Molecular cloning and expression of chondroitin 4-sulfotransferase.

Chondroitin 4-sulfotransferase (C4ST) catalyzes the transfer of sulfate from 3'-phosphoadenosine 5'-phosphosulfate to position 4 of N-acetylgalactosamine residue of chondroitin. The enzyme has been previously purified to apparent homogeneity from the serum-free culture medium of rat chondrosarcoma cells (Yamauchi, A., Hirahara, Y., Usui, H., Takeda, Y., Hoshino, M., Fukuta, M., Kimura, J. H., and Habuchi, O. (1999) J. Biol. Chem. 274, 2456-2463). The purified enzyme also catalyzed the sulfation of partially desulfated dermatan sulfate. We have now cloned the cDNA of the mouse C4ST on the basis of the amino acid sequences of peptides obtained from the purified enzyme by protease digestion. This cDNA contains a single open reading frame that predicts a protein composed of 352 amino acid residues. The protein predicts a Type II transmembrane topology. The predicted sequence of the protein contains all of the known amino acid sequence and four potential sites for N-glycosylation, which corresponds to the observation that the purified C4ST is an N-linked glycoprotein. The amino acid sequence of mouse C4ST showed significant sequence homology to HNK-1 sulfotransferase. Comparison of the sequence of mouse C4ST with human HNK-1 sulfotransferase revealed approximately 29% identity and approximately 48% similarity at the amino acid level. When the cDNA was introduced in a eukaryotic expression vector and transfected in COS-7 cells, the sulfotransferase activity that catalyzes the transfer of sulfate to position 4 of GalNAc residue of both chondroitin and desulfated dermatan sulfate was overexpressed. Northern blot analysis showed that, among various mouse adult tissues, 5.7-kilobase message of C4ST was mainly expressed in the brain and kidney.

Chondroitin sulfate proteoglycans are found in various tissues as molecules having divergent molecular architecture (1,2). Chondroitin sulfate chains attached to chondroitin sulfate proteoglycans appear to play important roles in the formation and maintenance of cartilage tissue, because undersulfation of chondroitin sulfate resulted from the defective synthesis of PAPS 1 (3,4) or defective sulfate transport (5) was found to cause underdevelopment of skeleton. Various chondroitin sulfate proteoglycans have been reported to be present in the brain (6,7) and to function in the regulation of neurite outgrowth and neural cell adhesion (8 -12), neuronal migration (13), and the survival of neurons (14). Chondroitin sulfate chains are also shown to be involved in the interaction with CD44 (15,16), phospholipase A 2 (17), Plasmodium falciparuminfected erythrocytes (18), and L-selectin (19). Chondroitin sulfates have sulfate group at various positions of the component sugars; position 6 and/or 4 of GalNAc residues and position 2 or 3 of GlcA residues. The pattern of sulfation of chondroitin sulfate chains varies with the source of the proteoglycans, development of animal (20 -22), and malignant change (23), suggesting that sulfate moieties attached to the specific position of the component sugars may be related to the specific function of the glycosaminoglycan. Characterization and molecular cloning of sulfotransferases, which participate in the formation of the defined structure of chondroitin sulfate, are important to clear the functional roles of chondroitin sulfate.
We have purified chick chondroitin 6-sulfotransferase (C6ST), which catalyzes transfer of sulfate to position 6 of N-acetylgalactosamine residue of chondroitin from PAPS (24), and cloned its cDNA (25). Uronosyl 2-sulfotransferase, which sulfates iduronyl and glucuronyl residues in dermatan/chondroitin sulfate, has been cloned (26). We have previously purified chondroitin 4-sulfotransferase (C4ST), which catalyzes the transfer of sulfate from PAPS to position 4 of N-acetylgalactosamine residue of chondroitin, to apparent homogeneity (27). This enzyme was also found to catalyze the sulfation of partially desulfated dermatan sulfate. In this paper we report the cloning of the cDNA encoding mouse C4ST and the expression of it in COS-7 cells.

Purification of Chondroitin 4-Sulfotransferase and Amino Acid Sequencing of the Intact Purified Protein and Peptides Obtained by En-
* This work was supported by the Grant-in-aid for Scientific Research on Priority Areas No. 10178102 from the Ministry of Education, Science, Sports and Culture of Japan, Grants-in-aid of Mizutani Foundation for Glycoscience, and by a special research fund from Seikagaku Corp. 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 reported in this paper has been submitted to the DDBJ/GenBank TM   doproteinase Digestion-C4ST was purified from the serum-free culture medium of rat chondrosarcoma cells as described previously (27). A portion of the purified C4ST (15 g as protein) was subjected to SDSpolyacrylamide gel electrophoresis (10% gel) according to the method of Laemmli (28) after reduction and denaturation in loading buffer containing 5% (v/v) 2-mercaptoethanol and then electrotransferred onto ProBlott polyvinylidene difluoride membrane (Applied Biosystems, Inc.). The transferred protein was visualized with Ponceau S. The band of the 60-kDa protein were excised and subjected to reduction and S-carboxymethylation on the polyvinylidene difluoride membrane according to Iwamatsu (29). After digestion with N-glycosidase F (recombinant, Roche Molecular Biochemicals), the membrane was digested sequentially with trypsin (modified, sequencing grade, Roche Molecular Biochemicals) and then endoproteinase Asp-N (Roche Molecular Biochemicals) followed by endoproteinase Lys-C (Promega) according to Iwamatsu (29). Each endoproteinase digest was subjected to reversephase capillary HPLC and eluted with a gradient of 2-100% solvent A (80% acetonitrile in 0.052% trifluoroacetic acid). The amino acid sequences of the separated peptides were analyzed with a model 476A protein sequencer (Applied Biosystems, Inc.). Another portion of the purified C4ST (5 g as protein) was used for NH 2 -terminal amino acid sequencing after separation with SDS-polyacrylamide gel electrophoresis and blotting onto polyvinylidene difluoride membrane as described above.
Polymerase Chain Reaction and Preparation of a Probe for Screening-We found that peptide 3 in Table I matched a mouse EST cDNA (accession number AA218236). We designed oligonucleotide primers from the sequence of the clone. The forward primer (primer F) was TGCCTACCGCAACAAGTTCACG, and the reverse primer (primer R) was AAGAACTCCGTGGTCATCTCGTCG. The first strand of cDNA was synthesized by the reverse transcriptase reaction using poly(A) ϩ RNA prepared from rat chondrosarcoma cells as a template. The reverse transcriptase reaction mixture contained, in a final volume of 20 l, 2 g of poly(A) ϩ RNA, 0.5 g of oligo(dT) or random primers, 10 mM dithiothreitol, 0.5 mM each of four deoxynucleoside triphosphates, 30 units of RNase inhibitor (Takara), and 400 units of reverse transcriptase (Superscript II, Life Technologies, Inc.). The reaction was carried out at 42°C for 50 min. After the reaction was stopped by heating at 70°C for 15 min, 0.5 unit of RNase H (Life Technologies, Inc.) was added and incubated for 20 min at 37°C. The PCR reaction was carried out in a final volume of 50 l containing 50 pmol each of the oligonucleotide primers, 2 l of the reverse transcriptase reaction mixture in which the first strand cDNA was synthesized, 0.2 mM each of four deoxynucleoside triphosphates, 1.5 units of AmpliTaq polymerase (Perkin-Elmer). Amplification was carried out by 40 cycles of 94°C for 45 s, 59°C for 1.5 min, and 72°C for 1 min.
Reaction products were subjected to agarose gel electrophoresis ( Fig.  1). The amplified DNA band (indicated by an arrowhead in Fig. 1) was cut out, and the DNA fragment was recovered from the gel using Jetsorb (Genomed), and cycle-sequenced using primer F or primer R as sequence primers. The radioactive probe for screening of the cDNA library was prepared from the PCR product by the random oligonucleotide-primed labeling method (30) using [␣-32 P]dCTP (Amersham Pharmacia Biotech) and a DNA random labeling kit (Takara Shuzo).
Screening of gt11 Library-Approximately 5 ϫ 10 5 plaques from the mouse brain cDNA library (CLONTECH) were screened. Hybond N ϩ nylon membrane (Amersham Pharmacia Biotech) replicas of the plaques from the gt11 cDNA library were fixed by the alkali fixation method recommended by the manufacturer, prehybridized in a solution containing 50% formamide, 5 ϫ SSPE, 5 ϫ Denhardt's solution, 0.5% SDS, 0.04 mg/ml of denatured salmon sperm DNA for 3.5 h at 42°C. Hybridization was carried out in the same buffer containing 32 P-labeled probe for 16 h at 42°C. The filters were washed at 55°C in 1 ϫ SSPE, 0.1% SDS, and subsequently in 0.1 ϫ SSPE, 0.1% SDS, and positive clones were detected by autoradiography.
DNA Sequence Analysis-DNA from gt11 positive clones were isolated and cut with EcoRI, which excised the cDNA insert. The fragments were inserted into pBluescript II vector (Stratagene). The complete nucleotide sequence was determined independently on both strands by the dideoxy chain termination method using a DNA sequencer (Applied Biosystem model 373A). DNA sequences were compiled and analyzed using the MacVector computer programs (Oxford Molecular Group PLC).
Construction of pCXNC4ST-To construct the plasmid containing the C4ST cDNA named pCXNC4ST, the EcoRI fragment with 1581 nucleotides indicated in Fig. 2A was excised from the pBluescript plasmid and ligated into the EcoRI site of pCXN2 expression vector (pCXN2 vector was constructed by Dr. Jun-ichi Miyazaki, Department of Dis-ease-related Gene Regulation, Faculty of Medicine, University of Tokyo (31) and a gift from Dr. Yasuhiro Hashimoto, Tokyo Metropolitan Institute of Medical Sciences). Escherichia coli XL1-Blue cells were transformed with the ligation mixture and plated on LB ampicillin plates. Recombinant plasmids were cut with XbaI to confirm the correct orientation of pCXNC4ST. The recombinant plasmids used for the transfection were purified with Qiagen Plasmid kit (Qiagen). The plasmid, which contained the cDNA fragment in the reversed orientation, was named as pCXNC4ST2 and used for control experiments.
Transient Expression of Chondroitin 4-Sulfotransferase cDNA in COS-7 Cells-COS-7 cells (obtained from Riken Cell Bank, Tsukuba, Japan) were plated in 100-mm culture dishes at a density of 8 ϫ 10 5 cells/dish. Volume of the medium was 10 ml. The medium used was DMEM containing penicillin (100 units/ml), streptomycin (50 g/ml), and 10% fetal bovine serum (Life Technologies, Inc.), and cells were grown at 37°C in 5% CO 2 , 95% air. When the cell density reached 3 ϫ 10 6 cells/dish (48 h after plating), COS-7 cells were transfected with pCXNC4ST, pCXNC4ST2, or vector alone. The transfection was performed using the DEAE-dextran method (32). 5 ml of the prewarmed DMEM containing 10% NuSerum (Collaborative Biomedical Products) was mixed with 0.2 ml of phosphate-buffered saline containing 10 mg/ml DEAE-dextran plus 2.5 mM chloroquine solution. 15 g of the recombinant plasmid was mixed with the solution, and the mixture was added to the cells. The cells were incubated for 4 h in a CO 2 incubator. The medium was then replaced with 5 ml 10% dimethyl sulfoxide in phosphate-buffered saline. After the cells were left at room temperature for 2 min, the dimethyl sulfoxide solution was aspirated and 25 ml of DMEM containing penicillin (100 units/ml), streptomycin (50 g/ml), and 10% fetal bovine serum was added. The cells were incubated for 67 h, washed with DMEM alone, scraped, and extracted with 10 mM Tris-HCl, pH 7.2, 10 mM MgCl 2 , 2 mM CaCl 2 , 0.5% Triton X-100, and 20% glycerol for 30 min on a rotatory shaker. The extracts were centrifuged at 10,000 ϫ g for 20 min, and the sulfotransferase activities in the supernatant fractions were measured using chondroitin or desulfated dermatan sulfate as acceptors. The sulfotransferase activities were assayed by the method described previously (27). The standard reaction mixture contained 2.5 mol of imidazole HCl, pH 6.8, 1.25 g of protamine chloride, 0.1 mol of dithiothreitol, 25 nmol (as glucuronic acid) chondroitin, 50 pmol [ 35 S]PAPS (about 5.0 ϫ 10 5 cpm) (33), and enzyme in a final volume of 50 l. When 25 nmol (as GalNAc) partially desulfated dermatan sulfate was used as acceptor, the amount of protamine chloride was increased to 10 g. Partially desulfated dermatan sulfate was prepared from pig skin dermatan sulfate as described (27). The degree of the desulfation was calculated as 83% from the proportion of ⌬Di-0S to the total unsaturated disaccharides formed after chondroitinase ABC digestion. When the desulfated dermatan sulfate was digested with chondroitinase ACII, the yield of ⌬Di-0S was only 5% of the total unsaturated disaccharides formed after chondroitinase ABC digestion (27). The reaction mixtures were incubated at 37°C for 20 min, and the reaction was stopped by immersing the reaction tubes in a boiling water bath for 1 min. After the reaction was stopped, 35 S-labeled glycosaminoglycans were isolated by the precipitation with ethanol followed by gel chromatography with a Fast Desalting Column as described previously and radioactivity was determined. For determining the incorporation into position 4 and position 6 of GalNAc residues, 35 S-labeled chondroitin and 35 S-labeled desulfated dermatan sulfate were digested with chondroitinase ACII and chondroitinase ABC, respectively. The resulting unsaturated disaccharides (⌬Di-4S and ⌬Di-6S) were separated with paper chromatography (24), and their radioactivities were measured.
Preparation of a FLAG-C4ST Fusion Protein-Recombinant C4ST was also expressed as a fusion protein with FLAG peptide. A DNA fragment, which codes for full open reading frame, was amplified by PCR using mouse C4ST cDNA as a template. The 5Ј and 3Ј primers were CGCAAGCTTATGAAGCCGGCGCTGCTGGAAGT and CAG-GAATTCCCTAATCCAACTTCAGGTAGT, respectively. At 5Ј-end of the oligonucleotide primers, restriction enzyme recognition sites were introduced: HindIII site for the sense primer and EcoRI site for the antisense primer. The PCR product was digested with EcoRI and Hin-dIII and subcloned into these sites of pFLAG-CMV-2 plasmid (Eastman Kodak Co.). The resulting plasmid was transfected in COS-7 cells, and the fusion protein was extracted from the cells as described above. The extracts were applied to an anti-FLAG monoclonal antibody-conjugated affinity column (Kodak) equilibrated with the buffer used for the extraction. The absorbed materials were eluted with FLAG peptide under the conditions recommended by the manufacturer.
HPLC Separation of 35 S-Labeled Products after Digestion with Chondroitinase ACII or Chondroitinase ABC-Separation of the degradation products formed from 35 S-labeled chondroitin and 35 S-labeled desulfated dermatan sulfate after digestion with chondroitinase ACII or chondroitinase ABC was carried out by HPLC using a Whatman Partisil 10-SAX column (4.5 mm x 25 cm) equilibrated with 10 mM KH 2 PO 4 . The column was developed with gradient elution as shown in Fig. 4. The flow rate was 1 ml/min, and the column temperature was 40°C. 0.5-ml fractions were collected.
Northern Blot Hybridization-Mouse Multiple Tissue Northern blot Filters (CLONTECH) (on which 2 g of poly(A) ϩ RNAs from various adult mouse tissues were blotted) were prehybridized in ExpresHyb solution (CLONTECH) at 68°C. Hybridization was carried out in the same solution containing 32 P-labeled probe for 1 h at 68°C. The radioactive probe was prepared from the cDNA fragment excised from the pBluescript II plasmid by the random oligonucleotide-primed labeling method (30) using [␣-32 P]dCTP. The filters were washed at room temperature in 2 ϫ SSC, 0.05% SDS, and subsequently in 0.1 ϫ SSC, 0.1% SDS at 50°C. The membrane was exposed to x-ray film at Ϫ80°C with an intensifying screen.  Asp-N DLVQK

RESULTS
Generation of PCR Probe to Screen cDNA Library-We obtained amino acid sequences of six peptides from the purified C4ST (Table I). We also determined amino-terminal amino acid sequence of the intact protein. When the sequence of peptide 3 was used for the homology search, we found that the peptide matched a mouse EST cDNA containing 505 nucleotides (accession number AA218236). Examination of the sequence of the cDNA indicated the presence of the nucleotide sequences corresponding to peptide 1, 2, 4, 5, and 6, although the reading frame was not the same. We thought that the clone must be a cDNA fragment encoding a partial sequence of C4ST. From the sequence of the EST cDNA, oligonucleotide primers were designed to amplify a DNA fragment with 374 nucleotides. The sequence of primer R was found in both the mouse EST cDNA and an another rat EST cDNA (accession number AI044878), but the rat sequence corresponding to primer F was not available; therefore we used the sequence of the mouse EST cDNA to design primer F. PCR was carried out using the first strand cDNA of poly(A) ϩ RNA prepared from rat chondrosarcoma cells as a template (Fig. 1). A DNA fragment with about 350-base pair was amplified in the presence of both forward and reverse primers and the first strand cDNA (indicated by an arrowhead in Fig. 1). The sequence of this DNA fragment was almost identical to that of the mouse EST cDNA (data not shown); therefore, we used the PCR product for the preparation of a probe for the screening.
Screening of gt11 Library-The above described 350-base pair PCR product was labeled with [␣-32 P]dCTP by the random oligonucleotide-primed labeling method and used as a probe to screen a gt11 mouse brain cDNA library. About 50 positive clones were observed from 4 ϫ 10 5 plaques. From the nucleotide sequence, a cDNA clone with 1581 base pairs was found to contain whole open reading frame.
cDNA and Predicted Protein Sequence of the Chondroitin 4-Sulfotransferase-The nucleotide sequence of the C4ST cDNA and the predicted amino acid sequence are shown in Fig.  2A. The amino-terminal sequence upstream from the putative transmembrane domain contains three in-frame ATG codons. A single open reading frame beginning at the first ATG codon predicts a protein of 352 amino acid residues with a molecular mass of 41,629 Da with four potential N-linked glycosylation sites. To determine the location of any transmembrane domain, a hydropathy plot was generated from the translated sequence. Analysis of the plot revealed one prominent hydrophobic segment in the amino-terminal region, 16 residues in length, that extends from amino acid residues 18 -33 (Fig. 2B). The NH 2terminal amino acid sequence of the purified C4ST was found in amino acid residues 58 -66. The molecular mass of the predicted protein truncated at the amino acid residue 58 was 34,962 Da, which agreed well with the molecular mass of the protein formed after N-glycanase digestion. All the amino acid sequences that have been obtained from the purified protein were found in the predicted protein sequences, although two amino acid residues among these peptides (Arg in peptide 5 and Gln in peptide 6) did not agree with the predicted sequence (Gln 291 and Gly 277 , respectively), confirming that the cDNA clone encodes the purified C4ST protein.
The amino acid sequence of C4ST showed no significant homology with other sulfotransferase so far cloned, except for HNK-1 sulfotransferase (34,35). Comparison of the sequence of C4ST with human HNK-1 sulfotransferase revealed ϳ29% identity and ϳ48% similarity at the amino acid level (Fig. 3).
Expression of Chondroitin 4-Sulfotransferase cDNA in COS-7 Cells-Direct evidence demonstrating that the isolated cDNA encodes the chondroitin 4-sulfotransferase protein was obtained by expressing it in COS-7 cells. COS-7 cells were transfected with the pCXNC4ST, a recombinant plasmid containing the isolated cDNA in the mammalian expression vector pCXN2. The transfected cells were scraped at 67 h after transfection, extracted with gentle shaking in a buffer containing 0.5% Triton X-100, and centrifuged. Sulfotransferase activities in the supernatant fractions were determined using chondroitin or partially desulfated dermatan sulfate (27) as acceptors. Control experiments without vector, and with vector containing the cDNA in the reversed orientation (pCXNC4ST2), were also done. As shown in Table II, when the vector containing the isolated cDNA was used, the sulfotransferase activity, which transfers sulfate to position 4 of GalNAc residue of chondroitin, and the sulfotransferase activity, which transfers sulfate to position 4 of GalNAc residue of partially desulfated dermatan sulfate, were increased 25-and 6-fold, respectively, above those of controls. In contrast, sulfotransferase activity, which transfers sulfate to position 6 of GalNAc residue, was not increased at all. These results demonstrate that the isolated cDNA encodes a protein with sulfotransferase activity identical to the purified C4ST. In Table II, C6ST activity was detected when chondroitin or partially desulfated dermatan sulfate was added to the reaction mixtures. C6ST activity observed in the presence of these acceptors appears to be due to the endogenous enzyme derived from COS-7 cells, since C6ST activity was not overexpressed. To confirm this, we prepared FLAG-C4ST fusion protein and purified with an anti-FLAG monoclonal antibody-conjugated affinity column. The affinity-purified FLAG-C4ST fusion protein was incubated with chondroitin or partially desulfated dermatan sulfate and [ 35 (hHNK-1). The predicted amino acid sequences were aligned using MacVector computer program. White letters in black boxes and boxed letters indicate the identical and similar amino acid residues, respectively, between the two sequences.
ACII and chondroitinase ABC, respectively, radioactivity was detected only at the position of ⌬Di-4S (Fig. 4A and B). When 35 S-labeled partially desulfated dermatan sulfate was digested with chondroitinase ACII, about 50% of the total radioactivity was recovered in ⌬Di-4S and the remainder appeared in minor peaks with higher retention time (Fig. 4C). These results were identical to those observed previously in the purified C4ST (27). The minor peaks with higher retention time observed when digested with chondroitinase ACII appear to be oligosaccharides as shown previously (27). These results clearly indicate that the expressed FLAG-C4ST fusion protein is able to transfer sulfate only to position 4 of GalNAc residue of chondroitin or partially desulfated dermatan sulfate.
Northern Blot Analysis-Among various mouse tissues, C4ST mRNA with 5.7 kilobases was expressed mainly in the brain and kidney (Fig. 5). A weak expression was also observed in the heart, spleen, and lung. A small fragment observed in the testis may not be an intact mRNA, but a degradation product. DISCUSSION We have cloned a cDNA that encodes the C4ST. The predicted sequence of the protein contains all of the known amino acid sequence and four potential N-linked glycosylation sites. When the cDNA was introduced into a eukaryotic expression vector and transfected in COS-7 cells, the C4ST activity was overexpressed. The predicted protein showed no significant sequence homology with other sulfotransferases except for HNK-1 sulfotransferase. C4ST transfers sulfate to position 4 of GalNAc residue, whereas HNK-1 sulfotransferase transfers sulfate to position 3 of nonreducing terminal GlcA residue. Although acceptor substrate specificity is quite different from each other, C4ST and HNK-1 sulfotransferase seem to belong to a common sulfotransferase family.
Two putative PAPS biding sites were found by x-ray crystallography of estrogen sulfotransferase performed in the presence of adenosine 3Ј, 5Ј-bisphosphate (36): one is KSGT for 5Ј-phosphosulfate binding site and another is RX 7 S for 3Јphosphate binding site. These motifs were observed in every glycosaminoglycan sulfotransferases so far cloned except for heparan sulfate N-deacetylase/N-sulfotransferase, in which the 3Ј-phosphate binding site is not RX 7 S but IX 7 S (37). In the predicted mouse C4ST, these motifs were observed; K 125 VACT for 5Ј-phosphosulfate binding site and R 186 EPFERLVS for 3Јphosphate binding site. It is of note that Cys is present in 5Ј-phosphosulfate binding site. C4ST was found to require sulfhydryl compounds for the activity. The presence of Cys in the 5Ј-phosphosulfate binding motif may relevant to the requirement for sulfhydryl compounds.
Some inconsistency in amino acid sequence was observed between peptides obtained from the purified rat C4ST and the predicted mouse C4ST; Arg in peptide 5 and Gln in peptide 6 were changed to Gln and Gly, respectively, in the predicted mouse C4ST. Since peptide 3 adjacent to the carboxyl side of peptide 5 was obtained after trypsin digestion, Arg in peptide 5 must be correct, and the change of Arg to Gln in peptide 5 may reflect the difference between rat C4ST and mouse C4ST. In contrast, it is unlikely that change of Gln to Gly in peptide 6 is due to the species variation, because this change should not be attained after single point mutation. It is most probable that we made an error in the assignment of amino acid residue of peptide 6.

TABLE II
Overexpression of chondroitin 4-sulfotransferase in COS-7 cells COS-7 cells were transfected as described under "Experimental Procedures" with a plasmid containing the C4ST cDNA (pCXNC4ST), a plasmid containing the C4ST cDNA in the reversed orientation (pCXNC4ST2), and the plasmid alone (pCXN2). Sulfotransferase activity was determined as described under "Experimental Procedure" using chondroitin or desulfated dermatan sulfate as acceptor. Incorporation of 35  The amino-terminal sequence contains three in-frame ATG codons. When the sequence surrounding the first ATG codon is compared with the eukaryotic consensus translation sequence (38), the purine at Ϫ3, is conserved, but Gly in position ϩ4 is not conserved. The sequence surrounding the second and the third ATG codon (Met 9 and Met 11 in Fig. 2A) also partially fits with the consensus sequence; position Ϫ3 of the ATG codon is Gly, whereas position ϩ4 is not Gly but Ala. It remains to be determined which ATG codons could function as the initiation codon.
The amino-terminal amino acid sequence of the purified rat C4ST was found in the predicted stem domain. Although we have not determined the whole sequence of the rat C4ST yet, this observation suggests that the purified rat C4ST might be released from the chondrosarcoma cells after proteolytic cleavage at the site, because the amino acid sequence of the mouse C4ST agreed well with that of the rat C4ST as far as determined. C6ST (25) and heparan sulfate 6-sulfotransferase (39) were also cleaved at the NH 2 -terminal region before secretion in the culture medium. The mechanism as well as physiological significance of the proteolytic cleavage of these sulfotransferases remains to be determined.
The message of C4ST was expressed mainly in the brain and kidney. Various chondroitin sulfate proteoglycans have been reported to be present in the brain such as neurocan (40), brevican (41), RPTP/␤ (42), NG2-PG (43), and neuroglycan C (44). On the other hand, rat mesangial cells were shown to synthesize various proteoglycans, including versican, biglycan, and decorin (45). Ligand for L-selectin in the kidney was shown to be versican (19). C4ST may be involved in the control of the function of these proteoglycans via the sulfation of glycosaminoglycan chains of these proteoglycans.
Although both C4ST and C6ST utilize chondroitin as a common acceptor substrate, no significant homology in the amino acid sequence was observed between the two sulfotransferases. In addition, expression pattern of the cDNAs of these two sulfotransferases in the mouse tissues are quite different; C4ST expressed mainly in the brain and kidney, whereas a strong message of C6ST was observed in the lung and spleen (46). It is thus possible that biosynthesis of chondroitin 4-sulfate and chondroitin 6-sulfate may be controlled by different mechanisms by which the expression of these two sulfotransferases are regulated. 4-Sulfated glycosaminoglycans (chondroitin 4-sulfate and dermatan sulfate) are synthesized in most tissues, whereas C4ST mRNA was expressed mainly in the brain and kidney among various mouse tissues, suggesting that multiple forms of C4ST may be present in the tissues. Unexpectedly, C4ST showed significant homology in the amino acid sequence with HNK-1 sulfotransferase. Oligosaccharides containing 3-O-sulfated GlcA, which was structurally related to the proteoglycan linkage region, have been found in the culture medium of human fibroblast (47) and human urinary thrombomodulin (48). It remains to be determined whether C4ST may share not only amino acid sequence but also substrate specificity with HNK-1 sulfotransferase.