Glycobiology Advance Access originally published online on February 22, 2007
Glycobiology 2007 17(6):631-645; doi:10.1093/glycob/cwm021
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Structural determination of novel sulfated octasaccharides isolated from chondroitin sulfate of shark cartilage and their application for characterizing monoclonal antibody epitopes
2 Department of Biochemistry, Kobe Pharmaceutical University, Higashinada-ku, Kobe 658-8558, Japan
3 Laboratory of Proteoglycan Signaling and Therapeutics, Graduate School of Life Science, Hokkaido University, Frontier Research Center for Post-Genomic Science and Technology, Nishi 11-choume, Kita 21-jo, Kita-ku, Sapporo, Hokkaido 001-0021, Japan
4 Department of Biotechnology, Faculty of Engineering, Kyoto Sangyo University, Kyoto 603-8558, Japan
1 To whom correspondence should be addressed; Tel: +81 11 706 9054; Fax: +81 11 706 9056; E-mail: k-sugar{at}sci.hokudai.ac.jp
Received on October 26, 2006; revised on January 24, 2007; accepted on February 19, 2007
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsSupplementary dataConflict of interest statementAcknowledgmentsReferences |
|---|
Twelve octasaccharide fractions were obtained from chondroitin sulfate C derived from shark cartilage after hyaluronidase digestion. Their sugar and sulfate composition was assigned by matrix-assisted laser desorption ionization time of flight mass spectrometry. The sequences were determined at low picomole amounts by a combination of enzymatic digestions with high-performance liquid chromatography, and were composed of disaccharide building units including O [GlcUAß1–3GalNAc], C [GlcUAß1–3GalNAc(6S)], A [GlcUAß1–3GalNAc(4S)], and/or D [GlcUA(2S)ß1–3GalNAc(6S)], where 2S, 4S, and 6S represent 2-O-, 4-O-, and 6-O-sulfate, respectively. As many as 24 different sequences including minor ones were revealed, exhibiting a high degree of structural diversity reflecting the enormous heterogeneity of the parent polysaccharides. Nineteen of them were novel, with the other four reported previously as unsaturated counterparts obtained after digestion with chondroitinase. Microarrays of these structurally defined octasaccharide fractions were prepared using low picomole amounts of their lipid-derivatives to investigate the binding specificity of four commercial anti-chondroitin sulfate antibodies CS-56, MO-225, 2H6, and LY111. The results revealed that multiple unique sequences were recognized by each antibody, which implies that the common conformation shared by the multiple primary sequences in the intact chondroitin sulfate chains is important as an epitope for each monoclonal antibody. Comparison of the specificity of the tested antibodies indicates that CS-56 and MO-225 specifically recognize octasaccharides containing an A–D tetrasaccharide sequence, whereas 2H6 and LY111 require a hexasaccharide as a minimum size for their binding, and prefer sequences with A- and C-units such as C-C-A-C (2H6) or C-C-A-O, C-C-A-A, and C-C-A-C (LY111) for strong binding but require no D-unit.
Key words: chondroitin sulfate / sugar sequencing / sulfation / antibody epitope / octasaccharides
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataConflict of interest statementAcknowledgmentsReferences |
|---|
Proteoglycans (PGs) are involved in various developmental processes such as cell migration, recognition, and morphogenesis (Bandtlow and Zimmermann 2000
; Lander and Selleck 2000; Perrimon and Bernfield 2000; Sugahara et al. 2003), and glycosaminoglycan (GAG) side chains with specific sequences selectively bind a wide range of proteins including growth factors, morphogens, and proteases, regulating their biological activities. Although most of these findings have been made for heparan sulfate PGs, increasing evidence suggests the involvement of chondroitin sulfate (CS)- or dermatan sulfate-containing PGs as well (Sugahara et al. 2003; Deepa et al. 2004), especially in brain development (Sugahara et al. 2003), axon regeneration in the central nervous system (Bradbury et al. 2002), the binding of growth factors and cytokines (Kawashima et al. 2000; Deepa et al. 2002; Bao et al. 2004, Bao, Mikami, et al. 2005; Bao, Muramatsu, et al. 2005), and signaling of growth factors in neurite outgrowth (Li et al. 2007). Although the importance of several unique structural elements including oversulfated "D" and "E" disaccharide units (Maeda et al. 1996; Clement et al. 1998, 1999; Nadanaka et al. 1998; Kawashima et al. 2000; Ueoka et al. 2000) and iduronic acid (Hikino et al. 2003; Bao et al. 2004, Bao, Mikami, et al. 2005, Bao, Muramatsu, et al. 2005; Bao, Pavão, et al. 2005; Nandini et al. 2004, 2005) have been proposed, the detailed sugar sequences have been largely unidentified. The backbone structure of CS chains is composed of repeating disaccharide units, consisting of D-glucuronic acid (GlcUA) and N-acetyl-D-galactosamine (GalNAc). Differences in the sulfation of these units give rise to a variety of structural units. The major forms of CS including CS-A, CS-C, CS-D, and CS-E, predominantly contain A-unit [GlcUAß1–3GalNAc(4S)], C-unit [GlcUAß1–3GalNAc(6S)], D-unit [GlcUA(2S)ß1–3GalNAc(6S)], and E-unit [GlcUAß1–3GalNAc(4S,6S)], respectively, where 2S, 4S, and 6S represent 2-O-, 4-O-, and 6-O-sulfate groups, respectively. The differential arrangement of these units results in the structural diversity of CS chains, which is the basis for their functional diversity. However, the high degree of heterogeneity displayed by the GAG chains and the lack of suitable analytical tools to decode the structures have hampered precise determination of the GAG structures.
The physiological importance of CS-PGs during development has been demonstrated by various studies including immunological studies using monoclonal antibodies (mAbs), which have shown that CS isoforms, differing in the sulfation pattern and degree, exhibit differential expression and distinct functions (Herndon and Lander 1990; Mark et al. 1990; Sorrell et al. 1993). Thus, the use of mAbs directed toward different parts of PGs offers a great potential in the immunohistochemical characterization of the PG subtypes distributed in various tissues for developmental studies of PGs (Faissner et al. 1994; Nadanaka et al. 1998). However, largely no detailed structural information is available regarding the epitopes recognized by these antibodies except for those directed against the unsaturated uronic acid residues generated by chondroitinase digestion (Couchman et al. 1984). Thus, the structural alteration of CS chains associated with functional changes of CS-PGs is only poorly understood.
To understand better the structure–function relationship of CS chains, oligosaccharides of varying chain lengths ranging from tri- to octa-saccharide have been isolated from various CS isoforms and sequenced by a combination of enzymatic and nuclear magnetic resonance (NMR) analyses, constructing an oligosaccharide library (Sugahara and Yamada 2000). A panel of such accumulated oligosaccharides was recently utilized to adapt the oligosaccharide microarray method developed for anti-carbohydrate antibodies (Fukui et al. 2002) to the study of anti-CS antibody epitopes (Ito et al. 2005), which demonstrated the effectiveness of the method in characterizing the antibody epitopes, for example, revealing the common A-D tetrasaccharide sequence in the epitopes of mAbs CS-56, MO-225, and 473HD (Ito et al. 2005). In addition, the minimum hexa- or octa-saccharide size for recognition was shown for MO-225 and 473HD or CS-56, respectively (Ito et al. 2005). In this study, a variety of novel octasaccharides were isolated from CS chains of shark cartilage and applied for characterization of the epitopes of several widely used anti-CS antibodies.
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataConflict of interest statementAcknowledgmentsReferences |
|---|
Previously we isolated tri- to octa-saccharides from various CS variants including CS-A, CS-B, CS-C, CS-D, CS-E, CS-H, and CS-K (for a review, see Sugahara and Yamada 2000
). In this study, CS-C from shark cartilage, which is recognized by several mAbs raised against mammalian CS-PGs, was used as a starting material to isolate octasaccharides likely to be recognized by at least some of these antibodies as recently demonstrated with mAbs CS-56 and MO-225 (Ito et al. 2005).
Preparation and isolation of octasaccharides from CS-C
Digestion of shark cartilage CS-C with mammalian hyaluronidase generates saturated even-numbered oligosaccharides in contrast to bacterial chondroitinases, which produce unsaturated di- and oligo-saccharides (Sugahara and Yamada 2000). In this study, a commercial CS-C preparation (100 mg) derived from shark cartilage was exhaustively digested with sheep testicular hyaluronidase, and the digest was size-fractionated by gel filtration on a Bio-Gel P-10 column as described in the Materials and methods section. The separated fractions were designated as fractions I–IX (Figure 1). The large peak which eluted at around fractions 80–90 was attributable to the buffer salts. Fractions were III–IX presumed to contain even-numbered oligosaccharide fragments ranging from hexadeca- to tetra-saccharides, respectively, based on the previous chromatographic data of a hyaluronidase digest of CS-D (Sugahara, Tanaka, et al. 1996). A portion (1 µmol corresponding to 2.0 mg of fraction VII or 13.9 mg of CS-C) of fraction VII, which is expected to contain octasaccharides, was subfractionated into fractions VIIa–i by anion-exchange high-performance liquid chromatography (HPLC) on an amine-bound silica column as shown in Figure 2. Each fraction was desalted by gel filtration on a Sephadex G-25 column, and quantified by the carbazole method. The total recovery of oligosaccharides after fractionation by HPLC and subsequent desalting was approximately 90%. The yields of the isolated oligosaccharide fractions are listed in Table I.
|
|
|
The apparent purity of each oligosaccharide fraction was checked by anion-exchange HPLC after labeling the oligosaccharides individually with a fluorophore 2AB as described in the Materials and methods section (data not shown). The results indicated that all fractions except fractions VIIb, VIIh, and VIIi were sufficiently pure for sequencing (Table I). Fraction VIIb could be subfractionated by anion-exchange HPLC only after 2AB-derivatization into three distinct peaks in a molar ratio of 8:32:60 as shown in Figure 3. Fraction VIIh was further purified by rechromatography into subfractions VIIh1 and h2 (Figure 4A) in a molar ratio of 27:73 and the purity of each fraction was examined after 2AB-labeling. Fraction VIIh1 was nearly homogeneous for sequencing (approximately 91% pure), while fraction VIIh2 was rather heterogeneous, and hence it was rechromatographed to obtain fractions VIIh2.1, VIIh2.2, VIIh2.3, and VIIh2.4 using a shallow gradient as shown in Figure 4C (see the inset for an enlarged view). The major fraction VIIh2.2 appeared to contain at least two components in equal proportions as shown by anion-exchange HPLC after 2AB-labeling (data not shown). Further purification of fraction VIIh2.2 was not possible and it was used for sequencing as a mixture. Fraction VIIi was separated by rechromatography into fractions VIIi1, VIIi2, and VIIi3 (Figure 4B) in a molar ratio of 71:24:5. The apparent purity of the 2AB-derivatized fractions VIIi1 and VIIi2 was 83 and 94%, respectively (Table I), and they were subjected to sequencing. Due to the small amount, fraction VIIi3 was not characterized. Thus, 11 separate fractions were obtained in total (Table I) and were used for the sequence analysis as described below. Here, we describe the sequencing of the major component in each fraction and the sequencing data of the minor components have been provided as Supplementary data.
|
|
Sequencing of the oligosaccharides in fraction VIIi2
First, sequencing of the major component in fraction VIIi2 is described here in detail as a representative. Sequencing of the isolated oligosaccharides is achieved by digestion of 2AB-labeled oligosaccharide with chondroitinases ABC, AC-II, and AC-I. While chondroitinases ABC and AC-I cleave 2AB-labeled octasaccharide into one tetrasaccharide unit labeled with 2AB and two disaccharide units, chondroitinase AC-II cleaves 2AB-octasaccharide into four disaccharide units.
The sugar composition and the number of sulfate groups in fraction VIIi2 was determined by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF/MS) (Figure 5) using a basic peptide (Arg-Gly)15 to prepare a noncovalent complex (Juhasz and Biemann; 1995; Yamada et al. 1998). The molecular mass of the detected octasaccharide was calculated to be 1932.17 by subtracting the measured m/z value of the protonated peptide (m/z 3218) from that of the protonated peptide/octasaccharide complex (m/z 5152.17), which corresponded to a pentasulfated octasaccharide (Table II). Disaccharide composition analysis by chondroitinase AC-II digestion followed by 2AB-labeling of the released disaccharides showed 2AB-derivatives of C [GlcUA–GalNAc(6S)], 
C [HexUA–GalNAc(6S)], A [HexUA–GalNAc(4S)], and D [HexUA(2S)–GalNAc(6S)] as major products in a molar ratio of 1.0:1.3:1.1:1.1 (Table I), by anion-exchange HPLC, where HexUA stands for 4,5-unsaturated hexuronic acid. This indicates that the fraction VIIi2 is composed of one A-unit, one D-unit and two C-units, one of them residing at the nonreducing terminus.
|
|
To determine the sequential arrangement of these disaccharides in the octasaccharide, the octasaccharide was tagged with 2AB at the reducing terminus. The 2AB-labeled octasaccharide was divided into three portions and each portion (30 pmol) was digested with chondroitinase AC-II, ABC, or AC-I, separately, and the digest was analyzed by anion-exchange HPLC as outlined in Scheme 1. Chondroitinase AC-II degrades a 2AB-labeled octasaccharide into four disaccharides: one saturated disaccharide derived from the nonreducing terminus, two unsaturated disaccharides from internal positions and one unsaturated disaccharide tagged with 2AB from the reducing terminus (Kinoshita and Sugahara 1999
). In contrast, chondroitinases ABC and AC-I cleave a 2AB-labeled octasaccharide into one saturated disaccharide from the nonreducing terminus, one 2AB-labeled tetrasaccharide from the reducing terminus, and one unsaturated disaccharide from the internal position as will be discussed below in the last paragraph of this section. Hence, digestion with chondroitinase AC-II will reveal the reducing terminal disaccharide unit, while chondroitinases ABC and AC-I will give information about the reducing terminus as well as the disaccharide unit adjacent to the reducing terminus disaccharide unit. The chondroitinase AC-II digest of 2AB-labeled VIIi2 showed 2AB-labeled C (86.2%) as the major product from the reducing terminus (Table III). In addition, A (8.9%) was released from the reducing terminus derived from a minor compound in fraction VIIi2. 2AB-labeled A-D-C hexasaccharide marked by an asterisk (4.9%) was also detected (Figure 6C and Table III), which could be an incompletely digested product from the major octasaccharide.
|
|
|
Chondroitinase ABC digestion of fraction VIIi2-2AB released 2AB-labeled
D-C (70.7%) from the reducing terminus (Table III), indicating that the residue adjacent to the reducing terminal C-unit is a D-unit. C-C-2AB (21.6%) and C-A-2AB (7.7%) (Figure 6A) were released from the reducing termini of minor components (Table III), suggesting again that fraction VIIi2 was approximately 70% pure and contain at least two other pentasulfated octasaccharides. A chondroitinase AC-I digest of the 2AB-labeled VIIi2 (Figure 6B) showed a major peak identified as a 2AB-labeled hexasaccharide A-D-C (77.4%), with two minor peaks of C-C-2AB (14.1%), and C-A-2AB (8.5%). Here, it should be remembered that chondroitinase AC-I cannot cleave the galactosaminidic linkage bound to the "D" disaccharide unit (Yoshida et al. 1993), which probably caused the production of a hexasaccharide-2AB rather than a tetrasaccharide-2AB. Taken together, it was concluded that the sequence of the major compound in fraction VIIi2, which represents 70% of the total octasaccharides in fraction VIIi2, is C-A-D-C [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)ß1–4GlcUA(2S)ß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)] (Table IV). Additional supportive data for the major sequence and for sequencing of two minor octasaccharides in fraction VIIi2 are provided as Supplementary data.
|
Using the same strategy, fractions VIIa, c, d, e, f, g, h1, h2.2, and i1 were also sequenced. Di-, tetra- and hexa-saccharide sequences released by digestions with chondroitinases AC-II, ABC, and AC-I from the reducing termini of the octasaccharides in each fraction are summarized in Table III. Digestion of the 2AB-labeled oligosaccharides with chondroitinase AC-II generated an unsaturated disaccharide from the reducing terminus, except for fractions VIIh1 and VIIh2.2, which showed an unsaturated tetrasaccharide containing D-unit on the reducing side. In contrast, digestion of the 2AB-labeled oligosaccharide with either chondroitinase ABC or AC-I generated unsaturated tetrasaccharides for all the octasaccharide fractions and hexasaccharides in some instances such as fractions VIIi1 and VIIi2. Thus, the results are in good agreement with the previous findings (Kinoshita and Sugahara 1999
) that while chondrotinase ABC is unable to digest 2AB-derivatized tetrasaccharide into two disaccharide components, chondroitinase AC-II is able to do so. In this study, it was further demonstrated that chondroitinase AC-I cannot digest 2AB-derivatized tetrasaccharide, and that even chondroitinase AC-II cannot to do so when tetrasaccharides contain a "D" unit on the reducing end attached to 2AB. The minor components released by chondroitinase ABC, AC-II, and AC-I digestions of all the sequences are summarized in Table III and Supplementary Table SI.
Sequencing of the oligosaccharides in fraction VIIa
Disaccharide composition analysis of a chondroitinase AC-II digest showed C, O, and C in a molar ratio of 1.0:0.76:2.0 (Table I). Therefore, the major compound in fraction VIIa is composed of one O-unit and three C-units, one of the C-unit residing at the nonreducing terminus, although MALDI-TOF/-MS gave no signal for the peptide–oligosaccharide complex for fraction VIIa for unknown reasons (Table II). Chondroitinase AC-II digestion of 2AB-labeled fraction VIIa released O (75.7%) as a major product, while chondroitinase ABC and AC-I digestions released C-O from the reducing terminus (Table III). On the basis of these results and the data shown in Supplementary Table SI, it was concluded that the sequence of the major compound, which represents 80–81% of fraction VIIa, is C-C-C-O [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc] (Table IV). Data for sequencing of two minor octasaccharides in fraction VIIa are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIb
It was not possible to isolate the major component in fraction VIIb by rechromatography even with a shallower gradient of NaH2PO4. Therefore, fraction VIIb was derivatized with 2AB and analyzed by anion-exchange HPLC, which clearly separated the 2AB derivatives into three peaks in a molar ratio of 8:32:60 (Figure 3). The major component (60%), designated fraction VIIb3, was turned out to be C-C-C-C (data not shown), which is the major component in fraction VIIc (see Sequencing of the oligosaccharide in fraction VIIc). The second major component (32%), designated fraction VIIb2, was isolated and subjected to structural analysis. Anion-exchange HPLC of a chondroitinase AC-II digest of the 2AB-labeled fraction VIIb2 showed O to be the sole 2AB-labeled product derived from the reducing terminus of the compound (Table III). A portion of the digest of the 2AB-labeled fraction VIIb2 obtained with chondroitinase AC-II was derivatized again with 2AB to tag the resultant disaccharides derived from the nonreducing terminal and internal positions of the parent octasaccharide, and analyzed by anion-exchange HPLC after purification by paper chromatography. It yielded 2AB-derivatives of C, C, A, and O in a ratio of 1.00:1.11:0.84:1.16 (Supplementary Table SI). Thus, the compound in fraction VIIb2 consists of two C-units, one A-unit, and one O-unit, one of the C-units being located at the nonreducing terminus position. These results suggest that the major compound in fraction VIIb2 has an octasaccharide sequence of C-C-A-O or C-A-C-O.
Digestion of 2AB-labeled fraction VIIb2 with chondroitinase ABC or AC-I showed a 2AB-derivative of a monosulfated unsaturated tetrasaccharide (data not shown). A portion of each digest was also derivatized again with 2AB to tag the resultant disaccharides derived from the nonreducing terminus and internal positions of the parent octasaccharide. These digests showed indistinguishable chromatographic patterns, yielding 2AB-derivatives of C, C, and a monosulfated unsaturated tetrasaccharie in a molar ratio of 1.00:0.92:1.31 (Supplementary Table SI). The chondroitinase ABC or AC-I digest of 2AB-labeled fraction VIIb2 was cochromatographed with a chondroitinase AC-I digest of 2AB-labeled fraction VIIa, which contained a 2AB-derivative of C-O. However, the 2AB-labeled tetrasaccharide derived from fraction VIIb2 was not coeluted with the 2AB-derivative of C-O (results not shown), indicating that the tetrasaccharide structure of the reducing terminus of the compound in fraction VIIb2 is not C-O, but rather A-O. These results together suggest that the sequence of the major compound in VIIb2 is C-C-A-O [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)ß1–4GlcUAß1–3GalNAc]. Likewise, sequencing of the other minor octasaccharide in fraction VIIb, designated VIIb1, was carried out, and revealed the structure of the compound unambiguously as C-A-C-O [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc]. The data are provided as Supplementary data.
Sequencing of the oligosaccharide in fraction VIIc
MALDI-TOF/MS analysis indicated the average molecular weight to be 1854.46, corresponding to a tetrasulfated octasaccharide (Table II). Disaccharide composition analysis of a chondroitinase AC-II digest showed only C and C in a molar ratio of 1:2.3 (Table I), suggesting that the major compound in fraction VIIc is composed of four C-units. The analytical data obtained by digestion of 2AB-labeled fraction VIIc with chondroitinase AC-II, ABC, or AC-I showed 2AB-derivatives of C, C-C, or C-C, respectively (Table III). These results together suggest that the sequence of the major compound in VIIc is C-C-C-C [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)] (Table IV). It is reasonable that this sequence is the most abundant octasaccharide isolated from CS-C rich in C-unit (Table I).
Sequencing of the oligosaccharides in fraction VIId
MALDI-TOF/MS analysis suggested that the major compound is a tetrasulfated octasaccharide (Table II). Chondroitinase AC-II digestion of fraction VIId showed 2AB-labeled C, C, and A in a molar ratio of 1.0:1.9:0.81 (Table I). Anion-exchange HPLC of the chondroitinase AC-II digest of fraction VIId-2AB showed 2AB-labeled C (96%), while chondroitinase ABC or AC-I digest showed A-C-2AB (75–78%) (Table III) from the reducing terminus. These results suggest that the sequence of the major compound in fraction VIId is C-C-A-C [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)ß1–4GlcUAß1–3GalNAc(6S)] (Table IV). Data for sequencing of one minor octasaccharide in fraction VIId are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIe
MALDI-TOF/MS analysis suggested that the major compound is a tetrasulfated octasaccharide (Table II). Chondroitinase AC-II digest of fraction VIIe showed C, C, and A in a molar ratio of 1.0:1.7:0.86 (Table I), suggesting a C-unit at the nonreducing terminus and two additional C-units and one A-unit on the reducing terminal side. Chondroitinase AC-II digest of fraction VIIe-2AB showed 2AB-labeled A (96%), whereas the chondroitinase ABC or AC-I digest showed C-A-2AB (94%) (Table III) from the reducing terminus. On the basis of these results and the data shown in Supplementary Table SI, it was concluded that the sequence of the major compound in fraction VIIe, which represents 94% of fraction VIIe, is C-C-C-A [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)] (Table IV). Data for sequencing of one minor octasaccharide in fraction VIIe are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIf
MALDI-TOF/MS analysis suggested that the major compound is a tetrasulfated octasaccharide (Table II). Chondroitinase AC-II digest of fraction VIIf showed C, C, and A in a molar ratio of 1.0:1.2:1.6 (Table I), suggesting a C-unit at the nonreducing terminus, and one C-unit and two A-units at the internal position and/or reducing terminus. Chondroitinase AC-II digest of fraction VIIf-2AB showed 2AB-labeled A (79.4%), whereas chondroitinase ABC or AC-I digest showed A-A-2AB (88.5%) (Table III) from the reducing terminus. On the basis of these results and the data shown in Supplementary Table SI, it was concluded that the sequence of the major compound (88.5%) in fraction VIIf is C-C-A-A [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)ß1–4GlcUAß1–3GalNAc(4S)] (Table IV). Data for sequencing of one minor octasaccharide in fraction VIIf are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIg
Chondroitinase AC-II digestion of fraction VIIg showed 2AB-labeled D and C in a molar ratio of 1.0:3.3 (Table I) and MALDI-TOF/MS analysis suggested that the major compound is a pentasulfated octasaccharide, implying a D-C-C-C (Table II). "D" accounted for 59% of the resulting saturated disaccharides (Table I), reflecting the purity of the major component. Anion-exchange HPLC of the chondroitinase AC-II digest of fraction VIId-2AB released C-2AB (86.5%), whereas the chondroitinase ABC or AC-I digest released C-C-2AB (91–92%) (Table III) from the reducing terminal. On the basis of these results and the data presented in Supplementary Table SI, it was concluded that the sequence of the major compound representing 61–65% of fraction VIIg is D-C-C-C [GlcUA(2S)ß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)] (Table IV). Data for sequencing of three minor octasaccharides in fraction VIIg are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIh1
MALDI-TOF/MS analysis indicated a pentasulfated octasaccharide (Table II). Chondroitinase AC-II digestion of fraction VIIh1 showed 2AB-labeled C, C, and D in a molar ratio of 0.6:1.7:1.0 (Table I). The chondroitinase AC-II, ABC, and AC-I released 2AB-C-D (82–89%) from the reducing terminal (Table III). Unlike the reducing terminus tetrasaccharides described so far for other fractions, 2AB-derivatized C-D was resistant to the action of chondroitinase AC-II. On the basis of these results and the data shown in Supplementary Table SI, it was concluded that the sequence of the major compound (82–89%) in VIIh1 is C-C-C-D [GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUA(2S)ß1–3GalNAc(6S)] (Table IV). Data for sequencing of two minor octasaccharides in fraction VIIh1 are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIh2.2
MALDI-TOF/MS analysis suggested that the major compound is a pentasulfated octasaccharide (Table II). Unlike other fractions described so far, fraction VIIh2.2 turned out to be a mixture of at least two major compounds in a molar ratio of 0.8:1.0, which could be well separated by anion-exchange HPLC after 2AB-labeling (data not shown). Chondroitinase AC-II digest showed C, A, D, C, A, and D in a molar ratio of 0.52:0.31:0.17:1.15:0.33:0.64 taking the sum of C, A, and D as 1.00 (Table I). Anion-exchange HPLC of the chondroitinase AC-II digest of 2AB-labeled fraction VIIh2.2 released C-2AB (52.9%) and A-D-2AB (47.1%), while chondroitinase ABC or AC-I digest released C-C-2AB (38.9 or 36.6%), A-D-2AB (50.0 or 52.3%), and A-C (11.1%) from the reducing terminal (Table III). On the basis on these findings and the data shown in Supplementary Table SI, it was concluded that the sequences of the two major compounds in VIIh2.2 are C-C-A-D[GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)ß1–4GlcUA(2S)ß1–3GalNAc(6S)] (50–52%) and A-D-C-C [GlcUAß1–3GalNAc(4S)ß1–4GlcUA(2S)ß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)] (30–39%) (Table IV). Data for sequencing of one minor octasaccharide in fraction VIIh2.2 are provided as Supplementary data.
Sequencing of the oligosaccharides in fraction VIIi1
MALDI-TOF/MS analysis suggested that the major compound is a pentasulfated octasaccharide (Table II). Chondroitinase AC-II digest of fraction VIIi1 showed D, C, and A in a molar ratio of 1.0:2.2:1.2 (Table I). Chondroitinase AC-II digest of fraction VIIi1-2AB showed 2AB-labeled A (65.5%) as the major product, while chondroitinase ABC or AC-I digest showed C-A-2AB (76.3 or 61.9%) (Table III) from the reducing terminus. On the basis of these results and the data shown in Supplementary Table SI, it was concluded that the sequence of the major compound (62–76%) in VIIi1 is D-C-C-A [GlcUA(2S)ß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(6S)ß1–4GlcUAß1–3GalNAc(4S)] (Table IV). Data for sequencing of three minor octasaccharides in fraction VIIi1 are provided as Supplementary data.
Reactivity of structurally defined octasaccharides with CS-56, MO-225, 2H6, and LY111
Interactions between the structurally defined CS-C octasaccharides and anti-CS antibodies were studied using the oligosaccharide microarray method (Fukui et al. 2002). The lipid-linked oligosaccharides arrayed on a nitrocellulose membrane are effective probes, permitting sensitive and high-throughput detection of ligands. Four widely used commercial anti-CS antibodies (CS-56, MO-225, 2H6, and LY111) were selected for the present study. The mAb CS-56 was developed against the ventral membranes of gizzard fibroblasts and recognizes mainly CS-A and CS-C (Avnur and Geiger 1984). MO-225 was produced against the chick embryo limb bud PG-M and its determinant contains the D-unit (Yamagata et al. 1987). mAb 2H6 was raised against soluble CS-PG purified from 10-day-old rat brains (Oohira et al. 1994) and reacts with CS-A, CS-C and CS-D. LY111 was developed against chicken type IX collagen containing CS-A (Yada et al. 1992) and recognizes CS-A and CS-D. Recently, we demonstrated that CS-56 and MO-225 recognized A-D-containing sequences (Ito et al. 2005), and extended the study here. CS-56 reacts with octa- or larger oligo-saccharides, whereas MO-225 reacts with hexa- or larger oligo-saccharides as recently shown (Ito et al. 2005). In this study, 2H6 and LY111 were demonstrated to react with hexa- or larger oligo-saccharides derived from CS-C (Figure 7). It should be noted, however, that the reducing terminal GalNAc in each sequence has a linear form but not a ring form since it is conjugated with L-
-dipalmitoyl phosphatidyl ethanolamine (DPPE), and may be only partially recognized by the antibodies.
|
The reactivity of these mAbs was tested at three different concentrations (10, 5, and 2.5 pmol per spot) of each immobilized octasaccharide fraction. Although most of these fractions are not 100% pure (Table I), reactive sequences could be analyzed. CS-56 recognized only two fractions, VIIh2.2 (major components: C-C-A-D and A-D-C-C) and VIIi2 (a major component: C-A-D-C), at 10, 5, and 2.5 pmol of the tested octasaccharides (Figure 8A1 and 3), with faint reactivity at 2.5 pmol of the immobilized samples (Figure 8A3). The sequences common to the oligosaccharides in fractions VIIh2.2 and VIIi2 are A-D-C and C-A-D, whereas such a sequence is absent in the other fractions tested. Hence, it was concluded that CS-56 specifically recognizes the A-D-C and/or C-A-D sequence. The recognition of the A-D-C sequence is in agreement with the recent findings (Ito et al. 2005
) that CS-56 showed reactivity with various unsaturated octasaccharide sequences, C-A-D-C, A-A-D-C, A-A-D-A, and C-A-D-A in this order, which contained the A-D tetrasaccharide sequence in common. While it is clear that CS-56 reacts with A-D-C-containing octasaccharides, it remains to be clarified whether it can react with the C-A-D-containing octasaccharide sequences.
|
MO-225 specifically recognized only fraction VIIi2 (C-A-D-C) at all the tested concentrations (Figure 8B1, 2, and 3), whereas larger amounts (10 and 5 pmol) of fractions VIIh2.2 (major components: C-C-A-D and A-D-C-C) and VIIi1 (a major component: D-C-C-A) also reacted weakly (Figure 8B1 and 2). The compounds in fractions VIIh2.2 and VIIi2 share the A-D-C and C-A-D sequences, being consistent with the recognition of the A-D-containing sequences as reported (Ito et al. 2005
). The reactivity of MO-225 with fraction VIIi1 is probably not due to the main component D-C-C-A without the A-D sequence, but is likely to be due to the minor components containing the A-D sequence. Since fraction VIIi1 is only 62–76% pure and may contain neighboring fractions VIIh2.2 and VIIi2 as contaminants (Figures 2, 4B and C), they may explain the reactivity of fraction VIIi1 with MO-225. Indeed, the presence of a minor octasaccharide sequence C-A-D-C was deduced (Supplementary data and Table IV). The reactivity of MO-225 with fraction VIIi2 was so strong that it reacted even with 2.5 pmol of fraction VIIi2 (Figure 8B), which is consistent with the strong reactivity with the unsaturated counterpart reported previously (Nadanaka et al. 1998). Although MO-225 recognizes not only the A-D sequence but also an E-D sequence derived from shark fin (Yamagata et al. 1987), this fraction contains five sulfate groups and should not contain the E-D sequence. The reactivity of 2H6 with the CS-C octasaccharides indicated that the antibody cross-reacted with 10 pmol of all the fractions except fractions VIIb and VIIf (Figure 8C1). On decreasing the concentration of the samples to 5 pmol, only fraction VIId, which contains C-C-A-C and C-C-C-C sequences, strongly reacted (Figure 8C2), and this reactivity was detectable even with 2.5 pmol of the oligosaccharide (Figure 8C3). The recognition of the multiple sequences at high doses by 2H6 is probably because 2H6 binds hexa- or larger saccharides (Figure 7), and presumably recognizes the conformations of a C-C-containing hexasaccharides common to the positive octasaccharides (Table V). Hence, the weak reactivity of fraction VIIi2 may be due to the minor components C-E-C-C and/or D-C-C-A, but not due to the major component C-A-D-C. Further investigation along this line remains to be carried out.
|
Likewise, LY111 also cross-reacted with multiple sequences at larger amounts of the oligosaccharides. It strongly reacted with 10 pmol of fractions VIIb, VIId, and VIIf, whereas many other fractions showed slight reactivity (Figure 8D1). Upon decreasing the amounts of the oligosaccharides to 5 pmol, it still showed strong reactivity with fractions VIIb, VIId, and VIIf, while the reactivity with the other fractions almost disappeared (Figure 8D2). At 2.5 pmol, only fraction VIIf showed the reactivity. The sequence of the major compound in fraction VIId is C-C-A-C, and that in fraction VIIf is C-C-A-A. Those of the major compounds in fraction VIIb are C-C-C-C (60%) and C-C-A-O (32%). The reactivity of fraction VIIb is most likely to be attributable to C-C-A-O, because fraction VIIc, which contains C-C-C-C as the major component, did not react with LY111. The reactivity of C-C-A-O itself could not be tested due to difficulty in its isolation (see Sequencing of the oligosaccharides in fraction VIIb). The sequence common to the compounds in fractions VIId and VIIf is C-C-A. Even though such a sequence is also present in fractions VIIe, VIIh2.2, and VIIi1, these fractions were not recognized by LY111, indicating that C-C-A is not the epitope sequence. Rather, C-A-O, C-A-C, and C-A-A, which reside on the reducing terminal sides of the major compounds in fractions VIIb, VIId, and VIIf, may form similar conformations without being adjacent to the D-unit and be recognized, with C-A-A being more reactive than C-A-C or C-A-O as indicated by the intensity of the reaction (Figure 8D3).
Comparison of the specificity of the tested antibodies clearly suggests that CS-56 and MO-225 require the D-unit in addition to the C- and A-unit in certain sequential order (see footnotes to Table V), while an octasaccharide devoid of D-unit was not recognized by either antibody. The present findings obtained with naturally occurring saturated CS-C octasaccharide sequences, which were prepared by hydrolytic hyaluronidase, confirmed the previous findings obtained with unsaturated CS-D octasaccharides prepared with eliminative chondroitinase. Namely, the main reactive structure for both CS-56 and MO-225 is A-D, the reactivity of which is differentially modified by adjacent units (Ito et al. 2005). On the contrary, 2H6 and LY111 prefer sequences with A- and C-units, while the D-unit is not a requirement for the antibody recognition. The results from the specificity studies altogether suggest that multiple sequences are recognized by each antibody, implying that the common conformation shared by the multiple primary sequences in the intact CS chains is important as an epitope. Recent studies (Blanchard et al. 2007) on the conformation of CS-D-derived four unsaturated octasaccharides, C-A-D-C, C-A-D-A, A-A-D-C, and A-A-D-A (Table V), which are recognized by MO-225 and CS-56, using 1H NMR and molecular modeling have revealed the importance of the 2-O-sulfate group of GlcUA and an exo-cyclic negative tail in C and C disaccharides on the core A-D tetrasaccharide sequence, and support the above hypothesis.
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataConflict of interest statementAcknowledgmentsReferences |
|---|
The 12 octasaccharide fractions isolated in this study unexpectedly contained as many as 24 different sequences, exhibiting a high degree of structural diversity reflecting the enormous heterogeneity of the parent CS-C polysaccharides of shark cartilage, which are composed of major disaccharide units C (73.8%), A (15.7%), and D (8.8%) with minor O (1.0%) and E (0.7%) (Kinoshita and Sugahara 1999
). Although the unsaturated counterparts of the sequences including C-C-C-C (fraction VIIc), C-C-C-A (fraction VIIe), D-C-C-A (fraction VIIi1), and C-A-D-C (fraction VIIi2) were previously isolated after partial digestion of CS-D with chondroitinase ABC (Chai et al. 1996, 1998; Nadanaka et al. 1998), the other 19 sequences are novel. Even though a small proportion of O-unit has been reported (Yamagata et al. 1987), O-unit-containing sequences such as the octasaccharides in fractions VIIa, VIIb1, and VIIb2 were isolated for the first time. The sequences isolated in this and previous studies (Chai et al. 1996, 1998; Sugahara, Nadanaka, et al. 1996; Sugahara K, Tanaka, et al. 1996; Nadanaka and Sugahara 1997; Nadanaka et al. 1998) support the notion that the sequences obtained after hyaluronidase digestion are, largely if not exclusively, the natural sequences rather than transglycosilation reaction products even though the hyaluronidase exhibits transglycosylation activity (Weissmann 1955; Highsmith et al. 1975) as discussed previously (Nadanaka and Sugahara 1997). Oversulfated CS variants, CS-D and CS-E, do not serve as donor substrates for transglycosylation reactions under the established optimum conditions (Saitoh et al. 1995; Takagaki and Endo 2006), and therefore the sequences containing a D- or E-unit are probably natural sequences. In addition, CS-C is an inefficient donor substrate as compared with nonsulfated chondroitin and CS-A (Takagaki and Endo 2006). Hence, the sequences with a C-unit at the nonreducing terminus are most likely naturally occurring, although a the possibility cannot be completely excluded that a few minor sequences such as those containing an O-unit were formed by transglycosylation reactions.
In fractions VIIa, b, c, d, e, and f, which were devoid of a D-unit, the nonreducing terminal residue was a C-unit, while a C-, A-, or O-unit constituted the reducing terminus, most likely reflecting the specificity of hyaluronidase. While the presence of consecutive C-units is a common feature of many of the isolated oligosaccharides, consecutive A-units were also found even in CS-C. In contrast, no consecutive D-units were observed as in the case of the octasaccharides isolated from CS-D (Nadanaka et al. 1998). Either a C- or an A-unit was found on the immediate nonreducing terminal side of the D-unit, forming a C-D or A-D sequence, whereas a C-unit was often found on the immediate reducing side of a D-unit, generating a D-C sequence, and an A-unit was also detected on this side but only in fraction VIIh2.2, forming a relatively rare D-A sequence. While the A-D tetrasaccharide sequence was a common feature of the CS-D octasaccharides (Nadanaka et al. 1998), it was detected only in relatively minor sequences in fractions VIIg, VIIh2.2, and VIIi2 from CS-C. Thus, CS-C-derived sequences share some structural features with CS-D-derived oligosaccharides but are considerably different from the latter, suggesting similar yet distinct structural features of the parent CS-C and CS-D chains that should reflect the difference in the assembly of the biosynthetic machinery of the tissues.
Sequencing of the octasaccharides was accomplished after derivatization with 2AB, as reported earlier for hexasaccharides from CS and heparan sulfate (Kinoshita and Sugahara 1999). The results unveiled the substrate specificity of chondroitinases towards 2AB-derivatized oligosaccharides. Chondroitinase ABC cleaved all the 2AB-derivatized octasaccharides into two disaccharide units and a 2AB-tagged tetrasaccharide, which is resistant to chondroitinase ABC, being consistent with previous findings (Kinoshita and Sugahara 1999). Chondroitinase AC-I, which was not used in the previous study (Kinoshita and Sugahara 1999), also digested each 2AB-derivatized octasaccharide into two disaccharides and a 2AB-labeled tetrasaccharide, except for C-A-D-C in fraction VIIi2, which was digested into one disaccharide and a 2AB-labeled hexasaccharide. This particular hexasaccharide was resistant to the endolytic action of chondroitinase AC-I probably due to the D-unit. In contrast, chondroitinase AC-II cleaved most of the 2AB-derivatized octasaccharide into four disaccharide units as reported (Kinoshita and Sugahara 1999). Interestingly, however, 2AB-derivatives of C-C-C-D in fraction VIIh1 and C-C-A-D in fraction VIIh2.2 were digested into two disaccharide units and a 2AB-labeled tetrasaccharide, suggesting that a 2AB-lebeled tetrasaccharide with a D-unit at the reducing terminus is resistant to the action of chondroitinase AC-II due to the D-unit. These properties of the chondroitinases provide useful information for sequencing. Our strategy, which requires less than 100 pmol of a purified oligosaccharide and a few days work, can sequence even an incompletely purified fraction containing two to four structurally similar yet distinct octasaccharides (Table IV). The strategy has recently been successfully applied to a series of octa- and deca-saccharides isolated from squid cartilage CS-E (Deepa et al. 2007). However, it needs improvements for sequencing oligosaccharides containing iduronic acid by using chondroitinase B or those containing rare GlcUA(3-O-sulfate) by developing a novel cleavage method, since such oligosaccharides cannot be digested by chondroitinase AC-I or AC-II, but can be degraded by chondroitinase ABC (Sugahara, Tanaka, et al. 1996; Sugahara and Yamada 2000).
The oligosaccharides isolated in this study were successfully utilized to obtain information on the structure of the oligosaccharide epitopes of the anti-CS mAbs, which have been widely used for the immunohistochemical localization of CS-PGs. For example, mAbs including CS-56, 2H6, and MO-225 have been used for studying a receptor-type protein-tyrosine phosphatase 
and its major secreted soluble form (phosphacan), which are abundantly expressed in the brain as a CS-PG (Maeda et al. 1992, 1994; Maurel et al. 1994, Garwood et al. 1999). Immunohistochemical studies using these mAbs demonstrated that the CS structure of phosphacan changes dramatically during mouse brain development, with the 2H6 and CS-56 epitopes highly expressed in the cerebral cortex of the postnatal day 7 (P7) mice, but with the MO-225 epitope intensely expressed in the cerebellum (Maeda et al. 2003). The requirement of a D-unit for MO-225 reactivity has been established (Yamagata et al. 1987; Ito et al. 2005) and correlates with the presence of a D-unit in the P20 mouse brain (Maeda et al. 2003). Our findings indicate that MO-225 specifically recognizes the C-A-D-C sequence among other sequences, and the same sequence is also recognized by CS-56, which, however, also reacted with C-C-A-D and/or A-D-C-C more strongly than MO-225 (Table V). This difference partly explains the different staining patterns of the postnatal mouse brain section with these mAbs (Maeda et al. 2003). The present findings also strengthen the concept that multiple sequences with similar conformations can be recognized by these antibodies. Hence, not only the primary sequence but also the conformation adopted by the epitope in the intact CS chain also appears to play a pivotal role in the antibody recognition.
Even though 2H6 reacts with whale cartilage CS-A (Oohira et al. 1994), immunohistochemical staining of the postnatal brain with 2H6 has indicated that there is no direct correlation between the content of A-unit in the phosphacan preparation and their reactivities with 2H6. PG from P7 with the lowest content of A-unit (64%) exhibits a higher reactivity with 2H6 than PG from P20 with a higher content of A-unit (86%). Since CS-A from whale cartilage has appreciable proportions of C-unit (25%) along with A-unit (72%) (Kinoshita and Sugahara 1999), it probably contains the sequence C-C-A-C (VIId) strongly recognized by 2H6.
Immunohistochemical staining of sheep forestomach mucosa with CS-56 and 2H6 has indicated that they give similar staining patterns (Yamamoto et al. 1995). Preabsorption studies of CS-56 have indicated that the staining with CS-56 is completely abolished by a low concentration (0.5 mg/mL) of CS-C and CS-D, and a much higher concentration (8 mg/mL) of CS-A, but is not affected by CS-E, indicating that CS-56 recognizes sequences common to CS-C and CS-D, while such structures are not so abundant in CS-A and are absent from CS-E. The reactivity appears to reflect the abundance of the A-D-containing sequences (Table V) recognized by CS-56. While A- and C-units are present in CS-A, CS-C, CS-D, and CS-E, albeit in varying proportions, the D-unit is present only in CS-A, CS-C, and CS-D, but not in CS-E. On the other hand, preabsorption studies of 2H6 have indicated that the epitope structure is present in all the four CS variants with higher concentrations in CS-C and CS-D than in CS-A and CS-E, being in good agreement with our finding that the major epitope structure of 2H6 is C-C-A-C.
GAG chains are not random polymers of disaccharides, but consist of domains that are assembled in specific combinations. Characterization of the carbohydrate ligands for their interaction with proteins is a challenge because of the structural diversity present in a single molecule. This task was made easier in this study by obtaining the relatively pure octasaccharides with defined sequences, with the help of the sensitive technology that allows the detection of samples at the picomole level (Fukui et al. 2002; Ito et al. 2005). Most of the isolated fractions contained a few minor components, reflecting the difficulty of the isolation of relatively long pure sequences from CS chains with a high degree of heterogeneity; the proposed epitope structures should be confirmed using pure synthetic oligosaccharides, as recently exemplified by the synthetic E-E terasaccharide used for confirming neurite outgrowth-promoting activity (Borman 2004; Tully et al. 2004). They will also be useful probes to investigate the recognition sequences for CS-binding functional proteins including growth factors and cytokines (Sugahara et al. 2003).
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsSupplementary dataConflict of interest statementAcknowledgmentsReferences |
|---|
Materials
CS-C derived from shark cartilage (super special grade), CS-disaccharide standards, chondroitinase ABC from Proteus vulgaris (EC 4.2.2.4 [EC] ), chondroitinase AC-I from Flavobacterium heparinum (EC 4.2.2.5 [EC] ), chondroitinase AC-II from Arthrobacter aurescens (EC 4.2.2.4 [EC] ), and the mAbs CS-56, MO-225, 2H6, and LY111 were purchased from Seikagaku Corp., Tokyo, Japan. Sheep testicular hyaluronidase (EC 3.2.1.35 [EC] ) and gentisic acid were obtained from Sigma (St. Louis, MO). Bio-Gel P-10 resin and nitrocellulose membrane (Trans-Blot® Transfer membrane, 0.45 µm) were obtained from Bio-Rad Laboratories (Hercules, CA). Bovine serum albumin (BSA Fraction V, chemical grade) was from Serological Proteins, Inc. (Kankakee, IL). Peroxidase-conjugated AffiPure goat anti-mouse IgG and IgM (H and L) were purchased from Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA). A peptide (Arg-Gly)15 was custom synthesized by Peptide Institute, Inc. (Osaka, Japan; Yamada et al. 1998
).
Preparation of oligosaccharide fractions by digestion of CS-C with hyaluronidase
A commercial CS-C of shark cartilage (100 mg) was digested with 10 mg (approximately 15 000 National formulary units) of sheep testicular hyaluronidase in a total volume of 2.0 mL of 50 mM sodium phosphate buffer, pH 6.0, containing 150 mM NaCl (1 National formulary unit corresponds to the amount of enzyme that hydrolyzes 74 µg of hyaluronate/min) (Poh et al. 1992; Sugahara et al. 1992) at 37°C. Exhaustive digestion of the sample was observed at 18 h, but the reaction was allowed to proceed for 36 h to achieve the complete digestion. Proteins were precipitated by the addition of 0.42 mL of 30% trichroloacetic acid and removed by c