Composition of Perineuronal Net Extracellular Matrix in Rat Brain

A DIFFERENT DISACCHARIDE COMPOSITION FOR THE NET-ASSOCIATED PROTEOGLYCANS*
      We developed a method to extract differentially chondroitin sulfate proteoglycans (CSPGs) that are diffusely present in the central nervous system (CNS) matrix and CSPGs that are present in the condensed matrix of perineuronal nets (PNNs). Adult rat brain was sequentially extracted with Tris-buffered saline (TBS), TBS-containing detergent, 1 m NaCl, and 6 m urea. Extracting tissue sections with these buffers showed that the diffuse and membrane-bound CSPGs were extracted in the first three buffers, but PNN-associated CSPGs remained and were only removed by 6 m urea. Most of the CSPGs were extracted to some degree with all the buffers, with neurocan, brevican, aggrecan, and versican particularly associated with the stable urea-extractable PNNs. The CSPGs in stable complexes only extractable in urea buffer are found from postnatal day 7–14 coinciding with PNN formation. Disaccharide composition analysis indicated a different glycosaminoglycan (GAG) composition for PGs strongly associated with extracellular matrix (ECM). For CS/dermatan sulfate (DS)-GAG the content of nonsulfated, 6-O-sulfated, 2,6-O-disulfated, and 4,6-O-disulfated disaccharides were higher and for heparan sulfate (HS)-GAG, the content of 6-O-sulfated, 2-N-, 6-O-disulfated, 2-O-, 2-N-disulfated, and 2-O-, 2-N-, 6-O-trisulfated disaccharides were higher in urea extract compared with other buffer extracts. Digestions with chondroitinase ABC and hyaluronidase indicated that aggrecan, versican, neurocan, brevican, and phosphacan are retained in PNNs through binding to hyaluronan (HA). A comparison of the brain and spinal cord ECM with respect to CSPGs indicated that the PNNs in both parts of the CNS have the same composition.
      In the adult brain, ECM
      The abbreviations used are: ECM, extracellular matrix; PG, proteoglycan; CS, chondroitin sulfate; HS, heparin sulfate; DS, dermatan sulfate; HA, hyaluronan; GAG, glycosaminoglycan; GlcUA, d-glucuronic acid; GalNAc, N-acetyl-d-galactosamine; TN-R, tenascin-R; CNS, central nervous system; PNN, perineuronal net; RPTPβ, receptor-type protein-tyrosine phosphatase β; PBS, phosphate-buffered saline; 2-AB, 2-aminobenzamide; WFA, Wisteria floribunda agglutinin; HABP, hyaluronan-binding protein; HPLC, high performance liquid chromatography; PFA, paraformaldehyde; NHS, normal horse serum; TRU, turbidityreducingunit; GPI, glycosylphosphatidylinositol; CSPG, chondroitin sulfate proteoglycans.
      2The abbreviations used are: ECM, extracellular matrix; PG, proteoglycan; CS, chondroitin sulfate; HS, heparin sulfate; DS, dermatan sulfate; HA, hyaluronan; GAG, glycosaminoglycan; GlcUA, d-glucuronic acid; GalNAc, N-acetyl-d-galactosamine; TN-R, tenascin-R; CNS, central nervous system; PNN, perineuronal net; RPTPβ, receptor-type protein-tyrosine phosphatase β; PBS, phosphate-buffered saline; 2-AB, 2-aminobenzamide; WFA, Wisteria floribunda agglutinin; HABP, hyaluronan-binding protein; HPLC, high performance liquid chromatography; PFA, paraformaldehyde; NHS, normal horse serum; TRU, turbidityreducingunit; GPI, glycosylphosphatidylinositol; CSPG, chondroitin sulfate proteoglycans.
      is mainly present in the intercellular spaces between neurons and glial cells. Whereas most of this matrix is amorphous, there are specialized structures of dense organized matrix called PNNs around many neurons with holes at the sites of synaptic contacts (
      • Hockfield S.
      • McKay R.D.
      ,
      • Celio M.R.
      • Blumcke I.
      ,
      • Celio M.R.
      • Spreafico R.
      • De Biasi S.
      • Vitellaro-Zuccarello L.
      ). PNNs are composed of CSPGs versican, brevican, neurocan, cat-301 antigen (aggrecan), phosphacan (DSD-1-PG), HA, tenascin-C, tenascin-R (TN-R), and link proteins (
      • Jaworski D.M.
      • Kelly G.M.
      • Hockfield S.J.
      ,
      • Asher R.A.
      • Scheibe R.J.
      • Keiser H.D.
      • Bignami A.
      ,
      • Yamaguchi Y.
      ,
      • Oohashi T.
      • Hirakawa S.
      • Bekku Y.
      • Rauch U.
      • Zimmermann D.R.
      • Su W.D.
      • Ohtsuka A.
      • Murakami T.
      • Ninomiya Y.
      ,
      • Bekku Y.
      • Su W.D.
      • Hirakawa S.
      • Fassler R.
      • Ohtsuka A.
      • Kang J.S.
      • Sanders J.
      • Murakami T.
      • Ninomiya Y.
      • Oohashi T.
      ,
      • Carulli D.
      • Rhodes K. E. Brown D.J.
      • Bonnert T. P. Pollack S.J.
      • Oliver K. Strata P.
      • Fawcett J.W.
      ). There is evidence that these structures are involved in the regulation of neuronal plasticity (
      • Hockfield S.
      • Kalb R.G.
      • Zaremba S.
      • Fryer H.
      ,
      • Pizzorusso T.
      • Medini P.
      • Berardi N.
      • Chierzi S.
      • Fawcett J.W.
      • Maffei L.
      ), in neuroprotection (
      • Bruckner G.
      • Hausen D.
      • Hartig W.
      • Drlicek M.
      • Arendt T.
      • Brauer K.
      ,
      • Morawski M.
      • Bruckner M.K.
      • Riederer P.
      • Bruckner G.
      • Arendt T.
      ), and in support of ion homeostasis around highly active neurons (
      • Bruckner G.
      • Brauer K.
      • Hartig W.
      • Wolff J.R.
      • Rickmann M.J.
      • Derouiche A.
      • Delpech B.
      • Girard N.
      • Oertel W.H.
      • Reichenbach A.
      ,
      • Bruckner G.
      • Hartig W.
      • Kacza J.
      • Seeger J.
      • Welt K.
      • Brauer K.
      ,
      • Hartig W.
      • Derouiche A.
      • Welt K.
      • Brauer K.
      • Grosche J.
      • Mader M.
      • Reichenbach A.
      • Bruckner G.
      ).
      The CSPGs in the brain can be grouped into secreted, membrane-bound, and GPI-anchored forms. Versican, aggrecan, neurocan, and brevican constitute a family of HA binding PGs called lecticans (
      • Ruoslahti E.
      ,
      • Oohira A.
      • Matsui F.
      • Tokita Y.
      • Yamauchi S.
      • Aono S.
      ,
      • Rauch U.
      ). The lecticans share a common structure with an N-terminal hyaluronan binding domain, C-terminal lectin binding domain, and GAG-carrying middle portion of variable length (
      • Yamaguchi Y.
      ,
      • Oohira A.
      • Matsui F.
      • Tokita Y.
      • Yamauchi S.
      • Aono S.
      ,
      • Bandtlow D.R.
      • Zimmermann C.E.
      ). Except for the GPI-anchored splice variant of brevican, all lecticans are secreted. Phosphacan is a secreted CSPG, representing the entire extracellular domain of receptor-type protein-tyrosine phosphatase β (RPTPβ), which is a transmembrane PG in the brain (
      • Maurel P.
      • Rauch U.
      • Flad M.
      • Margolis R.K.
      • Margolis R.U.
      ). NG2 and neuroglycan C are two other transmembrane CSPGs expressed in the brain (
      • Levine J.M.
      • Nishiyama A.
      ,
      • Watanabe E.
      • Maeda N.
      • Matsui F.
      • Kushima Y.
      • Noda M.
      • Oohira A.
      ). HA and the extracellular matrix glycoprotein TN-R are two ligands of lecticans (
      • Bignami A.
      • Asher R.
      • Perides G.
      ,
      • Aspberg A.
      • Miura R.
      • Bourdoulous S.
      • Shimonaka M.
      • Heinegard D.
      • Schachner M.
      • Ruoslahti E.
      • Yamaguchi Y.
      ). Much of the binding and biological activity of the CSPGs depends on the CS-GAGs, composed of repeating disaccharide units of d-glucuronic acid (GlcUA) and N-acetyl-d-galactosamine (GalNAc). In mammals, the CS disaccharides can be disulfated in the positions 2 and 4 of GlcUA and GalNAc, respectively (CS-D) and in positions 4 and 6 of GalNAc (CS-E) or monosulfated in positions 4 or 6 of GalNAc (CS-A or CS-C, respectively) or nonsulfated (
      • Sugahara K.
      • Yamada S.
      ). The differential arrangement of these units results in the structural diversity of CS chains and defines the charge patterns that give the GAGs their binding properties (
      • Deepa S.S.
      • Umehara Y.
      • Higashiyama S.
      • Itoh N.
      • Sugahara K.
      ,
      • Maeda N.
      • He J.
      • Yajima Y.
      • Mikami T.
      • Sugahara K.
      • Yabe T.
      ).
      CSPGs in the CNS can interact with various growth factors and cell adhesion molecules, playing a significant role in development (
      • Oohira A.
      • Matsui F.
      • Tokita Y.
      • Yamauchi S.
      • Aono S.
      ,
      • Bandtlow D.R.
      • Zimmermann C.E.
      ). They mostly have an inhibitory effect toward neurite outgrowth and regeneration, either via their CS chains or core proteins (
      • Rhodes K.E.
      • Fawcett J.W.
      ). They are up-regulated after CNS injury (
      • Asher R.A.
      • Morgenstern D.A.
      • Fidler P.S.
      • Adcock K.H.
      • Oohira A.
      • Braistead J.E.
      • Levine J.M.
      • Margolis R.U.
      • Rogers J.H.
      • Fawcett J.W.
      ,
      • Asher R.A.
      • Morgenstern D.A.
      • Shearer M.C.
      • Adcock K.H.
      • Pesheva P.
      • Fawcett J.W.
      ,
      • Thon N.
      • Haas C.A.
      • Rauch U.
      • Merten T.
      • Fassler R.
      • Frotscher M.
      • Deller T.
      ,
      • Jones L.L.
      • Yamaguchi Y.
      • Stallcup W.B.
      • Tuszynski M.H.
      ), and enzymatic removal of GAG chains from CSPGs with chondroitinase ABC improves axon regeneration and functional recovery (
      • Moon L.D.
      • Asher R.A.
      • Rhodes K.E.
      • Fawcett J.W.
      ,
      • Bradbury E.J.
      • Moon L.D.
      • Popat R.J.
      • King V.R.
      • Bennett G.S.
      • Patel P.N.
      • Fawcett J.W.
      • McMahon S.B.
      ). Degradation of CSPGs induces sprouting of Purkinje axons in the cerebellum (
      • Corvetti L.
      • Rossi F.
      ) and promotes retinal fiber sprouting after denervation of the superior colliculus in adult rats (
      • Tropea D.
      • Caleo M.
      • Maffei L.
      ).
      PNNs form late in development, surrounding particular classes of neurons. Their time of appearance corresponds with the termination of plasticity at the end of critical periods in many parts of the CNS, and their appearance in the visual cortex can be delayed by dark rearing, which prolongs plasticity. Chondroitinase digestion of CSPGs in the PNNs in the adult visual cortex reactivates plasticity after the critical period (
      • Pizzorusso T.
      • Medini P.
      • Berardi N.
      • Chierzi S.
      • Fawcett J.W.
      • Maffei L.
      ). This and other evidence has led to the hypothesis that PNNs are involved in the control of plasticity in the CNS. The identity of the CSPGs in PNNs that are responsible for the control of plasticity has not been determined, but investigations of animals lacking brevican and neurocan have revealed abnormalities in one form of plasticity, long term potentiation (
      • Brakebusch C.
      • Seidenbecher C.I.
      • Asztely F.
      • Rauch U.
      • Matthies H.
      • Meyer H.
      • Krug M.
      • Bockers T.M.
      • Zhou X.
      • Kreutz M.R.
      • Montag D.
      • Gundelfinger E.D.
      • Fassler R.
      ).
      Because enhancement of plasticity promotes recovery after damage to the CNS, development of methods to control neural plasticity is a priority. We have therefore investigated the composition and properties of PNNs. We have developed a method to differentially extract the CSPGs that are diffusely present in CNS parenchyma and those that are present in the condensed matrix of PNNs. This enabled us to examine the composition of PNNs and changes in development; to analyze the GAG chain composition of the different extracts in the adult brain; to study how the various components are retained within the net structure and to compare the properties of the rather different PNNs from brain and spinal cord.

      EXPERIMENTAL PROCEDURES

      Materials—Chondroitinase ABC from Proteus vulgaris (EC 4.2.2.4) and CS disaccharides were purchased from Seikagaku Corp., Tokyo, Japan. Pronase was obtained from Kaken Pharmaceutical Co., Tokyo, Japan. Hyaluronidase (EC 4.2.2.1) from Streptomyces hyalurolyticus was purchased from Calbiochem, La Jolla, CA. Protease-free preparation of chondroitinase ABC from P. vulgaris (EC 4.2.2.4) and biotinylated HABP were obtained from Seikagaku, Falmouth, MA. Hiprep 16/10 DEAE-FF column, Cy3 streptavidin, and ECL™ Western blotting detection reagents were obtained from Amersham Biosciences UK Ltd., Chalfont St.Giles, UK. Heparinase I (EC 4.2.2.7) and III (EC 4.2.2.8) from Flavobacterium heparinum were obtained from IBEX Pharmaceuticals Inc., Montreal, Canada. BCA protein assay kit was from Pierce. Peroxidase-conjugated anti-mouse IgG, biotinylated horse anti-mouse IgG and Vectastain ABC elite kit were from Vector, Peterborough, UK. Biotinylated Wisteria floribunda agglutinin (WFA) and bisbenzimide fluorescent dye (Hoechst 33342) were purchased from Sigma. DPX-mounting medium was purchased from Lamb, Eastbourne, UK. Sodium pentobarbitone was from Rhone Merieux, Harlow, UK. 2-Aminobenzamide (2AB) was from Nacalai Tesque, Kyoto, Japan and amine-bound silica PA-03 column was from YMC Co., Kyoto, Japan. Complete protease inhibitor mixture tablets and pepstatin were from Roche Applied Science, Mannheim, Germany.
      Extraction of PGs from Rat Brain and Spinal Cord—Brains or spinal cords from 3-month-old Sprague-Dawley rats (Charles River, Margate, UK) were homogenized with a tight-fitting Potter glass homogenizer using 5 ml/brain or 3 ml/spinal cord extraction buffer containing 50 mm TBS, pH 7.0, 2 mm EDTA, 10 mm N-ethylmaleimide, and 2 mm phenylmethylsulfonyl fluoride as protease inhibitors (buffer 1). The homogenate was centrifuged at 15,000 rpm at 4 °C for 30 min. The supernatant was collected, and the pellet was re-extracted with buffer 1 twice, and the supernatants from the first and second extractions were pooled together (extract 1), whereas the supernatant from third extraction was discarded. The pellet obtained after centrifugation was further extracted with buffer 2 (buffer 1 containing 0.5% Triton X-100) three times to obtain extract 2, followed by extraction with buffer 3 (buffer 2 containing 1 m NaCl) and buffer 4 (buffer 2 containing 6 m urea) to obtain extracts 3 and 4, respectively. Extracts 3 and 4 were dialyzed against phosphate-buffered saline (PBS), and the protein content in all the four extracts was quantified by BCA protein assay kit.
      To investigate the release of CSPGs from PNNs, adult rat brain was sequentially extracted with buffers 1, 2, and 3. The precipitate obtained after three washes in buffer was resuspended in the digestion buffers for chondroitinase ABC (0.1 m Tris-HCl, pH 8.0 containing 0.03 m sodium acetate) and hyaluronidase (20 mm sodium acetate, pH 6.0 containing 0.15 m NaCl), separately and treated with 0.1 international unit (IU) of protease-free preparation of chondroitinase ABC at 37 °C for 3 h or 20 turbidity reducing unit (TRU)/ml of hyaluronidase containing protease inhibitors and 2 μg/ml pepstatin at 37 °C for 3 h, separately. The released CSPGs were analyzed by Western blotting as described later. Four separate extractions and analyses were performed on brain tissue, two on spinal cord.
      Partial Purification of PGs and Analysis of CS/DS and HS-GAG Disaccharide Composition—Brains from two 3-month-old rats were sequentially extracted with buffers 1, 2, 3, and 4 and after dialyzing the extracts against 50 mm Tris, pH 7.5, 2 m urea, 0.2 m NaCl (wash buffer), the samples were separately applied to a Hiprep 16/10 DEAE-FF column (1.6 × 10 cm) pre-equilibrated with wash buffer. The unbound fraction was eluted with the wash buffer whereas the bound fraction was eluted with the same buffer containing 1 m NaCl. The eluted fractions were dialyzed against PBS, and the PGs were precipitated with 95% ethanol containing 1.3% potassium acetate for 16 h at 4 °C, and the precipitate was dried.
      Each buffer extract (2–3 mg as ethanol precipitate) was treated with Pronase at 60 °C for 20 h to degrade proteins. After incubation and subsequent treatment with 5% trichloroacetic acid, the GAG-containing fractions were recovered by ethanol precipitation. An aliquot of the resultant GAG preparation derived from each buffer extract was digested exhaustively with chondroitinase ABC or a mixture of heparinase I and III. The reaction mixtures were lyophilized and derivatized with a fluorophore, 2AB. After removal of excess 2AB reagent by repeated extraction with a water-chloroform 1:1 mixture (v/v), the water phase was dried. An aliquot of each 2AB-derivative was analyzed by anion exchange HPLC on an amine-bound silica column (
      • Kinoshita A.
      • Sugahara K.
      ) to identify and quantify the resultant 2AB-labeled unsaturated CS or HS disaccharides.
      Western Blotting—The extracted proteoglycan fractions corresponding to 200 μg of protein was precipitated with 95% ethanol containing 1.3% potassium acetate for 1 h in ice, and the precipitate was recovered by centrifugation at 15,000 rpm for 10 min and digested with 10 international milliunits of a protease-free preparation of chondroitinase ABC in 0.1 m Tris-HCl, pH 8.0 containing 0.03 m sodium acetate for 3 h at 37 °C. The digest containing the core proteins were subjected to SDS-PAGE in a 5% gel under reducing conditions (except for versican, for which non-reducing condition was applied). The separated proteins were electrotransferred onto a nitrocellulose membrane for 16 h at a constant current of 175 mA at 4 °C. All the following steps were performed at room temperature. After washing with TBS-T (TBS containing 0.05% Tween-20), the membrane was incubated with anti-CSPG antibodies in TBS-T for 1 h. The antibodies used are listed in Table 1. After washing with TBS-T for 5 min five times, the blot membrane was treated with a 1:10,000 dilution of peroxidase-conjugated anti-mouse IgG for 1 h. The blots were developed using chemiluminescent substrate.
      TABLE 1Antibodies used for Western blotting and immunohistochemistry
      AntibodyAntigenHost speciesSource
      a DSHB-Developmental Studies Hybridoma Bank (University of Iowa); Chemicon (Temecula, CA); BD Biosciences (Oxford, UK).
      Working dilution
      WB
      b WB, Western blotting.
      IHC
      c IHC, immunohistochemistry.
      1F6NeurocanMouseDSHB1:1001:5
      Neurocan-N
      3F8PhosphacanMouseDSHB1:50
      RPTPβ
      2B49PhosphacanMouseDSHB1:2
      RPTPβ
      Cat-301AggrecanMouseChemicon1:10001:500
      12C5VersicanMouseDr. R. A. Asher (
      • Asher R.A.
      • Morgenstern D.A.
      • Shearer M.C.
      • Adcock K.H.
      • Pesheva P.
      • Fawcett J.W.
      )
      1:201:2
      Anti-BrevicanBrevicanMouseBD-Biosciences1:1000
      D31.10NG2MouseDr. J. Levine (
      • Stallcup W.B.
      • Beasley L.
      • Levine J.
      )
      1:20
      Anti-Tenascin-RTenascin-RMouseDr. P. Pesheva (
      • Probstmeier R.
      • Stichel C.C.
      • Muller H.W.
      • Asou H.
      • Pesheva P.
      )
      1:500
      a DSHB-Developmental Studies Hybridoma Bank (University of Iowa); Chemicon (Temecula, CA); BD Biosciences (Oxford, UK).
      b WB, Western blotting.
      c IHC, immunohistochemistry.
      Immunohistochemistry—Postnatal (P) 7, 14, 21, and adult female Sprague-Dawley rats, three at each time point, were terminally anesthetized with an intraperiteonal overdose of sodium pentobarbitone and perfused through the heart with 200 ml of PBS prewash (pH 7.4) followed by 200 ml of 4% paraformaldehyde (PFA). The brains were post-fixed overnight at 4 °C, transferred to 30% sucrose and then sectioned into 40-μm sagittal sections. Sections were quenched (10% methanol and 3% H2O2) for 5 min and blocked in 3% normal horse serum (NHS) in PBS with 0.02% Triton X-100 (TXPBS) for 1 h at room temperature. Sections were incubated overnight in anti-neurocan, anti-phosphacan, or anti-versican antibodies (see Table 1 for details) with 1% NHS in TXPBS at 4 °C, and then in biotinylated horse anti-mouse IgG (1:200) for 1 h at room temperature. Then they were incubated in ABC solution (Vectastain ABC elite kit) for 1 h at room temperature, and the staining was revealed using diaminobenzamide as a chromogen.
      For immunofluorescence on fresh frozen sections of adult cerebellum, adult rats were sacrificed by decapitation, and the cerebellum was immediately frozen on dry ice. Sagittal sections (16-μm thick) were cut on a cryostat, collected on SuperFrost® Plus slides, and air-dried at room temperature for 20–30 min. The slides were sequentially washed with buffers 1, 2, 3, and 4, each wash lasting 1 h, and then were fixed in 4% PFA for 10 min. WFA and HABP histochemistry were performed by incubating sections in biotinylated WFA (20 μg/ml) or biotinylated HABP (10 μg/ml) in TXPBS overnight, then in Cy3 streptavidin (1:500) for 1 h at room temperature. For CSPG immunostaining, sections were incubated overnight in primary antibodies (see Table 1) with 1.5% NHS in TXPBS at 4 °C, then in biotinylated horse anti-mouse IgG (1:200) in TXPBS containing 1.5% NHS for 1 h at room temperature and then in a solution made of Cy3 streptavidin and 10 μg/ml bisbenzimide fluorescent dye for 1 h at room temperature. In control experiments primary antibodies were omitted, yielding unstained sections. The slides were examined on a Leitz DMRD microscope. Digital images were produced using a Lucia imaging program (Nikon, Kingston upon Thames, UK) with a Nikon DXM 1200 digital camera. The images were imported into Adobe Photoshop (Adobe Systems Inc., San Jose, CA) in which size, contrast, and brightness were adjusted when necessary.
      For enzymatic digestions sagittal sections of adult rat cerebellum (16-μm thickness) was equilibrated either with 0.1 m Tris, pH 8.0 containing 0.03 m sodium acetate for chondroitinase or 20 mm sodium acetate, pH 6.0 containing 0.15 m NaCl for hyaluronidase digestions. A protease-free preparation of chondroitinase ABC (0.1 international units/ml) and hyaluronidase (10 TRU/ml) containing protease inhibitors and 2 μg/ml pepstatin were added to the slides separately and incubated at 37 °C for 60 min and 37 °C for 120 min, respectively. After digestion, the slides were washed with PBS and fixed in 4% PFA in PBS and the PNN components were visualized by immunofluorescence as described earlier.

      RESULTS

      To extract and analyze CSPGs loosely associated in the ECM, those associated with membranes and those bound in stable ternary complexes in PNNs, adult rat brain was sequentially extracted with four different buffers as described under “Experimental Procedures.” The CSPGs were then characterized in each fraction. Concurrently we extracted CSPGs from tissue sections using the same sequential buffers so that we could visualize directly the locations of the CSPGs extracted by each step. In the present study we use the term PNN based on the evidence from immunohistochemical staining. However there is also condensed matrix containing CSPGs around the nodes of Ranvier in a perineuronal net-like structure (
      • Oohashi T.
      • Hirakawa S.
      • Bekku Y.
      • Rauch U.
      • Zimmermann D.R.
      • Su W.D.
      • Ohtsuka A.
      • Murakami T.
      • Ninomiya Y.
      ). The components of these structures are probably extracted in the same conditions as PNNs.
      Types and Location of CSPGs Extracted by the Sequential Buffers—Extraction of adult rat brain with buffer 1, which is normal saline, released most of the neurocan, brevican, phosphacan, and part of the aggrecan, but only a small proportion of versican V2 and NG2 (Fig. 1, lane 1). The monoclonal antibody 1F6 detects two forms of neurocan: the full-length neurocan (240-kDa core protein) and the proteolytically cleaved N-terminal neurocan, neurocan-N (130-kDa core protein), and both these forms were present in the saline extract. All the full-length neurocan was extracted in saline, because none appeared in the other buffers, but the proteolytically cleaved neurocan-N was bound to other structures, and appeared in the later buffers. Phosphacan was detected as a 400-kDa core protein, and there was a single isoform of brevican corresponding to 145-kDa core protein. Two strong bands were detected for the soluble form of aggrecan (>500-kDa core protein) together with three or more faintly stained bands (between 450 and 250 kDa). Among the three spliced variants of versican, only versican V2 (400 kDa protein) was detected in the soluble extract. For NG2, a 290-kDa core protein was detected.
      Figure thumbnail gr1
      FIGURE 1Sequential extraction of CSPGs from adult rat brain and their characterization by Western blotting. Adult rat brain was sequentially extracted with: buffer 1 (TBS 50 mm Tris, 2 mm EDTA, 10 mm NEM, 2 mm phenylmethylsulfonyl fluoride, pH 7.0) (lane 1), buffer 2 (buffer 1 + 0.5% Triton X-100) (lane 2), buffer 3 (buffer 2 + 1 m NaCl) (lane 3), and buffer 4 (buffer 2 + 6 m urea) (lane 4). Fractions corresponding to 200μg of protein were precipitated with 95% ethanol containing 1.3% potassium acetate and digested with 10 international milliunits of a protease-free preparation of chondroitinase ABC for 3 h at 37 °C. The digest was run on a 5% gel and electrotransferred onto a nitrocellulose membrane. The membrane was incubated with antibodies for neurocan, phosphacan, brevican, aggrecan, versican, and NG2 (see for antibodies and dilutions), followed by peroxidase-conjugated antimouse IgG and finally developed using chemiluminescent substrate.
      To monitor anatomically the location of the matrix molecules extracted by saline we performed immunohistochemistry on the deep cerebellar nuclei, the PNNs of which have been investigated in detail (
      • Carulli D.
      • Rhodes K. E. Brown D.J.
      • Bonnert T. P. Pollack S.J.
      • Oliver K. Strata P.
      • Fawcett J.W.
      ). The sections were fixed after buffer extraction to prevent further removal of matrix during staining. Before extraction there was staining for neurocan, aggrecan, versican, and phosphacan in the PNNs and also diffusely in the cerebellar ECM (Fig. 2, A–D). Brevican was not studied since the anti-brevican antibody gave high background on fresh frozen sections. Washing with normal saline resulted in a decrease in neurocan, aggrecan, and phosphacan staining in the diffuse ECM, whereas the staining in PNNs remained unchanged. WFA, which is a marker of PNNs did not show any decrease in the ECM staining after saline wash (Fig. 2E). The staining intensity of HABP and TN-R (Fig. 2, F and G) in the general ECM was decreased by saline wash, whereas their staining in PNNs remained unaltered.
      Figure thumbnail gr2
      FIGURE 2Sequential extraction of CSPGs from cerebellum, showing that PNNs are only extracted by 6 m urea. Sagittal sections (16-μm thickness) of fresh frozen cerebellum from adult rat on a superfrost slide were sequentially washed with buffers 1, 2, 3, and 4 and fixed with 4% paraformaldehyde for 10 min. The sections were stained for neurocan (A), aggrecan (B), versican (C), phosphacan (D), WFA (E), HABP (F), and TN-R (G). The panels show PNNs in the deep cerebellar nuclei. Scale bar:25 μm.
      Extraction with buffer 2 (detergent buffer), released neurocan-N, phosphacan/RPTPβ, brevican, and aggrecan in smaller amounts than was released by saline buffer (Fig. 1, lane 2). Large amounts of the membrane-attached CSPG NG2 (300-kDa core protein) was released along with a small proportion of 260-kDa core protein. Versican V2 was released in very small amounts by detergent buffer. The results from immunostaining of cerebellar sections indicated that the PNN staining for neurocan, aggrecan, versican, HABP, WFA, and TN-R (Fig. 2, A–C, E, F, and G), remained unchanged whereas phosphacan staining (Fig. 2D) was completely removed by detergent wash.
      Extraction with buffer 3 (saline-containing detergent and 1 m NaCl), released small amounts of versican and aggrecan as shown in Fig. 1 (lane 3). Large amounts of neurocan comparable to detergent wash were released by high salt buffer, but barely detectable amounts of NG2, phosphacan, and brevican were released. Immunohistochemical staining of sections showed that the ECM staining for neurocan, aggrecan, versican, and HABP was further reduced (Fig. 2, A–C and F), but their staining in PNNs remained more or less the same. PNN staining for TN-R was slightly decreased by NaCl buffer wash (Fig. 2G).
      Extraction with buffer 4 (saline-containing detergent and 6 m urea), released large amounts of neurocan-N, brevican, versican, and aggrecan (Fig. 1, lane 4). No NG2 was released by this buffer, whereas a very small amount of phosphacan was detected. Immunohistochemical staining indicated that urea wash completely removed all CSPGs from the sections, including PNN components (Fig. 2). Small amounts of versican was still present showing a scattered punctuate pattern (Fig. 2C, lane 4). In conclusion, PNNs containing CSPGs of the lectican family, HA and TN-R remained within the tissue during extraction by the first three buffers, and were only removed by urea buffer. Phosphacan was released from PNNs by detergent buffer suggesting that it is bound to them in a way different from other CSPGs.
      The CS/DS and HS Disaccharide Composition of GAG Chains Differs in the Four Buffer Extracts—Our results show that some CSPGs are tightly bound to PNNs, whereas some are membrane-associated and some are free floating (see Fig. 1), and most CSPG core proteins are found in both the free floating and PNN matrix. Hence we investigated whether the GAG composition of PGs varies with their location in the matrix. The PGs sequentially extracted with the four buffers from adult rat brain were partially purified and the GAG chains in the samples were released by Pronase digestion. The disaccharide composition was analyzed after digesting the GAG chains with chondroitinase ABC or heparinase I and III followed by 2AB-labeling and anion-exchange HPLC as described under “Experimental Procedures.” The results obtained for CS/DS-GAG disaccharide composition are summarized in Table 2. Whereas buffer 1 and 2 extracts accounted to 72.3 and 20.4%, buffer 3 and 4 extracts accounted to only 5.9 and 1.3% of the total CS/DS-GAGs. The degree of sulfation was higher for the CSPGs extracted with saline (0.99), whereas it was lower for urea-extracted CSPGs (0.94). The CS/DS disaccharide composition of the four extracts differed from each other. In all the extracts, 4-O-sulfated disaccharide unit (A-unit) was the predominant form. Whereas the content of nonsulfated (O-unit), 6-O-sulfated (C-unit), and 2,6-O-disulfated disaccharide units (D-unit) increased, A-unit gradually decreased from saline to urea extract. The content of O-, C-, D-, and 4,6-O-disulfated disaccharide units (E-unit) in urea extract was increased by 49, 36, 26, and 20%, respectively, whereas A-unit was decreased by 6% in comparison to the saline extract. On the other hand the E-unit content of detergent extract was increased by 22% as compared with saline extract and remained unchanged for NaCl and urea extracts. The amount of 2,4-O-disulfated disaccharide unit (B-unit) was the same in all the extracts. Taken together, this suggests that the GAG sulfation patterns of PNN-associated CSPGs differed from soluble and membranous CSPGs.
      TABLE 2Disaccharide composition of CS/DS-GAG chains in buffer 1, 2, 3, and 4 extracts from adult rat brain
      Buffer 1 extractBuffer 2 extractBuffer 3 extractBuffer 4 extract
      nmol/g brain
      ΔDi-0S
      a ΔDi-0S, ΔHexUA-GalNAc; ΔDi-6S, ΔHexUA—GalNAc(6-O-sulfate); ΔDi-4S, ΔHexUA-GalNAc(4-O-sulfate); ΔDi-diSD, ΔHexUA(2-O-sulfate)-GalNAc(6-O-sulfate); ΔDi-diSB, ΔIdoUA(2-O-sulfate)-GalNAc(4-O-sulfate); ΔDi-diSE, ΔHexUA-GalNAc(4,6-O- disulfate); ΔDi-TriS, ΔHexUA(2-O-sulfate)-GalNAc(4,6-O-sulfate).
      2.51 (3.6%)1.29 (6.6%)0.39 (6.8%)0.13 (10.4%)
      ΔDi-6S1.57 (2.3%)0.54 (2.8%)0.22 (3.9%)0.06 (4.9%)
      ΔDi-4S62.90 (91.2%)16.97 (87.0%)4.87 (85.5%)1.03 (80.6%)
      ΔDi-diSD0.45 (0.7%)0.15 (0.8%)0.06 (1.0%)0.02 (1.2%)
      ΔDi-diSB0.48 (0.7%)0.10 (0.5%)0.03 (0.6%)0.01 (0.8%)
      ΔDi-diSE0.98 (1.4%)0.43 (2.2%)0.13 (2.3%)0.03 (2.1%)
      ΔDi-TriS0.06 (0.1%)0.02 (0.1%)ND
      b ND, not detected.
      ND
      Total disaccharides68.95 (72.3%)
      c The values in parentheses represent the percentage of disaccharides in the four extracts, taking the sum of CS/DS disaccharides found in all four extracts as 100%.
      19.50 (20.4%)5.7 (5.9%)1.28 (1.3%)
      Sulfation degree
      d Sulfation degree was calculated as the average number of sulfate groups per disaccharide unit.
      0.990.970.970.94
      a ΔDi-0S, ΔHexUA-GalNAc; ΔDi-6S, ΔHexUA—GalNAc(6-O-sulfate); ΔDi-4S, ΔHexUA-GalNAc(4-O-sulfate); ΔDi-diSD, ΔHexUA(2-O-sulfate)-GalNAc(6-O-sulfate); ΔDi-diSB, ΔIdoUA(2-O-sulfate)-GalNAc(4-O-sulfate); ΔDi-diSE, ΔHexUA-GalNAc(4,6-O- disulfate); ΔDi-TriS, ΔHexUA(2-O-sulfate)-GalNAc(4,6-O-sulfate).
      b ND, not detected.
      c The values in parentheses represent the percentage of disaccharides in the four extracts, taking the sum of CS/DS disaccharides found in all four extracts as 100%.
      d Sulfation degree was calculated as the average number of sulfate groups per disaccharide unit.
      The HS-GAG disaccharide composition of the four buffer extracts was also analyzed, and the results are summarized in Table 3. Soluble and membrane-bound HS-GAG accounted for 40.1 and 39.3% whereas ionically bound and ECM-bound HS-GAG accounted for 15.6 and 5.0% of the total HS disaccharides. The degree of sulfation was higher for urea extract (0.67) than saline (0.54), detergent (0.59) or NaCl extracts (0.57). Nonsulfated disaccharide unit (HS-0S) was the major disaccharide in all the extracts, and its content was higher in the saline extract (62.4%) whereas lower in urea extract (55.7%). The second major disaccharide in all the extracts was 2-N-sulfated disaccharide (HS-NS), and its content remained unaltered in all the extracts. The content of HS-0S was decreased by 5.6% in the urea extract when compared with the saline extract and the contents of 6-O-sulfated (HS-6S), 2-N-,6-O-disulfated (HS-diS1), 2-O-,2-N-disulfated (HS-diS2), and 2-O-,2-N-,6-O-trisulfated HS-TriS) disaccharides were increased by 31, 16, 12, and 16%, respectively, in the urea extract when compared with the saline extract. These results suggest that the PNN-associated HSPGs contain highly sulfated GAG and its disaccharide composition differed from those in soluble and membrane-bound HSPGs.
      TABLE 3Disaccharide composition of HS-GAG chains in buffer 1, 2, 3, and 4 extracts from adult rat brain
      Buffer 1 extractBuffer 2 extractBuffer 3 extractBuffer 4 extract
      nmol/g brain
      ΔDiHS-0S
      a ΔDiHS-0S, ΔHexUA-GlcNAc; ΔDiHS-6S, ΔHexUA-GlcNAc(6-O-sulfate); ΔDiHS-NS, ΔHexUA-GlcNAc(2-N-sulfate); ΔDiHS-diS1, ΔHexUA-GlcNAc(2-N-, 6-O- disulfate); ΔDiHS-diS2, ΔHexUA(2-O-sulfate)-GlcNAc(2-N-sulfate); ΔDiHS-TriS, ΔHexUA(2-O-sulfate)-GlcNAc(2-N-, 6-O-disulfate).
      2.65 (62.4%)2.48 (59.8%)0.99 (59.9%)0.29 (55.7%)
      DDiHS-6S0.16 (3.9%)0.21 (5.0%)0.07 (4.4%)0.04 (7.4%)
      DDiHS-NS0.89 (21.0%)0.85 (20.5%)0.38 (23.0%)0.11 (20.2%)
      DDiHS-diS10.12 (2.8%)0.15 (3.6%)0.05 (3.0%)0.02 (3.9%)
      ΔDiHS-diS20.26 (5.9%)0.27 (6.6%)0.09 (5.8%)0.04 (7.5%)
      ΔDiHS-TriS0.16 (3.9%)0.19 (4.4%)0.07 (3.9%)0.03 (5.4%)
      Total disaccharides4.24 (40.1%)
      b The values in parentheses represent the percentage of disaccharides in the four extracts, taking the sum of HS disaccharides found in all four extracts as 100%.
      4.15 (39.3%)1.65 (15.6%)0.53 (5.0%)
      Sulfation degree
      c Sulfation degree was calculated as the average number of sulfate group per disaccharide unit.
      0.540.590.570.67
      a ΔDiHS-0S, ΔHexUA-GlcNAc; ΔDiHS-6S, ΔHexUA-GlcNAc(6-O-sulfate); ΔDiHS-NS, ΔHexUA-GlcNAc(2-N-sulfate); ΔDiHS-diS1, ΔHexUA-GlcNAc(2-N-, 6-O- disulfate); ΔDiHS-diS2, ΔHexUA(2-O-sulfate)-GlcNAc(2-N-sulfate); ΔDiHS-TriS, ΔHexUA(2-O-sulfate)-GlcNAc(2-N-, 6-O-disulfate).
      b The values in parentheses represent the percentage of disaccharides in the four extracts, taking the sum of HS disaccharides found in all four extracts as 100%.
      c Sulfation degree was calculated as the average number of sulfate group per disaccharide unit.
      The content of CS/DS-GAG in brain was about 9 times higher than HS-GAGs, and PNN-associated GAGs accounted for only a small proportion of the total CS/DS- and HS-GAGs, i.e. 1.3 and 5.0%, respectively. In PNNs, the proportion of CS/DS-GAG was higher (71% of total GAG) than HS-GAG (29%).
      CSPGs in Stable Complexes Only Extractable with Urea Buffer Appear at the Same Time as PNN Formation—The matrix compartments containing neurocan, phosphacan, and versican in rat brain during development were studied by Western blotting after sequential extractions in saline, detergent, and 6 m urea buffers. The developmental stages chosen were embryonic day 18 (E18), P7, P14, and P21. At E18, almost all the CSPGs were extracted in saline (E18, lane 1 in Fig. 3), except for NG2, almost all of which is membrane-associated at this stage of development and is therefore extracted by detergent buffer (E18, lane 2). All the CSPGs were extracted by saline or detergent leaving no CSPGs to appear in the 6 m urea buffer, indicating that none were associated with PNNs at this stage. Unlike later developmental stages, the V0 and V1 splice variants of versican were seen at E18 in addition to V2.
      Figure thumbnail gr3
      FIGURE 3Sequential extraction of CSPGs from rat brain during development and their characterization by Western blotting. CSPGs were sequentially extracted with buffers 1, 2, and 4 (lanes 1, 2, and 3) from E18, P7, P14, and P21 rat brains and analyzed by Western blotting for neurocan, phosphacan, versican, and NG2 as described in the legend for .
      At P7, the main change from E18 was the appearance of phosphacan and versican in more stable structures, extracted by detergent and 6 m urea (P7, lanes 2 and 3, respectively) and the presence of NG2 extractable in saline (P7, lane 1) and therefore presumably shed from the membrane. Whereas the amounts of versican V0, V1, and V2 were in equal proportions in the soluble extract (P7, lane 1), V2 became the predominant form in the detergent extract. Immunohistochemical studies of P7 rat brain showed faint neurocan staining in PNNs of the brainstem, such as in the pontine reticular nucleus (data not shown), and phosphacan/RPTPβ and versican staining started to appear in PNNs of a few brain areas, such as the vestibular nucleus (Figs. 4, B and C). These results are consistent with the Western blot results showing extraction of phosphacan/RPTPβ and versican by urea buffer at P7, but not neurocan.
      Figure thumbnail gr4
      FIGURE 4Immunostaining for neurocan, phosphacan, and versican in PNNs of rat brain during development. Sagittal sections (40-μm thickness) of perfused brain from rat were stained with anti-neurocan (A), phosphacan (B), and versican (C) antibodies. A shows neurocan-immunoreactive PNNs in P14 (left, arrows) and P21 (right) deep cerebellar nuclei (DCN). B shows phosphacan-positive PNNs in P7 (left, arrows) and P21 (right) vestibular nuclei (VN). In C, versican-immunoreactive PNNs in P7 (left, arrows) and P21 (right) VN are shown. Scale bar:25 μm.
      At P14, the main change was the further increase in CSPG that could only be extracted in 6 m urea (P14, lane 3). A significant amount of neurocan-N was incorporated into this stable pool of matrix together with versican and phosphacan. At this time point all the versican V0 and V1 was absent. There was also a further increase in the amount of shed NG2 (P14, lane 1). Immunohistochemical studies showed the appearance at this time point of neurocan staining in the PNNs of the deep cerebellar nuclei of P14 rats (Fig. 4A), coinciding with their extraction by urea buffer.
      At P21, the results were similar to those at P14, but with a further increase in the amount of CSPG now in the stable pool that can only be extracted with 6 m urea (P21, lane 3). In sections PNN staining for neurocan, phosphacan/RPTPβ, and versican became intense at P21 (Figs. 4, A–C) coinciding with the extraction of large amounts of these CSPGs with urea buffer.
      In conclusion, the evidence from immunohistochemistry indicates that neurocan, phosphacan and versican start to associate with PNNs from P7–14. Coincident with the histological appearance of PNNs, the biochemical extraction data show that there is the appearance of a pool of CSPG that cannot be extracted by saline or detergent but can be extracted in 6 m urea.
      Streptomyces Hyaluronidase Released Aggrecan, Versican, Neurocan-N, Phosphacan, and Brevican from PNNs—To investigate how the CSPGs in PNNs interact with HA, adult rat brain was sequentially extracted with saline, detergent, and NaCl buffers and the precipitate left after washing was digested either with a protease-free preparation of chondroitinase ABC or hyaluronidase in the presence of protease inhibitors, and the released products were analyzed by Western blotting (Fig. 5). Chondroitinase ABC digests the CS-GAG as well as HA (
      • Yamagata T.
      • Kawamura Y.
      • Suzuki S.
      ), whereas hyaluronidase from Streptomyces selectively digests HA (
      • Ohya T.
      • Kaneko Y.
      ). Chondrotinase ABC and hyaluronidase digestions of the brain homogenate after NaCl buffer wash released neurocan, phosphacan, aggrecan, versican, and brevican (lanes 1 and 2, respectively). Because saline, detergent, and NaCl buffer washes are unable to release CSPGs from PNNs (as shown previously by immunohistochemistry), the CSPGs released by chondroitinase ABC and hyaluronidase after NaCl buffer wash should be those from PNNs. Digestion with chondroitinase ABC after hyaluronidase digestion released small amounts of all the CSPGs (lane 3), indicating that small proportions of CSPGs are still retained after hyaluronidase digestion.
      Figure thumbnail gr5
      FIGURE 5Release of CSPGs from PNNs by chondroitinase ABC and hyaluronidase digestions. Adult rat brain was sequentially washed with buffers 1, 2, and 3, and the precipitate obtained after centrifugation was equilibrated with the digestion buffers of chondroitinase ABC and Streptomyces hyaluronidase, separately. The precipitate was resuspended in 3 ml of digestion buffers and treated with a 0.1 international units of protease-free preparation of chondroitinase ABC at 37 °C for 3 h or 60 TRU of Streptomyces hyaluronidase containing protease inhibitors at 37 °C for 3 h, separately. The digest was centrifuged, and the pellet was re-digested with chondroitinase ABC as before. The released CSPGs were precipitated with 95% ethanol/1.3% potassium acetate and Western-blotted for aggrecan, versican, neurocan, brevican, and phosphacan (see for antibodies and dilutions) as described in the legend to . Lanes 1 and 2 represent the CSPGs released by chondroitinase and hyaluronidase digestions, respectively. Lane 3 represents the CSPGs released by chondroitinase digestion after hyaluronidase digestion.
      Immunohistochemistry was also performed to see how the PNN components are affected by chondroitinase ABC and hyaluronidase digestions. The results of digestions are shown in Fig. 6 and are summarized in Table 4. Staining for WFA in PNNs was completely abolished by both chondroitinase and hyaluronidase digestions (Fig. 6A). Chondroitinase digestion was able to remove only neurocan and versican staining (Fig. 6, B and C) whereas the staining for aggrecan and phosphacan/RPTPβ (Fig. 6, D and E) remained even after the digestion. On the other hand, neurocan, versican, aggrecan and phosphacan/RPTPβ did not show any staining after hyaluronidase digestion. Because antibrevican antibody gave high background in immunostaining, brevican was not tested in the present study. HABP and TN-R staining (Fig. 6, F and G) disappeared after hyaluronidase digestion, but their staining was only partly removed by chondroitinase digestion.
      Figure thumbnail gr6
      FIGURE 6Release of PNN components from adult rat cerebellum by chondroitinase ABC or hyaluronidase digestions. Sagittal sections (16-μm thickness) of fresh adult rat cerebellum were digested either with a protease-free preparation of chondroitinase ABC (0.1 international units/ml) or hyaluronidase (10 TRU/ml). After washing with PBS, the sections were fixed using 4% PFA and immunostained for WFA (A), neurocan (B), versican (C), aggrecan (D), phosphacan (E), HABP (F), and TN-R (G) as described under “Experimental Procedures.” The antibodies used for staining are listed in . Scale bar:25 μm.
      TABLE 4Release of PNN components from adult rat cerebellum by chondroitinase ABC or Streptomyces hyaluronidase digestions
      PNN componentPNN staining after digestion with
      Chondroitinase ABC
      a Chondroitinase ABC (protease-free) digestion was performed at 37 °C for 60 min in 0.1 m Tris-HCl, pH 8.0, containing 0.03 m sodium acetate.
      Hyaluronidase
      b Streptomyces hyaluronidase digestion was performed at 37 °C for 120 min in 20 mm sodium acetate, pH 6.0 containing 0.15 m NaCl.
      WFADisappearDisappear
      NeurocanDisappearDisappear
      VersicanDisappearDisappear
      AggrecanNot removedDisappear
      Phosphacan/RPTPβNot removedDisappear
      HABPPartly removedDisappear
      TN-RPartly removedDisappear
      a Chondroitinase ABC (protease-free) digestion was performed at 37 °C for 60 min in 0.1 m Tris-HCl, pH 8.0, containing 0.03 m sodium acetate.
      b Streptomyces hyaluronidase digestion was performed at 37 °C for 120 min in 20 mm sodium acetate, pH 6.0 containing 0.15 m NaCl.
      In conclusion, the staining for all PNN components disappeared after hyaluronidase digestion indicating that HA could be the backbone structure of PNNs and hence degradation of HA abolished the staining of PNN components. The results obtained from chondroitinase digestion are in agreement with the staining results.
      Extraction and Characterization of CSPGs from Adult Rat Spinal Cord—Compared with brain, the gray matter of the spinal cord contains much denser matrix that stains strongly for most of the CNS CSPGs. We therefore examined the results of sequential extractions in buffers on the removal of CSPGs from adult spinal cord tissue. The CSPGs from adult rat spinal cord was sequentially extracted with saline, detergent, NaCl, and urea buffers as described under “Experimental Procedures.” Extraction of adult rat spinal cord with saline released large amounts of neurocan-N, phosphacan, brevican, and aggrecan, a small amount of versican V2 but no NG2 (Fig. 7, lane 1). Detergent buffer mainly extracted NG2, which is a membrane-bound CSPG (lane 2). In addition, detergent buffer extract contained small proportions of neurocan-N, brevican, aggrecan, and versican V2, but no phosphacan/RPTPβ (lane 2). NaCl buffer failed to extract any of the CSPGs (lane 3), whereas urea buffer extracted neurocan-N, brevican, aggrecan, and versican V2 (lane 4) which are PNN-associated CSPGs in brain.
      Figure thumbnail gr7
      FIGURE 7Sequential extraction of CSPGs from adult spinal cord and their characterization by Western blotting. Adult spinal cord was sequentially extracted with buffers 1, 2, 3, and 4 and fractions corresponding to 200 μg of protein were precipitated with 95% ethanol containing 1.3% potassium acetate, digested with 10 international milliunits of a protease-free preparation of chondroitinase ABC, and the digest was run on a 5% gel and electrotransferred onto a nitrocellulose membrane. The membrane was incubated with antibodies for neurocan, phosphacan, brevican, aggrecan, versican, and NG2 (see for antibodies and dilutions), followed by peroxidase-conjugated antimouse IgG and finally developed using chemiluminescent substrate.
      There were significant differences between brain and spinal cord. A comparison of the CSPGs extracted under the different conditions from adult rat brain and spinal cord are presented in Table 5. In brief, brain and spinal cord contain similar CSPGs but their extractability differed in these tissues. Interestingly, the cleaved products of aggrecan detected by cat-301 antibody differed in brain and spinal cord (see Figs. 1 and 7). In brain, much of the neurocan is in the intact uncleaved form, but this form was not seen in the spinal cord. About 40% of total neurocan-N was extracted with saline from brain, whereas it was 65% from spinal cord. A comparison of the extractability of other CSPGs by saline is given in Table 5. Whereas the ionically bound forms of neurocan-N, brevican, and aggrecan exist in brain, such a form is not present in the spinal cord. Urea buffer extracted neurocan-N, brevican, aggrecan, and versican V2 from both brain and spinal cord. The proportion of PNN-associated neurocan-N, brevican, and versican V2 were comparable in brain and spinal cord, but such a comparison could not be made for aggrecan because of the presence of different isoforms.
      TABLE 5Comparison of extractability of CSPGs from adult rat brain and spinal cord
      CSPGBuffer 1 extractBuffer 2 extractBuffer 3 extractBuffer 4 extract
      %%%%
      Neurocan-NBrain40
      a The values represent the percentage of each PG in the four extracts, taking the sum of one particular PG in all four extracts as 100%. The calculations were performed using the Image J software (Rasband, W. S. (1997-2005) National Institutes of Health, Bethesda, MD).
      181626
      Spinal cord6510-25
      PhosphacanBrain6331-6
      Spinal cord100---
      BrevicanBrain4527-28
      Spinal cord5322-25
      AggrecanBrain
      b The percentage of the major band, which is common in all the extracts.
      31212127
      Spinal cord
      c The bands in the buffer 1 and 2 extracts were different from the one in the buffer 4 extract (Fig. 7); hence a comparison cannot be made.
      ----
      VersicanBrain1491760
      Spinal cord1115-74
      NG2Brain20719-
      Spinal cord1387--
      a The values represent the percentage of each PG in the four extracts, taking the sum of one particular PG in all four extracts as 100%. The calculations were performed using the Image J software (Rasband, W. S. (1997-2005) National Institutes of Health, Bethesda, MD).
      b The percentage of the major band, which is common in all the extracts.
      c The bands in the buffer 1 and 2 extracts were different from the one in the buffer 4 extract (Fig. 7); hence a comparison cannot be made.
      To check whether the CSPGs extracted with urea buffer are from PNN in spinal cord, immunohistochemical staining was done after sequential washes of tissue sections with saline, detergent, NaCl, and urea buffers. Neurocan was randomly chosen as a candidate for this experiment. Whereas saline, detergent, and NaCl buffers failed to remove PNNs, urea wash removed the nets (data not shown), and these results are comparable with those obtained from brain.

      DISCUSSION

      Histologically it is possible to identify three types of extracellular matrix in the CNS. The first is the diffuse matrix that exists throughout the CNS. The second is defined by cell surface-associated matrix molecules such as the CSPG NG2, and also surface-associated HSPGs. The third is the dense organized matrix of PNNs. The objective of this experiment was to examine the CSPG composition of these different matrix compartments. We developed a method for the differential extraction of the various matrix compartments from brain and spinal cord tissue. Analysis of these extracts showed that there are different amounts of the various CSPGs in these compartments, and that CSPGs that are associated with PNNs have a different GAG disaccharide composition. Moreover, by means of chondroitinase and hyaluronidase digestions we could further understand how various CSPGs interact with PNNs.
      The Different ECM Compartments Contain Different CSPGs—The majority of the brain CSPGs are in the easily extracted fraction that is soluble in saline and are therefore unbound or loosely attached to the ECM. We were able to visualize removal of these CSPGs in tissue sections treated in the same way. Of the lecticans, which are secreted molecules, much of the neurocan and brevican was present in this compartment as reported previously (
      • Rauch U.
      • Gao P.
      • Janetzko A.
      • Flaccus A.
      • Hilgenberg L.
      • Tekotte H.
      • Margolis R.K.
      • Margolis R.U.
      ,
      • Seidenbecher C.I.
      • Richter K.
      • Rauch U.
      • Fassler R.
      • Garner C.C.
      • Gundelfinger E.D.
      ,
      • Perides G.
      • Rahemtulla F.
      • Lane W.S.
      • Asher R.A.
      • Bignami A.
      ). Interestingly, all the intact form of neurocan was removed in the first saline extraction, whereas the proteolytically cleaved form, neurocan-N, was also found in the detergent, high salt, and urea extractable fractions. This would imply that it is only after proteolytic cleavage that the N-terminal part of neurocan becomes incorporated into PNNs. Aggrecan also partly appeared in the soluble fraction as did versican. Most of the phosphacan was also present in this soluble form, as was some NG2 which must have been shed from the membrane. The second buffer containing detergent dissolves membrane lipids and hence extracts proteins bound to the membrane and also releases intracellular molecules. The presence of a proportion of all the lecticans in the detergent extract is intriguing. These CSPGs are not membrane-linked, so their presence in the detergent extract indicates either that are linked to molecules in the plasma membrane or are present intracellularly. Neurocan-N binds various cell adhesion molecules (
      • Grumet M.
      • Flaccus A.
      • Margolis R.U.
      ,
      • Friedlander D.R.
      • Milev P.
      • Karthikeyan L.
      • Margolis R.K.
      • Margolis R.U.
      • Grumet M.
      ,
      • Retzler C.
      • Gohring W.
      • Rauch U.
      ,
      • Oleszewski M.
      • Beer S.
      • Katich S.
      • Geiger C.
      • Zeller Y.
      • Rauch U.
      • Altevogt P.
      ) and is also present inside neurons (
      • Rauch U.
      • Gao P.
      • Janetzko A.
      • Flaccus A.
      • Hilgenberg L.
      • Tekotte H.
      • Margolis R.K.
      • Margolis R.U.
      ,
      • Aquino D.A.
      • Margolis R.U.
      • Margolis R.K.
      ,
      • Aquino D.A.
      • Margolis R.U.
      • Margolis R.K.
      ) and in the cytoplasm of glial processes (
      • Matsui F.
      • Nishizuka M.
      • Yasuda Y.
      • Aono S.
      • Watanabe E.
      • Oohira A.
      ). Brevican has a GPI-anchored form but is also reported to be membrane-associated through a different mechanism (
      • Seidenbecher C.I.
      • Smalla K.H.
      • Fischer N.
      • Gundelfinger E.D.
      • Kreutz M.R.
      ), and is also associated with the glial membrane in the cerebellar glomeruli (
      • Yamada H.
      • Fredette B.
      • Shitara K.
      • Hagihara K.
      • Miura R.
      • Ranscht B.
      • Stallcup W.B.
      • Yamaguchi Y.
      ). The detergent-released phosphacan fraction may be mostly the full-length RPTPβ. There is some evidence that this molecule may also be on the membrane of net-bearing neurons (
      • Engel M.
      • Maurel P.
      • Margolis R.U.
      • Margolis R.K.
      ,
      • Hayashi N.
      • Miyata S.
      • Yamada M.
      • Kamei K.
      • Oohira A.
      ,
      • Miyata S.
      • Nishimura Y.
      • Hayashi N.
      • Oohira A.
      ). Interaction of phosphacan with neuronal cell adhesion molecules has been reported (
      • Maurel P.
      • Rauch U.
      • Flad M.
      • Margolis R.K.
      • Margolis R.U.
      ,
      • Milev P.
      • Friedlander D.R.
      • Sakurai T.
      • Karthikeyan L.
      • Flad M.
      • Margolis R.K.
      • Grumet M.
      • Margolis R.U.
      ,
      • Milev P.
      • Maurel P.
      • Haring M.
      • Margolis R.K.
      • Margolis R.U.
      ,
      • Falk J.
      • Bonnon C.
      • Girault J.A.
      • Faivre-Sarrailh C.
      ) and hence one of the isoforms released by detergent wash may be phosphacan bound in this way. As expected, the majority of the cell surface-associated CSGP NG2 appeared in the detergent extract. The third buffer containing 1 m NaCl releases CSPGs bound to the ECM through ionic interactions. All the CSPGs except NG2 appeared in small amounts in this fraction but the binding partners for these CSPGs are unknown. The fourth buffer containing urea is capable of solubilizing PGs that are tightly bound to the ECM and therefore extracts CSGPs that are bound to the PNNs as ternary complexes with HA and TN-R. We were able to visualize in tissue sections treated with the same solutions that the first three buffers remove most of the CPSGs from the brain, but leave PNNs almost untouched. The presence of neurocan-N, brevican, versican V2, and aggrecan in the fraction that requires 6 m urea for solubilization indicates their strong association with PNN components. In agreement with our results, it has been reported that about 25% of brevican is insoluble under physiological conditions (
      • Seidenbecher C.I.
      • Smalla K.H.
      • Fischer N.
      • Gundelfinger E.D.
      • Kreutz M.R.
      ).
      Comparing Brain and Spinal Cord ECM—Sequential extraction of CSPGs from spinal cord indicated that all the same ECM-associated CSPGs that are found in brain (neurocan, brevican, aggrecan and versican) except phosphacan are also associated with the stable spinal cord PNN ECM. The PNNs in the spinal cord in rat have previously been reported to contain aggrecan (
      • Kalb R.G.
      • Hockfield S.
      ), neurocan, brevican (
      • Beggah A.T.
      • Dours-Zimmermann M.T.
      • Barras F.M.
      • Brosius A.
      • Zimmermann D.R.
      • Zurn A.D.
      ), and HA (
      • Bignami A.
      • Asher R.
      • Perides G.
      ). Histology shows that the gray matter of spinal cord is almost like a single continuous PNN, unlike the brain where PNNs are found on single neurons with loose soluble matrix between them.
      The Appearance of PNNs in Development Coincides with the Appearance of Urea-extractable CSPGs—In the brain tissue PNNs started to appear around P7 and increased in number and density from then onwards. Before P7 almost all the CSPGs were readily soluble in saline, except for membrane-associated NG2. CSPGs that needed urea for solubilization started to appear at P7–14, at the same time as PNN formation. In embryos and newborns the two larger splice forms of versican (V0 and V1) were present, but only versican V2 was detected in PNNs from P7 to adulthood, where versican V2 is the only form present (
      • Asher R.A.
      • Morgenstern D.A.
      • Shearer M.C.
      • Adcock K.H.
      • Pesheva P.
      • Fawcett J.W.
      ,
      • Schmalfeldt M.
      • Bandtlow C.E.
      • Dours-Zimmermann M.T.
      • Winterhalter K.H.
      • Zimmermann D.R.
      ,
      • Milev P.
      • Maurel P.
      • Chiba A.
      • Mevissen M.
      • Popp S.
      • Yamaguchi Y.
      • Margolis R.K.
      • Margolis R.U.
      ). The different functions of the various versican isoforms are not fully known, but there may be differential effects on axon growth, cell death, and differentiation (
      • Wu Y.
      • Sheng W.
      • Chen L.
      • Dong H.
      • Lee V.
      • Lu F.
      • Wong C.S.
      • Lu W.Y.
      • Yang B.B.
      ).
      CSPGs Interact with HA in PNNs to Form Ternary Complexes—Lecticans are able to link to HA through their N-terminal HA binding domain and to TN-R through their C-terminal lectin domain forming HA-lectican-TN-R complexes (
      • Yamaguchi Y.
      ,
      • LeBaron R.G.
      • Zimmermann D.R.
      • Ruoslahti E.
      ,
      • Rauch U.
      • Hirakawa S.
      • Oohashi T.
      • Kappler J.
      • Roos G.
      ,
      • Seyfried N.T.
      • McVey G.F.
      • Almond A.
      • Mahoney D.J.
      • Dudhia J.
      • Day A.J.
      ). In the cartilage, aggrecan forms link protein-stabilized complexes with HA, and may form similar complexes in PNNs (
      • Day A.J.
      • Prestwich G.D.
      ,
      • Toole B.P.
      ). Complexes of this type require buffers such as 6 m urea for solubilization. A requirement for the formation of PNNs is probably the existence of a pericellular HA coat, produced by the expression of hyaluronan synthases in those neurons with PNNs (
      • Carulli D.
      • Rhodes K. E. Brown D.J.
      • Bonnert T. P. Pollack S.J.
      • Oliver K. Strata P.
      • Fawcett J.W.
      ).
      To evaluate how CSPGs are retained in PNNs, we first extracted tissue with saline, detergent, and high salt buffers to remove all the CSPGs except those in PNNs. We then treated with hyaluronidase and with chondrotinase and analyzed the CSPGs that were released. The release of aggrecan, versican, neurocan, phosphacan, and brevican from PNNs by hyaluronidase indicates that the core protein of these CSPGs is involved in the interaction with HA. Chondroitinase also released all the CSPGs and this could be attributed to its known hyaluronidase activity. The release of CSPGs by chondroitinase after hyaluronidase digestion might indicate that the CS chains also play a role in the binding. Removal of neurocan and WFA staining from PNNs after chondroitinase digestion has also been reported by Koppe et al. (
      • Koppe G.
      • Bruckner G.
      • Brauer K.
      • Hartig W.
      • Bigl V.
      ) and Pizzorusso et al. (
      • Pizzorusso T.
      • Medini P.
      • Berardi N.
      • Chierzi S.
      • Fawcett J.W.
      • Maffei L.
      ). In our experiment there was some disagreement between observations on removal of CSPGs in tissue sections and from dissociated tissue. In our sections aggrecan and phosphacan/RPTPβ were not completely released by chondroitinase digestion, whereas there was release from dissociated tissue. It may be that removal from both tissues is partial. Partial removal of TN-R staining by chondroitinase may be because its binding partner CSPGs have been removed, or may suggest a GAG-mediated interaction of TN-R (
      • Woodworth A.
      • Pesheva P.
      • Fiete D.
      • Baenziger J.U.
      ) with its binding partners.
      The CS/DS- and HS-GAG Composition of PNN-associated CSPGs and HSPGs Differs from Membrane-associated or Free Floating CSPGs—Our disaccharide composition analysis has demonstrated that CS/DS- and HS-GAGs strongly associated with ECM in PNNs have a different composition from those of soluble and transmembrane PGs suggesting a different function for PNN associated GAGs. Many of the CSPGs within PNNs are produced by the neurons themselves, but some are made by surrounding glia (
      • Carulli D.
      • Rhodes K. E. Brown D.J.
      • Bonnert T. P. Pollack S.J.
      • Oliver K. Strata P.
      • Fawcett J.W.
      ). It is possible that these neurons produce GAGs with different sulfation patterns from other neurons and glia, but this needs to be investigated. Sulfation is necessary for CS-GAG to be inhibitory (
      • Smith-Thomas L.C.
      • Stevens J.
      • Fok-Seang J.
      • Faissner A.
      • Rogers J.H.
      • Fawcett J.W.
      ) and among the various subtypes CS-C (6-sulfated GAG) is particularly inhibitory (
      • Snow D.M.
      • Lemmon V.
      • Carrino D.A.
      • Caplan A.I.
      • Silver J.
      ,
      • Gilbert R.J.
      • McKeon R.J.
      • Darr A.
      • Calabro A.
      • Hascall V.C.
      • Bellamkonda R.V.
      ). Recent studies have demonstrated an up-regulation of nonsulfated, 6-sulfated, and 4,6-disulfated CS disaccharides after cortical stab injury (
      • Properzi F.
      • Carulli D.
      • Asher R.A.
      • Muir E.
      • Camargo L.M.
      • van Kuppevelt T.H.
      • ten Dam G.B.
      • Furukawa Y.
      • Mikami T.
      • Sugahara K.
      • Toida T.
      • Geller H.M.
      • Fawcett J.W.
      ) and CS-E has been reported to be both a potent inhibitor and promoter of neurite extension (
      • Gilbert R.J.
      • McKeon R.J.
      • Darr A.
      • Calabro A.
      • Hascall V.C.
      • Bellamkonda R.V.
      ,
      • Clement A.M.
      • Nadanaka S.
      • Masayama K.
      • Mandl C.
      • Sugahara K.
      • Faissner A.
      ,
      • Clement A.M.
      • Sugahara K.
      • Faissner A.
      ). It is possible that the main action of GAGs is to bind and present molecules that affect synaptic plasticity, dendrite growth, and axon growth. Specific interactions between CS-GAG chains and semaphorins (sema5A and sema3A) have been recently reported (
      • Kantor D.B.
      • Chivatakarn O.
      • Peer K.L.
      • Oster S.F.
      • Inatani M.
      • Hansen M.J.
      • Flanagan J.G.
      • Yamaguchi Y.
      • Sretavan D.W.
      • Giger R.J.
      • Kolodkin A.L.
      ,
      • De Wit J.
      • De Winter F.
      • Klooster J.
      • Verhaagen J.
      ) and, interestingly, CS-GAGs convert Sema5A from an attractive to an inhibitory guidance cue for developing axons (
      • Kantor D.B.
      • Chivatakarn O.
      • Peer K.L.
      • Oster S.F.
      • Inatani M.
      • Hansen M.J.
      • Flanagan J.G.
      • Yamaguchi Y.
      • Sretavan D.W.
      • Giger R.J.
      • Kolodkin A.L.
      ). This hypothesis correlates with the reports showing that enzymatic removal of CS-GAG chains is associated with axon sprouting and functional recovery (
      • Pizzorusso T.
      • Medini P.
      • Berardi N.
      • Chierzi S.
      • Fawcett J.W.
      • Maffei L.
      ,
      • Moon L.D.
      • Asher R.A.
      • Rhodes K.E.
      • Fawcett J.W.
      ,
      • Bradbury E.J.
      • Moon L.D.
      • Popat R.J.
      • King V.R.
      • Bennett G.S.
      • Patel P.N.
      • Fawcett J.W.
      • McMahon S.B.
      ,
      • Corvetti L.
      • Rossi F.
      ,
      • Tropea D.
      • Caleo M.
      • Maffei L.
      ).
      We showed strong association of HSPGs with PNN ECM molecules, as indicated by with the fact that their extraction required urea buffer. This suggests specialized functions for HSPGs in nets. Highly sulfated HS sequences are important for the binding of various molecules that could affect synaptic plasticity. These include growth factors (
      • Kreuger J.
      • Jemth P.
      • Sanders-Lindberg E.
      • Eliahu L.
      • Ron D.
      • Basilico C.
      • Salmivirta M.
      • Lindahl U.
      ) and sema 5A (
      • Kantor D.B.
      • Chivatakarn O.
      • Peer K.L.
      • Oster S.F.
      • Inatani M.
      • Hansen M.J.
      • Flanagan J.G.
      • Yamaguchi Y.
      • Sretavan D.W.
      • Giger R.J.
      • Kolodkin A.L.
      ) and heparin enhances sema3A binding to neuropilin-1 and potentiates its growth cone collapsing activity (
      • De Wit J.
      • De Winter F.
      • Klooster J.
      • Verhaagen J.
      ). HS is also involved in the slit1 and slit2 mediated repulsion and collapse of olfactory axons in explant cultures (
      • Hu H.
      ) and it is reported that heparin bind netrins and their receptor DCC (
      • Kappler J.
      • Franken S.
      • Junghans U.
      • Hoffmann R.
      • Linke T.
      • Muller H.W.
      • Koch K.W.
      ).

      Acknowledgments

      We thank Dr. R. Asher for helpful discussions and D. Story for technical assistance.

      References

        • Hockfield S.
        • McKay R.D.
        Proc. Natl. Acad. Sci. U. S. A. 1983; 80: 5758-5761
        • Celio M.R.
        • Blumcke I.
        Brain Res. Brain Res. Rev. 1994; 19: 128-145
        • Celio M.R.
        • Spreafico R.
        • De Biasi S.
        • Vitellaro-Zuccarello L.
        Trends Neurosci. 1998; 21: 510-515
        • Jaworski D.M.
        • Kelly G.M.
        • Hockfield S.J.
        Cell Biol. 1994; 125: 495-509
        • Asher R.A.
        • Scheibe R.J.
        • Keiser H.D.
        • Bignami A.
        Glia. 1995; 13: 294-308
        • Yamaguchi Y.
        Cell Mol. Life Sci. 2000; 57: 276-289
        • Oohashi T.
        • Hirakawa S.
        • Bekku Y.
        • Rauch U.
        • Zimmermann D.R.
        • Su W.D.
        • Ohtsuka A.
        • Murakami T.
        • Ninomiya Y.
        Mol. Cell. Neurosci. 2002; 19: 43-57
        • Bekku Y.
        • Su W.D.
        • Hirakawa S.
        • Fassler R.
        • Ohtsuka A.
        • Kang J.S.
        • Sanders J.
        • Murakami T.
        • Ninomiya Y.
        • Oohashi T.
        Mol. Cell. Neurosci. 2003; 24: 148-159
        • Carulli D.
        • Rhodes K. E. Brown D.J.
        • Bonnert T. P. Pollack S.J.
        • Oliver K. Strata P.
        • Fawcett J.W.
        J. Comp. Neurol. 2006; 494: 559-577
        • Hockfield S.
        • Kalb R.G.
        • Zaremba S.
        • Fryer H.
        Cold Spring Harb. Symp. Quant. Biol. 1990; 55: 505-514
        • Pizzorusso T.
        • Medini P.
        • Berardi N.
        • Chierzi S.
        • Fawcett J.W.
        • Maffei L.
        Science. 2002; 298: 1248-1251
        • Bruckner G.
        • Hausen D.
        • Hartig W.
        • Drlicek M.
        • Arendt T.
        • Brauer K.
        Neuroscience. 1999; 92: 791-805
        • Morawski M.
        • Bruckner M.K.
        • Riederer P.
        • Bruckner G.
        • Arendt T.
        Exp. Neurol. 2004; 188: 309-315
        • Bruckner G.
        • Brauer K.
        • Hartig W.
        • Wolff J.R.
        • Rickmann M.J.
        • Derouiche A.
        • Delpech B.
        • Girard N.
        • Oertel W.H.
        • Reichenbach A.
        Glia. 1993; 8: 183-200
        • Bruckner G.
        • Hartig W.
        • Kacza J.
        • Seeger J.
        • Welt K.
        • Brauer K.
        J. Neurocytol. 1996; 25: 333-346
        • Hartig W.
        • Derouiche A.
        • Welt K.
        • Brauer K.
        • Grosche J.
        • Mader M.
        • Reichenbach A.
        • Bruckner G.
        Brain Res. 1999; 842: 15-29
        • Ruoslahti E.
        Glycobiology. 1996; 6: 489-492
        • Oohira A.
        • Matsui F.
        • Tokita Y.
        • Yamauchi S.
        • Aono S.
        Arch. Biochem. Biophys. 2000; 374: 24-34
        • Rauch U.
        Cell Mol. Life Sci. 2004; 61: 2031-2045
        • Bandtlow D.R.
        • Zimmermann C.E.
        Physiol. Rev. 2000; 80: 1267-1290
        • Maurel P.
        • Rauch U.
        • Flad M.
        • Margolis R.K.
        • Margolis R.U.
        Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2512-2516
        • Levine J.M.
        • Nishiyama A.
        Perspect. Dev. Neurobiol. 1996; 3: 245-259
        • Watanabe E.
        • Maeda N.
        • Matsui F.
        • Kushima Y.
        • Noda M.
        • Oohira A.
        J. Biol. Chem. 1995; 270: 26876-26882
        • Bignami A.
        • Asher R.
        • Perides G.
        Brain Res. 1992; 579: 173-177
        • Aspberg A.
        • Miura R.
        • Bourdoulous S.
        • Shimonaka M.
        • Heinegard D.
        • Schachner M.
        • Ruoslahti E.
        • Yamaguchi Y.
        Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 10116-10121
        • Sugahara K.
        • Yamada S.
        Trends Glycosci. Glycotechnol. 2000; 12: 321-349
        • Deepa S.S.
        • Umehara Y.
        • Higashiyama S.
        • Itoh N.
        • Sugahara K.
        J. Biol. Chem. 2002; 277: 43707-43716
        • Maeda N.
        • He J.
        • Yajima Y.
        • Mikami T.
        • Sugahara K.
        • Yabe T.
        J. Biol. Chem. 2003; 278: 35805-35811
        • Rhodes K.E.
        • Fawcett J.W.
        J. Anat. 2004; 204: 33-48
        • Asher R.A.
        • Morgenstern D.A.
        • Fidler P.S.
        • Adcock K.H.
        • Oohira A.
        • Braistead J.E.
        • Levine J.M.
        • Margolis R.U.
        • Rogers J.H.
        • Fawcett J.W.
        J. Neurosci. 2000; 20: 2427-2438
        • Asher R.A.
        • Morgenstern D.A.
        • Shearer M.C.
        • Adcock K.H.
        • Pesheva P.
        • Fawcett J.W.
        J. Neurosci. 2002; 22: 2225-2236
        • Thon N.
        • Haas C.A.
        • Rauch U.
        • Merten T.
        • Fassler R.
        • Frotscher M.
        • Deller T.
        Eur. J. Neurosci. 2000; 12: 2547-2558
        • Jones L.L.
        • Yamaguchi Y.
        • Stallcup W.B.
        • Tuszynski M.H.
        J. Neurosci. 2002; 22: 2792-2803
        • Moon L.D.
        • Asher R.A.
        • Rhodes K.E.
        • Fawcett J.W.
        Nat. Neurosci. 2001; 4: 465-466
        • Bradbury E.J.
        • Moon L.D.
        • Popat R.J.
        • King V.R.
        • Bennett G.S.
        • Patel P.N.
        • Fawcett J.W.
        • McMahon S.B.
        Nature. 2002; 416: 636-640
        • Corvetti L.
        • Rossi F.
        J. Neurosci. 2005; 25: 7150-7158
        • Tropea D.
        • Caleo M.
        • Maffei L.
        J. Neurosci. 2003; 23: 7034-7744
        • Brakebusch C.
        • Seidenbecher C.I.
        • Asztely F.
        • Rauch U.
        • Matthies H.
        • Meyer H.
        • Krug M.
        • Bockers T.M.
        • Zhou X.
        • Kreutz M.R.
        • Montag D.
        • Gundelfinger E.D.
        • Fassler R.
        Mol. Cell. Biol. 2002; 21: 7417-7427
        • Kinoshita A.
        • Sugahara K.
        Anal. Biochem. 1999; 269: 367-378
        • Probstmeier R.
        • Stichel C.C.
        • Muller H.W.
        • Asou H.
        • Pesheva P.
        J. Neurosci. Res. 2000; 60: 21-36
        • Yamagata T.
        • Kawamura Y.
        • Suzuki S.
        Biochim. Biophys. Acta. 1966; 115: 250-252
        • Ohya T.
        • Kaneko Y.
        Biochim. Biophys. Acta. 1970; 198: 607-609
        • Rauch U.
        • Gao P.
        • Janetzko A.
        • Flaccus A.
        • Hilgenberg L.
        • Tekotte H.
        • Margolis R.K.
        • Margolis R.U.
        J. Biol. Chem. 1991; 266: 14785-14801
        • Seidenbecher C.I.
        • Richter K.
        • Rauch U.
        • Fassler R.
        • Garner C.C.
        • Gundelfinger E.D.
        J. Biol. Chem. 1995; 270: 27206-27212
        • Perides G.
        • Rahemtulla F.
        • Lane W.S.
        • Asher R.A.
        • Bignami A.
        J. Biol. Chem. 1992; 267: 23883-23887
        • Grumet M.
        • Flaccus A.
        • Margolis R.U.
        J. Cell Biol. 1993; 120: 815-824
        • Friedlander D.R.
        • Milev P.
        • Karthikeyan L.
        • Margolis R.K.
        • Margolis R.U.
        • Grumet M.
        J. Cell Biol. 1994; 125: 669-680
        • Retzler C.
        • Gohring W.
        • Rauch U.
        J. Biol. Chem. 1996; 271: 27304-27310
        • Oleszewski M.
        • Beer S.
        • Katich S.
        • Geiger C.
        • Zeller Y.
        • Rauch U.
        • Altevogt P.
        J. Biol. Chem. 1999; 274: 24602-24610
        • Aquino D.A.
        • Margolis R.U.
        • Margolis R.K.
        J. Cell Biol. 1984; 99: 1117-1129
        • Aquino D.A.
        • Margolis R.U.
        • Margolis R.K.
        J. Cell Biol. 1984; 99: 1130-1139
        • Matsui F.
        • Nishizuka M.
        • Yasuda Y.
        • Aono S.
        • Watanabe E.
        • Oohira A.
        Brain Res. 1998; 790: 45-51
        • Seidenbecher C.I.
        • Smalla K.H.
        • Fischer N.
        • Gundelfinger E.D.
        • Kreutz M.R.
        J. Neurochem. 2002; 83: 738-746
        • Yamada H.
        • Fredette B.
        • Shitara K.
        • Hagihara K.
        • Miura R.
        • Ranscht B.
        • Stallcup W.B.
        • Yamaguchi Y.
        J. Neurosci. 1997; 17: 7784-7795
        • Engel M.
        • Maurel P.
        • Margolis R.U.
        • Margolis R.K.
        J. Comp. Neurol. 1996; 366: 34-43
        • Hayashi N.
        • Miyata S.
        • Yamada M.
        • Kamei K.
        • Oohira A.
        Neuroscience. 2005; 131: 331-348
        • Miyata S.
        • Nishimura Y.
        • Hayashi N.
        • Oohira A.
        Neuroscience. 2005; 136: 95-104
        • Milev P.
        • Friedlander D.R.
        • Sakurai T.
        • Karthikeyan L.
        • Flad M.
        • Margolis R.K.
        • Grumet M.
        • Margolis R.U.
        J. Cell Biol. 1994; 127: 1703-1715
        • Milev P.
        • Maurel P.
        • Haring M.
        • Margolis R.K.
        • Margolis R.U.
        J. Biol. Chem. 1996; 271: 15716-15723
        • Falk J.
        • Bonnon C.
        • Girault J.A.
        • Faivre-Sarrailh C.
        Biol. Cell. 2002; 94: 327-334
        • Kalb R.G.
        • Hockfield S.
        J. Neurosci. 1988; 8: 2350-2360
        • Beggah A.T.
        • Dours-Zimmermann M.T.
        • Barras F.M.
        • Brosius A.
        • Zimmermann D.R.
        • Zurn A.D.
        Neuroscience. 2005; 133: 749-762
        • Bignami A.
        • Asher R.
        • Perides G.
        Exp. Neurol. 1992; 117: 90-93
        • Schmalfeldt M.
        • Bandtlow C.E.
        • Dours-Zimmermann M.T.
        • Winterhalter K.H.
        • Zimmermann D.R.
        J. Cell Sci. 2000; 113: 807-816
        • Milev P.
        • Maurel P.
        • Chiba A.
        • Mevissen M.
        • Popp S.
        • Yamaguchi Y.
        • Margolis R.K.
        • Margolis R.U.
        Biochem. Biophys. Res. Commun. 1998; 247: 207-212
        • Wu Y.
        • Sheng W.
        • Chen L.
        • Dong H.
        • Lee V.
        • Lu F.
        • Wong C.S.
        • Lu W.Y.
        • Yang B.B.
        Mol. Biol. Cell. 2004; 15: 2093-2104
        • LeBaron R.G.
        • Zimmermann D.R.
        • Ruoslahti E.
        J. Biol. Chem. 1992; 267: 10003-10010
        • Rauch U.
        • Hirakawa S.
        • Oohashi T.
        • Kappler J.
        • Roos G.
        Matrix Biol. 2004; 22: 629-639
        • Seyfried N.T.
        • McVey G.F.
        • Almond A.
        • Mahoney D.J.
        • Dudhia J.
        • Day A.J.
        J. Biol. Chem. 2005; 280: 5435-5448
        • Day A.J.
        • Prestwich G.D.
        J. Biol. Chem. 2002; 277: 4585-4588
        • Toole B.P.
        Nat. Rev. Cancer. 2004; 4: 528-539
        • Koppe G.
        • Bruckner G.
        • Brauer K.
        • Hartig W.
        • Bigl V.
        Cell Tissue Res. 1997; 288: 33-41
        • Woodworth A.
        • Pesheva P.
        • Fiete D.
        • Baenziger J.U.
        J. Biol. Chem. 2004; 279: 10413-10421
        • Smith-Thomas L.C.
        • Stevens J.
        • Fok-Seang J.
        • Faissner A.
        • Rogers J.H.
        • Fawcett J.W.
        J. Cell Sci. 1995; 108: 1307-1315
        • Snow D.M.
        • Lemmon V.
        • Carrino D.A.
        • Caplan A.I.
        • Silver J.
        Exp. Neurol. 1990; 109: 111-130
        • Gilbert R.J.
        • McKeon R.J.
        • Darr A.
        • Calabro A.
        • Hascall V.C.
        • Bellamkonda R.V.
        Mol. Cell. Neurosci. 2005; 29: 545-558
        • Properzi F.
        • Carulli D.
        • Asher R.A.
        • Muir E.
        • Camargo L.M.
        • van Kuppevelt T.H.
        • ten Dam G.B.
        • Furukawa Y.
        • Mikami T.
        • Sugahara K.
        • Toida T.
        • Geller H.M.
        • Fawcett J.W.
        Eur. J. Neurosci. 2005; 21: 378-390
        • Clement A.M.
        • Nadanaka S.
        • Masayama K.
        • Mandl C.
        • Sugahara K.
        • Faissner A.
        J. Biol. Chem. 1998; 273: 28444-28453
        • Clement A.M.
        • Sugahara K.
        • Faissner A.
        Neurosci. Lett. 1999; 269: 125-128
        • Kantor D.B.
        • Chivatakarn O.
        • Peer K.L.
        • Oster S.F.
        • Inatani M.
        • Hansen M.J.
        • Flanagan J.G.
        • Yamaguchi Y.
        • Sretavan D.W.
        • Giger R.J.
        • Kolodkin A.L.
        Neuron. 2004; 44: 961-975
        • De Wit J.
        • De Winter F.
        • Klooster J.
        • Verhaagen J.
        Mol. Cell. Neurosci. 2005; 29: 40-55
        • Kreuger J.
        • Jemth P.
        • Sanders-Lindberg E.
        • Eliahu L.
        • Ron D.
        • Basilico C.
        • Salmivirta M.
        • Lindahl U.
        Biochem. J. 2005; 389: 145-150
        • Hu H.
        Nat. Neurosci. 2001; 4: 695-701
        • Kappler J.
        • Franken S.
        • Junghans U.
        • Hoffmann R.
        • Linke T.
        • Muller H.W.
        • Koch K.W.
        Eur. J. Neurosci. 1997; 9: 306-318
        • Stallcup W.B.
        • Beasley L.
        • Levine J.
        Cold Spring Harb. Symp. Quant. Biol. 1983; 48: 761-774