INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

During development of the nervous system, molecules of the extracellular matrix (ECM)¹ (1) play an important regulatory role in processes such as cell migration, cell differentiation, neurite outgrowth, and the establishment of synaptic connections (for review see Refs. 1 and 2). ECM molecules interact with various components, e.g. growth factors or other constituents of the ECM, which are thus concentrated in topographically restricted territories and arranged in pericellular superstructures (3, 4). Several types of receptors have been described to mediate cell-surface interactions with the ECM, most notably the integrins, heterodimeric transmembrane proteins, and some proteoglycans (2, 5). Considering the biochemical subclasses, much more is known about structure and functions of the well characterized ECM glycoproteins laminin, fibronectin, and tenascin-C (6-8) than about the functions of proteoglycans (PGs) during neural development and regeneration.

PGs constitute a heterogeneous class of molecules, which comprise a protein core and at least one covalently linked sulfated glycosaminoglycan (GAG) chain (9). Each GAG chain consists of repeating disaccharide units, and the composition of these dimers and their sulfation patterns determine distinct GAG types. The resulting structural diversity is reflected in the different functions, which have been attributed to PGs during development and regeneration (for review see Refs. 1, 10, and 11). Chondroitin sulfate PGs like neurocan and phosphacan (12, 13) are reported to have primarily inhibitory properties in cell adhesion and neurite outgrowth, and some of these effects are thought to be mediated through the core protein (14-16). Yet, the core protein of an as yet unidentified PG (17) and the integral DSD-1-PG have been shown to promote neurite outgrowth (18) and the latter to be expressed in areas of axonal growth (19). Several lines of investigations suggest, however, that not only the core proteins but also the GAGs are involved in neural development. Thus, it has been reported that digestion of chondroitin sulfate by injection of chondroitinase ABC in vivo modifies the patterning and differentiation of retinal ganglion cells and their axonal projections during chick retina development (20). Along those lines, other studies suggest an influence of soluble GAGs on the establishment of neuronal polarity (21, 22), the promotion of neuronal survival (23), and the attachment of dopaminergic neurons in wounded adult striatum (24). The role of chondroitin sulfate in neurite outgrowth is controversial. An inhibitory influence of chondroitin sulfates on neurite outgrowth (25-29) is consistent with studies showing an enrichment of chondroitin sulfate in glial boundaries supposed to restrain neurite growth (30, 31) and in lesions of the central nervous system (32-36). In contrast to those findings, other studies have described either no or a stimulatory influence of chondroitin sulfate on neurite outgrowth (18, 22, 37, 38). These results are in agreement with the findings that chondroitin sulfates are up-regulated after lesion during the period of sciatic nerve regeneration (39, 40) and required for the regeneration of retinal goldfish axons (41). The divergence of results might have been caused by differences of the presentation of chondroitin sulfate and/or chondroitin sulfate PGs and by the use of distinct neuronal cell types. Additionally, chondroitin sulfates bear a considerable structural variability, which is presently not well understood. The basic disaccharide, which consists of glucuronic acid and N-acetylgalactosamine, can be modified by ester sulfation reactions at various positions. The chemical heterogeneity and the formation of defined structured motifs on singular chondroitin sulfate chains have been documented by several detailed studies (38, 42, 43). In addition, studies with monoclonal antibodies (mAbs) directed against specific epitopes on chondroitin sulfate chains revealed restricted spatio-temporal patterns of expression in various tissues (44-46). The regulation of chondroitin sulfate isoforms, which vary both in the degree and the positions of sulfation, is compatible with the mediation of distinct functions during development (46).

One of these structures is the DSD-1 epitope, a chondroitin sulfate modification, which has been identified on the glial-derived DSD-1-PG using the specific mAb 473HD. The corresponding epitope could be detected in chondroitin sulfate C (CS-C) and CS-D preparations from shark cartilage but not in those of the other members of the chondroitin sulfate family (18, 38). Substrate-bound DSD-1-PG promotes neurite outgrowth of several neuronal types including hippocampal neurons. Perturbation studies using mAb 473HD or chondroitinase ABC indicated that this functional activity requires the presence of the DSD-1 epitope, suggesting that this epitope represents a functional GAG structure (18).

To explore this concept, the structure-function relationship of the DSD-1 epitope, which is predominantly expressed in the nervous system, was investigated. We show here that chondroitin sulfate fractions containing the DSD-1 epitope promote neurite outgrowth of embryonic day 18 (E18) hippocampal neurons. This effect is directly correlated with the amount of substrate bound DSD-1 epitope-containing GAGs. Furthermore, recognition of the the DSD-1 epitope by mAb 473HD and the neurite outgrowth promotion by CS-D are dependent on the sulfation of the GAGs. This opens the possibility that sulfation patterns regulate the functional potential of chondroitin sulfates.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Animals-- For preparation of rat embryonic day 18 (E18) hippocampal neurons or mixed postnatal mouse cerebellar cultures, SD rats or NMRI mice were used, respectively. The animals were kept at the local facility (animal house of the University of Heidelberg, Zentralbereich Theoretikum, Heidelberg, Germany).

Materials-- Chondroitin sulfates preparations from the folowing biological sources were purchased from Seikagaku Corp. (Tokyo, Japan) or from Sigma (Deisenhofen, Germany): CS-A, bovine trachea; CS-B, bovine mucosa; CS-C and CS-D, shark cartilage; CS-E, squid cartilage. Proteinase-free preparations of chondroitinase ABC were obtained from Boehringer Mannheim (Mannheim, Germany). Other chemicals, including cell culture media, were purchased from Sigma, Merck (Darmstadt, Germany), Life Technologies, Inc. (Eggenstein, Germany), or Boehringer Ingelheim Bioproducts (Heidelberg, Germany) unless specified otherwise.

Cell Culture and Immunocytochemistry-- Rat hippocampal neuron cultures were established from E18 animals as described by Goslin and Banker (47), with some modifications (48). Cells were plated on glass coverslips at low density (8,000-10,000 cells/cm²) and cultivated in minimal Eagle's medium (MEM) supplemented with the N2 mixture, namely 5 mg/ml insulin, 20 nM progesterone, 100 µM putrescine, and 30 nM selenite (49), and 0.1 mM pyruvate, 0.1% (w/v) ovalbumin, and 0.01% (w/v) apo-transferrin. The cultures were kept in a humidified atmosphere with 5% CO₂ at 36 °C.

Morphometric Analysis and Statistics-- The stained hippocampal neurons were analyzed with a morphometric station (Leica, Bensheim, Germany; invert microscope, camera, Quantimed 500 MC). The length of the longest neurite was determined by drawing this process in the interactive mode of the program. Only cells with neurites longer than one neuronal cell body diameter were counted as neurite-bearing cells. The fractions of process forming neurons and the distribution of the longest neurites obtained under different experimental conditions were compared using the non-parametric Mann-Whitney U test in the SigmaStat program (SPSS Inc., Chicago).

Enzyme-linked Immunosorbent Assay (ELISA)-- Purified DSD-1-PG was absorbed overnight on polyvinylpyrrolidone (Falcon) at 0.5 µg/ml as GlcUA in 0.1 M NaHCO₃, pH 8.1, 100 µl/well. The wells were washed three times with PBS and blocked with 1% (w/v) BSA in PBS including 0.05% (v/v) Tween 20 (PBST) for 1 h at RT. Subsequently, the polyvinylpyrrolidone plates were incubated with mAb 473HD (1-5 µg/ml final concentration) in blocking reagent for 1 h at RT. For competition studies, mAb 473HD was preincubated with different GAGs for 2 h at 37 °C at various concentrations (see "Results") and added to the DSD-1-PG in the presence of soluble competitors. The plates were washed three times with PBST and incubated for 1 h at RT with specific anti-rat IgM secondary antibodies derivatized with horseradish peroxidase diluted 1:5,000 in blocking buffer. After three washes, the plates were developed with ABTS (52). The colored reaction product was quantified with an ELISA reader (Titertek multiscan; Flow Laboratories, Meckenheim, Germany) at OD405 nm. Experiments were performed in triplicate. The percent inhibition was calculated as follows: % inhibition = ((OD_test  OD_control)/Od_control) × 100.

Biosynthetic Labeling of Cell Cultures and Immunoprecipitation-- For biosynthetic labeling of proteins or carbohydrates with [³⁵S]methionine/cysteine (250 µCi/ml) or Na₂³⁵SO₄ (300 µCi/ml) (Amersham Buchler GmbH, Braunschweig, Germany), postnatal cerebellar cultures or Oli-neu cells were cultured in Na₂SO₄-free and methionine/cysteine-free Dulbecco's modified Eagle's medium for 45 min. Thereafter, 7 mM sodium chlorate was added to the medium, and after a further 15 min either [³⁵S]methionine/cysteine (250 µCi/ml) together with unlabeled Na₂SO₄ (10 µg/ml) or Na₂³⁵SO₄ (300 µCi/ml) together with unlabeled methionine and cysteine (each 30 µg/ml) was added to the cultures for 4 h. Subsequently, the supernatants were collected, and a mixture of proteinase inhibitors was added (benzamidine and phenylmethylsulfonyl fluoride at 1 mM; aprotinin, iodoacetamide, and pepstatin at 1 µg/ml). The cells were lysed for 20 min on ice in 0.15 M NaCl, 20 mM Tris-HCl, 1 mM EDTA, 1 mM EGTA, 1% (v/v) Triton X-100, 1% BSA (w/v), pH 7.4, with added proteinase inhibitors. Supernatants and detergent extracts were cleared by consecutive rounds of centrifugation at 800 × g, 4 °C for 10 min and at 100,000 × g, 4 °C for 45 min. The immunoprecipitation was carried out as described (53). In brief, 1 ml of the cleared lysates or cell culture supernatants were mixed with 20 µg/ml mAb 473HD or 200 µg/ml pDSD-1-PG for at least 1 h or overnight at 4 °C. With the polyclonal antibody the aliquots were incubated for 1 h at 4 °C with 150 µl of preswollen protein A-Sepharose conjugate, whereas with the mAb 473HD aliquots were incubated with 150 µl of protein A-Sepharose preincubated with goat anti-rat IgM antibodies for 1 h at 4 °C (20 µg/ml final concentration). After several washes the Sepharose beads were resuspended in 100 µl of chondroitinase ABC digestion buffer (40 mM sodium acetate, 0.1% (w/v) BSA, pH 8.0, with protease inhibitors). Half of the material was supplemented with the same volume of 2-fold concentrated SDS sample buffer and separated on a 4-10% gradient SDS-polyacrylamide gel. The radioactive signals were monitored on an x-ray sensitive film (Amersham Buchler GmbH, Braunschweig) and developed with a Kodak x-ray developing machine (Kodak M35 X-Omat processor) or detected on a phosphorimager plate (Fuji, MacBas 100; RayTest, Straubenhardt, Germany).

Preparation of Tissue Extracts-- Different tissues of adult NMRI mice were homogenized in 50 mM Tris-HCl, 50 mM sodium acetate, 60 mM octyl glucoside, pH 8.0, supplemented with a mixture of proteinase inhibitors (soybean trypsin inhibitor, antipain, aprotinin at 10 µM; phenylmethylsulfonyl fluoride at 5 µM; pepstatin and leupeptin at 2 µM). The homogenate was extracted for 1 h at 4 °C with gentle stirring and cleared by centrifugation at 100,000 × g, 4 °C for 1 h. The protein content of the extract was determined with the Bio-Rad protein assay using BSA as a standard (Bio-Rad, München, Germany). For Western blot analysis 100 µg of protein were loaded per lane.

SDS-PAGE and Western Blots-- SDS-PAGE was performed on 4-10% gradient gels (54). Proteins were transferred according to standard protocols (55) on a nylon membrane (polyvinylidene difluoride; Millipore, Eschborn, Germany). Membranes were blocked with 4% (w/v) milk powder in PBST for 30 min at RT. MAb 473HD (1-3 µg/ml final concentration) diluted in blocking buffer was incubated for 1 h at RT. To detect the bound antibodies the membranes were incubated with horseradish peroxidase-conjugated goat anti-rat IgM diluted in blocking buffer for 1 h at RT. After washing (five times for 7 min with PBST), bound antibodies were visualized by enhanced chemiluminescence (ECL Kit, Amersham-Buchler).

Determination of Uronic Acid Equivalents-- Glucuronic acid (GlcUA) concentrations were determined with a colorimetric assay using m-hydroxydiphenyl (56) or carbazole (57). CS-C or GlcUA were used as standards.

Desulfation of CS-D-- CS-D was desulfated as described (58) by treating the pyridinium salt of the polysaccharide with 90% (v/v) dimethyl sulfoxide for 1, 2.5, and 5 h at 80 °C. The fractions were analyzed by HPLC. The efficiency of the reaction is shown in Table II.

Radioabeling of Chondroitin Sulfates-- The radiolabeling of chondroitin sulfate chains was conducted by ³H-acetylation of N-deacetylated galactosamine residues by treating CS-D polysaccharides successively with hydrazine and then [³H]acetic anhydride (59). N-Deacetylation was carried out as described (60). CS-D (1 mg) was mixed with 0.2 ml of anhydrous hydrazine and 28 mg of hydrazine sulfate. The tube was sealed and heated at 96 °C for 6 h. At the end of the reaction period, the mixture was dessicated. The N-deacetylated chondroitin sulfate chains were recovered by precipitation with 80% (v/v) ethanol. The resultant precipitate was dissolved in 200 µl of 10% methanol containing 0.05 M Na₂CO₃. The solution was kept on ice, and 2.5 mCi of [³H]acetic anhydride was added. The reaction was agitated, and the pH was kept between 7.0 and 7.5 by intermittent additions of 10% (v/v) methanol saturated with Na₂CO₃. The reaction was continued for a total period of 1 h, with repeated additions of 2.5 mCi of [³H]acetic anhydride every 20 min. Unlabeled acetic anhydride was then added to the sample to complete the acetylation. During a 1-h period, three portions of 1 µl of acetic anhydride were added, and the pH was maintained neutral as above. After completion of acetylation, the reaction mixture was applied to a column (1 × 47 cm) of Sephadex G-50 (fine grade), which was eluted with 0.25 M NH₄HCO₃ containing 7% (v/v) propyl alcohol. Labeled chondroitin sulfate chains excluded from the gel were pooled and dessicated. The hydrazinolysis conditions corresponded to those used for N-deacetylation but not for release of N-glycans from core peptides (61). No significant structural alterations were caused by this treatment, as evidenced by the observation that complete degradation to the expected disaccharides could be obtained by chondroitinase ABC digestion.

Immunoaffinity Chromatography-- The following procedures were performed at 4 °C. ³H-Acetylated CS-D (5 × 10⁵ cpm/260 µg) was applied to a column (3.5-ml bed volume) of Sepharose 4B, which had been coupled with mAb 473HD and equilibrated with PBS. The resin contained the antibody at a concentration of 1.4 mg/ml. The column was washed with PBS. The absorbed materials were eluted with 0.1 M diethylamine, 0.1 M NaCl, 1 mM EDTA, and 1 mM EGTA, pH 11.5. One-ml fractions were collected and monitored by scintillation counting. For inhibition experiments on the mAb 473HD affinity matrix [³H]CS-D (3000 cpm) was mixed with unlabeled CS-A or CS-D (50 µg), and the mixture was applied to a column (3.5 ml bed volume) of 473HD-Sepharose as above. The bound and unbound fractions were quantified by scintillation counting to estimate inhibition of the binding of the labeled CS-D by unlabeled CS isoforms.

Analysis of Chondroitin Sulfate Chains Attached to DSD-1-PG-- DSD-1-PG (50.3 nmol as GlcUA) was digested using 25 milliunits of chondroitinase ABC lyase as described previously (62). The digest corresponding to 8.4 nmol of GlcUA was analyzed by HPLC on an amine-bound silica PA03 column (4.6 × 250 mm; YMC Co., Kyoto, Japan) as described (63, 64). HPLC was performed in an LC-10AS system (Shimadzu Co., Kyoto, Japan) using a linear gradient from 16 to 530 mM NaH₂PO₄ over a 60-min period at a flow rate of 1.0 ml/min at RT. Eluates were monitored by UV absorbance at 232 nm. Identification and quantification of the resulting disaccharide units were achieved by comparison with chondroitin sulfate-derived authentic unsaturated disaccharides and by enzymatic digestion using hexuronate-2-sulfatase, chondro-4- or 6-sulfatases as reported (62).

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

The DSD-1 Epitope Is Predominantly Expressed in the Nervous System-- It has been previously shown that mAb 473HD recognizes an epitope primarily expressed on the neural chondroitin sulfate proteoglycan DSD-1-PG, which is produced by immature glial cells (18, 65). To investigate whether the DSD-1 epitope is specifically expressed in the nervous system, a Western blot analysis of adult mouse tissues was performed (Fig. 1). When equal amounts of protein were loaded, mAb 473 HD reactivity was only found in the octyl glucoside extracts of the cerebellum and the residual brain but not in extracts of several other organs, such as liver, spleen, kidney, thymus, lung, and heart. The data suggest a tissue-related restriction with a predominant expression of the DSD-1 epitope in the adult mouse brain.

View larger version (57K): [in this window] [in a new window]   Fig. 1.   Western blot analysis of different mouse tissue extracts by mAb 473HD. Western blots of detergent extracts of adult mouse tissues as indicated were performed. mAb 473HD predominantly recognized the DSD-1-PG expressed in nervous tissue as a polydisperse smear in the high molecular weight range.

The DSD-1 Epitope Is Dependent on the Sulfation of CS-- mAb 473HD recognizes an epitope on GAGs of the chondroitin sulfate family (18). Recently it has been shown that the DSD-1 epitope is localized in CS-C and CS-D preparations but not in those of CS-A, CS-B, or CS-E (38). The main difference between these carbohydrates resides in the sulfation pattern, whereas the common structure is based on disaccharides composed of glucuronic acid (or the epimer iduronic acid in case of CS-B) and N-acetylgalactosamine. To test whether sulfation is critical for the recognition of the DSD-1 epitope by mAb 473HD, the addition of sulfate groups to newly synthesized polymers was blocked in vitro. To this end Oli-neu cells, a DSD-1-PG-producing oligodendroglial precursor cell line (51, 65), were grown in the presence of sodium chlorate, a reagent that suppresses the addition of sulfate groups to proteins and carbohydrates by up to 95% (66). Sodium chlorate blocks the ATP-sulfurylase, the first enzyme in the biogenesis of phosphoadenosine-phosphosulfate, which is the ubiquitous co-substrate for sulfation. The protein biosynthesis is not affected by sodium chlorate (Fig. 2A). Immunoprecipitation and immunocytochemistry of the treated versus the untreated cells with mAb 473HD showed that the DSD-1 epitope was not present on the cells treated with sodium chlorate. In contrast, the polyclonal antibody pDSD-1-PG still recognized DSD-1-PG, indicating that the core protein was correctly synthesized by the chlorate-treated Oli-neu cells (Fig. 2B and Fig. 3). Analogous results were obtained when the experiment was carried out with mixed cerebellar cultures, which also synthesize DSD-1-PG (18).²

View larger version (50K): [in this window] [in a new window]   Fig. 2.   Immunoprecipitation of DSD-1-PG from Oli-neu culture supernatants. The expression of the DSD-1-PG in sodium chlorate-treated Oli-neu cells was studied by immunoprecipitation with mAb 473HD and the polyclonal serum pDSD-1-PG. A, shows a fluorograph of a 4-10% SDS gradient gel with supernatants of Oli-neu cells cultured in the presence of Na235SO4 or [35S]methionine/cysteine. Sodium chlorate suppressed the incorporation of Na235SO4, whereas the [35S]methionine/cysteine incorporation was not affected. B, the autoradiograph of a 4-10% SDS gradient gel resolving immunoprecipitates of [35S]methionine/cysteine-labeled Oli-neu culture supernatants shows that the DSD-1-PG was precipitated by mAb 473HD from cells cultured in the absence but not in the presence of sodium chlorate. pDSD-1-PG did recognize the DSD-1-PG when expressed by sodium chlorate-treated and -untreated Oli-neu cells.

View larger version (107K): [in this window] [in a new window]   Fig. 3.   Expression of the DSD-1 epitope by sodium chlorate-treated Oli-neu cells. Oli-neu cells were cultured without (A, B, E, and F) or in the presence of sodium chlorate (C, D, G, and H) and stained with mAb 473HD (A and C) and pDSD-1-PG (E and G) as described under "Experimental Procedures." Notably, the staining for mAb 473HD is less prominent on the chlorate-treated as compared with the cells cultured in the absence of sodium chlorate. The staining for pDSD-1-PG is not affected in either condition. Bar, 50 µm.

2

DSD-1 Epitope-bearing Carbohydrates Can Be Enriched on a mAb 473HD Affinity Column-- To enrich for DSD-1 epitope-bearing carbohydrates, tritiated CS-D ([³H]CS-D) was loaded on a mAb 473HD affinity column. After washing with loading buffer, a substantial fraction of the applied activity was retained on the column. The binding of [³H]CS-D could be blocked by the addition of unlabeled CS-D (Fig. 4B) but not with unlabeled CS-A (Fig. 4A), underlining that DSD-1 epitope-bearing carbohydrates but not other chondroitin sulfates are specifically enriched by the mAb 473HD affinity matrix. These results support the previous conclusion that the DSD-1 epitope is contained in CS-D preparations (38). To examine whether the DSD-1 epitope is comprised in all carbohydrate polymers of CS-D mixtures either commercially available CS-C or CS-D were repetitively loaded on the mAb 473HD affinity column. The eluates of three consecutive chromatography steps and the final unbound material were analyzed by competition ELISA using purified neural DSD-1-PG (CS-C affinity retained fraction of the first, second, and third chromatographies were named C-FR 1, C-FR 2, and C-FR3, respectively; C-FR UB represents the unbound fraction; likewise the corresponding fractions with CS-D were named D-FR 1, D-FR 2, D-FR 3, and D-FR UB, respectively) (Fig. 4, C and D). Carbohydrates of all fractions displayed comparable sizes of at least 10 kDa as shown by size exclusion chromatography during the purification procedure. The fraction with the highest affinity for mAb 473HD was the eluate of the first round of chromatography (C-FR 1 in case of CS-C; D-FR 1 in case of CS-D), whereas the unbound carbohydrates of the third cycle of purification (C-Fr UB; D-FR UB) did not interfere with the binding of the mAb 473HD to the immobilized DSD-1-PG. The disaccharide analysis of the DSD-1 epitope-enriched fractions revealed that the D disaccharide content was highest in the most enriched fractions (Table I). These findings reinforce the notion that DSD-1 epitope-bearing GAGs are specifically enriched on the mAb 473HD affinity column. They also suggest that the DSD-1 epitope is localized on subpopulations of carbohydrate polymers contained in CS-C or CS-D preparations and that the D disaccharide unit contributes to the DSD-1 epitope, although the sequential arrangement of the disaccharide units has yet to be determined.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Enrichment of DSD-1 epitope-bearing carbohydrates on a mAb 473HD affinity matrix. [3H]CS-D was loaded on an mAb 473HD affinity column in the presence of non-labeled CS-A. A, most of the radioactivity was bound to the column and eluted with pH 11.5. B, non-labeled CS-D competed the binding of [3H]CS-D to the column. The arrows in A and B indicate the change to pH 11.5 buffer. To enrich DSD-1 epitope-bearing carbohydrates, CS-C and CS-D were loaded repetitively on a mAb 473HD affinity column as described under "Experimental Procedures" (CS-C affinity retained fraction of the first, second, and third chromatographies were named C-FR 1, C-FR 2, and C-FR3, respectively; C-FR UB represents the unbound fraction; likewise the corresponding fractions with CS-D were named D-FR 1, D-FR 2, D-FR 3, and D-FR UB, respectively). The eluted fractions were analyzed by competition ELISA using different concentrations of competitors. C shows the analysis of the eluted fractions of CS-C and D of CS-D. The highest DSD-1 epitope-enriched fractions interfered with the binding of mAb 473HD to the DSD-1-PG, whereas the lowest affinity fractions did not. Note, the CS-D fractions inhibited the binding at lower concentrations as compared with CS-C fractions. The percent inhibition was calculated as follows: % inhibition = [(ODtest  ODcontrol)/ODcontrol ] × 100.

Analysis of Chondroitin Sulfate Chains Attached to DSD-1-PG-- Chondroitinase ABC lyase digestion of the purified DSD-1-PG yielded four disaccharide units (a) HexA1-3GalNAc, (b) HexA1-3GalNAc(6-O-sulfate), (c) HexA1-3GalNAc(4-O-sulfate), and (d) HexA(2-O-sulfate)1-3GalNAc(6-O-sulfate) and (e) an unidentified compound with recoveries of 2.3, 23.2, 67.6, 5.0, and 2.0%, respectively, as quantified by HPLC (Fig. 5A). Furthermore, the structure of each peak was identified by digestion with sulfatases. Peak b [HexA1-3GalNAc(6-O-sulfate)] and peak d [HexA(2-O-sulfate)1-3GalNAc(6-O-sulfate)] in Fig. 5A were shifted to the position of peaks a [HexA1-3GalNAc] and f [HexA(2-O-sulfate)1-3GalNAc], respectively, upon chondro-6-sulfatase digestion (Fig. 5B), confirming that the GalNAc residues of the compounds in peaks b and d were sulfated at the C-6 position. The nature of peak e, which was resistant to chondro-6-sulfatase, hence is not [HexA1-3GalNAc(4, 6-di-O-sulfate)], remains to be determined. Peak d [HexA(2-O-sulfate)1-3GalNAc(6-O-sulfate)] was shifted to the position of peak b [HexA1-3GalNAc(6-O-sulfate)] upon hexuronate-2-O-sulfatase digestion, confirming that the HexA residue of the compound in peak d was sulfated at C-2 position (Fig. 5C). The peaks observed before peak a in Fig. 5, AC, were attributable to buffer salts. The results altogether indicate that the DSD-1-PG bears chondroitin sulfate chains that are composed of disaccharide units including GlcUA-GalNAc (unsulfated disaccharides), GlcUA-GalNAc(6S) (CS-C), GlcUA-GalNAc(4S) (CS-A), and GlcUA(2S)-GalNAc(6S) (CS-D). Thus, the rare CS-D unit was clearly demonstrated in the DSD-1-PG being consistent with the higher proportion of the D disaccharide in the high affinity fractions (D-FR 1, D-FR 2, and C-FR 1 in Table I).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   HPLC analysis of GAGs of the DSD-1-PG. The DSD-1-PG was incubated with chondroitinase ABC (A), chondroitinase ABC, and then chondro-6-sulfatase (B), or chondroitinase ABC and then hexuronate-2-sulfatase (C), and each reaction mixture was analyzed by HPLC on an amine-bound silica column as described under "Experimental Procedures." Peak a, HexA1-3GalNAc; peak b, HexA1-3GalNAc(6S); peak c, HexA1-3GalNAc(4S); peak d, HexA(2S)1-3GalNAc(6S); peak e, unidentified material; peak f, HexA(2S)1-3GalNAc. The elution positions of the standard disaccharides are indicated in A as follows: 1, HexA1-3GalNac; 2, HexA1-3GalNac(6-O-sulfate); 3, HexA1-3GalNAc(4-O-sulfate); 4, HexA(2-O-sulfate)1-3GalNAc(6-O-sulfate); 5, HexA1-3GalNAc(4,6-O-disulfate); 6, HexA(2-O-sulfate)1-3GalNAc(4,6-O-disulfate).

GAGs Containing the DSD-1 Epitope Promote Neurite Outgrowth When Presented as Substrate-- It has been shown previously that the DSD-1 structure is involved in the neurite outgrowth-promoting capacity of substrate-bound DSD-1-PG (18). To test whether DSD-1 epitope-carrying GAG chains are able to promote neurite outgrowth, E18 hippocampal neurons were cultured in low density on glass coverslips coated with PORN and different carbohydrate preparations. As shown previously, the fraction of neurite-bearing cells increased on coverslips coated with CS-D (61%) compared with the number of neurons with processes cultured on CS-A, CS-B, and CS-C (45, 42, and 46%, respectively) or on PORN (38%) alone (38). The present morphometric analysis showed that not only the number of neurite-bearing cells but also the length of the longest neurites was enhanced on CS-D versus the other members of the chondroitin sulfate family (Figs. 6 an 7A). Although the promotion of neurite length on CS-D was not as strong as the effect obtained with the integral DSD-1-PG, it was significantly inhibited by mAb 473HD when added to the culture medium (Fig. 7B). This indicates that GAGs are able to promote neurite outgrowth of hippocampal neurons although the carbohydrate chains are not linked to a protein such as the DSD-1-PG core protein and that the neurite outgrowth-promoting activity of CS-D is linked to the DSD-1 epitope.

View larger version (84K): [in this window] [in a new window]   Fig. 6.   Neurite outgrowth promotion by different chondroitin sulfate isoforms. E18 hippocampal neurons were cultured on glass coverslips coated with PORN and CS-A (A), CS-B (B), CS-C (C), and CS-D (D), and on PORN alone (E). The cells were stained for tubulin after 24 h in culture. Note the increased neurite lengths on CS-D. Bar, 35 µm.

View larger version (24K): [in this window] [in a new window]   Fig. 7.   Length distribution of neurites of hippocampal neurons cultured on chondroitin sulfate substrates. Neurites were morphometrically analyzed after hippocampal neurons had been cultured on different chondroitin sulfate isoforms without (A) or in the presence of mAb 473HD for 24 h (B). The lengths of the longest neurites of 100 neurite-bearing cells were measured in three independent experiments. Graph A shows the percent increase of the mean length of the longest neurites versus the PORN control. CS-D promoted neurite outgrowth as compared with the Porn control. Graph B shows the percent increase of the mean length of the longest neurites in the presence of mAb 473HD versus the length in the absence of antibody. mAb 473HD inhibited the neurite outgrowth-promoting effect of CS-D. In both experiments percentages were calculated as follows: % increase = [(Ltest  Lcontrol)/Lcontrol] × 100 (where L indicates length). Statistical analysis of the pooled data was evaluated with the Mann-Whitney U test. n.s., not significant; *0.01 p < 0.05; ***, p > 0.001.

The DSD-1 Epitope Is Sufficient to Promote Neurite Outgrowth-- To study whether the DSD-1 structure is sufficient to promote neurite outgrowth, fractions enriched for the DSD-1 epitope were tested as substrates for hippocampal neurons, and the number of neurite-bearing cells was determined. The DSD-1 epitope-enriched fractions (C-FR 1, C-FR 2, C-FR 3, and C-FR UB; D-FR 1, D-FR 2, D-FR 3, and D-FR UB) rather than the chondroitin sulfate mixtures were used, because the biochemical analysis had revealed varying concentrations of the DSD-1 epitope in the latter. DSD-1 epitope-enriched fractions obtained from CS-C and CS-D promoted neurite outgrowth more efficiently than the carbohydrate fractions that were not retained on the mAb 473HD affinity column (Fig. 8, A and B). In the case of CS-C, the number of neurite-bearing cells on the DSD-1 epitope-enriched fraction C-FR 1 was statistically higher than on the unprocessed CS-C (59 versus 46%). However, the fraction of neurite-bearing cells on CS-D preparations was not statistically different when compared with the DSD-1 epitope-enriched fractions D-FR 1 and D-FR 2 (61 versus 61 and 63%, respectively) obtained therefrom, indicating that the content of DSD-1 epitope in CS-D and D-FR 1 and D-FR 2 is at saturation with respect to the neurite outgrowth-promoting effect under these conditions (coating 16 µg/ml as total weight or 5 µg/ml as GlcUA). These results show that the neurite outgrowth-promoting capacity of both CS-C and CS-D correlates with the content of CS-D units and the DSD-1 epitope in these preparations.

View larger version (24K): [in this window] [in a new window]   Fig. 8.   Promotion of neurite outgrowth of hippocampal neurons by chondroitin sulfates bearing the DSD-1 epitope. DSD-1 epitope-enriched fractions (see Fig. 4) were tested as substrates for hippocampal neurons as described in Fig. 6, and the percentage of neurite-bearing cells was determined. A, C-FR 1 promoted neurite outgrowth significantly stronger than C-FR 2, C-FR 3, C-Fr UB, and CS-C. The effects of these fractions were not significantly different. B, fractions D-FR 1 and D-FR 2 could not promote neurite outgrowth more extensively than the original CS-D preparation. In contrast, the effect of D-FR UB was significantly less as compared with CS-D, D-FR1, and D-FR 2. The number of experiments is given in parentheses. 100 cells were counted per experiment. Statistical analysis was evaluated with the Mann-Whitney U test. Only the significant differences are indicated. *, 0.01 < p < 0.05; **, 0.001 < p < 0.01.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In the present study the structural and functional properties of the DSD-1 epitope, a chondroitin sulfate variant recognized by mAb 473HD, were examined. The DSD-1 epitope displays a restricted tissue distribution and is preferentially expressed in the central nervous system of the adult mouse. Structurally, it is associated with the D motif and depends on sulfation of carbohydrate polymers that are comprised in CS-C and CS-D preparations. Culture substrates prepared with carbohydrate fractions enriched for the DSD-1 epitope promote neurite outgrowth of hippocampal neurons, a property that is neutralized by mAb 473HD. We propose that carbohydrates of the chondroitin sulfate family contain discrete structural motifs with functional properties, such as the DSD-1 epitope. To exert their physiological functions, these domains might interact with specific cellular receptors.

The DSD-1-PG has initially been identified using the mAb 473HD and shown to be expressed by immature glial cells of the central nervous system (18, 65). Western blotting experiments conducted with octyl glucoside extracts of several adult mouse organs indicated that the DSD-1 epitope is mainly expressed in the cerebellum and the telencephalon but not in various non-neural tissues, different from other mAbs such as CS-56 which recognize chondroitin sulfate epitopes (67). In view of the involvement in the neurite outgrowth-promoting effect of substrate-bound DSD-1-PG (18) and the lineage-related, restricted expression (18, 65), the structural characteristics of the DSD-1 epitope were examined. Confirming earlier studies, mAb 473HD recognized epitopes on CS-C and CS-D but not on CS-A, CS-B, and CS-E (18, 38). These GAG structures differ by the relative positions of sulfate groups in the dimeric building blocks. A sulfate dependence of the DSD-1 epitope is highlighted by the finding that the inhibition of sulfation of newly synthesized GAG chains in vitro and the chemical de-sulfation of CS-D both abolished the binding site of mAb 473HD. It was possible to enrich DSD-1 epitope-bearing sugars from CS-C and CS-D by affinity chromatography, which suggests that not all but a subset of carbohydrates contain the epitope. These DSD-1 epitope-enriched fractions were also rich in D disaccharide units. Other than in shark cartilage, the D unit is rarely detected in animal tissues (68, 69). Interestingly, a trace of D disaccharides has been found in the developing rat cerebellum (70). Recent structural studies provided evidence that CS-D polymers are composed of defined sequences of disaccharide building blocks, such as -A-D-A- (38, 42, 43). So far, it was not possible to link one of the identified CS-D carbohydrate sequences to the DSD-1 epitope, because mAb 473HD recognizes only polymers longer than tetradecasaccharides.³ These results in conjunction with the present data support the view that the DSD-1 epitope represents a particular structure detectable in CS-C and CS-D preparations which is associated with the D unit. Consistent with this notion, the rare D unit was detected in DSD-1-PG preparations, although the concentration proved lower than in the DSD-1 epitope-enriched fractions obtained from other sources. This could be due to a confinement of D units to the DSD-1 epitope of DSD-1-PG, whereas these might be distributed to several other domains in CS-C and CS-D polymers. The sequence of disaccharide units which constitutes the DSD-1 epitope in conjunction with the D motif remains to be clarified.

Because indirect evidence suggested that the DSD-1 epitope is required for the neurite outgrowth-promoting activity of substrate-bound DSD-1-PG (18), the functional activities of chondroitin sulfates were examined in a neurite outgrowth assay. The carbohydrate CS-D, but not CS-A, CS-B, and CS-C, which possess the same carbohydrate backbone structure with sulfation patterns different from that of CS-D, significantly increased the fraction of neurite-bearing cells (38) and the mean length of the longest neurites. In addition CS-E, which was not identified in DSD-1-PG preparations, stimulated neurite extension in this assay. Yet, mAb 473HD neutralized the neurite outgrowth-stimulating effect of CS-D but did not affect the neurons cultured on the other GAGs, suggesting that CS-E acts in a DSD-1 epitope-independent manner.² Although both CS-D and CS-E species contain di- and/or tri-sulfated disaccharide units (61, 62), we assume that the arrangement of sulfate groups, but not merely charge density on the carbohydrate backbone, is important for the neurite outgrowth-promoting effect. This implies that a particular arranged pattern of sulfate groups on a larger backbone is crucial for the DSD-1 epitope, in agreement with the sulfate dependence of the latter. Furthermore, the existence of sulfate-based neurite growth-promoting motifs other than the DSD-1 epitope, e.g. in CS-E, seems possible. In support of this interpretation, the relative concentration of the DSD-1 structure in chondroitin sulfates was found to vary, with a stronger reactivity of mAb 473HD for CS-D as compared with CS-C. The DSD-1 epitope-enriched fractions derived from CS-C and CS-D by affinity chromatography displayed a concentration-dependent activity in that the highest affinity fractions promoted neurite growth more efficiently than the unbound ones. The active carbohydrates contained an elevated proportion of D disaccharide units, which supports the conclusion that D units are of critical significance for the constitution of the neurite outgrowth-promoting DSD-1 epitope. In view of the asserted functional significance of sulfation, it is of interest to consider potential neuritogenic influences of heparin and heparan sulfate, which carry comparable sulfate charges than chondroitin sulfates. It has been shown that these GAGs exert influences on neurite outgrowth patterns of several types of central nervous system neurons, yet the effects were clearly distinct from those obtained with chondroitin sulfates. For example, heparin favored the extension of axons, whereas CS-B or derived fragments sustained the growth of both dendrites and axons (21, 22). This finding emphasizes that chondroitin sulfate effects on neurite growth are selective and correlated with particular carbohydrate motifs, which are defined by specific sulfation patterns.

Several studies concluded that chondroitin sulfates are organized in domains. This assumption is based on structural analysis, for example of CS-D (38, 42, 43), and on the immunohistological analysis of mAbs generated against chondroitin sulfates (44, 45, 71, 72). These different mAbs yielded distinct spatial and temporal staining profiles in developing and in lesioned neural tissues (39, 40, 46). One further example is provided by the neurite outgrowth-promoting DSD-1 epitope, which is expressed in regions of intense axon growth in vivo, e.g. the inter-rhombomeric boundaries of the chicken hindbrain (73). This property contrasts the proposed inhibitory action of chondroitin sulfates reported in other systems (25-29). Most of the latter studies, however, were carried out with CS-A, CS-B, and CS-C but not with CS-D preparations that contain a sufficient concentration of DSD-1 epitope to promote neurite growth. After lesions of the central nervous system chondroitin sulfates including the DSD-1 epitope are up-regulated in the region of the glial scar, where these carbohydrates are supposed to contribute to the prevention of regeneration (30, 32-36). Recently, a DSD-1 epitope-bearing chondroitin sulfate proteoglycan with neurite growth inhibiting properties related to phosphacan was described as a ligand for tenascin-R (74). On the other hand, the DSD-1 epitope is up-regulated in the regenerating lesioned peripheral nervous system (40). Considering its concentration dependence, it is conceivable that other inhibitory molecules or motifs of the same molecule override the neurite outgrowth-promoting activity of the DSD-1 epitope.

In order to substantiate the hypothesis that chondroitin sulfate domains are organized in functional domains, the enzymes and biosynthetic pathways generating structures such as the DSD-1 epitope have to be uncovered. Another implication of the concept resides in the necessity of complementary receptors that would decipher specific information contained in GAG chains and mediate selective cellular responses. In the case of heparan sulfate, specific ligands of characterized domain structures have been defined (Ref. 75; for review see Ref. 76). It has been reported that heparin by itself activates a receptor tyrosine kinase (77). Thus, it is conceivable that also the DSD-1 epitope might elicit second messenger responses which finally lead to a stimulation of neurite growth.