Molecular Cloning and Expression of Two Distinct Human Chondroitin 4- O -Sulfotransferases That Belong to the HNK-1 Sulfotransferase Gene Family*

Using an expression cloning strategy, the cDNA encoding the human HNK-1 sulfotransferase (HNK-1ST) has been cloned. During this cloning we found that HNK-1ST and other Golgi-associated sulfotransferases cloned before share homologous sequences including the RDP motif (Ong, E., Yeh, J.-C., Ding, Y., Hindsgaul, O., and Fukuda, M. (1998) J. Biol. Chem . 223, 5190–5195). Using this conserved sequence in HNK-1ST as a probe, we identified two expressed sequence tags in EST data base which have 31.6 and 30.7% identity with HNK-1ST at the amino acid levels. Expression of these two full-length cDNAs failed to form HNK-1 glycan nor to add sulfate to CD34 or NCAM. Surprisingly, proteins expressed by these cDNAs transferred sulfate to the C-4 position of N -acetylgalactosamine in chondroitin and desulfated dermatan sulfate, thus we named these two enzymes, chondroitin 4- O -sulfotransferase 1 and -2 (C4ST-1 and C4ST-2). Both C4ST-1 and C4ST-2, however, did not form 4,6-di- O -sulfated N -acetylgalactosamine when chondroitin sulfate C was used as an acceptor.

Sulfate groups in carbohydrates play important roles in conferring highly specific functions on glycoproteins, glycolipids, and proteoglycans (1)(2)(3). Expression of certain sulfated carbohydrates is spatially and temporally controlled, thereby provid-ing developmental regulation of those functions displayed by such sulfation. One of these sulfated glycans is the HNK-1 glycan (4,5). The functional significance of HNK-1 glycan was first recognized as an antigen involved in peripheral demyelinative neuropathy. The structural analysis of glycolipids reactive with the autoantibodies led to the discovery that the HNK-1 epitope is sulfo33GlcA␤1 1 33Gal␤134GlcNAc␤3 R (6,7).
Subsequently, HNK-1 glycan has been found in a number of neural cell adhesion molecules, including NCAM, myelin-associated glycoprotein, L1, contactin, and Po (5, 8 -11). Using monoclonal antibodies or isolated carbohydrates, various laboratories reported that HNK-1 glycan is involved in cell-cell and cell-substratum interactions (12,13). In one study, a non-sulfated form of HNK-1 precursor glycan did not facilitate neurite outgrowth as opposed to a functional, intact HNK-1 glycan (13). These results, combined together, suggest that HNK-1 glycan plays critical roles in development, in particular during neural cell development.
The HNK-1 carbohydrate is synthesized by the addition of a sulfate to ␤1,3-glucuronylated N-acetyllactosamine, GlcA-␤133Gal␤134GlcNAc3 R (4). Recently, we and others cloned the cDNA encoding HNK-1 sulfotransferase using an expression cloning strategy (14,15). During this cloning, we discovered that the newly cloned HNK-1 sulfotransferase and other Golgi-associated sulfotransferases cloned before share a common sequence motif, which includes ZZRDPXXXZ, where X and Z denote any amino acid and hydrophobic amino acids, respectively (14). Subsequently, it was revealed that this sequence motif is a part of the binding site for 3Ј-phosphate group of the donor substrate, 3Ј-phosphoadenosine 5Ј-phosphosulfate (PAPS) (16,17). Most recent studies showed that the arginine residue (Arg) in the RDP motif is involved in hydrogen bonding to 3Ј-phosphate group while aspartic acid (Asp) and proline (Pro) residues participate in the core structure of the 3Ј-phosphate-binding site by residing in a tight turn of the polypeptides (17,18). In addition, the amino acid sequences responsible for binding to 5Ј-phosphosulfate are conserved among different sulfotransferases (17).
The presence of the above weak but discernible similarity among different sulfotransferases suggested a possibility that other sulfotransferases may be identified by their similarity to sulfotransferases cloned already. In fact, we and others cloned the cDNA encoding L-selectin ligand sulfotransferases that add a sulfate to the 6-position of N-acetylglucosamine, which is eventually converted to 6-sulfo sialyl Lewis X, NeuNAc␣233-Gal␤134[sulfo36(Fuc␣133)GlcNAc]␤136R (19 -21). This cloning was achieved by searching the EST data base for cDNAs related to chondroitin sulfate 6-O-sulfotransferase (22) and keratan sulfate Gal-6-O-sulfotransferase (23).
In the present study, we first describe the isolation of two isoforms of cDNAs by screening the EST data base for cDNAs related to the human HNK-1 sulfotransferase (14). The expression of full-length cDNAs unexpectedly revealed that these cDNAs encode novel chondroitin 4-O-sulfotransferases, adding a sulfate to 4-position of N-acetylgalactosamine residues in chondroitin and desulfated dermatan sulfate. Moreover, we found that these two chondroitin 4-sulfotransferases exhibit diverse tissue distribution, indicating that these two enzymes play complementary roles in different tissues.

Isolation of cDNAs Encoding Chondroitin 4-O-Sulfotransferases-In
HNK-1ST, the conserved motif, IVRDPFERL residues in amino acid residues 187-195 (14). The amino acid sequence of residues 165-230, which includes the above motif, was thus used as a probe to search dbEST using the TBLASTN program. Initially, two query genes AA310375 and AA233362 were identified, which had 50% in 50 amino acids and 69% in 26 amino acids identity with HNK-1ST, respectively.
After blast search for a sequence homologous to AA310375, AA744877 was identified. AA744877 is a cDNA prepared from germinal center B lymphocytes. Sequence analysis of this cDNA, obtained from Genome Systems (St.Louis, MS), revealed that this cDNA encodes a protein consisting of 352 amino acids. The cDNA also contains 5Јuntranslated sequence (150 base pairs) and 3Ј-untranslated sequence (330 base pairs). The cDNA insert was digested with HindIII and XhoI and cloned into the same sites of pcDNA3.1/Hygro (Invitrogen, Carlsbad, CA), resulting in pcDNA3.1-C4ST-1 (the name of C4ST-1 was given after the determination of acceptor specificity).
The second gene was initially identified in AA233362 and AA777237 derived from the human NT2 cell line and SS20w fetal liver/spleen. Since these two clones lacked the 5Ј-region, 5Ј-rapid amplification of cDNA ends was carried out using poly(A) ϩ RNA from human lymph nodes (CLONTECH, Palo Alto, CA). However, a new EST sequence, AA182585, which contained the full coding sequence was released in the meantime. The cDNA was thus excised from AA182585 with BamHI and XhoI and cloned into the same sites of pcDNA3.1/Hygro, resulting in pcDNA3.1-C4ST-2.
Chondroitin sulfate 4-and 6-O-sulfotransferase and heparan sulfate sulfotransferase activities were assayed as described previously (26). Briefly, the reaction mixture (50 l) contained 50 mM imidazole-HCl, pH 6.8, 0.005% protamine chloride, 2 mM dithiothreitol, 50 g of acceptor glycosaminoglycans, 2 M [ 35 S]PAPS (about 5 ϫ 10 5 cpm), and 25 l of an enzyme solution. After incubation for 1 h at 37°C, the reaction mixture was boiled for 2 min, then 0.1 volume of 4 M potassium acetate and 3 volumes of ethanol were added. The reaction products were precipitated by brief centrifugation, and subjected to Sephadex G-25 gel filtration in 0.1 M NH 4 HCO 3 to separate high molecular weight products from the remaining unreacted [ 35 S]PAPS and degradation products.
Chondroitin sulfate A (whale cartilage), chondroitin sulfate C (shark cartilage), completely desulfated and N-sulfated heparin, completely desulfated and N-acetylated heparin, and keratan sulfate were purchased from Seikagaku Corp. (Tokyo). Dermatan sulfate (porcine intestinal mucosa) was purchased from Calbiochem (San Diego, CA). Dermatan sulfate was subjected to chemical desulfation (27) before use as an acceptor. Desulfated dermatan sulfate produced ⌬Di-0S in more than 95% of total unsaturated disaccharides after chondroitinase ABC digestion, confirming that more than 95% of sulfate group was removed.
Analysis of Enzymatic Reaction Products-Enzymatic reaction products were analyzed after digestion with chondroitinase ABC (28) (Seikagaku Corp.), AC I Flavo (29) (Calbiochem), or chondroitinase B (Refs. 30 and 31, Calbiochem) and analyzed by HPLC or Bio-Gel P-4 gel filtration. Briefly, 35 S-labeled products were digested with 25 milliunits of chondroitinase ABC for 1 h at 37°C, 50 milliunits of chondroitinase AC I for 16 h at 37°C, or 25 milliunits of chondroitinase B for 20 h at 30°C. The resultant oligosaccharides were separated by HPLC using a Whatman Partisil SAX-10 column (4.6 mm ϫ 25 cm) (Whatman, Clifton, NJ) equilibrated with 35 mM KH 2 PO 4 at room temperature. The elution condition was modified from that published before (26). The column was eluted with a linear gradient from 35 mM KH 2 PO 4 to 135 mM KH 2 PO 4 in the first 20 min, then to 335 mM KH 2 PO 4 in the next 20 min. Finally, the elution was linearly increased to 535 mM KH 2 PO 4 in the additional 10 min. The column was then re-equilibrated with 35 mM KH 2 PO 4 . The flow rate was 1 ml/min and each fraction contained 0.5 or 1 ml. The products from [ 35 S]sulfate-labeled dermatan sulfate were also applied to a column (10 mm ϫ 120 cm) of Bio-Gel P-4 (Bio-Rad, Hercules, CA) equilibrated with 0.1 M NH 4 CH 3 CO 2 as described previously (20).
Chromosome Mapping-DNA samples were prepared from 83 radiation hybrids of human X rodent somatic cell hybrids containing human minichromosomes of the Stanford Human Genome Center G3 RH panel A (32, 33) (Research Genetics, Huntsville, AL). To determine the C4ST-1 and C4ST-2 loci, these DNA samples were analyzed by PCR. The PCR primers used to amplify the sequences corresponded to nucleotides 384 -802 for C4ST-1 and nucleotides 528 -797 for C4ST-2.
The PCR conditions were 10 cycles at 94°C for 1 min, 60°C for 1 min, and 72°C for 30 s followed by 25 cycles at 94°C for 1 min, 63°C for 1 min, and 72°C for 30 s. The maximum likelihood estimation was carried out by submitting the results to the RH server at the Stanford Genome Center and NCBI Gene Map '98, as described previously (33).

Isolation of cDNAs Encoding Chondroitin 4-O-Sulfotransferases (C4STs)-
Comparison of the amino acid sequences of cloned sulfotransferases demonstrated that there is a weak but discernible homologous sequence motif among Golgi-associated sulfotransferases (14). In particular, the RDP sequence motif was conserved among those sulfotransferases compared. By searching the EST data base for a novel cDNA related to HNK-1ST, two distinct cDNA sequences were found to have homology to the HNK-1ST sequence. The first cDNA (AA744877 in dbEST) encodes an open reading frame of 1,059 base pairs, predicting a protein of 352 amino acid residues (41,488 Da), which we subsequently termed C4ST-1 (Fig. 1). The second cDNA (AA182585 in dbEST) encodes an open reading frame of 1,245 base pairs, predicting a 414-amino acid residue protein (48,348 Da), which we subsequently termed C4ST-2 (Fig. 2). The cDNAs encoding C4ST-1 and C4ST-2 were cloned into pcDNA3.1/Hygro, resulting in pcDNA3.1-C4ST-1 and pcDNA3.1-C4ST-2, respectively.
The comparison of the amino acid sequences of C4ST-1 and C4ST-2 with HNK-1ST reveals the following points (Fig. 3). The sequences of the cytoplasmic segment and the transmembrane/anchoring domain are not strongly similar among these sulfotransferases, while the sequences are highly homologous to each other in the catalytic domains. There are four regions where sulfotransferases are highly homologous. The first two are the 5Ј-phosphosulfate-binding and 3Ј-phosphate-binding sites, respectively (Fig. 3). The third and fourth regions (A and B in Fig. 3) have not been reported before, but probably corresponds to two ␣-helical domains near the carboxyl-terminal ends (34).
When the activity of C6ST was assayed, C6ST incorporated [ 35 S]sulfate into chondroitin, chondroitin sulfate A, chondroitin sulfate C, and keratin sulfate, as expected (22). On the other hand, HNK-1ST did not show any detectable activity toward these glycosaminoglycan acceptors (Fig. 4). These results indicate that newly cloned C4ST-1 and C4ST-2 are sulfotransferases that add sulfate(s) to chondroitin, chondroitin sulfate, and desulfated dermatan sulfate.
Identification of Reaction Products-The above results showed that both C4ST-1 and C4ST-2 utilized almost identical acceptors, but did not show if C4ST-1 and C4ST-2 added sulfate to the 4-or 6-position of N-acetylgalactosamine or the 2-position of D-glucuronic acid.
To determine the structures of the sulfated products derived from chondroitin, we took advantage of the fact that isomers of sulfated disaccharide units produced by chondroitinase ABC can be separated by SAX-10 HPLC. As shown in Fig. 5A, almost all of the products by C4ST-1 eluted at the position of ⌬Di-4S.
The peak corresponding to ⌬Di-4S released sulfate after treatment with chondro-4-sulfatase (Fig. 5B), but not with chondro-6-sulfatase (Fig. 5C). These results, combined together, indicate that C4ST-1 incorporated a sulfate to the 4-position of N-acetylgalactosamine in chondroitin. The products derived from chondroitin sulfate A or C showed a prominent peak corresponding to ⌬Di-4S after chondroitinase ABC digestion, but did not contain any disulfated disaccharide (Fig. 5D), indicating that C4ST-1 adds a sulfate group only when neither glucuronic acid nor N-acetylgalactosamine in the acceptors contain a sulfate group. The amount of ⌬Di-6S was almost the same as that observed in control experiments, indicating that 6-O-sulfation was due to an endogenous enzyme (Fig. 5A). The products from C4ST-2 were analyzed in an identical manner. The results are very similar to those described for C4ST-1 (Fig.  5, E-H).
Sulfation of Dermatan Sulfate by C4ST-1 and C4ST-2-Both C4ST-1 and C4ST-2 incorporated [ 35 S]sulfate to desulfated dermatan sulfate (Fig. 4). To determine how C4ST-1 and C4ST-2 act on dermatan sulfate, 35 S-labeled products obtained from desulfated dermatan sulfate were digested with chondroitinase ABC. HPLC analysis of the digested material showed that C4ST-1 produced ⌬Di-4S, indicating that C4ST-1 added a sul-  (Fig. 6A). No disulfated disaccharide was detected. Almost identical results were obtained for C4ST-2 (data not shown). To further delineate the acceptor specificity of C4ST-1, the 35 S-labeled products were digested with chon-droitinase AC I, which cleaves only N-acetylgalactosaminyl linkage to D-glucuronic acid. The results demonstrated that approximately one-fourth of the total radioactivity was released as ⌬Di-4S and the rest eluted in later fractions (Fig. 6B). After digestion with chondro-4-sulfatase, the peak correspond- ing to ⌬Di-4S disappeared and a prominent free sulfate ion peak appeared instead (Fig. 6C). However, no significant change in larger 35 S-labeled oligosaccharides, eluted after 24.5 min, was observed, indicating that chondro-4-sulfatase did not release sulfate from oligosaccharides larger than disaccharides. Digestion of the same material by chondro-6-sulfatase, on the other hand, barely changed the elution profile (Fig. 6D), being consistent with the above conclusions that C4ST-1 incorporated [ 35 S]sulfate to the 4-position of N-acetylgalactosamine.
To determine the nature of larger oligosaccharides obtained after chondroitinase AC I treatment, the same sample analyzed in Fig. 6D was subjected to Bio-Gel P-4 gel filtration. The results showed that approximately one-fourth of the total radioactivity eluted at ⌬Di-4S and approximately 10% of total radioactivity eluted as tetrasaccharide and hexasaccharide (Fig. 6E). Chondroitinase AC I can release ⌬Di-4S only from GlcA3 GalNAc(4S) that is flanked by GlcA3 GalNAc units (29). These results suggest that [ 35 S]sulfate was incorporated into GlcA3 GalNAc unit.
To corroborate the above experiments, intact [ 35 S]sulfatelabeled dermatan sulfate was directly digested by chondroiti-nase B, which cleaves a sulfated N-acetylgalactosaminyl linkage to iduronic acid flanked by IdoA3 GalNAc units (30,31). The results demonstrated no release of 35 S-labeled ⌬Di-4S or 35 S-labeled oligosaccharides (Fig. 6F). These results combined together indicate that C4ST-1 and most likely C4ST-2 preferentially incorporate a sulfate at the 4-position of N-acetylgalactosamine in GlcA3 GalNAc than in the IdoA3 GalNAc unit.
C4ST-1 and C4ST-2 Are Differentially Expressed in Various Tissues-To determine the expression of C4ST-1 and C4ST-2 transcripts in various tissues, Northern and dot blot analysis was carried out. Gel fractionated blot (Fig. 7) and dot blot (Fig.  8) analyses show that the C4ST-1 transcript is highly expressed in spleen, thymus, peripheral blood leukocytes, lymph node, bone marrow, lung, and placenta. In contrast, the transcripts of C4ST-2 are expressed more ubiquitously (Fig. 7), but significantly more in spinal cord, heart, thyroid, pituitary gland, adrenal gland, small intestine, spleen, peripheral blood leukocytes, thymus, lung, fetal kidney, fetal spleen, and fetal lung on the dot blot (Fig. 8). These results show that C4ST-1 is mostly expressed in leukocytes and hematopoietic tissues, while C4ST-2 is widely expressed in various tissues, including and C4ST-2 genes, PCR analysis was carried out using the Stanford G3 RH panel. PCR primers were chosen from coding regions of C4ST-1 and C4ST-2 genes, and based on the criteria that PCR products showed the same molecular weight when C4ST-1 or C4ST-2 cDNA or genomic DNA was used as a template, but not using hamster genomic DNA. This analysis placed C4ST-1 between D12S1607 and D12S360, thus mapping the gene to the q23 region of chromosome 12. Similarly, the C4ST-2 gene was placed between D7S2563 and D7S2521, mapping the gene to the p22 region of chromosome 7.

DISCUSSION
The present study describes the isolation of novel cDNAs encoding chondroitin 4-O-sulfotransferase by searching the EST data bases for cDNAs homologous to the human HNK-1ST (14). HNK-1ST adds a sulfate to the 3-position of glucuronic acid, which in turn is attached to the 3-position of galactose in N-acetyllactosamine. C4ST, on the other hand, adds a sulfate to the 4-position of N-acetylgalactosamine, which is in turn attached to the 4-position of glucuronic acid. These results are striking since these two acceptors are rather dissimilar. The ␤-glucuronyl residue in HNK-1 glycan is at the nonreducing terminal. In contrast, C4ST apparently acts on an already elongated chondroitin chain since no preferential addition to shorter acceptors has been noticed when the products were analyzed by gel filtration (data not shown, see also Ref. 26). The hydroxyl groups in both C-3 of glucuronic acid and C-4 of N-acetylgalactosamine are projected above their respective pyranose rings in their normal conformations (37). It is tempting to speculate that the active sites of both HNK-1ST and C4ST may approach the acceptor from above the plane of GlcA␤133Gal␤134GlcNAc␤13 R (for HNK-1ST) and

GlcA␤133GalNAc␤134GlcA␤13 R (for C4ST).
It is noteworthy that C4ST-1 and C4ST-2 share only 41.8% identity at the amino acid levels, but share a common catalytic property. C4ST-1 and C4ST-2, however, are much more homologous to each other in the vicinity of 5Ј-phosphosulfate and 3Ј-phosphate binding sites (Fig. 3). Moreover, C4ST-1 and C4ST-2 apparently share common structural domains toward the carboxyl-terminal regions (A and B in Fig. 3). These regions do not share homology with other sulfotransferases (34) and further studies are necessary to determine their roles. Fig. 9 illustrates the phylogenetic relationship of cloned Golgi-associated sulfotransferases that add a sulfate on carbohydrate acceptors. The results clearly indicate that C4ST-1, C4ST-2, and HNK-1ST form a gene family distinct from the rest of the sulfotransferase gene families. The members within the same gene family depicted in Fig. 9 catalyze identical or similar reactions, except for one case. LSST, I-GlcNAc6ST, GlcNAc6ST, C6ST, and KSST form a gene family whose acceptor specificities are not clearly related to each other. The cDNAs (GlcNAc6ST, LSST, and I-GlcNAc6ST) encoding a sulfotransferase that adds a sulfate to the 6-position of N-acetylglucosamine at the nonreducing terminal were identified in EST data base for their homology to C6ST or KSST (19 -21, 38). In contrast, C6ST and KSST add a sulfate on the 6-position of N-acetylgalactosamine or galactose on already elongated substrates (22,23). These results, combined together with the results obtained in the present study, indicate that it is possible to identify cDNAs encoding enzymes that utilize very different acceptors from those utilized by a protein whose cDNA was used as a probe. Further studies will be significant to determine the three-dimensional structures of these sulfotransferases bound to acceptors in order to test if these enzymes approach their acceptors from above the plane of the acceptors.
The present study demonstrated that both C4ST-1 and C4ST-2 act much more efficiently on non-sulfated chondroitin or desulfated dermatan sulfate than chondroitin sulfate A, chondroitin sulfate C, or dermatan sulfate (Fig. 4). No disulfated disaccharide was released after chondroitinase ABC digestion of reaction products derived from chondroitin sulfate C (Fig. 5). These results indicate that C4ST-1 and C4ST-2 add sulfate only on unsulfated N-acetylgalactosamines.
The present study also demonstrated that C4ST-1 and C4ST-2 add a sulfate to the 4-position of N-acetylgalactosamine residues in dermatan sulfate, which had been chemically desulfated (Figs. 4 and 6). The detailed analysis of dermatan sulfated by C4ST-1 revealed the following points. Even though glucuronic acid residues are minor components in the dermatan sulfate, at least one-fourth of the total radioactivity was detected in ⌬Di-4S when released by chondroitinase AC I digestion. In this case, ⌬Di-4S was released only when 4-sulfated N-acetylgalactosamine are positioned between two glucuronic acids. This finding suggests that C4ST-1 acts on N-acetylgalactosamine residues next to glucuronic acid. If C4ST-1 transfers a sulfate to N-acetylgalactosamine linked to iduronic acid as efficiently as to N-acetylgalactosamine linked to glucuronic acid, more 35 S-labeled oligosaccharides would be released by chondroitinase B digestion than by chondroitinase AC I digestion. However, ⌬Di-4S was hardly released after chondroitinase B digestion (Fig. 6). These results, combined together, indicate that the GlcA3 GalNAc unit is a much better acceptor for C4ST-1 (and most likely for C4ST-2 as well) than the IdoA3 GalNAc unit.
The results obtained in the present study are similar to those obtained on C4ST purified from a rat chondrosarcoma cell line (26). However, the C4ST in that study added a sulfate more on desulfated dermatan sulfate (porcine skin) and the products were highly susceptible to chondroitinase AC II, which cleaves only a GlcA3 GalNAc unit flanked by GlcA3 GalNAc (39). This discrepancy is probably due to the difference in the source of dermatan sulfate and that the dermatan sulfate from pig skin probably contains more glucuronic acid residues which are clustered than the dermatan sulfate from porcine intestinal mucosa used in the present study.
The detailed biosynthetic steps of dermatan sulfate are currently unknown. Malström (40) showed that epimerization from glucuronic acid to iduronic acid takes place in unsulfated chondroitin. However, the conversion to iduronic acid in that report was only 15%, which is much lower than the actual iduronic acid content in nature (41). Moreover, this epimerization is reversible for unsulfated forms, but IdoA3 GalNAc(4S) is not converted to GlcA3 GalNAc(4S). On the other hand, Silbert et al. (42) showed that lower sulfation leads to lower epimerization, suggesting that sulfation at the 4-position of N-acetylgalactosamine precedes epimerization. If C4ST-1 and C4ST-2 involve mainly in the biosynthesis of dermatan sulfate, the expression of C4ST-1 or C4ST-2 should result in higher expression of dermatan sulfate containing large amounts of IdoA3 GalNAc(4S). However, if the same enzyme contributes to form both chondroitin sulfate and dermatan sulfate, several factors may determine the destination of these synthesized molecules to chondroitin sulfate or dermatan sulfate. When an epimerase is expressed, it converts a glucuronic acid to an iduronic acid, and this reaction is accelerated by the sulfation of N-acetylgalactosamine introduced by C4ST. In contrast, in the absence of an epimerase, no dermatan sulfate is formed. Thus, the presence of an epimerase should be a main regulator, but another possibility needs to be considered. The results of in vitro enzyme assay showed that the sulfotransferase activity of C4ST to chondroitin and dermatan reciprocally changes depending upon the concentration of protamine chloride in the reaction mixture (Ref. 26 and the present study). The concentration of protamine chloride also affects the activity of chondroitin 6-O-sulfotransferase including its substrate specificity (43). These findings suggest that the environmental factors affecting the activity of C4ST might contribute to the regulation of the chondroitin sulfate and dermatan sulfate biosynthesis. It is also possible that another sulfotransferase preferentially acting on IdoA3 GalNAc is involved in dermatan sulfate biosynthesis. Further studies will be significant to clarify these points.
While we were preparing this manuscript, the mouse counterpart of C4ST-1 was reported (44). The human and mouse C4ST-1 have 96.0% identity at the amino acid levels. The expression profile of the mouse C4ST is slightly different from that of human C4ST-1 in that it is mainly expressed in the brain and kidney. The mouse C4ST-1, however, apparently exhibits almost the same substrate specificity as the human C4ST-1 and C4ST-2 (44).
The transcripts of human C4ST-1 and C4ST-2 are differentially expressed in various tissues. The C4ST-1 transcript is predominantly expressed in peripheral blood leukocytes and hematopoietic tissues such as, bone marrow and spleen, while the C4ST-2 transcript is more widely expressed, including in the pituitary gland, adrenal gland, spinal cord, small intestine, spleen, and lung ( Fig. 7 and 8). These results indicate that C4ST-1 and C4ST-2 may play complementary roles in different tissues.
Chondroitin sulfate proteoglycans have been found in the brain and have been shown to play roles in neural cell adhesion and neurite outgrowth (45)(46)(47)(48), and neural cell migration (49). Chondroitin sulfate is also present in blood cells and has been shown to be involved in interaction with CD44 (50, 51) and L-selectin (52). C4ST-1 and C4ST-2 cloned in the present study will be powerful tools to determine the roles of chondroitin sulfate in these various biological systems.