A Functional Dermatan Sulfate Epitope Containing Iduronate(2-O-sulfate)α1–3GalNAc(6-O-sulfate) Disaccharide in the Mouse Brain

Oversulfated chondroitin sulfate (CS), dermatan sulfate (DS), and CS/DS hybrid structures bind growth factors, promote the neurite outgrowth of hippocampal neurons in vitro, and have been implicated in the development of the brain. To investigate the expression of functional oversulfated DS structures in the brain, a novel monoclonal antibody (mAb), 2A12, was generated against DS (An-DS) from ascidian Ascidia nigra, which contains a unique iD disaccharide unit, iduronic acid (2-O-sulfate)α1→3GalNAc(6-O-sulfate), as a predominant disaccharide. mAb 2A12 specifically reacted with the immunogen, and recognized iD-enriched decasaccharides as minimal structures. The 2A12 epitope was specifically observed in the hippocampus and cerebellum of the mouse brain on postnatal day 7, and the expression in the cerebellum disappeared in the adult brain, suggesting a spatiotemporally regulated expression of this epitope. Embryonic hippocampal neurons were immunopositive for 2A12, and the addition of the antibody to the culture medium significantly reduced the neurite growth of hippocampal neurons. In addition, two minimum 2A12-reactive decasaccharide sequences with multiple consecutive iD units were isolated from the An-DS chains, which exhibited stronger inhibitory activity against the binding of various growth factors and neurotrophic factors to immobilized embryonic pig brain CS/DS chains (E-CS/DS) than the intact E-CS/DS, suggesting that the 2A12 epitope at the neuronal surface acts as a receptor or co-receptor for these molecules. Thus, we have selected a unique antibody that recognizes iD-enriched oversulfated DS structures, which are implicated in the development of the hippocampus and cerebellum in the central nervous system. The antibody will also be applicable for investigating structural alterations in CS/DS in aging and pathological conditions.

implicated in the development of the hippocampus and cerebellum in the central nervous system. The antibody will also be applicable for investigating structural alterations in CS/DS in aging and pathological conditions.
Chondroitin sulfate (CS) 1 and dermatan sulfate (DS), as well as heparan sulfate (HS), are glycosaminoglycans, which are synthesized as carbohydrate side chains covalently attached to a core protein of proteolgycan (PG) (for reviews, see Ref. [1][2][3]. CS/DS-PGs are present at cell surfaces and in the extracellular matrices of most tissues, and are significant components in the mammalian brain, where they participate in neural development by regulating neuronal adhesion and migration, neurite formation, and axonal guidance (for reviews, see Ref. 4 -8). The backbone of CS and DS consists of repeating disaccharide units of -GlcUA-GalNAc-and -IdoUA-GalNAc-, respectively, and hybrid chains composed of both units in varying proportions also exist (9). These disaccharide units are modified during chain elongation by specific sulfotransferases at C-2 of GlcUA/IdoUA and/or C-4 and/or C-6 of GalNAc in various combinations, thereby producing characteristic sulfation patterns and enormous structural diversity for CS, DS, and CS/DS hybrid chains. Recent studies have shown that the GlcUA/IdoUA ratio and sulfation pattern of the brain CS/DS change during development (10 -12), and subpopulations of the brain CS and DS chains with distinct structures play different roles in neuritogenesis (13)(14)(15). In addition, oversulfated CS and DS chains from various marine organisms promoted neurite outgrowth in murine hippocampal neurons in vitro (16 -19).
Brain CS/DS chains contain small yet significant amounts of oversulfated disaccharides (11,14,20). For example, embryonic * This work was supported in part by HAITEKU (2004HAITEKU ( -2008 from the Japan Private School Promotion Foundation, and Grants-in-aid for Exploratory Research 15659021 and Scientific Research-B 16390026 from MEXT. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. We recently demonstrated that these rare oversulfated disaccharides, in addition to IdoUA-containing disaccharides, are critical structural elements for the neuritogenic activity of E-CS/DS (11,15). Meanwhile, Tsuchida et al. (21) showed that the CS chains of appican, which is a CS-bearing form of amyloid precursor protein, produced by cultured glioma cells contain 14.3% E units, suggesting that some brain PGs may carry a cluster of oversulfated CS disaccharides. However, the presence of oversulfated DS structures in the mammalian brain has not been rigorously characterized or explored.
The DS chains from the body of ascidian Ascidia nigra (An-DS) consist of a major (ϳ80%) iD unit (IdoUA(2S)-GalNAc(6S)) in addition to a minor (ϳ20%) iC unit (IdoUA-GalNAc(6S)), where 2S represents 2-O-sulfate (22). The iD unit-containing structure is unique, and has not been detected in mammalian tissues. However, An-DS binds various growth factors including hepatocyte growth factor (23), fibroblast growth factor 2 (FGF2), pleiotrophin (PTN), midkine (MK), and heparin-binding epidermal growth factor, 2 and promotes neurite outgrowth toward hippocampal neurons in vitro (18), implying biological functions for the iD-containing DS structures. CS/DS hybrid chains isolated from shark skin also exhibited marked binding activity toward various growth factors including those described above and also strong neuritogenic activity, and contain small yet significant proportions of various disulfated disaccharide units including D or iD units (19).
To study the structure-function relationship of CS/DS chains and possible biological functions of iD-containing structures during the development of the brain, we generated a new monoclonal antibody (mAb) 2A12 against An-DS in the present study. Immunohistochemistry and immunocytochemistry using mAb 2A12 were performed, and two minimal decasaccharides recognized by this antibody were isolated. Interestingly, mAb 2A12 inhibited neurite growth, and the two decasaccharides inhibited the binding of several growth factors and neurotrophic factors to E-CS/DS. These results indicate that iDcontaining DS structures are present in the hippocampus and cerebellum of the mammalian brain, and appear to play specific roles in the development of the central nervous system.
Preparation of An-DS-The isolation of GAGs from A. nigra was carried out as described by Pavão et al. (22) with some modifications. Briefly, the lyophilized body (100 g) of A. nigra was treated with actinase E, and sulfated GAGs were recovered with a yield of 0.5% by precipitation with cetylpyridinium chloride. The GAG mixture was treated with nitrous acid at pH 1.5 to remove HS, and the resistant polymers were subjected to anion-exchange chromatography with an Acell QMA Plus cartridge, which was pre-equilibrated with 0.5 M NaClcontaining 0.3 M phosphate buffer, pH 6.0, using a stepwise elution with phosphate buffers containing 0.5, 1.0, and 1.5 M NaCl. A portion (60 mg) of the 1.5 M NaCl-eluted fraction (144 mg) was re-chromatographed under the same conditions, yielding 50.4 mg in the corresponding fraction. This sample contained 2.4% peptides but no detectable amounts of HS or hyaluronate as evaluated by high performance liquid chromatography (HPLC) analysis, and was used for immunization of the mice. For the screening of antibodies and structural and functional analyses, the DS-peptide fraction was further purified by C-18 hydrophobic chromatography to remove free peptides. The water-eluted DS-peptide fraction from the C-18 column was designated An-DS, which was completely digested with chondroitinase ABC, confirming the purity.
Cellulose Acetate Membrane Electrophoresis-An-DS (2 g each) was incubated with chondroitinase ABC (10 mIU), AC-I (4 mIU), or B (4 mIU) in an appropriate buffer at 37°C for 1 h. Each digest was spotted on a cellulose acetate membrane, and subjected to electrophoresis at a constant voltage of 100 V for 30 min. The membrane was thereafter stained with Alcian blue (0.1% in 0.1% acetic acid) for 10 min and then destained with 0.1% acetic acid for 20 min.
Enzyme-linked Immunosorbent Assay (ELISA)-If not specified, all the steps were performed at room temperature. For antibody screening and the evaluation of substrate specificity, streptavidin-coated plates were used. The reactivity of antibodies with various GAG subtypes was tested using ELISA as described previously (15). Briefly, various GAG preparations were individually biotinylated and immobilized on the plates. Wells were blocked with 1% bovine serum albumin (BSA), and incubated with hybridoma supernatants or ascitic fluid containing antibodies followed by alkaline phosphatase-linked goat anti-mouse Ig(GϩM) or IgM. The secondary antibody was detected using p-nitrophenyl phosphate, and the absorbance was measured at 415 nm. Negative controls received no An-DS and/or primary antibody. When mAb 473HD and commercial anti-CS mAbs, CS-56 and MO-556, were used as primary antibodies, a 50-fold dilution was employed.
For the inhibitory ELISA, a certain amount (0.5 g) of GAGs or the oligosaccharide fractions generated by digestion with chondroitinase ABC (see below) was incubated with mAb 2A12 (as ascitic fluid, diluted 400-fold) in a volume of 50 l at room temperature for 1 h before being applied to the plate. The inhibition was calculated from the reduced absorbance relative to that obtained from a control incubation without GAG.
To characterize the 2A12 epitopes in An-DS, which contains a disaccharide repeating region and an oligosaccharide-peptide linkage region, the purified An-DS (12 g) was incubated with chondroitinase ABC (10 mIU), chondroitinase B (4 mIU), or a mixture of chondroitinases AC-I (4 mIU) and AC-II (4 mIU) in their appropriate buffers at 37°C for 2 h. Incubation with heat-inactivated chondroitinase ABC was run as a control. The digests were added to a Maxisorp plate in a 0.1 M sodium bicarbonate buffer, pH 9.2. After incubation at 4°C for 16 h, the plate was washed with phosphate-buffered saline (PBS), followed by the steps described above. Absorbance was recorded 2 h after the addition of p-nitrophenyl phosphate.
Preparation and Selection of Antibodies-Monoclonal antibodies were raised by immunizing BALB/cA Jol mice using the DS-peptide preparation obtained from A. nigra. Briefly, the immunogen was injected subcutaneously into the back of mice at a dose of 200 g/injection every 2 weeks, and after the third injection the serum was screened against highly purified An-DS using ELISA. Spleen B lymphocytes of the positive mice after the fifth injection were isolated and fused with myeloma cells. The fused hybridoma cells were cultured, and the culture supernatant was screened by ELISA using biotinylated An-DS as 2 X. Bao and K. Sugahara, unpublished results.
antigen. Four positive clones (2A12, 3G11, 4B5, and 5F4) were selected. Each clone was injected intraperitoneally into mice to obtain ascitic fluid, which was used for evaluating substrate specificity. Among the antibodies selected, mAb 2A12 is of particular interest because of its high specificity and unique staining pattern in the mouse brain (see below).
An aliquot of 2A12 was purified using a mouse IgM purification kit according to the manufacturer's directions. The purified antibody was quantified with a BCA protein assay kit and used for assaying the inhibition of the neurite growth of cultured embryonic hippocampal neurons (see below).
Immunohistochemistry-P7 and adult ddY mice were anesthesized, and the brains were dissected, immediately frozen with dry ice, and kept at Ϫ80°C. Within 2-4 days of the dissection, the frozen brains were cut into sections 12 m thick, dehydrated by heating at 60°C for 1 h, and stored at Ϫ80°C for use. For immunostaining with the antibodies, the brain sections were fixed with acetone/methanol (1:1) and rehydrated with distilled water. The sections were then treated sequentially with the following solutions: 1) 2.5% hydrogen peroxide in PBS (10 mM, pH, 7.4) for 30 min; 2) 1% BSA, 4% normal goat serum in PBS for 60 min; 3) primary antibody in 1% BSA/PBS (diluted 200-fold for 2A12 and 100-fold for CS-56) at 4°C overnight; 4) biotinylated antimouse IgM ( chain) (8 ng/ml) in 1% BSA/PBS for 60 min; 5) Vectastain ABC solution in PBS (200-fold) for 60 min; and 6) 0.06% diaminobenzidine, 0.01% hydrogen peroxide in Tris-buffered saline (20 mM, pH, 7.6). Finally, sections were fixed with a series of ethanol solutions and mounted with a xylene-based mounting medium. As a control, mouse IgM was used as a primary antibody. To confirm the specificity of the staining with the antibodies, brain sections were pretreated with chondroitinase ABC protease-free (2 mIU/section) to remove CS and DS, and then processed for immunostaining as described above.
Cell Culture-Hippocampal cells were cultured using embryonic day 15.5 (E15.5) or 16.5 (E16.5) mouse brains as described previously (13,18). The hippocampi were obtained by microdissection and dissociated with a brief trypsin treatment. Dissociated cells were suspended in Eagle's modified essential medium containing an N2 supplement, seeded on P-ORN-coated coverslips, and cultured at 37°C in a humidified atmosphere containing 5% CO 2 .
Immunocytochemistry-E15.5 mouse hippocampal cells were cultured for 48 h, washed once with PBS, and then fixed with 4% paraformaldehyde in PBS for 30 min. Fixed cells were rinsed with PBS three times and blocked with 1% BSA, 4% goat normal serum in PBS. Cells were then incubated with mAb 2A12 (200-fold) in 1% BSA/PBS for 2 h, washed with PBS three times, permeated with 0.2% Triton X-100 in PBS for 30 min, and incubated with anti-MAP2 (200-fold) and anti-NF (200-fold) in 1% BSA/PBS at 4°C overnight. After three washes with PBS, the cells were incubated with Alexa 568-conjugated goat anti-mouse IgG (200-fold) and Alexa 488-conjugated goat anti-mouse IgM (200-fold) in 1% BSA/PBS for 1 h. The cells were washed with PBS three times, mounted, and observed using an Olympus confocal microscope. The procedure was carried out at room temperature if not specified otherwise.
Neurite Growth Assay-Neurite growth of hippocampal neurons from E16.5 mouse brains was assayed as described (13,18) with some modifications. Dissociated hippocampal cells were seeded on coverslips precoated with P-ORN at 10,000, 25,000, or 50,000 cells/cm 2 , and cultured for 2 h. After the pre-culture, the purified mAb 2A12 (10 -200 g/ml) was added to the culture medium. Mouse IgM was run as a control.
After a 60-h culture, cells were fixed with 4% paraformaldehyde and then immunostained with anti-MAP2 and anti-NF, and detection was carried out with a Vectastain ABC kit using diaminobenzidine as a chromagen as described above. Neurite growth was evaluated by determining the percentage of neurons bearing at least one neurite and the total length of neurites per cell. Neurites longer than the cell body were chosen for counting. About 300 isolated cells were randomly chosen for calculating the percentage, and 100 cells for measuring neurite length, per coverslip. This assay was performed three times in triplicate, and the results were expressed as the mean Ϯ S.E. The significance of the difference between means was evaluated with the Student's t test.
Delayed Extraction Matrix-assisted Laser Desorption Ionization Time-of-flight Mass Spectrometry (MALDI-TOF MS)-Dried oligosaccharides (5-10 pmol of purified decasaccharides, and 30 pmol of an oligosaccharide mixture) was first mixed with 1-2 l of (Arg-Gly) 15 (10 pmol) and then 1 l of gentisic acid (1 mg/ml) (24). Each mixture was spotted on a plate for MS analysis using a mixture of (Arg-Gly) 15 and gentisic acid as a control. The analysis was run in a positive mode using the HCD1001 method according to the manufacturer's instructions. The MS spectra were recorded on a Voyager DE-RP-Pro (PerSeptive Biosystems, Framingham, MA) using the linear mode.
Fragmentation of An-DS-An-DS (2 mg) was incubated with protease-free chondroitinase ABC (25 mIU) in 50 mM Tris-HCl buffer, pH 8.0, containing 60 mM sodium acetate in a volume of 100 l at 37°C for 32 min (25). After boiling at 100°C for 1 min, the digests were filtered and subjected to gel filtration on a column of Superdex Peptide (10 ϫ 300 mm) using 0.2 M NH 4 HCO 3 as an eluent at a flow rate of 0.3 ml/min. Each peak was collected, re-chromatographed under the same conditions, desalted by repeated evaporation, and quantified by the carbozole reaction (26). Size determination of each fraction was achieved by MS analysis as described above.
Fractionation of Decasaccharides and Analysis of Disaccharide Composition by Anion-exchange Chromatography-Separation of the decasaccharide fraction and the analysis of disaccharide composition were carried out by HPLC with an amine-bound silica PA-03 column (YMC Co., Ltd., Kyoto, Japan) using a linear gradient of NaH 2 PO 4 at a flow rate of 1 ml/min as described for the separation of CS-D octasaccharides (16) and 2-aminobenzide (2AB)-derivatized unsaturated CS disaccharides (27), respectively. The identification and quantification of the unsaturated disaccharides generated by bacterial chondroitinases were carried out by making comparisons with authentic unsaturated disaccharides in the CS and DS chains: Enzymatic Treatments-For the analysis of disaccharide composition, poly-or oligosaccharides (100 -500 pmol as disaccharides) were incubated with 5 mIU of chondroitinase ABC at 37°C for 2 h. The digests were derivatized with 2AB, and excess 2AB was removed by extraction with CHCl 3 (28). For examination of the digestibility of the decasaccharides isolated from An-DS (An-DS 10-a and 10-b) with various chondroitinases, each decasaccharide (50 pmol) was incubated with chondroitinase ABC (1 mIU), chondroitinase B (1 mIU), or a mixture of chondroitinases AC-I (0.5 mIU) and AC-II (0.5 mIU) at 37°C for 1 h, respectively. To identify tetrasaccharide structures at the reducing ends of the purified decasaccharides, each 2AB-labeled decasaccharide (ϳ10 pmol) was incubated with chondroitinase ABC (1 mIU). Half of the digest was analyzed by anion-exchange HPLC on a PA-03 column, and the other half was further incubated with ⌬hexuronate-2sulfatase (4 IU) in 20 mM sodium buffer, pH 6.5, containing 0.15% BSA at 37°C for 30 min in a volume of 30 l (16, 29). All the enzymatic reactions were terminated by heating at 100°C for 1 min. Each digest was subjected to anion-exchange HPLC analysis with or without 2AB labeling as described above.
Interaction Analysis-The inhibitory activity of An-DS 10-a and 10-b decasaccharides against the binding of growth factors and neurotrophic factors to immobilized E-CS/DS was examined using a BIAcore system (BIAcore AB, Uppsala, Sweden) (15). An E-CS/DS-immobilized sensor chip was prepared as reported previously (11). Various growth factors or neurotrophic factors (100 ng/per time) were mixed with polysaccharides or oligosaccharides (1.5 g/ml) and co-injected onto the surface of the sensor chip. The reaction continued for 2 min (the association phase), and the sensor was washed for at least 2 min (the dissociation phase). Response curves were recorded, and the inhibitory efficiency was expressed as the percentage relative to the response obtained without mixing with sugar chains.

Isolation and Structural Characterization of Ascidian
A. nigra DS-An-DS was isolated and purified from the bodies of A. nigra (see "Experimental Procedures"), with a yield of 0.12% from the dried tissue. The average molecular mass of this preparation was 6.3 ϫ 10 4 as evaluated by gel filtration (data not shown) according to a reported method (30). An analysis of disaccharide composition showed ⌬D disaccharide to be the predominant unit (76.6 mol%) with ⌬O, ⌬C, and ⌬T units (Table I) as minor components, which is consistent with a report by Pavão et al. (22). Interestingly, although An-DS is composed of almost exclusively, if not completely, IdoUA-containing disaccharides, it was totally resistant not only to chondroitinase AC-I but also to chondroitinase B (Fig. 1).
Selection and Evaluation of the Specificity of Anti-An-DS mAb 2A12-Monoclonal Abs were raised against An-DS in mice, and four mAbs were selected, among which 2A12 is of particular interest because of its high specificity (see below). In this study, only mAb 2A12 was characterized. The reactivity of 2A12 toward various GAG species was analyzed using ELISA ( Fig. 2A), in which biotinylated GAGs were individually immobilized onto a streptavidin-coated plate. 2A12 specifically reacted with An-DS, but not any other GAGs tested including CS-A, CS-B, CS-C, CS-D, CS-E, CS-H,and Hr-DS (DS from ascidian Halocynthia roretzi, which contains iB as a predominant disaccharide unit), 2 or heparin. It should be emphasized that 2A12 discriminated iD (IdoUA(2S)-GalNAc(6S))-containing An-DS from other oversulfated CS and DS including CS-D, which is rich (20 -21%) in D units (GlcUA(2S)-GalNAc(6S)) (16,18). This specificity was confirmed by inhibitory ELISA, in which only soluble An-DS significantly inhibited the binding of 2A12 to immobilized An-DS (data not shown). To investigate whether 2A12 recognizes the disaccharide repeating region or the oligosaccharide-peptide linkage region of the intact An-DS, An-DS was incubated with various chondroitinases individually, and subjected to ELISA using a Nunc Maxisorp plate, which has a greater ability to adsorb the linkage region than the disaccharide repeating region. 3 As shown in Fig. 2B, treatment of An-DS with chondroitinase ABC completely abolished the reactivity with 2A12, suggesting that 2A12 recognized a structure embedded in the disaccharide repeating region rather than the linkage region of An-DS. Incubation with chondroitinase AC-I, AC-II, or B had no effect on the reactivity of 2A12 with An-DS, supporting the notion that An-DS was resistant to these enzymes (Fig. 1). The mAb 2A12 was identified as IgM by ELISA, in which only anti-mouse IgM, but not anti-mouse IgG or IgA, bound to 2A12 (data not shown).
The reactivity of mAb 2A12 with E-CS/DS was also examined. Compared with mAbs 473HD (13), CS-56 (31), and MO-225 (32), all of which recognize A-D-tetrasaccharide-containing structures (33), 2A12 exhibited much less yet significant reactivity toward E-CS/DS (Fig. 2C), which is probably because of the low abundance of the unique 2A12 epitope in the E-CS/DS chains (see below). mAb 2A12 also reacted with DS isolated from shark skin (data not shown), the D or iD content of which was 3% (19). These results suggest that the 2A12 epitope may be present in certain animal tissues besides the body of the ascidian.
Immunohistochemical Detection and Localization of the 2A12 Epitope in the Mouse Brain-To investigate the expression of the unique 2A12 epitope during the development of the brain, immunohistochemical staining of mouse brain sections from postnatal day 7 (P7) and adult mice was performed using 2A12. The staining pattern of the sagittal sections from the brains of P7 and adult mice are shown in Fig. 3. Similar to the results reported by Maeda et al. (12), the CS-56 epitope was highly and ubiquitously expressed in the P7 brain (Fig. 3, panels B), and its expression decreased in the adult brain (Fig. 3, panels F). In contrast, 2A12 specifically stained the P7 hippocampus and cerebellum (Fig. 3, panels A), and the adult hippocampus (Fig.  3, panels E) as well. In the P7 cerebellum, 2A12 stained the granular cell layer and white matter; and in the hippocampus, the 2A12 epitope was expressed in the granular cell layer of the dentate gyrus and pyramidal cell layer. The staining completely disappeared on a treatment of the tissue sections from the P7 mice with chondroitinase ABC (Fig. 3, panels C), confirming the specificity of the staining. Interestingly, chondroitinase B also largely eliminated the staining in the P7 brain (Fig.  3, panels D), which implied that the 2A12 epitope is embedded in DS domains of DS or CS/DS hybrid chains. Using enzymatic treatments, similar results were also obtained from the adult brain (data not shown). These results suggest that the expression of the 2A12 epitope changes spatiotemporally in the mouse brain.
Detection of the 2A12 Epitope at the Surface of Hippocampal Neurons-To further verify the expression of 2A12 in the hippocampus, dissociated hippocampal neurons from an embryonic mouse were doubly stained with antibodies against the neuronal markers (anti-MAP2 and anti-NF) and 2A12. As shown in Fig. 4, cultured hippocampal neurons were stained positive with 2A12. The cell body and most neurites showed immunoreactivity to 2A12. Notably, expression of the 2A12 epitope is stronger in the neurite-sprouting regions than other regions of the cell body.
Effects of 2A12 on the Neurite Growth of Hippocampal Neurons-Hippocampal neurons formed multiple neurites after a 24-h culture on a substratum containing E-CS/DS and P-ORN at a density of 10,000 cells/cm 2 (15). In the absence of E-CS/DS chains in the substratum, a higher cell density (2.5-5-fold) and a longer culture period (2-3-fold) are required for the neurons to form elaborate neurites (Fig. 5A) compared with cells grown on a substratum containing E-CS/DS chains. To investigate 3 K. Kalayamamitra and K. Sugahara, unpublished observation.  whether the 2A12 epitope at the hippocampal neuronal surface is involved in the process of neurite formation and growth, 2A12 was added to the culture medium. 2A12 significantly suppressed neurite growth of the neurons grown on a substratum containing P-ORN only (Fig. 5B). This observation was confirmed by a statistical morphometric analysis, which revealed that 2A12 markedly inhibited the formation of neurites (Fig. 5C) and decreased the total length of neurites per cell (Fig. 5D) at concentrations over 50 g/ml. Control IgM showed no such activity even at a concentration of 200 g/ml (Fig. 5, C  and D). In contrast, squid cartilage CS-E-induced neurite outgrowth, which is mediated through a yet unidentified mechanism but not mediated by PTN unlike E-CS/DS-induced neurite outgrowth (15), was not influenced by 2A12 at a concentration of 100 g/ml (data not shown), indicating that the inhibitory activity of the antibody was specific and was not a cytotoxic effect.
Characterization of the Minimal Structure Required for the Recognition by 2A12-In view of the findings that the 2A12 epitope is expressed by hippocampal neurons and may be involved in the formation and growth of neurites, we characterized the minimal structure of An-DS required for the recogni- A, the reactivity of mAb 2A12 with various GAG species was examined using ELISA, where authentic GAGs and Hr-DS, the DS preparation from the ascidian H. roretzi, which contains iB as a predominant unit, 2 were included. Biotinylated GAGs (2 g each) were individually immobilized to wells of a streptavidin-coated plastic plate, and processed for incubation with as a primary antibody, mAb 2A12 (diluted 400-fold), followed by incubation with alkaline phosphataselinked goat anti-mouse Ig(G ϩ M) (5,000-fold). The secondary antibody was detected using p-nitrophenyl phosphate as a substrate. This assay was performed three times, and the results are shown as the mean Ϯ S.D. B, the effects of treatments with various chondroitinases (CSases) on the reactivity of An-DS with mAb 2A12 were evaluated. An-DS (12 g) was incubated with CSase ABC (10 mIU), CSase B (4 mIU), or a mixture of CSases AC-I (4 mIU) and AC-II (4 mIU) at 37°C for 2 h, and then each digest was added to the well of a Nunc Maxisorp plate in a 0.1 M sodium bicarbonate buffer, pH 9.2. After incubation at 4°C for 16 h, wells were washed with PBS and processed for ELISA as described above. In the negative control, An-DS was omitted. The positive control received An-DS and inactivated chondroitinase ABC. C, the reactivity of the embryonic pig brain-derived CS/DS chains (E-CS/DS) with mAb 2A12 and with three other anti-CS antibodies, CS-56, MO-225, and 473HD, was compared using ELISA as described as above. The negative control did not receive mAb 2A12. Note that mAb 2A12 exhibited much weaker yet significant reactivity with E-CS/DS than the other three anti-CS antibodies. *, 0.01 Ͻ p Ͻ 0.05, significant difference from the control. Elimination of the epitope by these enzymes was also observed in the adult tissues (data not shown). Solid arrows indicate the regions that were stained by mAb 2A12. Py, pyramidal cell layer; GrDG, granular cell layer of the dentate gyrus; WM, white matter; GL, granular cell layer. Note that the border (open arrows) between the CA1 region and the dentate gyrus was stained with CS-56, but not with 2A12. tion by this antibody. Intact An-DS was partially degraded by chondroitinase ABC treatment, and the digests were fractionated by gel filtration. The effluent fractions were collected as indicated in Fig. 6A, and the molecular mass of each fraction was determined by MALDI-TOF MS analysis. To evaluate which fraction contained the minimal structure required for recognition by 2A12, an equal amount of each fraction was tested as an inhibitor against the reactivity of 2A12 with immobilized An-DS in an ELISA. Fig. 6B shows that the decasaccharides and larger saccharides exhibited inhibitory activity, and the activity increased with molecular size, suggesting that the decasaccharide fraction derived from An-DS contained the minimal structure needed for recognition by 2A12.
Next, the decasaccharide fraction was separated by anionexchange HPLC, and two main oligosaccharides (An-DS 10-a and 10-b) were isolated (Fig. 6C). The molecular masses of An-DS 10-a and 10-b were 2613.0 and 2697.0 (Table II), respectively, indicating that the former is a decasaccharide with nine sulfate groups and the latter a decasaccharide with 10 sulfate groups. These two components exhibited comparable inhibitory effects on the reactivity of 2A12 to An-DS (Fig. 6D), suggesting that they are both recognized by the antibody. To precisely determine the sequences of both oligosaccharides, enzymatic treatment and an analysis of disaccharide composition were conducted. The results are summarized in Table II. Both were sensitive to chondroitinase ABC, but not to chondroitinase AC-I, AC-II, or B. Treatment of An-DS 10-b with chondroitinase ABC generated only ⌬D, suggesting that the decasaccharide has the structure ⌬HexUA(2S)␣133GalNAc(6S)␤13 4IdoUA(2S)␣133GalNAc(6S)␤134IdoUA(2S)␣133GalNAc-(6S)␤134IdoUA(2S)␣133GalNAc(6S)␤134IdoUA(2S)␣13 3GalNAc(6S) (⌬D-iD-iD-iD-iD). Treatment of An-DS 10-a with chondroitinase ABC generated ⌬C and ⌬D units in a molar ratio of 1:3.9. To locate the iC unit in the An-DS 10-a sequence, the reducing end of the oligosaccharide was labeled with 2AB, and the labeled oligosaccharide was treated with chondroitinase ABC and then with ⌬hexuronate-2-sulfatase. Chondroitinase ABC generates a 2AB-attached unsaturated tetrasaccharide irrespective of the structure of the parent oligosaccharide (27), and ⌬hexuronate-2-sulfatase removes a sulfate group only from the C-2 position of a ⌬HexUA located at the non-reducing terminus (29). As shown in Fig. 7, ⌬hexuronate-2-sulfatase treatment caused the position where the 2AB-labeled unsaturated tetrasaccharide generated by digestion with chondroitinase ABC of the labeled An-DS 10-a was eluted to shift, suggesting that a sulfate group was removed from the non-reducing end of the tetrasaccharide. Such a shift was also observed for the 2AB-labeled unsaturated tetrasaccharide derived from An-DS 10-b (data not shown). However, ⌬hexuronate-2-sulfatase had no effect on the sulfate group at the C-2 position of the ⌬HexUA residue located at the reducing end of a 2AB-labeled unsaturated tetrasaccharide, such as ⌬A-D-2AB used as a negative control (data not shown). Taken together, it was concluded that An-DS 10-a is composed of four consecutive iD units with an iC unit at the reducing end: These results suggested that mAb 2A12 recognizes iD-enriched decasaccharides.

Effects of the 2A12 Epitope on the Binding of Growth Factors and Neurotrophic Factors with E-CS/DS-Growth factors and
neurotrophic factors play important roles as regulators of cell fate and the formation of neurites (34,35). Our previous results have demonstrated that PTN and a subpopulation of E-CS/DS chains co-operate to promote the outgrowth of neurite from hippocampal neurons in vitro (15). Oversulfated disaccharides and IdoUA-containing disaccharides of the E-CS/DS chains are required for binding PTN and inducing the outgrowth of neurites. In the present study, to understand the mechanism through which 2A12 inhibited the neurite growth, An-DS 10-a and 10-b were used as inhibitors of the interaction of various growth factors and neurotrophic factors with immobilized E-CS/DS using a BIAcore system. Fig. 8 shows that, besides growth factors, two neurotrophic factors, GDNF (a member of the transforming growth factor-␤ superfamily) and BDNF (a member of the neurotrophin family), bound to E-CS/DS. In all the cases tested, An-DS 10-a and 10-b exhibited inhibitory activity that was comparable with or stronger than that of the intact E-CS/DS chains (Fig. 8). Interestingly, although An-DS 10-a and 10-b differ in the structure of the disaccharide unit at their reducing end only, the latter showed 5-and 3-fold stronger inhibition than the former against the binding of PTN or MK with E-CS/DS, respectively, suggesting an important role for the reducing end in the binding of oligosaccharides to these two particular growth factors. These results suggest that mAb 2A12 may directly mask the functional binding sites for various growth factors and neurotrophic factors, thereby inhibiting the signaling of these regulators.
In addition, the K d values of the binding of BDNF and GDNF with E-CS/DS were estimated to be 102.1 and 28.9 nM, respectively, which were comparable with those for the interaction of BDNF (K d ϭ 254 nM) and GDNF (K d ϭ 24 nM) with immobilized bovine intestinal mucosa HS (30). Considering that HS appears to be required for GDNF signaling (36), the comparable affinity for the binding of GDNF with HS and E-CS/DS may suggest that CS/DS chains in the brain also act as a co-receptor for GDNF-like HS.  An-DS was partially digested with chondroitinase ABC, and the digest was subjected to gel filtration on a column of Superdex Peptide (10 ϫ 300 mm). Fractions were collected as indicated by bars and arrows. Sizes of resolved peaks were determined by MALDI-TOF MS analysis, and are indicated by numbers of constituent monosaccharides: 2-12, di to dodecasaccharides. B, an equal amount (0.5 g) of each fraction obtained by gel filtration was tested by ELISA as an inhibitor against the reactivity of mAb 2A12 with immobilized An-DS. Note that the decasaccharide was the smallest fraction to show significant inhibitory activity. C, the decasaccharide fraction was separated by anion-exchange HPLC as described under "Experimental Procedures" to isolate two main components, 10-a and 10-b. *, a baseline shift. D, the two purified decasaccharides were subjected to an inhibitory ELISA as described above. Note that 10-a and 10-b showed similar inhibitory activity. The values in B and D represent the mean Ϯ S.D. of those obtained from two independent experiments.

DISCUSSION
In this study, we characterized a novel IgM mAb 2A12, which was raised against the DS from ascidian A. nigra and recognized a unique oversulfated DS epitope structure present in the mouse brain. The antibody specifically reacted with An-DS, but not any other typical CS/DS variants or heparin, as evaluated by ELISA ( Fig. 2A), suggesting that the essential residues are 2-O-sulfated ␣-IdoUA and 6-O-sulfated ␤-GalNAc, which are joined together to form 34IdoUA(2S)-␣133GalNAc(6S)␤13 and 33GalNAc(6S)␤134IdoUA-(2S)␣13. The An-DS decasaccharides were the smallest oligosaccharides to inhibit the binding of mAb 2A12 to An-DS, and larger oligosaccharides showed stronger inhibition (Fig.  6B), suggesting that not only the iD unit but also size is important for the recognition by this antibody. Two decasaccharides, ⌬D-iD-iD-iD-iD and ⌬D-iD-iD-iD-iC, which were isolated from the products of the partial enzymatic digestion of An-DS with chondroitinase ABC, inhibited the binding of 2A12 to immobilized An-DS in the ELISA, indicating that mAb 2A12 recognizes iD-enriched decasaccharides as minimal structures. The presence of iD in mammalian tissues has been implied (13,33), but not clearly demonstrated because of a low abundance and the lack of a specific enzyme differentiating the iD unit from the D unit. It was unambiguously demonstrated here that mAb 2A12 discriminated the iD unit from the D unit ( Fig. 2A and Table II) and will be a useful tool with which to study the distribution of the iD-containing structures.
Significant yet small proportions (typically, 1.5-5%) of D and/or iD have been found in the mouse brain (20), pig brain (11), and some isolated neuronal CS-PGs, such as DSD-1-PG/ phosphacan (12,14) and versican (37). Notably, mAb 2A12 showed reactivity with E-CS/DS (Fig. 2C) and stained restricted regions of the mouse brain (Fig. 3) and hippocampal neurons (Fig. 4), indicating the expression of iD-containing DS structures in the mammalian brain. Interestingly, pretreatment of the brain sections with chondroitinase ABC or B, largely abolished the staining by 2A12 (Fig. 3, C and D), suggesting that the 2A12 epitope in the mouse brain is embedded in DS domains rich in iD units. Although the 2A12 epitopes in the brain may contain several consecutive iD units as in the An-DS 10a and 10b decasaccharides isolated from An-DS, the possibility cannot be excluded that the 2A12 minimum epitope may be a decasaccharide with a core iD unit flanked by other disaccharides, or an oligosaccharide in which multiple iD units are scattered, considering the finding that An-DS and its decasaccharides were resistant to the action of chondroitinase B ( Fig. 1 and Table II). The isolation of 2A12-reactive structures from the brain-derived CS/DS chains will help clarify the structure of this unique epitope.
Immunohistochemical staining (Fig. 3) using mAb 2A12 revealed a highly specific spatiotemporally regulated expression in the mouse brain. In contrast to a strong and widespread expression of the CS-56 epitope consistent with a recent report by Maeda et al. (12), the expression of the 2A12 epitope was restricted to the cerebellum and hippocampus in the early postnatal period. The distinct staining patterns of these two antibodies were probably attributable to their different specificities: the CS-56 epitope includes the A-D (GlcUA-GalNAc(4S)-GlcUA(2S)-GalNAc(6S)) tetrasaccharide-containing octasaccharide sequences (33), whereas 2A12 recognizes iD-containing structures as discussed above. It is noteworthy that CS-56 strongly stained the border between the CA1 region and gyrus dentate in both developing and mature mouse hippocampus, whereas the 2A12 epitope was absent in both situations. Given the notion that the expression of the CS-56 epitope in this region inhibits the sprouting of axons from the CA1 region to the dentate (38), it is conceivable that the 2A12 to the E-CS/DS chains were evaluated using a BIAcore system. Each growth factor or neurotrophic factor was mixed with the decasaccharide 10-a or 10-b (1.5 g/ml) and injected onto the surface of the E-CS/DSimmobilized sensor chip. Co-injection of the growth factor or neurotrophic factor with the intact E-CS/DS chains (1.5 g/ml) was used for comparison. The inhibitory activity was calculated from the reduced maximal response compared with that obtained without mixing with GAGs. epitope plays a distinct role from the CS-56 epitope in axon guidance. In addition, the spatiotemporal expression of the 2A12 epitope in the developing cerebellum and hippocampus suggests that this epitope may play a role in the development of these two particular regions of the central nervous system, which was supported by the findings that the cultured embryonic hippocampal neurons were markedly stained with 2A12 and that the addition of this antibody to the culture medium inhibited the growth of neurites in the hippocampal neurons. In biosynthesis, the iD unit is generated by the 6-O-sulfation of GalNAc, epimerization of D-GlcUA to L-IdoUA, and 2-O-sulfation of IdoUA, which take place most likely in this order, because the 6-O-sulfation of GalNAc precedes the epimerization and 2-O-sulfation (39), and 2-O-sulfotransferase can transfer sulfate to both GlcUA and IdoUA (40). Biosynthetic regulation of the expression of the iD unit by 6-O-sulfotransferases, 2-O-sulfotransferase, and C5-epimerase may play critical roles in the development of the hippocampus and cerebellum.
CS/DS chains have recently been implicated in regulating the signaling of various growth factors, such as PTN (12,41,42), MK (43,44), hepatocyte growth factor (45), and FGF7 (46), like HS. Our recent study has shown that certain CS/DS chains co-operate with PTN to promote the outgrowth of neurites, and that oversulfated disaccharides and IdoUA-containing units are required for the interaction of PTN with the CS/DS chains (15). On the other hand, neurotrophic factors, such as GDNF and BDNF, are involved in the formation of myelin (47) and in preventing neurodegeneration in Parkinson disease (48), and the signaling of GDNF reportedly requires HS (36). Recently, Nandini et al. (19,30) demonstrated that both GDNF and BDNG bind CS-H from hagfish notochord and CS/DS hybrid chains from shark skin with higher affinity than they bind HS, suggesting that CS/DS chains act as a receptor or co-receptor for these two neurotrophic factors. This notion is supported by our finding that both of these neurotrophic factors bind E-CS/DS and HS with comparable affinity (Fig. 8). Interestingly, the 2A12-reactive decasaccharides isolated from An-DS exhibited inhibitory activity comparable with or stronger than that of the intact E-CS/DS chains against the binding of FGF2, PTN, MK, GDNF, and BDNF to the E-CS/DS chains, suggesting that the 2A12 epitope is involved in the binding and probably in the signaling of these molecules as well, thereby regulating the cell morphology and/or differentiation. The inhibitory activity of the oligosaccharides toward the binding of each growth factor to E-CS/DS differed, suggesting that the specificity of the binding varied. This finding also strongly implies that oligosaccharides with certain structures from CS/DS chains might be useful for developing therapeutic agents targeting to certain growth factors or neurotrophic factors for neuronal diseases and brain damage. To our knowledge, except for the two discrete functional decasaccharides isolated in this study, the isolation of structurally defined sulfated DS oligosaccharides has been limited to the hexasaccharide, IdoUA(2S)-GalNAc(4S)-IdoUA(2S)-GalNAc(4S)-IdoUA-(2S)-2,5-anhydrotalitol, which binds to heparin co-factor II (49).
The functions of CS-PGs and their carbohydrate moieties in the development of the brain have been a subject of debate. We and others have shown that oversulfated CS, DS, and CS/DS hybrid chains promote neurite outgrowth (11, 13-17, 19, 30; for a review, see Ref. 7), which is in contrast to the conventional concept that CS chains in the brain are intrinsic inhibitory components for the axonal growth and path finding of various neurons (50,51). Enzymatic removal of CS chains permits axonal regeneration after nigrostriatal tract axotomy and spinal cord injury (52)(53)(54). This seeming discrepancy is most likely attributable to the structural changes of the CS/DS chains and/or the expression of other relevant components, whose functions are regulated by these CS/DS chains in the microenvironment during development. Our antibody will be a useful tool with which to examine dynamic structural and functional alterations during development. It will be interesting to identify those PGs carrying such a unique iD-containing 2A12 epitope, which would provide valuable information on the functions of CS-PGs during the development of the central nervous system.