Molecular Cloning and Characterization of a Human Uronyl 2-Sulfotransferase That Sulfates Iduronyl and Glucuronyl Residues in Dermatan/Chondroitin Sulfate*

A partial-length human cDNA with a predicted amino acid sequence homologous to a previously described heparan sulfate iduronyl 2-sulfotransferase (Kobayashi, M., Habuchi, H., Yoneda, M., Habuchi, O., and Kimata, K. (1997) J. Biol. Chem. 272, 13980–13985) was obtained by searching the expressed sequence-tagged data bank. Northern blot analysis was performed using this homologous cDNA as a probe, which demonstrated ubiquitous expression of messages of 5.1 and 2.0 kilobases in a number of human tissues and in several human cancer cell lines. Since the human lymphoma Raji cell line had the highest level of expression, it was used to isolate a full-length cDNA clone. The full-length cDNA was found to contain an open reading frame that predicted a type II transmembrane protein composed of 406 amino acid residues. The cDNA in a baculovirus expression vector was expressed in Sf9 insect cells, and cell extracts were then incubated together with 3′-phosphoadenosine 5′-phospho[35S]sulfate and potential glycosaminoglycan acceptors. This demonstrated substantial sulfotransferase activity with dermatan sulfate, a small degree of activity with chondroitin sulfate, but no sulfotransferase activity with desulfated N-resulfated heparin. Analysis of [35S]sulfate-labeled disaccharide products of chondroitin ABC, chondroitin AC, and chondroitin B lyase treatment demonstrated that the enzyme only transferred sulfate to the 2-position of uronyl residues, which were preponderantly iduronyl residues in dermatan sulfate, but some lesser transfer to glucuronyl residues of chondroitin sulfate.

Dermatan sulfate is a glycosaminoglycan polysaccharide consisting of N-acetylgalactosamine (GalNAc) residues alternating with varying proportions of glucuronyl (GlcA) 1 and iduronyl (IdceA) residues that are formed from the GlcA by epimerization during polymerization and GalNAc 4-sulfation (1)(2)(3)(4). Thus dermatan sulfate can be considered as a variant of chondroitin 4-sulfate, containing some IdceA as well as GlcA, with the IdceA only found next to 4-sulfated GalNAc residues (5). In addition the IdceA of dermatan sulfate is frequently 2-sulfated (6). Some 2-sulfation of GlcA on chondroitin sulfate has also been found but only next to GalNAc 6-sulfate rather than GalNAc 4-sulfate. Proteoglycans containing dermatan sulfate are ubiquitously present in most tissues, where the dermatan sulfate portion may be involved in various biological activities presumably relating in great part to its fine structure. Activities include interaction with heparin cofactor II (7,8) requiring repeating 2-sulfated iduronyl-containing disaccharide units (7), hepatocyte growth factor/scatter factor (9), and promotion of fibroblast growth factor-2 during wound repair (10). Although there is little information concerning detailed biological activities based on the structural diversity of the dermatan sulfate, the 2-sulfation of IdceA would appear to be of special interest.
The only galactosaminoglycan sulfotransferase that has been cloned to date is a chondroitin 6-sulfotransferase (11). However, several glucosaminoglycan sulfotransferases have been cloned (12)(13)(14)(15)(16)(17)(18)(19), including an IdceA 2-sulfotransferase for heparan sulfate (15,20). It seemed likely that this enzyme would have similarities to IdceA 2-sulfotransferase for dermatan sulfate. Therefore, in an attempt to find such an IdceA 2-sulfotransferase, we employed the heparan sulfate IdceA 2-sulfotransferase sequence to obtain a related expressed sequence-tagged (EST) clone. This provided for the molecular cloning of a human cDNA which we found to encode a uronyl 2-sulfotransferase. We have found this enzyme to have no 2-sulfotransferase activity with heparan sulfate but to be involved in the sulfation of the IdceA residues of dermatan sulfate with some lesser activity in 2-sulfation of GlcA residues in chondroitin sulfate. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB020316.
‡ ‡ To whom correspondence and reprint requests should be addressed: Massachusetts Institute of Technology, 68-480, 77 Massachusetts Ave., Cambridge, MA 02139. Tel.: 617-253-8803; Fax: 617-258-6553. 1 The abbreviations used are: GlcA, D-glucuronic acid; PAPS, 3Јphosphoadenosine 5Ј-phosphosulfate; CHO, Chinese hamster ovary; EST, expressed sequence-tagged; PCR, polymerase chain reaction; SSC, sodium citrate-sodium chloride; HPLC, high performance liquid chromatography; CDSNS-heparin, completely desulfated and N-resulfatedheparin; ⌬HexA, 4-deoxy-␣-threo-hex-4-enepyranosyluronic acid; ⌬Di- Biotechnology Information (NCBI) data bank of I.M.A.G.E. Consortium (Lawrence Livermore National Laboratory) EST cDNA clones (21) was probed with a deduced CHO cell heparan sulfate 2-sulfotransferase sequence (15). Clone ID HE9MJ06 was obtained from the TIGR/ATCC Special Collection (ATCC). This was a partial-length EST clone from a 9-week-old human embryo, in which only 248 bp of sequence (positions 473-720, Fig. 1) was present in the data base (22). The cDNA from this clone was inserted into the EcoRI and XhoI sites of the Bluescript SKϪ plasmid and carried in an Escherichia coli host strain. The pBluescript plasmid DNA was purified from the host bacteria using QIAfilter plasmid kits (Qiagen) and used to prepare a PCR probe consisting of a 614-bp fragment at positions 473-1086 as shown in Fig. 1 using 20-bp oligonucleotides of both ends as primers. The PCR was carried out in a 50-l volume containing 0.5 M each primer, 40 ng of the template, 0.2 mM each dNTP, and 1.25 units of Taq2000 TM DNA polymerase (Stratagene) under the conditions of 30 cycles of denaturation at 94°C for 1 min, annealing at 56°C for 1 min, and extension at 72°C for 1 min. The products were subjected to agarose gel electrophoresis, and an amplified DNA band of ϳ600 bp was excised, recovered from the gel using QIAEX II (Qiagen), and radiolabeled for the probe using [␣-32 P]dATP and a Prime-It II random primer labeling kit (Stratagene).
Northern Blot Hybridization-To obtain full-length clones, we first scanned tissue-specific and cell type-specific expression of the cDNA by Northern blot hybridization using the 32 P-labeled PCR probe. Tissue type-specific expression of presumptive 2-sulfotransferase cDNA was analyzed with human RNA Master Blot TM and human Multiple Tissue Northern (MTN TM ) Blot (CLONTECH) membranes and cell type-specific expression with a human cancer cell line Multiple Tissue Northern (MTN TM ) Blot (CLONTECH) membrane. Each Northern analysis was carried out according to the manufacturer's protocol with some modifications. The Master Blot membrane was prehybridized in an Ex-pressHyb solution containing 100 g/ml denatured sheared salmon testes DNA (Sigma) at 65°C for 30 min, hybridized in the same solution containing the denatured 32 P-labeled probe, 6 g/ml human Cot-1 DNA (Boehringer Mannheim), and 0.2ϫ SSC at 65°C for 16 h, and washed four times with 2ϫ SSC, 1% SDS at 65°C for 20 min, twice with 0.1ϫ SSC, 0.5% SDS at 55°C for 20 min. The human and cancer cell line MTN Blot membranes were prehybridized in an ExpressHyb at 68°C for 30 min, hybridized in the same solution containing the 32 P-labeled probe at 68°C for 16 h, and washed twice with 2ϫ SSC, 0.05% SDS at 22°C for 20 min, twice with 0.1ϫ SSC, 0.1% SDS at 50°C for 20 min. The membranes were then exposed to x-ray film with an intensifying screen at Ϫ80°C.
Screening of cDNA Library-The human lymphoma 5Ј-STRETCH PLUS cDNA library (CLONTECH) was constructed from mRNA from Burkitt's lymphoma Raji cell line at EcoRI-cloning sites of the gt11 vector by the priming method with oligo(dT) and random primers. The host strain Y1090rϪ cells were infected with phage from the library, plated at 4 ϫ 10 5 plaque-forming units/dish, and approximately 1.2 ϫ 10 6 plaques were screened. Colony/Plaque screen TM (NEN Life Science Products) membrane replicas of the plaques were fixed by the rapid autoclave method recommended by the manufacturer, prehybridized in an ExpressHyb TM hybridization solution (CLONTECH) for 30 min at 68°C, and hybridized in the same solution containing the denatured 32 P-labeled probe at 68°C according to the manufacturer's protocol with a 16-h modification. The filters were washed twice with 2ϫ sodium citrate-sodium chloride (SCC), 0.05% sodium dodecyl sulfate (SDS) for 20 min at 22°C, and then twice with 0.1ϫ SCC, 0.1% SDS for 20 min at 50°C. The positive clones were detected by autoradiography.
Characterization of cDNA Clones-Plaque solutions from the positive clones were initially characterized with LD-insert screening amplimer sets (CLONTECH) according to the manufacturer's PCR protocol. The resultant PCR products were subjected to agarose gel electrophoresis to determine the sizes of the inserts, recovered from the gel, and sequenced. gt11 DNA of the clones was isolated from its plate lysate using a Qiagen lambda kit (Qiagen), subcloned into pcDNA3 vector (Invitrogen) at the EcoRI sites, and sequenced again to confirm sequence data. For DNA sequencing, the 5Ј and 3Ј insert regions were enzymatically sequenced from flanking primer sites of the respective PCR fragments or vectors. The remaining sequences of both strands were obtained with internally priming oligonucleotides. Primers were spaced no more than 400 bp apart with a 200-bp offset between sense and antisense strands. Automated fluorescence sequencing was performed with Perkin-Elmer Applied Biosystems models 373A and 477 DNA Sequencers. The DNA sequence files obtained were aligned and compiled with Sequencher (Gene Codes Corp.) and GENETYX-MAC (Software Development Corp.) computer programs. Sequence comparison searches were performed on the data bases of GenBank TM , EMBL, PDB, SwissProt, SPupdate, PIR, and dbEST.
Construction of Baculovirus Expression Vector-The PvuII-EcoRI fragment containing the coding region from positions 193 to 1,382 shown in Fig. 1 was excised from the gt11 cloning vector, blunted with T4 DNA polymerase, and ligated into the StuI site of the pFASTBAC TM HTa plasmid (Life Technologies, Inc.). The recombinant bacmid-sulfotransferase (presumptive) molecules were then produced by Tn7-mediated site-specific transposition when MAX EFFICIENCY DH10BAC TM competent cells (Life Technologies, Inc.) were transformed with the recombinant pFASTBAC HTa donor plasmid according to the manufacturer's instructions. The recombinant molecules were isolated and analyzed by agarose gel electrophoresis and PCR with vector and genespecific primers to confirm the presence of bacmid high molecular weight DNA and the correct orientation of the inserted cDNA. Recombinant bacmid-heparan sulfate 2-sulfotransferase DNA that contained the entire coding region of its cDNA (SacII-AflII fragment) (15) was also constructed and used for control experiments.
Expression of cDNA-cDNA was expressed using a BAC-TO-BAC TM HT baculovirus expression system (Life Technologies, Inc.) according to the manufacturer's instructions with slight modifications. Sf9 insect cells (Invitrogen) seeded onto 35-mm culture dishes containing 2 ml of SF-900 II SFM (Life Technologies, Inc.) were transfected with bacmidsulfotransferase (presumptive) or bacmid-heparan sulfate 2-sulfotransferase using CELLFECTIN TM reagent (Life Technologies, Inc.). The medium was replaced with Grace's insect medium (Invitrogen), 10% fetal bovine serum (JRH Biosciences), and the culture was continued for another 3 days at 27°C. The spent medium was centrifuged for 5 min at 500 ϫ g to obtain the virus-containing supernatant as viral stock. 150-mm Petri dishes of Sf9 cells were then infected by each recombinant viral stock and incubated at 27°C for 3 days. After collecting the spent medium, the 150-mm Petri dishes of infected Sf9 cells were washed with phosphate-buffered saline, scraped, and homogenized in a 3-ml solution of 10 mM Tris-HCl, pH 7.4, 0.5% (w/v) Triton X-100, 0.15 M NaCl, 10 mM MgCl 2 , 2 mM CaCl 2 , 20% (v/v) glycerol, and a mixture of protease inhibitors (5 M N␣-p-tosyl-L-lysine chloromethyl ketone, 3 M N-tosyl-L-phenylalanine chloromethyl ketone, 30 M phenylmethylsulfonyl fluoride, and 3 M pepstatin A) as described previously (15,20). After 1 h of gentle stirring at 4°C, the homogenate was centrifuged at 4°C for 30 min at 10,000 ϫ g. Sulfotransferase activities in the supernatant fractions (cell extracts) were measured as described below. Protein contents of the cell extracts were estimated by a micro-BCA protein assay reagent kit (Pierce) using bovine serum albumin as a standard.
Assay for Sulfotransferase Activity-Completely desulfated and Nresulfated heparin (CDSNS-heparin) and shark cartilage chondroitin sulfate C (4-sulfate:6-sulfate, 10:90) were obtained from Seikagaku; porcine skin dermatan sulfate was obtained from Sigma. Chemical desulfation (23) was used to obtain dermatan and chondroitin, which resulted in apparent complete desulfation of the chondroitin but left small amounts of 4-sulfate on the dermatan. A standard reaction mixture (25 l) contained 1.25 mol of imidazole HCl, pH 6.8, 1.88 g of protamine chloride, 12.5 g of glycosaminoglycan, 0.5 nmol (2.5 Ci/ nmol) of 3Ј-phosphoadenosine 5Ј-phospho[ 35 S]sulfate (NEN Life Science Products, Inc.), and 5 l of the cell extract. After incubation at 37°C for 30 min, the reaction mixtures were directly spotted on Whatman No. 1 paper and chromatographed in ethanol, 1 M ammonium acetate (5:2 (v/v)) overnight. The origins, which contained the sulfated products, were assayed for radioactivity as described previously (24).
Structural Analysis of 35 S-Labeled Products-In order to obtain sufficient labeled products for detailed analyses, higher specific activity 5Ј-phospho[ 35 S]sulfate (ϳ150 Ci/nmol) prepared as described previously (25) was used. After phenol:chloroform:isoamyl alcohol (25:24:1) treatment and subsequent ethanol precipitation, the 35 S-labeled glycosaminoglycans were digested with protease-free chondroitin ABC lyase (20 milliunits/g substrate), chondroitin AC I lyase (10 milliunits/g substrate), or chondroitin B lyase (100 milliunits/g substrate) (Seikagaku Corp.) for 16 h at 37°C (30°C for B lyase) and boiled at 100°C for 1 min to terminate the reaction. Products (ϳ9,000 cpm) were applied on Bio-Gel P-2 (Bio-Rad) columns (0.75 ϫ 200 cm) that were equilibrated and eluted with 0.1 M ammonium bicarbonate at a flow rate of 4 ml/h and assayed for radioactivity. Aliquots of the boiled ABC lyase degradation mixtures were then incubated for an additional 16 h at 37°C with (a) ⌬ 4,5 hexuronate-2-sulfatase (10 milliunits/g substrate ferases the NCBI Data Bank of EST cDNA clones was probed with the deduced amino acid sequence of CHO cell heparan sulfate IdceA 2-sulfotransferase cDNA (15). As described under "Experimental Procedures," a human partial-length cDNA clone, Clone ID HE9MJ06, was found, encoding a novel related species. The cDNA from this clone was 3,743-bp (positions 473-4215) in length as shown in Fig. 1. The 614-bp (positions 473-1086) PCR probe for library screening and Northern hybridization was generated as described under "Experimental Procedures." Approximately 1.2 ϫ 10 6 plaques of a gt11 human lymphoma Raji cell cDNA library were screened using this PCR fragment as a probe, resulting in 25 positive clones. Sixteen insert cDNAs of these clones were selected, amplified, and sequenced as described under "Experimental Procedures," but only one (1.4 kb) appeared to have the complete coding sequence of the presumptive sulfotransferase.
The amino-terminal sequence of this clone was found to contain three in-frame ATG codons and a TGA stop codon in frame at position 56 upstream from the first ATG codon. A single open reading frame beginning at the first ATG codon predicted a protein of 406 amino acid residues with a molecular mass of 47,672 Da with five potential N-linked glycosylation sites. Hydropathic analysis (28) of the predicted amino acid sequence of the presumptive sulfotransferase revealed that it had a prominent type II membrane protein hydrophobic segment in the amino-terminal region, 18 residues in length at positions 48 -65 ( Figs. 1 and 2).
Comparison of the sequence of this human presumptive sul-fotransferase with CHO cell heparan sulfate IdceA 2-sulfotransferase (GenBank TM accession number D88811) (15) revealed ϳ30% identity and ϳ50% similarity at the amino acid level. In particular, extensive homology existed across 210amino acid residues from 114 to 320 for a consensus sequence in the middle region of these enzymes, which included the 5Ј-phosphosulfate binding motif, its catalytic Lys, and 3Ј-phosphate binding motif corresponding to the reports of new algorithms using PAPS on substrates (29 -31) (Fig. 3). There was no significant homology at the nucleotide level. In addition there was considerable homology and identity with a protein from Caenorhabditis elegans (GenBank TM accession number Z81479) (32) and Drosophila melanogaster segregation distorter protein (GenBank TM accession number P25722) (33) (Fig. 3). It shared little overall sequence similarity and no common sequence elements with any glycosaminoglycan sulfotransferase previously reported (11-14, 16 -19) other than the heparan sulfate IdceA 2-sulfotransferase (15), indicating that it was also most likely an IdceA 2-sulfotransferase. (Table I)  Pileup was used to align amino acid sequence of the presumptive sulfotransferase (DS2ST) with sequences from heparan sulfate IdceA 2-sulfotransferase (HS2ST), C. elegans (Ce), and D. melanogaster (Dm). Consensus residues (shaded) are indicated for each position where at least three candidates exhibit identical or similar amino acids. Numeration is given for each and for a consensus sequence. The 5Ј-phosphosulfate binding motif (5ЈPSB) and 3Ј-phosphate binding motif (3ЈPB) for PAPS (29) are labeled, and the catalytic lysines are shown with a black dot (30). membrane analysis (Fig. 4A) demonstrated a major band of 5.1 kb and a minor band of 2.0 kb for human tissues. Analysis with the human cancer cell line MTN Blot (Fig. 4B) showed the same two bands except with promyelocytic leukemia HL-60 and chronic myelogenous leukemia K-562. Burkitt's lymphoma Raji cell line showed the greatest expression. For this reason the cDNA library of this lymphoma cell line was chosen as the cDNA source to isolate the present gene. 35 S-Labeled dermatan/chondroitin sulfate glycosaminoglycans were digested with chondroitin ABC lyase alone and with chondroitin ABC lyase immediately followed by disaccharide 2-sulfatase, 4-sulfatase, or 6-sulfatase. The digests were then analyzed by HPLC on a YMC-Pack Polyamine II column as described under "Experimental Procedures." The major 35 S-labeled disaccharide from dermatan sulfate was found to chromatograph with standard ⌬Di-2,4S (⌬HexA-2S-GalNAc-4S) (Fig. 5A). Following 2-sulfatase digestion, the radioactivity was shifted to the position of free sulfate (Fig. 5B); following 4-sulfatase it was shifted to the position of ⌬Di-2S (⌬HexA-2S-GalNAc) (Fig. 5C); but following 6-sulfatase it did not shift (Fig. 5D). These results established that the enzyme was a uronyl 2-sulfotransferase. Chondroitin B lyase, which degrades between GalNAc-4S and IdceA or IdceA-2S but not if there is GlcA or GlcA-2S, provided disaccharides as the only 35 S-labeled product (not shown). This confirmed that the enzyme was an IdceA 2-sulfotransferase.

Characterization of Dermatan [ 35 S]Sulfate and Chondroitin [ 35 S]Sulfate-
Desulfated dermatan sulfate and chondroitin sulfate C used similarly as potential [ 35 S]sulfate acceptors were degraded, and disaccharide products were characterized in the same fashion. Comparison with the disaccharide products from the dermatan sulfate 2-sulfation are shown (Table II). The predominant disaccharides from 35 S-labeled desulfated dermatan sulfate were shown to be ⌬Di-2S and ⌬Di-2,4S, indicating that 2-sulfation could take place on IdceA residues adjacent to non-   sulfated GalNAc as well as next to the small amount of Gal-NAc-4S that apparently had remained following chemical desulfation to prepare the dermatan. B lyase had no action on the chondroitin sulfate C, but there were 35 S-sulfated disaccharides equally produced by ABC lyase or AC lyase that were found to consist almost entirely of ⌬Di-2,6S (⌬HexA-2S-Gal-NAc-6S). This indicated that some sulfation of GlcA residues had taken place but essentially only if there were an adjacent GalNAc-6S. This was confirmed by showing that there was no incorporation of sulfate into GlcA residues of the desulfated chondroitin. It was of interest to note that the need for Gal-NAc-6S adjacent to a GlcA was in contrast to the sulfation of IdceA of dermatan sulfate that occurred mainly where there was an adjacent GalNAc-4S. We also used chondroitin sulfate A (Seikagaku Corp.) as an acceptor. However, chondroitin AC lyase did not degrade it completely, showing that it contained considerable dermatan residues as well as chondroitin 4-sulfate and chondroitin 6-sulfate. Analysis of the small amount of 2-sulfated AC lyase products indicated that there was mainly ⌬Di-2,6S, a minor amount of ⌬Di-2,4S, and no ⌬Di-2S (data not shown).
The amino-terminal sequence of the dermatan/chondroitin sulfate 2-sulfotransferase cDNA contains three in-frame ATG codons (Fig. 1). When the sequence surrounding the first ATG codon is compared with the eukaryotic consensus translation sequence (36,37), the purine G at position Ϫ3 is conserved, whereas G at position ϩ4 is not. The sequence surrounding the second and third ATG codons (Met-21 and Met-52 in Fig. 1) also partially fit the consensus sequence; the nucleotide at position Ϫ3 of these ATG codons is not a purine, whereas the nucleotide at position ϩ4 is G. The third ATG codon (Met-52), however, is unlikely to be an initiation site because of its location in the amino-terminal transmembrane domain (Fig. 1). It remains to be determined which ATG codons could function as the initiation codon.
Northern analysis showed two transcripts of 5.1 and 2.0 kb (Fig. 4), similar to heparan sulfate IdceA 2-sulfotransferase (5.0 and 3.0 kb) (15). Such multiple transcripts of different sizes are also observed in other glycosaminoglycan sulfotransferases (11-13, 18, 38) and are likely to be due to the difference in size and sequence of the untranslated regions. The possible existence of largely different untranslated regions may be important in the function and distribution of the transcripts (38).
The expressed dermatan/chondroitin sulfate uronyl 2-sulfotransferase catalyzed some 2-sulfation of the IdceA residues of IdceA-GalNAc and better sulfation of IdceA-GalNAc-4S. No 2-sulfation of dermatan-6S or 4,6S residues was found. The enzyme also had some activity in 2-sulfation of GlcA residues of GlcA-GalNAc-6S of chondroitin sulfate, but essentially no 2-sulfation of unsulfated disaccharide residues and little or no sulfation of 4S disaccharide residues (Table II). The activity on chondroitin sulfate raises the possibility that this enzyme functions in vivo for 2-sulfation of chondroitin 6-sulfate, but alternatively it is possible that this is due to a certain degree of nonspecificity. The results conform with the 2-sulfation of IdceA and GlcA found in connective tissue of many species, where IdceA-2S has not been described in the absence of Gal-NAc-4S, and GlcA-2S has not been described in the absence of GalNAc-6S. The differences in GalNAc sulfation specificities are apparently due to differences in the conformation of IdceA and GlcA with their positioning relative to the GalNAc-4S and GalNAc-6S, respectively. The results show that 2-sulfation of IdceA preferentially occurs next to GalNAc-4S rather than non-sulfated GalNAc, and 2-sulfation of GlcA requires Gal-NAc-6S with no 2-sulfation next to non-sulfated GalNAc. Thus the biosynthetic order apparently proceeds by prior GalNAc 4-sulfation for 2-sulfation of IdceA and prior GalNAc 6-sulfation for 2-sulfation of GlcA.
Tissue-specific patterns of epimerization and 4-and 6-sulfation in dermatan sulfate have not been reported in detail, and no preferred chain location or distribution of 2-sulfated IdceA has been reported. Comparison between decorin dermatan sulfate and biglycan dermatan sulfate was reported to show a greater correlation to tissue source than to the core protein (39). It is likely, however, that contiguous disaccharide sequences containing IdceA or IdceA-2S residues might account for the weak interaction reported with many heparin/heparan sulfate-binding proteins such as basic fibroblast growth factor (40), histidine-rich glycoprotein, platelet factor 4 (41), fibronectin (42), interleukin-7 (43), and protein C inhibitor (44). In contrast to these weak interactions, a comparable high affinity interaction of dermatan sulfate has been demonstrated with heparin cofactor II (7,8) and with hepatocyte growth factor/ scatter factor (9). Furthermore, the affinity for heparin cofactor II was shown to be dependent upon 2-sulfation of the IdceA residues (45). Sulfation profiles of chondroitin sulfate have been shown to change with concomitant specific spatiotemporal patterns in various tissues, suggesting that differences in sulfation position and degree might have distinct functions in development (46).
The GlcA 2-sulfated chondroitin has been shown to be expressed by immature glial cells of the central nervous system to promote neurite outgrowth (47) and to be expressed in the cerebellum and the telencephalon of adult mouse (48). This would suggest that our finding of IdceA/GlcA 2-sulfotransferase expression in brain and especially cerebellum (data not shown) could account for the 2-sulfation of chondroitin sulfate in brain. Disaccharide residues of GlcA-2S-GalNAc-6S have also been identified in mouse mast cells derived from immune lymph nodes and function as an important phenotypic marker distinguishing different mast cell subsets (49) and mouse tooth germ basement membrane (46). Characteristic oligosaccharide sequences including 2-sulfated IdceA or GlcA residues in DS and CS may serve as functional domain structures recognized by some protein ligands as in heparin/heparan sulfate.