Contactin-1 Is a Functional Receptor for Neuroregulatory Chondroitin Sulfate-E*

Chondroitin sulfate (CS) plays critical roles in central nervous system development and regeneration, and individual modifications of CS form a “sulfation code” that regulates growth factor signaling or neuronal growth. Although we have shown that CS-E polysaccharide, but not CS-A or -C polysaccharide, has an inherent ability to promote neurite outgrowth toward primary neurons, its molecular mechanism remains elusive. Here, we show the involvement of a plasma membrane-tethered cell adhesion molecule, contactin-1 (CNTN-1), in CS-E-mediated neurite extension in a mouse neuroblastoma cell line and primary hippocampal neurons. CS-E, but not CS-A, -C, or heparan sulfate, engaged CNTN-1 with significant affinity and induced intracellular signaling downstream of CNTN-1, indicating that CS-E is a selective ligand for a potential CS receptor, CNTN-1, leading to neurite outgrowth. Our data provide the first evidence that biological functions of CS are exerted through the CS receptor-mediated signaling pathway(s).

Chondroitin sulfate (CS) plays critical roles in central nervous system development and regeneration, and individual modifications of CS form a "sulfation code" that regulates growth factor signaling or neuronal growth. Although we have shown that CS-E polysaccharide, but not CS-A or -C polysaccharide, has an inherent ability to promote neurite outgrowth toward primary neurons, its molecular mechanism remains elusive. Here, we show the involvement of a plasma membrane-tethered cell adhesion molecule, contactin-1 (CNTN-1), in CS-E-mediated neurite extension in a mouse neuroblastoma cell line and primary hippocampal neurons. CS-E, but not CS-A, -C, or heparan sulfate, engaged CNTN-1 with significant affinity and induced intracellular signaling downstream of CNTN-1, indicating that CS-E is a selective ligand for a potential CS receptor, CNTN-1, leading to neurite outgrowth. Our data provide the first evidence that biological functions of CS are exerted through the CS receptor-mediated signaling pathway(s).
Chondroitin sulfate (CS) 2 is a representative sulfated glycosaminoglycan (GAG) that is covalently attached to a panel of core proteins to form proteoglycans (CSPGs), and is ubiquitously located in extracellular matrices and on cell surfaces in various tissues. CSPGs regulate diverse physiological phenomena such as cytokinesis, morphogenesis, and infections with viruses and bacteria (1)(2)(3)(4). In particular, the pathologic functions of CS moieties of CSPGs as major axon growth-inhibitory molecules in the injured adult central nervous system have attracted widespread attention and prompted research aimed at overcoming their barrier effects on neuronal regeneration processes. Although axonal regeneration is indeed improved by the removal of CS moieties around lesion sites (5,6), CS does not always impede neurite outgrowth. For example, several CS preparations serve as stimulatory substrata for neurite outgrowth of cultured primary neurons (2,7).
The apparently contradictory actions of CS in the central nervous system are thought to be attributable to its structural diversity. CS is a linear polysaccharide that contains repeating disaccharide units consisting of glucuronic acid (GlcUA) and N-acetyl-D-galactosamine (GalNAc). The building blocks can be substituted with sulfate groups at various positions, thereby producing characteristic "sulfation codes." CS polysaccharides are divided into subclasses based on their disaccharide composition. The major CS subclasses found in mammalian tissues contain monosulfated disaccharide units, A (GlcUA-GalNAc(4-O-sulfate)) and C (GlcUA-GalNAc(6-O-sulfate)). CS polysaccharides rich in A and C units (CS-A and CS-C, respectively) are poorly permissive for neurite extension (2), probably reflecting the inhibitory nature of typical mammalian CS. In contrast, squid cartilage-derived CS-E polysaccharide possesses strong neuritogenic activity toward primary hippocampal neurons (2). CS-E is characterized by the predominant disulfated disaccharide E unit, (GlcUA-GalNAc(4,6-O-disulfate)). Recent studies on chemically synthesized CS-E tetrasaccharides also support the structural importance of E unit for neurite outgrowth-promoting activity (8,9).
The inherent potential of CS-E is of special interest for therapeutic application to central nervous system injury. CS-E binds to several humeral factors, such as midkine (MK) and brain-derived neurotrophic factor (BDNF) (10,11), and stimulates neurite outgrowth through the activation of MK and BDNF signaling inputs to primary neurons (9), suggesting the possible roles of CS-E as a coreceptor and/or reservoir for neuritogenic factors. In contrast, although it has been postulated that functional receptor molecules receiving a sulfation code of CS reside on the neuronal membrane surface (12), identification of such potential molecules remains challenging.
Focusing on neuroregulatory roles of CS, we show here the involvement of a cell adhesion molecule, contactin-1 (CNTN-1), in CS-E-mediated neuritogenesis in a neuroblastoma cell line and primary hippocampal neurons. CS-E engaged CNTN-1 and induced intracellular signaling downstream of CNTN-1, indicating that CS-E is a ligand for a potential CS receptor, CNTN-1. Our data provide the first evidence for functional expression of CS through the CS receptor-mediated signaling pathway(s). grown in Dulbecco's modified Eagle's medium (Sigma) containing 10% fetal bovine serum. The cDNA encoding mouse CNTN-1 (GenBank TM accession no. NM_007727) was subcloned into a pCMV expression vector (Stratagene). The fidelity of the plasmid construct (pCMV/CNTN-1) was confirmed by DNA sequencing. N2a cells were transfected with pCMV/ CNTN-1 using FuGENE TM 6 transfection reagent (Roche Applied Science) according to the manufacturer's instructions. Clonal cell lines (N2a/CNTN-1 cells) were selected and grown in the presence of 800 g/ml G418 (Invitrogen). To avoid any side effects of G418 on neuronal differentiation, N2a/CNTN-1 cells were passaged and maintained in G418-free medium for at least 7 days before the following experiments. Single-cell suspensions of hippocampal cells were prepared from E16 mouse brains as described previously (13). For the culture of crudely purified neurons, hippocampal cells were precultured on Petri dishes in Neurobasal TM medium (Invitrogen) containing 10% fetal bovine serum for 3 h at 37°C to remove more adherent non-neuronal cells. After 3 h, non-adherent hippocampal cells were plated on poly-L-lysine-precoated plates and maintained in Neurobasal TM medium supplemented with B27 supplement (Invitrogen) for 3 days. Total RNA was extracted from parent N2a, N2a/CNTN-1, and crudely purified hippocampal cells using an RNeasy Mini Kit (Qiagen). Amplification of target cDNAs, including CNTN-1 (for comparative expression analysis) and glyceraldehyde-3-phosphate dehydrogenase (as an internal control), was initially conducted by RT-PCR. Quantitative real-time RT-PCR was performed using a FastStart DNA Master plus SYBR Green I in a LightCycler ST300 (Roche Applied Science). Primer sequences were as follows: CNTN-1, 5Ј-AAGCCATATCCTGCTGATATT-3Ј (forward) and 5Ј-CTGACGTGCTTATCTCGG-3Ј (reverse); glyceraldehyde-3phosphate dehydrogenase, 5Ј-CATCTGAGGGCCCACTG-3Ј (forward) and 5Ј-GAGGCCATGTAGGCCATGA-3Ј (reverse). To evaluate the cell-surface distribution of CNTN-1, the cultured cells were labeled with an anti-CNTN-1 polyclonal antibody (R & D Systems, goat IgG, final concentration of 2 or 10 g/ml) followed by Alexa Fluor 488 anti-goat IgG antibody (Invitrogen, 1:1000), and/or with an anti-MAP2 antibody (Leico Technologies Inc., mouse IgG, 1:20) followed by Alexa Fluor 568 anti-mouse IgG antibody (Invitrogen, 1:1000).
CS-mediated Neuritogenesis-GAG-precoated substrata were prepared as described previously (13), with slight modifications. Briefly, 8-well chamber slides (Nunc) were precoated with poly-DL-ornithine (Sigma, 1.5 g/ml) and then coated with individual GAG preparations (Seikagaku Corp., 2 g per well), including CS-A from whale cartilage (average molecular mass of 19 kDa), CS-C from shark cartilage (average molecular mass of 43 kDa), CS-E from squid cartilage (average molecular mass of 70 kDa), and heparan sulfate (HS) from bovine kidney (average molecular mass of 11 kDa). Disaccharide composition analysis revealed that the CS-E preparation used in this study consisted primarily of the disaccharide E unit (68.4%, supplemental Table S1). Dissociated hippocampal cells were plated at a cell density of 15,000 cells/cm 2 on each well precoated with a defined substrate, and cultured for 24 h. To elicit differentiation, parent N2a and N2a/CNTN-1 cells were resuspended with a serum-de-prived medium (Neurobasal TM medium supplemented with B27 supplement) and plated at a cell density of 6,000 cells/ cm 2 on the defined CS substrata, and maintained for 48 h. To neutralize CNTN-1, an anti-CNTN-1 polyclonal antibody (goat IgG, final concentration of 5 g/ml) was added to the medium 2 h after cell seeding. The cultured cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100, and labeled with anti-MAP2 (1:250) plus anti-neurofilament (Sigma, mouse IgG, 1:250) antibodies followed by development using an M.O.M. TM Immunodetection kit (Vector) with 3,3Ј-diaminobenzidine as a chromogen. Micrographic images of immunostained cells were analyzed using morphological analysis software (Mac SCOPE, Mitani Corp). In each condition, clearly isolated cells with at least one process longer than the cell body diameter were counted as a "neurite-bearing cell." The length of the longest neurite was measured by tracing the corresponding neurite of ϳ100 neurite-bearing cells. At least three independent experiments per condition were carried out.
Interaction Analysis-The binding of various GAGs to CNTN-1 was examined using a BIAcore J system (GE Healthcare) as described previously (14), with slight modifications. Briefly, recombinant human CNTN-1 (R & D Systems) was immobilized on a CM5 sensor chip (GE Healthcare) according to the manufacturer's instructions. GAGs in a series of concentrations ranging from 10 to 100 g/ml in running buffer were applied to flow cells, and changes in resonance units were recorded. Data were analyzed using BIAevaluation 3.0 software (GE Healthcare) using a 1:1 Langmuir binding model.
In Situ Binding of CS-E to the Cell Surface of N2a/CNTN-1 Cells-CS-E was biotinylated as described previously (11). Biotinylated CS-E (15 g) was incubated with Avidin-Fluorescein (R & D, 0.1 g) for 30 min. The fluorescein-labeled CS-E (final concentration, 100 g/ml) was added to serum-free culture medium (Dulbecco's modified Eagle's medium) of parent N2a or N2a/ CNTN-1 cells and incubated for 2 h. After washing with PBS, fluorescence on the cell surface was visualized on an all-in-one type fluorescence microscope BZ-8000 (Keyence, Osaka, Japan).
Kinase Assay-N2a/CNTN-1 cells grown in 100-mm dishes were incubated in 8 ml of serum-free Dulbecco's modified Eagle's medium for 24 h and further treated with CS-A, CS-C, or CS-E (final concentration, 50 or 100 g/ml) for 2 h. The cells were harvested and lysed in 500 l of M-PER Mammalian Protein Extraction Reagent (Pierce) containing 1 mM Na 3 VO 4 and protease inhibitor mixture (Nacalai Tesque). For immunoprecipitation, the cell lysate (300 g of protein) was incubated with anti-p59 fyn (FYN3, Santa Cruz Biotechnology) polyclonal antibody for 60 min at 4°C, then protein A-Sepharose (Pierce) beads were added, and the incubation continued overnight at 4°C. An aliquot of the immunoprecipitate was assayed for kinase activity as described previously (15), with slight modifications. Briefly, the kinase assay was performed in a 20-l reaction containing 10 mM Pipes (pH 7.0), 5 mM MnCl 2 , 0.5 mM dithiothreitol, 0.25 mM Na 3 VO 4 , and 1 nmol of ATP at 37°C for 10 min. The reaction mixture was mixed with sample-loading buffer and heated at 100°C. The reaction sample and the untreated aliquot of the immunoprecipitate were analyzed by Western blotting as described previously (16). For immunode-tection of endogenously phosphorylated/autophosphorylated and total Fyn, anti-phosphotyrosine (Upstate, 4G10), and anti-p59 fyn antibodies were used, respectively.

RESULTS AND DISCUSSION
For simplicity, we attempted to uncover the mechanisms underlying CS-E-mediated neurite extension using a neuro-blastoma cell line, Neuro2a (N2a) cells, instead of primary neurons. Because they differentiate and extend neurites in response to serum deprivation, the possible contributions of neuritogenic humoral factors, being derived from non-neuronal cells and/or sera, are largely excluded from this neuronal differentiation system. Surprisingly, when N2a cells were cultured on CS-E-precoated substrate under serum-free conditions in the presence of a serum replacement, neuritogenesis was not observed, even after 48-h culture (Fig. 1a). Furthermore, neither MK nor BDNF treatment improved the neuritogenesis of N2a cells on CS-E substrate (supplemental Fig. S1). These initial observations raised the possibility that the insensitivity of N2a cells to CS-E is due to defects (or relatively low expression) of particular cell surface molecules potentially sensing CS-E. In the search for candidate molecules by comparative gene expression analysis using cDNA libraries prepared from N2a and cultured hippocampal cells, we found that contactin-1 (CNTN-1), a glycosylphosphatidylinositol-anchored cell adhesion molecule of the immunoglobulin superfamily, was not expressed in N2a cells (Fig. 1b). The widely accepted features of CNTN-1 that modulate neurite outgrowth and interact with multiple macromolecules, including a CSPG RPTP␤/phosphacan (17)(18)(19)(20)(21), prompted us to validate the function of CNTN-1 as a neuronal cell surface receptor for CS-E.
To assess the involvement of CNTN-1 in CS-E-medicated neurite outgrowth, N2a cells were stably transfected with an expression vector carrying CNTN-1 cDNA. The significant expression of CNTN-1 and its cell surface distribution in established N2a clones (N2a/CNTN-1) were confirmed by quantitative RT-PCR and immunocytochemistry, respectively ( Fig. 1c and supplemental Fig. S2). Unlike parent N2a cells, vigorous neurite sprouting was observed in N2a/CNTN-1 cells when plated on CS-E substrate, as judged by the increased fraction of neurite-bearing cells (Fig. 1, a and d). In contrast, no such appreciable promotion was seen even in N2a/CNTN-1 cells when cultured on substrate precoated with CS-A, CS-C, or

Contactin-1, a Neuritogenic Chondroitin Sulfate Receptor
HS from bovine kidney, another class of representative sulfated GAGs (supplemental Fig. S3). Notably, application of a functionally blocking anti-CNTN-1 polyclonal antibody in culture medium effectively abrogated neuritogenesis in N2a/CNTN-1 cells on CS-E substrate ( Fig. 1d and supplemental Fig. S3). Additional quantitative analysis measuring the length of the longest neurite of individual N2a/CNTN-1 cells also revealed that CS-E displayed marked neurite outgrowth-promoting effects compared with the poly-DL-ornithine control, CS-A, CS-C, and HS (Fig. 1e). These findings clearly indicate that the functional expression of CNTN-1 is required for neuritogenesis in N2a cells grown on CS-E substrate. We next investigated whether CNTN-1 is also required for CS-E-induced neurite outgrowth of primary hippocampal neurons. In support of the mRNA expression in cultured hippocampal cells (Fig. 1a), immunoreactivity for CNTN-1 was detected in the majority of MAP2-positive neurons in the culture (Fig. 2a), enabling us to test the above-mentioned functional blocking assay. As reported previously (22), elongation of prominent long neurites was frequently observed in hippocampal neurons cultured on CS-E substrate (Fig. 2b). As observed in N2a/CNTN-1 cells, the neurite outgrowth-promoting effect of CS-E was also effectively inhibited by treatment with the CNTN-1neutralizing antibody (Fig. 2, b and  c). Notably, combined treatments with the CNTN-1-neutralizing antibody and either MK-or BDNF-neutralizing antibodies (9) completely abolished the CS-E-mediated neurite outgrowth promoting effects, when the poly-DL-ornithine was deemed as a basal control substrate (Fig. 2c). These results suggested that not only humoral factors, MK and BDNF, but also CNTN-1 is important in CS-E-mediated neuritogenesis in primary hippocampal neurons and also indicated that MK and BDNF might be largely dispensable for neurite outgrowth via CNTN-1.
To further clarify the receptor function of CNTN-1 specific for CS-E, the molecular interaction of CNTN-1 with various GAG preparations was examined using a surface plasmon resonance biosensor, BIAcore. Recombinant CNTN-1 was immobilized on the carboxymethyl-dextran sensor chip surface, and aqueous solutions of GAG preparations (80 g/ml) were individually injected over the sensor surface to detect direct binding. An overlaid sensorgram disclosed the clear binding of CS-E (Fig. 3a); in contrast, CS-A, CS-C, and HS showed less binding than did CS-E. This binding preference of CNTN-1 was positively correlated with the structure-dependent neuritogenic activity of CS preparations toward both N2a/ CNTN-1 cells and primary hippocampal neurons. Evaluation of kinetic parameters revealed that CS-E interacted with CNTN-1 with significant affinity (apparent equilibrium dissociation con- Fig. S4 and Table 1), which was comparable to measurement of carbohydrate-recognizing proteins, including animal lectins and their target glycoconjugates (23,24). In contrast, all the other GAGs examined had 2 or 3 orders of magnitude lower apparent affinity for CNTN-1 compared with CS-E (Table 1). Based on the kinetic data, the binding ability of CS-E to cell-surface CNTN-1 was examined by addition of a fluorescein-labeled CS-E to the culture medium. As shown in Fig. 3b, the fluorescent signal was detected on the cell surface of N2a/CNTN-1 cells, whereas no detectable signal was observed on parent N2a cells. Thus, the direct interaction of CS-E, but not CS-A, CS-C, and HS, with cell-surface CNTN-1 seemed to be a prerequisite for subsequent actions leading to neurite outgrowth on CS substrata.
Several lines of evidence suggest that CNTN-1-mediated signaling is expressed as the activation of cytoplasmic, non-  FEBRUARY 13, 2009 • VOLUME 284 • NUMBER 7

JOURNAL OF BIOLOGICAL CHEMISTRY 4497
receptor-type tyrosine kinase Fyn, which regulates neurite outgrowth (25)(26)(27); therefore, to evaluate the functional aspects of CS-E as a signaling molecule, we examined in vitro autophosphorylation activity of Fyn immunoprecipitated from lysates of N2a/CNTN-1 cells cultured in the presence or absence of CS preparations. Treatment with CS-E (50 or 100 g/ml) brought about a significant increase of Fyn kinase activity in a concentration-dependent manner, which was 168% or 245% of the autophosphorylation activity of nontreated cells, respectively (Fig. 3, c  and d). By contrast, treatment with CS-A or CS-C did not augment the basal kinase activity observed in non-treated cells (Fig.  3, c and d). In addition, no significant increment in kinase activity occurred in parent N2a cells, even when treated with the higher concentration (100 g/ml) of CS-E (supplemental Fig. S5). These results indicate that CS-E can act as an effective ligand for neuronal cell-surface receptor CNTN-1 that stimulates the intracellular signaling pathways involved in neuritogenesis.
Taking advantage of the observation that N2a cells were less responsive to the neuritogenic CS-E substrate, we described here how a glycosylphosphatidylinositol-linked cell adhesion molecule CNTN-1 can function as a neuronal cell-surface receptor for CS-E. This is the first report showing that CS sugar chains indeed behave as extracellular signaling molecules that can induce intracellular signaling in a sulfation pattern-dependent manner. Based on these findings, in conjunction with the capability of CS-E to bind humoral factors (10,11,28), we propose bidirectional regulatory modes of action for CS-E-mediated neurite outgrowth in cultured primary neurons as follows. CS-E substrate stimulates intracellular signaling cascades leading to neurite outgrowth by interacting directly with plasma membrane-residing CNTN-1. In addition, CS-E substrate concentrates soluble neuritogenic factors, derived from cultured cells and/or medium, onto its own surface and effectively presents them to their respective receptors on the extending neurites and/or neuronal cell soma (7,9), resulting in additive or synergistic acceleration of neurite extension (see Fig. 2c).
It has been reported that neuronal CNTN-1-mediated neurite outgrowth is induced via trans-heterophilic interaction with a glial CSPG, RPTP␤/phosphacan (18 -21). Although the contribution of CS moieties of RPTP␤/phosphacan to such neuron-glia interaction remains to be investigated, functional similarities between CS-E and RPTP␤/phosphacan in neurite

Contactin-1, a Neuritogenic Chondroitin Sulfate Receptor
outgrowth suggest that neuritogenic sulfation domains composed of CS-E-like structures may be embedded in CS moieties of RPTP␤/phosphacan expressed by particular glial cells. In fact, significant proportions of CS-E-like structures are found in the mammalian brain (2). Given that the sulfation modification of CS chains is a cell/tissue-specific and core protein-dependent process, the emergence of neuritogenic CSPGs in the central nervous system may be also spatiotemporally regulated; therefore, further identification of CSPGs carrying the CS-Elike structures, but not typical CS-A-and CS-C-like structures, is essential to elucidate the functional aspect of CS as a neuritogenic molecule. In addition, future studies focusing on the sulfotransferases responsible for synthesis of the CS-E-like structures will advance our understanding of the functional importance of the sulfation code in the central nervous system development and regeneration and will aid in the development of novel, additional therapeutic approaches for central nervous system injuries and neurodegenerative diseases.