N-Linked Oligosaccharides on Chondroitin 6-Sulfotransferase-1 Are Required for Production of the Active Enzyme, Golgi Localization, and Sulfotransferase Activity toward Keratan Sulfate*

We have shown previously that purified chondroitin 6-sulfotransferase-1 (C6ST-1) was a glycoprotein abundant in N-linked oligosaccharides and could sulfate both chondroitin (C6ST activity) and keratan sulfate (KSST activity); however, functional roles of the N-glycans have remained unclear. In the present study, we show essential roles of N-glycans attached to C6ST-1 in the generation of the active enzyme and in its KSST activity. Treatment with tunicamycin of COS-7 cells transfected with C6ST-1 cDNA totally abolished production of the active C6ST-1. A nearly complete removal of N-glycans of the recombinant C6ST-1 by peptide N-glycosidase F increased the C6ST activity but decreased the KSST activity. Among six potential N-glycosylation sites, deletion of the fourth or sixth site from the amino terminus inhibited production of the active C6ST-1, whereas deletion of the fifth site resulted in a marked loss of the KSST activity. Wild-type recombinant C6ST-1 showed a typical Golgi localization, whereas M-4 recombinant C6ST-1, in which the fourth N-glycosylation site was deleted, colocalized with calnexin, an endoplasmic reticulum-resident protein. Unlike wildtype recombinant C6ST-1, M-4 recombinant C6ST-1 showed a weak affinity toward wheat germ agglutinin and was converted completely to the nonglycosylated form by endoglycosidase H. These observations suggest that N-glycan attached to the fourth N-glycosylation site may function in the proper processing of N-glycans required for the Golgi localization, thereby causing the production of the active C6ST-1, and that N-glycan attached to the fifth N-glycosylation site may contribute to the KSST activity of C6ST-1.

units bearing sulfate groups on various positions of each sugar residue. Transfer of sulfate to the repeating unit is catalyzed by sulfotransferases having strict substrate specificity (1)(2)(3). To date, four sulfotransferase families involved in the biosynthesis of CS have been cloned, and their substrate specificities were characterized: chondroitin 6-sulfotransferase (C6ST) (4,5) and chondroitin 4-sulfotransferase (C4ST) (6 -9) transfer sulfate to position 6 and 4, respectively, of GalNAc residues; GalNAc(4SO 4 ) 6-O-sulfotransferase transfers sulfate to position 6 of GalNAc(4SO 4 ) residues (10); and uronosyl 2-O-sulfotransferase transfers sulfate to position 2 of GlcA or IdoUA (11,12). Of these sulfotransferases, mutations in mouse and human C6ST-1 are associated with a decrease in the naive T lymphocytes in the spleen (13) and chondrodysplasia (14), respectively, and a gene trap deletion of mouse C4ST-1 resulted in severe chondrodysplasia (15). Increasing data have been accumulated showing that these sulfotransferases recognize not only the position of the sugar residue to which sulfate is transferred but also sequences or sulfation pattern of the component sugar residues. C4ST-1 preferentially transferred sulfate to GalNAc residues adjacent to the reducing side of the GlcA residue but not the IdoUA residue (16). Sulfation of position 6 of the nonreducing terminal GalNAc(4SO 4 ) in the unique sequence found in the nonreducing terminal of CS, GalNAc(4SO 4 )-GlcA(2SO 4 )-GalNAc(6SO 4 )-, by GalNAc(4SO 4 ) 6-O-sulfotransferase was markedly stimulated by the penultimate GlcA(2SO 4 ) residue (17). Uronosyl 2-O-sulfotransferase selectively transferred sulfate to GlcA located in a unique sequence in CS, GalNAc(4SO 4 )-GlcA-GalNAc(6SO 4 ) (12). These observations strongly suggest that the order of participation of each sulfotransferase into the biosynthetic steps should be important to elaborate the functional fine structures in CS. Subcellular localization of these sulfotransferases is probably relevant to determining the order. The purified C6ST-1, C4ST-1, and GalNAc(4SO 4 ) 6-O-sulfotransferase (18 -20) as well as the recombinant C6ST-1, C4ST-1, GalNAc(4SO 4 ) 6-O-sulfotransferase, and uronosyl 2-O-sulfotransferase (12,16,17) have been found to be glycoproteins containing relatively abundant N-linked oligosaccharides. It has been reported that N-glycans contained in various glycoproteins are involved in protein folding, ER-associated protein degradation, intracellular trafficking, and production of functional proteins (21)(22)(23)(24)(25). We have previously shown that N-glycans attached to C4ST-1 are involved in the production and stability of the active enzyme (26). However, functional roles of N-glycans attached to these sulfotransferases has not been fully understood. In the present study, we examined the roles of N-glycans attached to C6ST-1 and found that N-glycans on C6ST-1 affected the intracellular localization and thereby the production of the active enzyme. We also found unexpectedly that one of the N-glycans is required for the sulfotransferase activity of C6ST-1 toward KS.
Expression Plasmids-A DNA fragment that codes for the full open reading frame of mC6ST-1 (29) was amplified by PCR. The first PCR was carried out using oligonucleotides mC6ST-1-F1 (CACCTGAAGAGGGCTGATTG) and mC6ST-1-R1 (TGGGTGTGAGGGAAG), which were synthesized according to the sequence of the mC6ST-1 cDNA clone (GenBank TM accession number AB008937), as primers and the QUICK-Clone TM cDNA derived from mouse adult spleen (Clontech) as a template. Amplification was carried out by 35 cycles of 95°C for 30 s, 50°C for 1 min, and 68°C for 5.5 min. The second PCR was carried out using oligonucleotides mC6ST-1-F2 (GAC-GAATTCGATGGAGAAAGGACTCGC) and mC6ST-1-R2 (GCAAAGCTTCTACGTGACCCAGAAGGTGCG) as primers and the first PCR mixture as the template. At the 5Ј-ends of oligonucleotides mC6ST-1-F2 and mC6ST-1-R2, restriction enzyme recognition sites were introduced, an EcoRI site for mC6ST-1-F2 and a HindIII site for mC6ST-1-R2. Amplification was carried out by 35 cycles of 95°C for 30 s, 55°C for 1 min, and 68°C for 3 min 10 s. The reaction products were subjected to agarose gel electrophoresis. The amplified DNA band was cut out, and the DNA fragment was recovered from the gel, digested with HindIII and EcoRI, and subcloned into these sites of pFLAG-CMV-2 plasmid. Recombinant mC6ST-1 was expressed in COS-7 cells as a fusion protein with FLAG peptide and was affinity-purified as described below. Site-directed mutagenesis was carried out by PCR using a QuikChange TM site-directed mutagenesis kit (Stratagene), according to the manufacturer's instructions. Putative N-glycosylation sites, the Asn-X-(Ser/Thr) motif, in which X is any amino acid except for Pro, were disrupted by substituting asparagine (Asn 63 , Asn 74 , Asn 96 , Asn 250 , Asn 413 , and Asn 457 ) with serine. The nucleotide sequences of all of the mutated cDNAs were confirmed by the sequencing. The sequences of all of the PCR primers used for the site-directed mutagenesis are shown in Table 1. All of the mC6ST-1 constructs have a FLAG tag at the N-terminal end.
Cell Culture and Transfection of COS-7 or Chinese Hamster Ovary Cells-COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 50 g/ml streptomycin, 100 units/ml penicillin, and 10% (v/v) fetal bovine serum. CHO-K1 cells were cultured in F-12 medium containing the antibiotics and fetal bovine serum as above. These cells were grown at 37°C in 5% CO 2 , 95% air. The plasmids prepared as above were transfected into COS-7 cells or CHO-K1 cells by the lipofection method using TransFast reagent (Promega) under serum-free conditions according to the manufacturer's instructions. At 48 h after the addition of cDNA, the transfectants were extracted with 0.15 M NaCl, 10 mM Tris-HCl, pH 7.2, 10 mM MgCl 2 , 2 mM CaCl 2 , 0.5% (v/v) Triton X-100, 20% (v/v) glycerol, and protease inhibitors (50 M TLCK, 30 M TPCK, 300 M phenylmethylsulfonyl fluoride, and 3 M pepstatin A) for 30 min on a rotatory shaker. The lysates were clarified by centrifugation at 10,000 ϫ g for 10 min, and the supernatants were applied to an anti-FLAG mAb-conjugated agarose column. The absorbed materials were eluted with a buffer containing FLAG peptide under the conditions recommended by the manufacturer. The purified protein solution (1.5 ml) was used to determine C6ST and KSST activity. Under the standard assay conditions, C6ST activity of the affinity-purified wild-type recombinant C6ST-1 ranged from 100 to 500 pmol/min/ml among different preparations. The KSST activity of each preparation was ϳ30% of C6ST activity. Treatment with Tunicamycin-To determine the effects of tunicamycin, tunicamycin was added to the culture medium after transfection at the final concentrations indicated in Fig. 1, and the culture was continued for a further 48 h. The cellular extracts were prepared as described above.
Western Blot Analysis-SDS-PAGE was carried out according to Laemmli (30). The affinity-purified proteins before or after PNGase F or Endo H digestion as described below were boiled in the sample buffer with 10% (w/v) 2-mercaptoethanol and then separated by 10% gels. Proteins were electorophoretically transferred to Hybond ECL membranes (Amersham Biosciences) at 60 V for 4 h. For immunoblotting, nonspecific binding sites were blocked with 5% (w/v) skim milk in TBST (20 mM Tris-HCl, pH 7.6, 130 mM NaCl, and 0.1% (v/v) Tween 20) at 4°C for 16 h. The FLAG-tagged recombinant proteins were stained with an anti-FLAG M2 mAb at a 1:10,000 dilution. The blots were developed with polyclonal anti-mouse IgG antibody coupled to horseradish peroxidase at a dilution of 1:8000 using an ECL detection kit and a Hyperfilm ECL (Amersham Biosciences). Molecular masses of polypeptides were calculated from the migration distance of a molecular mass standard (Sigma) or a prestained standard (Nakarai Tesque, Kyoto, Japan) run in every gel. Protein bands in Hybond ECL membrane were visualized by Amido Black staining. Amounts of FLAG-tagged fusion proteins were estimated by immunoblotting of these proteins together with FLAG-tagged bacterial alkaline phosphatase (Sigma) as a standard as described previously (31).
WGA-Agarose Chromatography-The affinity-purified protein was applied to a WGA-agarose column (bed volume 0.5 ml) equilibrated with 10 mM HEPES-NaOH, pH 7.2, 10 mM MgCl 2 , 2 mM CaCl 2 , 0.1% (v/v) Triton X-100, 20% glycerol, 150 mM NaCl. The column was washed with 2.5 ml of the same buffer. The absorbed materials were eluted with 2.5 ml of 0.5 M Glc-NAc in the buffer. The column was further eluted with 2.5 ml of 1.0 M NaCl in the buffer. All fractions were used to determine the C6ST activity.
Assay of Sulfotransferase Activity-C6ST activity and KSST activity were assayed by the method described previously (18,32). The standard reaction mixture contained, in a final volume of 50 l, 2.5 mol of imidazole-HCl, pH 6.8 (for C6ST activity) or pH 6.4 (for KSST activity), 1.25 g (for C6ST activity) or 3.75 g (for KSST activity) of protamine chloride, 0.1 mol of dithiothreitol, 25 nmol (as glucuronic acid) of chondroitin (for C6ST activity) or 25 nmol (as GlcNAc) of KS (for KSST activity), 50 pmol of [ 35 S]PAPS (ϳ5.0 ϫ 10 5 cpm), and enzymes. 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 precipitation with ethanol, followed by gel chromatography with a fast desalting column, as described previously (18), and radioactivity was determined. To determine C4ST and C6ST activity, 35 S-labeled chondroitin was digested with chondroitinase ACII, and the unsaturated disaccharides (2-acetamide-2-deoxy-3-O-(␤-Dgluco-4-enepyranosyluronic acid)-D-galactose and 2-acetamide-2-deoxy-3-O-(␤-D-gluco-4-enepyranosyluronic acid)-6-O-sulfo-D-galactose) were separated by paper chromatography, and the radioactivity was counted as described (18). To determine the V max for PAPS, chondroitin, and KS, the sulfotransferase activity of the recombinant proteins was normalized by the density of the protein bands on the Western blot detected by anti-FLAG antibody. To determine C6ST activity in the presence of KS, the 35 S-labeled glycosaminoglycans were digested with chondroitinase ABC, and the resulting unsaturated disaccharides were separated from KS by ethanol precipitation as described above. The radioactivity of the ethanol-soluble fractions was determined.
Digestion with PNGase F, Endo H, Chondroitinase ACII, Chondroitinase ABC, and Keratanase II-The affinity-purified recombinant C6ST-1 with or without mutations was precipitated with 3 volumes of ethanol containing 1.3% (w/v) potassium acetate. The precipitates were dissolved in 4 l of buffer 1 (0.5% SDS and 0.15 M Tris-HCl, pH 7.8) for PNGase F digestion or 4 l of buffer 2 (0.1 M sodium acetate, pH 5.2, 0.05% SDS, 0.2 M 2-mercaptoethanol) for Endo H digestion and heated at 100°C for 2 min. After cooling, 4 l of buffer 3 (3.2 mM phenylmethylsulfonyl fluoride, 4% Nonidet P-40, and 32 mM EDTA, pH 8.0) including 0.5 units of PNGase F or 5 milliunits of Endo H were added to the mixture and incubated at 37°C for 18 h. After digestion, the samples were subjected to SDS-PAGE as described above. Digestion of the recombinant protein with PNGase F under nondenaturing conditions was carried out at 37°C in the buffer used for the extraction of the cells as described above. Digestion with chondroitinase ACII or chondroitinase ABC was carried out for 4 h at 37°C in the reaction mixture containing, in a final volume of 25 l, 1.25 mol of Tris acetate buffer, pH 7.5, 2.5 g of bovine serum albumin, and 30 milliunits of chondroitinase ACII or chondroitinase ABC. Digestion with keratanase II was carried out for 24 h at 37°C in the reaction mixture containing, in a final volume of 50 l, 0.005 units of keratanase II and 2.5 mol of acetate buffer, pH 6.5 (33,34).
Immunocytochemistry-COS-7 cells were plated on the coverslips in 35-mm culture dishes at a density of 1.0 ϫ 10 5 cells/ dish. Cells were transfected as above. At 24 h after transfection, cells were fixed by ice-cold methanol for 30 min. After washing with phosphate-buffered saline (PBS) several times, the cells were blocked with 0.1% Tween 20 in PBS (PBST) containing 5% BSA. After washing with PBS, the cells were incubated with a mixture of anti-FLAG monoclonal antibody (1:500 dilution in 5% BSA-PBST) and anti-calnexin C terminus rabbit polyclonal

Effects of Tunicamycin on the Formation of the Active C6ST-1-
To determine the effects of N-glycans on the generation of the active C6ST-1, we cultured COS-7 cells transfected with C6ST-1 cDNA in the medium containing various concentration of tunicamycin. In the presence of 0.05 g/ml tunicamycin, only a protein band of 54 kDa that corresponds to the core protein without N-glycans was detected (Fig. 1B, lane 4), indicating that attachment of N-glycans to C6ST-1 was totally abolished. Under such conditions, C6ST activity was lost (Fig. 1A). These results indicate that attachment of N-glycans is essential for the generation of the active C6ST-1. In the presence of 0.01 g/ml tunicamycin, C6ST activity was fully retained, whereas the expressed protein exhibited multiple discrete bands on the Western blot (Fig. 1B, lane 3). These discrete protein bands are thought to correspond to different glycoforms of C6ST-1 bearing different numbers of N-glycans, because these protein bands were converted to the nonglycosylated form after the PNGase F digestion (data not shown). Partial loss of N-glycans in the presence of 0.01 g/ml of tunicamycin may not affect the production of the active C6ST-1 significantly.
Treatment of the Recombinant C6ST-1 with PNGase F-To obtain information about the effect of N-glycans attached to C6ST-1 on the enzyme activity, we digested the recombinant C6ST-1 with PNGase F under the nondenaturation conditions and determined C6ST activity ( Fig. 2A, open bar). The molecular mass of the recombinant C6ST-1 was also determined before or after PNGase F digestion (Fig. 2B). Before PNGase F  digestion, the glycosylated protein bands were detected as a single band at 67 kDa with a diffused band larger than 67 kDa (Fig. 2B, lane 1). These protein bands disappeared after PNGase F digestion, and a main protein band of 54 kDa and a faint smear larger than 54 kDa appeared (Fig. 2B, lane 2). Because the molecular mass of the recombinant C6ST-1 calculated from the cDNA is 53996, the main protein band detected after PNGase F digestion should correspond to the core protein without N-glycans, indicating that N-glycans attached to the recombinant C6ST-1 appeared to be removed nearly completely by PNGase F digestion under the nondenaturation conditions. On the other hand, C6ST activity was not decreased but rather increased after PNGase F digestion (Fig. 2A). These results suggest that N-glycans attached to C6ST-1 are not necessarily required for C6ST activity if the active enzyme is once generated. We previously reported that C6ST-1 could transfer sulfate not only to chondroitin but also to KS. Unlike C6ST activity, the activity toward KS (KSST activity) was decreased after PNGase F digestion ( Fig. 2A, closed bars). The roles of N-glycans on KSST activity will be discussed below.
Effects of Deletion of One N-Glycosylation Site-There are six potential N-glycosylation sites in C6ST-1, Asn 63 , Asn 74 , Asn 96 , Asn 250 , Asn 413 , and Asn 457 . To evaluate contribution of the individual N-glycan on these sites to the generation of the active C6ST-1, we constructed expression vectors in which the Asn residue of each N-glycosylation site was mutated to Ser and transfected into COS-7 cells. In Fig. 3, structures and names of these mutant cDNAs are shown. The C6ST activity of the expressed proteins was determined after the FLAG affinity purification (Fig. 4A, open bars). A marked decrease in the C6ST activity was observed in M-4 and M-6 recombinant proteins. When the recombinant proteins were detected with the anti-FLAG antibody after PNGase F digestion, only a protein band of 54 kDa was detected in every mutant recombinant protein at the same density as that of the wild-type recombinant protein (Fig. 4C), indicating that the introduction of mutation into any N-glycosylation site did not cause degradation of the recombinant proteins or alteration in the expression level of these proteins. Before PNGase F digestion, broad protein bands were detected at the migration position larger than 67 kDa in wild-type, M-1, M-2, M-3, and M-5 mutant proteins in addition to a main band of 67 kDa. The broad protein bands were not observed in M-4 and M-6 mutant protein (Fig. 4B), suggesting that the degree of microheterogeneity of N-glycans may be decreased in M-4 and M-6 mutant proteins. When the recombinant proteins were detected with the anti-FLAG antibody after Endo H digestion, the broad protein bands were still observed in wild-type, M-1, M-2, M-3, and M-5 mutant protein in addition to the 54-kDa band that corresponded to the core protein of C6ST-1. On the other hand, only the 54-kDa band was detected in M-4 and M-6 mutant proteins (Fig. 4D). Essentially the same results were obtained when CHO-K1 cells were used (data not shown). Thus, the decrease in C6ST activity of  M-4 and M-6 mutant proteins appears to be accompanied with the incomplete processing of N-glycans.
WGA-Agarose Chromatography-The Western blot analysis of the recombinant proteins described above suggest that structural alterations may have arisen in N-glycans of M-4 and M-6 mutant proteins. To confirm such a possibility, we examined the affinity of the recombinant proteins to WGA-agarose. We previously reported that C6ST-1 secreted to the culture medium of chick embryo chondrocytes was bound to WGAagarose and was eluted with a buffer containing GlcNAc. As shown in Fig. 5A, the wild-type recombinant protein was bound completely to the WGA-agarose and eluted with 0.5 M GlcNAc as observed in the purified chick C6ST-1. As shown in Fig. 5D, the pass-through fractions without C6ST activity contained a 54-kDa protein corresponding to the core protein, whereas the fractions eluted with 0.5 M GlcNAc contained a 67-kDa protein and diffused bands larger than 67 kDa. These observations clearly indicate that the active C6ST-1 bears N-glycans that were processed to the forms with affinity toward WGA. In contrast, when M-4 mutant protein was applied to the WGA-agarose column, more than 50% of the C6ST activity was recovered in the pass-through fractions (Fig. 5B), indicating that structures of N-glycans of M-4 mutant protein should be altered so that the affinity to WGA-agarose became lower. M-6 mutant protein behaved similarly to the wild-type recombinant proteins, albeit the proportion of the activity recovered in the pass-  through fractions was slightly larger than that of the wild-type recombinant protein (Fig. 5C).
Effects of Deletion of 3-6 N-Glycosylation Sites-To clarify the minimal requirement of N-glycans for the production of C6ST activity, N-glycans attached to C6ST-1 were deleted in several combinations (Fig. 6). It is evident from the Western blot shown in Fig. 6B that each N- C6ST activities of M-12356, M-12345, and M-123456 were not detected at all, whereas C6ST activity of M-1235 was 25% of C6ST activity of the wild-type recombinant C6ST-1 (Fig. 6A, open bar), indicating that the existence of N-glycans on both the N-4 and N-6 glycosylation sites is essential and sufficient for the production of the active C6ST-1.
As Effects of N-Glycans Attached to N-4 and N-6 Glycosylation Sites on the Intracellular Localization-It has been observed in various glycoproteins that intracellular localization of glycoproteins depended on the presence or absence of N-glycans (35)(36)(37). N-Glycans attached to the N-4 or N-6 glycosylation site of C6ST-1 may also affect the proper intracellular localization of C6ST-1 and thereby contribute to the production of the active C6ST-1. To investigate such a possibility, we stained the cells transfected with wild-type, M-4, or M-6 cDNA with anti-FLAG mAb (Figs. 7 and 8). The wild-type recombinant protein showed condensed perinuclear localization characteristic of Golgi-resident proteins and colocalized   -h), and M-6 (i-l) recombinant C6ST-1 were fixed and stained with a mixture of anti-FLAG monoclonal antibody and anti-calnexin C terminus rabbit polyclonal antibody followed by Alexa Fluorா 488-conjugated anti-mouse IgG goat antibody and TRITC-conjugated anti-rabbit IgG goat antibody. The nucleus of each cell was stained with Hoechst 33258. The panels show Nomarski differential interference contrast (a, e, and i ), FLAG immunofluorescence (b, f, and j ), calnexin immunofluorescence (c, g, and k), and merged images (d, h, and i ). Bar in a, 10 m.
Functional Role of N-Glycans Attached to C6ST-1 JULY 21, 2006 • VOLUME 281 • NUMBER 29 well with the Golgi marker protein, mannosidase II (Fig. 7, a-d) but only partially overlapped with the ER marker protein, calnexin (Fig. 8, a-d). On the other hand, M-4 mutant protein showed diffused localization around the nucleus and only partially colocalized with the Golgi marker protein (Fig. 7, e-h), but colocalized well with the ER marker protein (Fig. 8, e-h). These observations imply that the wild-type recombinant C6ST-1 may be localized at the Golgi apparatus, but the major part of the M-4 mutant protein may still be located at the ER. The intracellular localization of M-6 mutant protein was very similar to that of the wild-type recombinant C6ST-1, although the overlap with mannosidase II (Fig. 7, i-l) and with calnexin ( Fig. 8, i-l) was slightly less clear than that of the wild-type recombinant C6ST-1 and M-4 mutant protein, respectively.
Effects of N-Glycans on KSST Activity-We previously found that purified C6ST-1 transferred sulfate not only to position 6 of the GalNAc residue of chondroitin (C6ST activity) but also to position 6 of the Gal residue of KS (KSST activity) (18,32). We confirmed that the FLAG-tagged recombinant C6ST-1 also exhibited both C6ST activity and KSST activity (Fig. 9). The sulfated product obtained from chondroitin was degraded by chondroitinase ABC but not by keratanase II (Fig. 9, A-C), whereas the sulfated product obtained from KS was degraded by keratanase II but not by chondroitinase ABC (Fig. 9, D-F). As shown in Fig. 2A, KSST activity of the wild-type recombinant C6ST-1 was decreased after PNGase F digestion, suggesting that the presence of some N-glycans may be required for the KSST activity. To elucidate the contribution of each N-glycan to KSST activity, we compared KSST activity of the mutant recombinant proteins in which one of the N-glycosylation sites was mutated (Fig. 4A, closed bar). KSST activity of M-4 and M-6 mutant proteins were almost lost as observed in C6ST activity. Unlike C6ST activity, KSST activity of M-5 mutant protein was  To determine the importance of the N-glycan attached to the N-5 glycosylation site in KSST activity, we compared KSST activity among various mutant C6ST-1s in which 3-6 N-glycosylation sites were deleted (Fig. 6A, closed bar). As observed in C6ST activity, KSST activity of M-12356, M-12345, and M-123456 were not detected at all. KSST activity of M-123 was 33% of KSST activity of the wild-type recombinant C6ST-1 (Fig. 6), whereas KSST activity of M-1235 was almost lost. These results clearly indicate that N-glycan attached to the N-5 glycosylation site was essential for the KSST activity. To consider the reason why M-5 mutant recombinant protein exhibited low KSST activity, we determined the inhibitory effects of KS on C6ST activity (Fig.   10). M-5 mutant protein was inhibited by KS more strongly than the other recombinant proteins; in the presence of 7 nmol (as glucosamine) of KS, C6ST activity of M-5 mutant protein decreased to 20% of the control, whereas C6ST activity of wildtype and other mutant proteins were not affected at all, and in the presence of 14 nmol of KS, C6ST activity of M-5 mutant protein was less than 6% of control, whereas wild type and the mutant proteins other than M-5 mutant still retained about 50% of C6ST activity. These observations suggest that apparent low KSST activity of M-5 mutant recombinant protein may be attributable to the augmented sensitivity of M-5 mutant recombinant protein to the substrate inhibition.

Effects of Deletion of N-Glycans Attached to the Recombinant C6ST-1 on the Kinetics Parameters for C6ST
Activity and KSST Activity-To evaluate the effects of N-glycans on C6ST activity and KSST activity of the recombinant C6ST-1 quantitatively, we determined the K m and the V max of the mutant recombinant C6ST-1 for PAPS and for each acceptor (Table 2)

DISCUSSION
In this report, we examined functional roles of N-glycans attached to C6ST-1. Treatment of COS-7 cells transfected with C6ST-1 cDNA with tunicamycin abolished both N-glycosylation of the recombinant C6ST-1 and production of the active enzyme. On the other hand, C6ST activity of the recombinant C6ST-1 was not decreased but rather increased by PNGase F digestion under the conditions where N-glycans of the recombinant C6ST-1 were nearly completely removed. These observations indicate that N-glycans of C6ST-1 are essential for the generation of the active C6ST-1 in the cells but not required for the C6ST activity itself after the formation of the active protein.
Among six potential N-glycans, N-glycans attached to N-4 and N-6 glycosylation sites appeared to be necessary and sufficient for production of the active C6ST-1, because deletion of N-4 or N-6 glycosylation site resulted in marked decrease in C6ST activity, and M-1235 mutant protein still retained 25% of the C6ST activity of the wild-type recombinant C6ST-1.
The effects of N-glycans on KSST activity showed somewhat different aspects from the effects on C6ST activity. KSST activity of the recombinant C6ST-1 was decreased by the PNGase F digestion. When the N-5 glycosylation site was deleted, C6ST activity was hardly affected, but a marked loss of the KSST activity was observed. Unlike C6ST activity, KSST activity of M-1235 mutant protein almost disappeared. These results indicate that KSST activity of the recombinant C6ST-1 appears to be supported by the presence of N-glycan attached to the N-5 glycosylation site in addition to N-glycans attached to the N-4 and N-6 glycosylation sites. At present, the mechanism by which N-glycan attached to the N-5 glycosylation site could stimulate KSST activity is not clear. It has been shown that, in the corneal KS used for the substrate, almost all GlcNAc residues and about a half of the Gal residues are sulfated at position 6 (38). Because C6ST-1 transfers sulfate to position 6 of the Gal residue of KS, the corneal KS contains structures acting as both the substrate and the reaction product of C6ST-1. The structure corresponding to the reaction products in which Gal(6SO 4 ) is abundantly contained may inhibit the sulfotransferase activity of C6ST-1. The presence of N-glycan at the N-5 glycosylation site might block the binding of Gal(6SO 4 ) residues of KS and thereby relieve the inhibition by KS. In Fig. 10, we showed that the degree of inhibition of C6ST activity by KS was higher in M-5 mutant protein than in the wild-type recombinant C6ST-1. The observed low KSST activity of M-5 mutant protein may be due partly to the increased inhibition by KS used for the acceptor.
In C6ST-1, N-glycan attached to the N-4 glycosylation site appears to be involved in the Golgi localization, because M-4 mutant protein colocalized with calnexin, an ER-resident protein, although the wild-type recombinant C6ST-1 showed a typical Golgi localization (Figs. 7 and 8). M-4 mutant protein was shown to bear no Endo H-resistant glycans (Fig. 4), suggesting that N-4 glycan may be required for the interaction with ER-resident chaperons, such as calnexin, to form the correctly folded protein. Deletion of N-4 glycan may thus result in the accumulation of the misfolded protein in the ER, thereby inhibiting processing of the N-glycans to the Endo H-resistant form that occurs in the Golgi apparatus. To clarify the functional role of N4-glycan in the folding of the C6ST-1, structural characterization of N-4 glycan should be important. On the other hand, M-6 mutant protein also had no Endo H-resistant glycans, although intracellular localization of M-6 mutant protein was similar to that of wild-type recombinant C6ST-1. M-6 mutant protein may be located in the Golgi subcompartment different from the subcompartment where wild-type recombinant protein resides, and hence processing of the N-glycans to the Endo H-resistant form might be ablated. It remains to be studied whether the Golgi subcompartments where M-6 mutant protein and wild-type recombinant protein reside are the same or different.
On Western blot, the wild-type protein showed a 54-kDa core protein band, a 67-kDa slightly broad band, and higher smeared ones. In Fig. 4, the higher smeared bands appeared to correlate to C6ST activity, suggesting that the smeared bands correspond to the active enzyme. To determine if such a possibility is the case, we analyzed fractions eluted from the WGAagarose by Western blot (Fig. 5D). However, because the 67-kDa band and the higher smeared one behaved identically in WGA-agarose and both peaked at the peak position of C6ST activity, it remains unclear whether the smeared bands correspond to the active enzyme.
At present, it is not clear whether C6ST-1 is relevant to the sulfation of KS in vivo. In the brain of C6ST-1-deficient mice, synthesis of KS does not appear to be affected, because the staining pattern of the brain tissues of the mice with 5D4 antibody could not be distinguished from that of the wild mice (13). Because 5D4 antibody is able to stain the highly sulfated KS (39,40), involvement of C6ST-1 in the synthesis of the lower sulfated KS would not be thoroughly excluded. Functional roles of N-5 glycan on C6ST-1 in the synthesis of KS in vivo remain to be studied.
M-4 and M-6 mutant proteins were devoid of the Endo H-resistant N-glycans, whereas N-4 and N-6 glycans themselves seem not to be the Endo-H-resistant forms, because M-12356 and M-12345 mutant proteins were both sensitive to Endo-H. It is thus possible that N-4 and N-6 glycans may affect the processing of the other N-glycans by an as yet undefined manner. We have previously reported that deletion of the N-glycan attached to the C-terminal region of C4ST-1 also affected microheterogeneity of N-glycans attached to different sites of C4ST-1 (26). There may exist a common mechanism by which proper processing of N-glycans is influenced by the presence or absence of one N-glycan.
We have shown previously that KS Gal-6-sulfotransferase transfers sulfate to position 6 of the Gal residue in KS (41). As far as we know, only two sulfotransferases, C6ST-1 and KS Gal-6-sulfotransferase, are able to sulfate position 6 of the Gal residue in KS. ClustalW alignment of these genes showed that the N-glycosylation site corresponding to the N-4 site is the only one that is conserved between these genes and that the amino acid sequence around the N-4 site is well conserved between these proteins (Fig. 11), suggesting that