cerebellar external granule cells are attached to the basal lamina from the onset of migration up to the end of their Adams, J. C. and Watt, F. M. (1993). Regulation of development and proliferative activity. J. Comp. Neurol. 241, 50-62. differentiation by the extracellular matrix. Development 117, 1183- Hayashi, M. and Yamada, K. M. (1983). Domain structure of the carboxyl- 1198. terminal half of human plasma fibronectin. J. Biol. Chem. 258, 3332-3340. Aukhil, I., Joshi, P., Yan, Y. and Erickson, H. P. (1993). Cell- and heparin- Humphries, M. J., Komoriya, A., Akiyama, S. K., Olden, K. and Yamada, binding domains of the hexabrachion arm identified by tenascin expression K. M. (1987). Identification of two distinct regions of the type III connecting proteins. J. Biol. Chem. 268, 2542-2553. segment of human plasma fibronectin that promote cell type-specific Bartsch, S., Bartsch, U., Dorries, U., Faissner, A., Weller, A., Ekblom, P. adhesion. J. Biol. Chem. 262, 6886-6892. and Schachner, M. (1992). Expression of tenascin in the developing and Humphries, M. J., Akiyama, S. K., Komoriya, A., Olden, K. and Yamada, adult cerebellar cortex. J. Neurosci. 12, 736-749. K. M. (1988). Neurite extension of chicken peripheral nervous system Bartsch, U., Pesheva, P., Raff, M. and Schachner, M. (1993). Expression of neurons on fibronectin: relative importance of specific adhesion sites in the janusin (J1-160/180) in the retina and optic nerve of the developing and adult central cell-binding domain and the alternatively spliced type III connecting mouse. 9, 57-69. segment. J. Cell Biol. 106, 1289-1297. Bolton, A. E. and Hunter, W. M. (1973). The labelling of proteins to high Husmann, K., Faissner, A. and Schachner, M. (1992). Tenascin promotes specific radioactivity by conjugation to a 125I-containing acylating agent. cerebellar granule cell migration and neurite outgrowth by different domains Biochem. J. 133, 529-536. in the fibronectin type-III repeats. J. Cell Biol. 116, 1475-1486. Bourdon, M. A. and Ruoslahti, E. (1989). Tenascin mediates cell attachment Hynes, R. O. and Lander, A. D. (1992). Contact and adhesive specificities in through an RGD-dependent receptor. J. Cell Biol. 108, 1149-1155. the associations, migrations, and targeting of cells and axons. Cell 68, 303- Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of 322. microgram quantities of protein using the principle of protein-dye binding. Lefcort, F., Venstrom, K., McDonald, J. A. and Reichardt, L. F. (1992). Anal. Biochem. 72, 248-254. Regulation of expression of fibronectin and its receptor, α5β1, during Bristow, J., Kian Tee, M., Gitelman, S. E., Mellon, S. H. and Miller, W. L. development and regeneration of peripheral nerve. Development 116, 767- (1993). Tenascin-X: a novel extracellular matrix protein encoded by the 782. human XB gene overlapping P450c21B. J. Cell Biol. 122, 265-278. Lightner, V. A. and Erickson, H. P. (1990). Binding of hexabrachion Tenascin-R and tenascin-C block fibronectin-mediated adhesion 2333

(tenascin) to the extracellular matrix and substratum and its effect on cell Prieto, A. L., Jones, F. S., Cunningham, B. A., Crossin, K. L. and Edelman, adhesion. J. Cell Sci. 95, 263-277. G. M. (1990). Localization during development of alternatively-spliced Lochter, A., Vaughan, L., Kaplony, A., Prochiantz, A., Schachner, M. and forms of cytotactin by in situ hybridization. J. Cell Biol. 111, 685-698. Faissner, A. (1991). J1/tenascin in substratum-bound and soluble form Prieto, A. L., Andersson-Fisone, C. and Crossin, K. L. (1992). displays contrary effects on neurite outgrowth. J. Cell Biol. 113, 1159-1171. Characterization of multiple adhesive and counteradhesive domains in the Martini, R., Schachner, M. and Faissner, A. (1990). Enhanced expression of extracellular matrix protein cytotactin. J. Cell Biol. 119, 663-678. the extracellular matrix molecule J1/tenascin in the regenerating adult mouse Probstmeier, R., Kühn, K. and Schachner, M. (1990a). Fibroblast-derived J1 sciatic nerve. J. Neurocytol. 19, 601-616. adhesion glycoproteins show binding properties to extracellular matrix McCarthy, J. B., Hagen, S. T. and Furcht, L. T. (1986). Human fibronectin constituents different from those of central nervous system origin. J. contains distinct adhesion and motility promoting domains for metastatic Neurochem. 54, 1016-1019. melanoma cells. J. Cell Biol. 102, 179-188. Probstmeier, R., Martini, R. and Schachner, M. (1990b). Expression of J1 McCarthy, J. B., Chelberg, M. K., Mickelson, D. J. and Furcht, L. T. tenascin in the crypt villus unit of adult mouse small intestine - implications (1988). Localization and chemical synthesis of fibronectin peptides with for its role in epithelial cell shedding. Development 109, 313-321. melanoma adhesion and heparin binding activities. Biochemistry 27, 1380- Rathjen, F. G. and Schachner, M. (1984). Immunocytological and 1388. biochemical characterization of a new neuronal cell surface component (L1 McCarthy, J. B., Skubitz, A. P. N., Zhao, Q., Yi, X.-y., Mickelson, D. J., antigen) which is involved in cell adhesion. EMBO J. 3, 1-10. Klein, D. J. and Furcht, L. T. (1990). RGD-independent cell adhesion to the Rathjen, F. G., Wolff, J. M. and Chiquet-Ehrismann, R. (1991). Restrictin: carboxy-terminal heparin-binding fragment of fibronectin involves heparin- an extracellular matrix glycoprotein involved in cell attachment co-purifies dependent and -independent activities. J. Cell Biol. 110, 777-787. with the cell recognition molecule F11. Development 113, 151-164. Mendler, M., Priest, A. and Bourdon, M. A. (1991). Tenascin is recognized Reichardt, L. F. and Tomaselli, K. J. (1991). Extracellular matrix molecules by two different integrin receptors, αβ1-group integrin (αβ1) and the and their receptors: functions in neural development. Annu. Rev. Neurosci. receptor (αvβ3). J. Cell Biol. 115, 137a-795. 14, 531-570. Nörenberg, U., Wille, H., Wolff, J. M., Frank, R. and Rathjen, F. G. (1992). Saga, Y., Yagi, T., Ikawa, Y., Sakakura, T. and Aizawa, S. (1992). Mice The chicken neural extracellular matrix molecule restrictin: similarity with develop normally without tenascin. Dev. 6, 1821-1831. EGF-, fibronectin type III-, and fibrinogen-like motifs. Neuron 8, 849-863. Sanes, J. R., Schachner, M. and Covault, J. (1986). Expression of several Oakley, B. R., Kirsch, D. and Morris, R. (1980). A simplified ultrasensitive adhesive macromolecules (N-CAM, L1, J1, NILE, uvomorulin, laminin, silver stain for detecting proteins in polyacrylamide gels. Anal. Biochem. fibronectin, and a heparan sulfate proteoglycan) in embryonic, adult, and 105, 361-363. denervated adult skeletal muscle. J. Cell Biol. 102, 420-431. Pantazis, N. J. and Jensen, R. (1988). Nerve growth factor, not laminin, is the Schneider-Schaulies, J., von Brunn, A. and Schachner, M. (1990). major neurite-promoting component in medium conditioned by mouse L929 Recombinant peripheral myelin protein Po confers both adhesion and neurite fibroblast cells. Dev. Brain Res. 40, 123-137. outgrowth-promoting properties. J. Neurosci. Res. 27, 286-297. Peden, K. W. C., Rutkowski, J. L., Gilbert, M. and Tennekoon, G. I. (1990). Schwarzbauer, J. E. (1991). Identification of the fibronectin sequences Production of Schwann cell lines using a regulated oncogene. Ann. NY Acad. required for assembly of a fibrillar matrix. J. Cell Biol. 113, 1463-1473. Sci. 605, 286-293. Seilheimer, B. and Schachner, M. (1987). Regulation of neural cell adhesion Pesheva, P., Juliano, R. L. and Schachner, M. (1988). Expression and molecule expression on cultured mouse Schwann cells by nerve growth localization of the fibronectin receptor in the mouse nervous system. J. factor. EMBO J. 6, 1611-1616. Neurosci. Res. 20, 420-430. Stallcup, W. B., Pytela, R. and Ruoslahti, E. (1989). A neuroectoderm- Pesheva, P., Spiess, E. and Schachner, M. (1989). J1-160 and J1-180 are associated ganglioside participates in fibronectin receptor-mediated oligodendrocyte-secreted nonpermissive substrata for cell adhesion. J. Cell adhesion of germinal cells to fibronectin. Dev. Biol. 132, 212-229. Biol. 109, 1765-1778. Wehrle, B. and Chiquet, M. (1990). Tenascin is accumulated along Pesheva, P., Probstmeier, R., Spiess, E. and Schachner, M. (1991). Divalent developing peripheral nerves and allows neurite outgrowth in vitro. cations modulate the inhibitory substratum properties of murine glia-derived Development 110, 401-415. J1-160 and J1-180 extracellular matrix glycoproteins for neuronal adhesion. Zisch, A. H., D’Alessandri, L., Ranscht, B., Falchetto, R., Winterhalter, K. Eur. J. Neurosci. 3, 356-365. H. and Vaughan, L. (1992). Neuronal cell adhesion molecule contactin/F11 Pesheva, P., Gennarini, G., Goridis, C. and Schachner, M. (1993). The binds to tenascin via its immunoglobulin-like domains. J. Cell Biol. 119, F3/11 cell adhesion molecule mediates the repulsion of neurons by the 203-213. extracellular matrix glycoprotein J1-160/180. Neuron 10, 69-82. Pierschbacher, M. D., Hayman, E. G. and Ruoslahti, E. (1981). Location of the cell attachment site in fibronectin with monoclonal antibodies and proteolytic fragments of the molecule. Cell 26, 259-267. (Received 17 January 1994 - Accepted 25 April 1994)