Molecular Characterization and Expression of Heparan-sulfate 6-Sulfotransferase

Heparan-sulfate 6-sulfotransferase (HS6ST) catalyzes the transfer of sulfate from 3′-phosphoadenosine 5′-phosphosulfate to position 6 of the N-sulfoglucosamine residue of heparan sulfate. The enzyme was purified to apparent homogeneity from the serum-free culture medium of Chinese hamster ovary (CHO) cells (Habuchi, H., Habuchi, O., and Kimata, K. (1995)J. Biol Chem. 270, 4172–4179). From the amino acid sequence data of the purified enzyme, degenerate oligonucleotides were designed and used as primers for the reverse transcriptase-polymerase chain reaction using poly(A)+ RNA from CHO cells as a template. The amplified cDNA fragment was then used as a probe to screen a cDNA library of CHO cells. The cDNA clone thus obtained encoded a partial peptide sequence composed of 236 amino acid residues that included the sequences of six peptides obtained after endoproteinase digestion of the purified enzyme. This cDNA clone was applied to the screening of a human fetal brain cDNA library by cross-hybridization. The isolated cDNA clones contained a whole open reading frame that predicts a type II transmembrane protein composed of 401 amino acid residues. No significant amino acid sequence identity to any other proteins, including heparan-sulfate 2-sulfotransferases, was observed. When the cDNA for the entire coding sequence of the protein was inserted into a eukaryotic expression vector and transfected into COS-7 cells, the HS6ST activity increased 7-fold over the control. The FLAG fusion protein purified by anti-FLAG affinity chromatography showed the HS6ST activity alone. Northern blot analysis revealed the occurrence of a single transcript of 3.9 kilobases in both human fetal brain and CHO cells. The results, together with the ones from our recent cDNA analysis of heparan-sulfate 2-sulfotransferase (Kobayashi, M., Habuchi, H., Yoneda, M., Habuchi, O., and Kimata, K. (1997) J. Biol. Chem. 272, 13980–13985), suggest that at least two different gene products are responsible for 6- and 2-O-sulfations of heparan sulfate.

Heparan sulfate proteoglycans are ubiquitously present on the cell surface and in the extracellular matrix, including the basement membrane (1,2). Heparan sulfate proteoglycans are known to interact with a variety of proteins such as heparinbinding growth factors, extracellular matrix components, selectins, protease inhibitors, and lipoprotein lipase (3)(4)(5)(6)(7) and are thereby implicated not only in various dynamic cellular behaviors including cell proliferation, differentiation, adhesion, migration, and morphological regulation during development (8,9), but also in various physiological phenomena such as blood coagulation (10,11), viral and bacterial infection (12,13), and tumor cell malignancy (14). The interactions of heparan sulfate with various ligands seem to be mediated by specific regions of heparan sulfate with different, albeit overlapping, structures. Heparan sulfate is required for high affinity binding of fibroblast growth factors to their receptors as well as for oligomerization of fibroblast growth factors themselves (15)(16)(17). The structure of the fibroblast growth factor 2-binding region in heparan sulfate is represented by a cluster of IdceA(2SO 4 )-GlcNSO 3 units (18 -22). An additional structural region of heparan sulfate, containing a IdceA(2SO 4 )-GlcNSO 3 (6SO 4 ) unit, is required for the signal transduction with fibroblast growth factor 2 (23,24). Hepatocyte growth factor is a heparin-binding polypeptide and has also been shown to have affinity for the regions of heparan sulfate that contain at least two repeating IdceA(2SO 4 )-GlcNSO 3 (6SO 4 ) units/octasaccharide (25,26). The binding of hepatocyte growth factor to heparan sulfate may modulate the mitogenic activity of hepatocyte growth factor (27)(28)(29).
The ability of cells to produce heparan sulfate with the specific structural regions described above depends on the specific mechanism of heparan sulfate biosynthesis (30,31). Especially, sulfation plays important roles in the mechanism. Several kinds of sulfotransferases involved in the biosynthesis were purified (32)(33)(34)(35)(36)(37), and some of the enzymes have been cloned recently (38 -42). Two kinds of N-deacetylase/N-sulfotransferases, one (N-deacetylase/N-sulfotransferase-1) from heparan sulfate-producing rat liver (32,38) and another (N-deacetylase/N-sulfotransferase-2) from heparin-producing mouse mastocytoma (33,40) and MST cells (39), have been purified and cloned. The specificities of these two kinds of enzymes appear to be different with respect to the extent of N-sulfation. Heparan-sulfate glucosaminyl-3-O-sulfotransferase was also purified and cloned and found to exist as three isoforms, the substrate specificities of which appear to be different from each other (31,37,42). The occurrence of such isoforms with different substrate specificities may be responsible for further structural diversity of heparan sulfate.
We previously purified heparan-sulfate 6-O-sulfotransferase (HS6ST) 1 (35) and heparan-sulfate 2-O-sulfotransferase * This work was supported in part by special coordination funds from the New Energy and Industrial Technology Development Organization and from the Science and Technology Agency of the Japanese Government; by a grant-in-aid for research at the Division of Matrix Glycoconjugates, Research Center for Infectious Disease, Aichi Medical University from the Ministry of Education, Science, Sports, and Culture of Japan; 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 sequences reported in this paper have been submitted to the DNA Data Bank of Japan and the GenBank™/EMBL Data Bank with accession numbers AB006179 (human mRNA) and AB006180 (hamster mRNA).
‡ To whom correspondence should be addressed. (HS2ST) (36) to apparent homogeneity from the medium and cell layer of a CHO cell culture, respectively. The sequence information from the purified HS2ST has recently enabled us to perform the cloning of the hamster HS2ST cDNA (41). The transfection of the vector containing the cloned cDNA into COS-7 cells resulted in the overexpression of the HS2ST activity alone. In contrast, Wlad et al. (34) reported the presence of a 60-kDa enzyme in mouse mastocytoma tissue that contained both HS2ST and HS6ST activities. The molecular cloning of HS6ST may be helpful to assess the above discrepancy between our results and those of Wlad et al. and to understand the regulatory mechanisms of the 6-and 2-O-sulfations in heparan sulfate biosynthesis. In this study, we cloned the cDNAs containing the entire and partial coding sequences of human and hamster HS6ST, respectively, and further showed that the protein expressed in COS-7 cells by transfection with the expression vector containing the cloned human cDNA had the HS6ST activity alone.

EXPERIMENTAL PROCEDURES
Amino Acid Sequence of Peptides Derived from Purified Protein-HS6ST was purified from the serum-free culture medium of CHO cell as described previously (35). The purified HS6ST was subjected to SDSpolyacrylamide gel electrophoresis according to the method of Laemmli (43) and then electrotransferred onto ProBlott polyvinylidene difluoride membrane (Applied Biosystems, Inc.). The transferred protein was visualized with Ponceau S. The bands of the 45-and 52-kDa proteins, which had the HS6ST activity (35), were excised, and a part of the membrane was used for N-terminal amino acid sequencing. The remaining membrane of the 45-kDa protein was subjected to reduction and S-carboxymethylation. After digestion with N-glycanase, the membrane was digested sequentially with endoproteinase Lys-C and then endoproteinase Asp-N followed by trypsin according to Iwamatsu (44). Each endoproteinase digest was subjected to reverse-phase 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 sequenator (Applied Biosystems, Inc.).
Oligonucleotides and Polymerase Chain Reaction-Based on the amino acid sequences of peptides derived from the purified HS6ST, degenerate primers of both sense and antisense strands were designed with deoxyinosine substitutions as indicated in Fig. 1A. A reverse transcriptase reaction was performed at 50°C for 1 h using poly(A) ϩ RNA from CHO cells as a template and a mixture of oligo(dT) and random hexamer as a primer. The PCR was carried out in a final volume of 50 l containing 100 pmol of each primer, 5-l aliquots of reverse transcriptase reaction mixture, 0.25 mM each dNTP, and 1.25 units of AmpliTaq polymerase (Perkin-Elmer) under the conditions of 35 cycles of denaturation at 95°C for 1 min, annealing at 46 -56°C for 1 min, and extension at 72°C for 2 min. PCR products were subjected to agarose gel electrophoresis, and DNA fragments were excised and subcloned into pBluescript II vector (Stratagene). Subclones were characterized by sequencing.
Construction of gt11 Library-Total RNA was prepared from the cultured CHO cells using an RNA extraction kit (Amersham Pharmacia Biotech), and poly(A) ϩ RNA was purified with oligo(dT)-latex (Oligotex dT30, Hoffmann-La Roche). Synthesis of cDNA and ligation of the cDNA to EcoRI-digested gt11 (Amersham Pharmacia Biotech) were carried out using a TimeSaver cDNA synthesis kit (Amersham Pharmacia Biotech). Random oligonucleotide primers and oligo(dT) were used for the reverse transcriptase reaction. The ligated DNA was packaged in vitro using Gigapack II packaging extract (Stratagene). The library was used for cDNA screening without further amplification.
Screening of gt11 Library-Approximately 8 ϫ 10 5 plaques from the CHO cell cDNA library were screened using the above PCR product as a probe. 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 and prehybridized in a solution containing 50% formamide, 5ϫ SSPE (SSPE ϭ sodium chloride/sodium phosphate/EDTA buffer), 5ϫ Denhardt's solution, 0.5% SDS, and 0.05 mg/ml denatured salmon sperm DNA for 3 h at 42°C. Hybridization was carried out at 42°C in the same buffer containing 32 P-labeled probe for 16 h. One positive clone was detected by autoradiography, and the inserted cDNA was subcloned into pBluescript by EcoRI digestion. Approximately 8 ϫ 10 5 plaques from the human fetal brain cDNA library (CLONTECH) were screened using the above inserted cDNA as a probe under the same conditions used for screening the CHO cell cDNA library. Positive clones were detected by autoradiography.
DNA Sequence Analysis-The subcloned DNAs were sequenced on both strands by the dideoxy chain termination method using Taq polymerase (dye terminator cycle sequencing; Perkin-Elmer) with a DNA sequencer (Applied Biosystems 373A) and using deaza-GTP kits with Sequenase Version 2.0 (U. S. Biochemical Corp.). The DNA sequences thus obtained were compiled and analyzed using GENETYX-MAX computer programs (Software Development Co., Ltd., Tokyo, Japan).
Construction of pFLAG-CMV-2hHS6ST-For the construction of pFLAG-CMV-2hHS6ST, the PCR product containing the open reading frame of 1206 base pairs from positions 139 to 1344 was prepared using ATGGTTGAGCGCGCCAGCAAG for the 5Ј-primer and TACCACTTCT-CAATGATGTG for the 3Ј-primer and was ligated into the blunted HindIII/EcoRI site of the pFLAG-CMV-2 expression vector (Eastman Kodak Co.). The direction of the inserted cDNA was determined by restriction mapping.
Transient Expression of Heparan-sulfate 6-Sulfotransferase cDNA in COS-7 Cells-COS-7 cells (5 ϫ 10 5 ) precultured for 48 h in a 60-mm culture dish were transfected with 4 g of pFLAG-CMV-2hHS6ST or pFLAG-CMV-2 alone. The transfection was performed using the DEAEdextran method (45). After incubation in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and antibiotics for 72 h, the spent medium was collected, and the cell layers (1.5 ϫ 10 6 cells) were washed with Dulbecco's modified Eagle's medium alone, scraped, and homogenized in 1.5 ml of 10 mM Tris-HCl, pH 7.2, 0.5% (w/v) Triton X-100, 0.15 M NaCl, 10 mM MgCl 2 , and 2 mM CaCl 2 . The homogenates were stirred for 1 h and then centrifuged at 10,000 ϫ g for 30 min. FLAG fusion proteins in the supernatant (cell extract) and the spent medium were isolated, respectively, by anti-FLAG M2 (Kodak) affinity chromatography according to the method described by the manufacturer. The activities of HS6ST and HS2ST in the supernatant (cell extract), in the spent medium, and in the anti-FLAG affinity-bound fractions were measured as described previously (35). There was no significant difference in the total cell numbers and protein concentrations among the transfectants under the transfection conditions used.
Northern Blot Analysis-Poly(A) ϩ RNA samples (5 g) prepared from cultured CHO cells and human fetal brain (CLONTECH) were denatured, electrophoresed on 1% agarose gel containing 6% formaldehyde, and transferred to a Hybond N ϩ nylon membrane. These membranes were prehybridized in a solution containing 50% formamide, 5ϫ SSPE, 5ϫ Denhardt's solution, 0.5% SDS, and 100 g/ml denatured salmon sperm DNA at 42°C for 3 h and then were hybridized in the same solution containing 32 P-labeled probes (1 ϫ 10 6 cpm/ml) at 42°C for 16 h. The probes used were a 1343-base fragment at positions 88 -1430 and a 710-base fragment for Northern analysis of human tissues and CHO cells, respectively, and human glyceraldehyde-3-phosphate dehydrogenase cDNA, which were labeled with [␣-32 P]dCTP by random oligonucleotide priming (Ready-to-Go DNA labeling kit, Amer-ster ovary; HPLC, high performance liquid chromatography; PCR, polymerase chain reaction. peptides derived from in situ protease digestion The amino acid sequences of different peptides after sequential in situ digestion of purified CHO cell HS6ST with endoproteinases Asp-N and Lys-C and trypsin were obtained as described under "Experimental Procedures." The amino-terminal amino acid sequences of 45-kDa intact HS6ST and 52-kDa intact HS6ST were also determined. X represents an unidentified amino acid residue.

Samples
Amino acid sequence Intact HS6ST (45-kDa protein) ( L/Y )P G P X L X L G A P X X R A P Intact HS6ST (52-kDa protein) sham Pharmacia Biotech). The membranes were washed twice with 1ϫ SSPE and 0.1% SDS at 50°C and subsequently with 0.1ϫ SSPE and 0.1% SDS at 50°C and then subjected to a FUJIX BAS-2000 II bioimaging analyzer (Fuji Photo Film Co.).

RESULTS
Amino Acid Sequence of Chinese Hamster Heparan-sulfate 6-Sulfotransferase-HS6ST purified from the serum-free culture medium of CHO cells was immobilized on a polyvinylidene difluoride membrane after SDS-polyacrylamide gel electrophoresis. The amino-terminal amino acid sequence of the 45-kDa protein was identical to that of the 52-kDa protein as far as the identification was possible (Table I). We also analyzed the internal amino acid sequence of the 45-kDa protein. The protein prepared by SDS-polyacrylamide gel electrophoresis was immobilized on a polyvinylidene difluoride membrane and digested sequentially with endoproteinase Lys-C and then endoproteinase Asp-N followed by trypsin. The digests were separated on a reverse-phase column by HPLC and analyzed with a protein sequencer. Sequence was determined from five different peptides (Table I).
PCR Probe and Screening of CHO Cell cDNA Library-Based on the amino acid sequences of peptides 2 and 5, sense and antisense primers were designed as shown in Fig. 1A (sense primers 2s and 5s and antisense primers 2a and 5a). When primers 2s and 5a were used in a PCR with the first strand cDNA of CHO cell poly(A) ϩ RNA as a template, a major PCR product of 330 base pairs was obtained only in the presence of both primers (Fig. 1B). When the 330-base pair PCR product was subcloned and sequenced, it contained the nucleotide sequences encoding peptide 4 in addition to those extended from peptides 2 and 5 used as primers. The 330-base pair PCR product was used as a probe to screen the CHO cell cDNA library. We obtained only one positive clone containing the 0.7-kilobase insert after several screenings. Nucleotide sequence analysis indicated that the clone contained 710 nucleotides and encoded a polypeptide composed of 236 amino acid residues (Fig. 2). The predicted amino acid sequence contained a partial N-terminal amino acid sequence of the purified HS6ST and all four internal peptides obtained by endoproteinase digestion of the enzyme (Fig. 2, underlined). However, this clone (clone CHO-HS6ST-3) did not cover all of the open reading frame.
cDNA Cloning and Predicted Protein Sequence of Human Heparan-sulfate 6-Sulfotransferase-The EcoRI fragment from the clone CHO-HS6ST-3 was used as a probe for screening the human fetal brain cDNA library. Seventeen positive clones were obtained from 8 ϫ 10 5 plaques. Seven independent clones were isolated and subcloned into pBluescript II KS. The nucleotide sequence determined from two overlapping clones (1.7 and 2.0 kilobases) encoded the entire open reading frame of human HS6ST as shown in Fig. 3A. The amino-terminal sequence contains two in-frame ATG codons. According to the consensus eukaryotic translation initiation sequence (46), the second ATG codon, but not the first one, appeared to possess the preferred nucleotides in the crucial positions for a translation initiation site. A single open reading frame beginning at  2. Partial nucleotide sequence of the hamster HS6ST cDNA and the predicted amino acid sequence. The predicted amino acid sequence is shown below the nucleotide sequence. Peptides from which amino acid sequence data were obtained are underlined; the one used as a PCR probe is double-underlined. the second ATG codon predicts a protein of 401 amino acid residues with a molecular mass of 47,075 Da and two potential N-linked glycosylation sites. To determine the location of the transmembrane domain, if any, a hydropathy plot was generated from the translated sequence. Analysis of the plot revealed one prominent hydrophobic segment in the amino-terminal region, 15 residues in length, that extends from amino acid residues 8 to 22 (Fig. 3B). The amino acid sequence predicted from the human HS6ST cDNA showed 97% identity to the sequence of the hamster HS6ST cDNA as far as the available sequence data were compared. The amino acid sequence near the carboxyl-terminal end of the transmembrane domain, which started at amino acid residue 25 (see triangle in Fig. 3A), was almost identical to the amino-terminal sequence of HS6ST purified from the culture medium of CHO cells (see Table I). If the protein were truncated at this site near the transmembrane domain, the molecular mass of the product would be calculated as 44,389 Da, which matches the molecular mass of the CHO-HS6ST protein formed after N-glycanase digestion.
Expression of Human Heparan-sulfate 6-Sulfotransferase cDNA in COS-7 Cells-To demonstrate that the isolated cDNA encodes heparan-sulfate 6-sulfotransferase, COS-7 cells were transfected with pFLAG-CMV-2hHS6ST, a recombinant plasmid containing the isolated cDNA in the mammalian expression vector pFLAG-CMV-2, or with the vector alone as a control. The HS6ST activity in the cell extract prepared from the transfectants was increased 7.5-fold over the control, but there was no increase in the HS2ST activity in the transfectants compared with the control (Table II). When the cell extract from the transfectants was purified on the anti-FLAG antibody column, the HS6ST activity alone was recovered in the affinitypurified fraction. The HS2ST activity contained in the cell extract was completely removed by this chromatography (Table  II). The results clearly indicate that the isolated cDNA encodes a protein for HS6ST. Since HS6ST was found to be secreted from CHO cells to the culture medium (35), we also examined if the HS6ST activity in the transfectants was secreted to the culture medium. The HS6ST activity in the culture medium of the transfectants was increased 3.7-fold over the control. However, when the medium fraction was applied to the anti-FLAG antibody column, no activity was recovered in the affinity-purified fraction (Table II). The loss of FLAG sequence from the recombinant HS6ST might have happened during the secretion.
Northern Analysis-When a poly(A) ϩ RNA sample was prepared from the cultured CHO cells and hybridized with a radioactive probe prepared from an EcoRI fragment of the clone CHO-HS6ST-3, a single transcript of 3.9 kilobases was observed (Fig. 4, lane 1). When a poly(A) ϩ RNA sample from human fetal brain was hybridized with the radioactive probe prepared from the open reading frame of human HS6ST cDNA, a single transcript of the identical size was also detected (Fig. 4, lane 2). DISCUSSION The following lines of evidence suggest that the cloned cDNA encodes a human homologue of the hamster HS6ST protein that we previously purified from CHO cell culture medium (35). The polypeptide encoded by the hamster cDNA fragment contained all of the partial amino acid sequences obtained from the purified hamster HS6ST. The human cDNA clones obtained by cross-hybridization with the above hamster cDNA fragment encoded the protein that contained a segment with 97.4% amino acid sequence identity to the polypeptide encoded by the hamster cDNA fragment. When the human cDNA was introduced into a eukaryotic expression system, a preferential increase in the HS6ST activity was observed without any significant change in the HS2ST activity. In addition, the HS6ST activity alone was found to be associated with the fusion protein after the anti-FLAG affinity purification. The enzyme properties of the purified fusion protein were the same as those of the purified Chinese hamster HS6ST protein that we reported previously (35) (for instance, the inhibitory effect of dithiothreitol and the number of and sites for N-linked glycosylation) (data not shown).
Two in-frame ATG codons were found in the amino-terminal sequence of the human HS6ST cDNA, but only the sequence around the second ATG codon matched the consensus eukaryotic translation initiation sequence (46,47). However, the possibility that the first ATG codon might function as the initiation codon could not be ruled out because Kozak's rule does not explain all eukaryotic translation sequences and there are a few exceptions such as some hormones or lymphokines (48).
The predicted amino acid sequence of human HS6ST showed no significant homology to that of any other sulfotransferase (38 -40, 49, 50), including HS2ST (41). It contained neither the sequence of the putative 3Ј-phosphoadenosine 5Ј-phosphosulfate-binding site found in arylsulfotransferase IV, LEKCGR (51), nor the so-called "P-loop," GXXGXXK(R) (52), which is found in most sulfotransferases and thought to be a possible ATP-or GTP-binding site. Thus, human HS6ST may have a tertiary structure for the 3Ј-phosphoadenosine 5Ј-phosphosulfate-binding site that is constructed from the amino acid sequence distinct from those of other sulfotransferases.
The hydropathy plot (Fig. 3B) suggests that human HS6ST has a hydrophobic domain at the amino-terminal region and belongs to type II membrane protein. Since the amino acid sequence near the carboxyl-terminal end of the transmembrane domain was almost identical to the amino-terminal sequence of hamster HS6ST purified from the culture medium, the purified HS6ST might have been released from the CHO cells after proteolytic cleavage at the putative site near the Overexpression of heparan-sulfate 6-sulfotransferase in COS-7 cells COS-7 cells were transfected as described under "Experimental Procedures" with vector alone (pFLAG-CMV-2) and with a plasmid containing the human HS6ST cDNA (pFLAG-CMV-2hHS6ST). The enzyme activities of HS6ST and HS2ST in the cell extract, the spent medium, and their FLAG affinity fractions, which were derived from 1.5 ϫ 10 6 cells, were determined as described under "Experimental Procedures." Values represent means Ϯ S.D. of triplicate cultures. The activities in the absence of exogenous sulfate acceptors were usually less than 0.1 pmol/min/1.5 ϫ 10 6 cells.  transmembrane domain. The proteolytic cleavage was also suggested in the secretion of chondroitin 6-sulfotransferase from chick chondrocytes (49). The secretion of HS6ST may occur not only in cultured cells, but also in tissues in vivo since the HS6ST activity was found in the serum collected from adult mice. As observed in the cultured CHO cells, the serum contained little, if any, HS2ST activity. 2 Northern analysis showed only one transcript of 3.9 kilobases in poly(A) ϩ RNA samples prepared from both cultured CHO cells and human fetal brain (Fig. 4). Since the purified preparation of hamster HS6ST gave two protein bands of different sizes (52 and 45 kDa) that have the identical N-terminal amino acid sequence (Table I), it is likely that, in addition to the proteolytic cleavage described above, further post-translational modification or processing may be involved in the synthesis and secretion of HS6ST. In contrast to HS6ST, multiple transcripts of different sizes are observed in other sulfotransferases (38, 39, 41, 49 -54). However, as far as investigated, they have the same open reading frame, and the differences are detected only in the size and sequence of the untranslated region (52)(53)(54).
Together with our previous study on the HS2ST cDNA (41), this study suggests that HS6ST and HS2ST may be products from two different and independent transcripts because they are encoded by the different cDNAs without any common sequence and the sizes of their transcripts were different from each other (5.0 and 3.0 kilobases for HS2ST and 3.9 kilobases for HS6ST). The above possibility is also supported by the report of Bai and Esko (21) on a mutant CHO cell line that is defective only in heparan sulfate 2-O-sulfation. It has been shown recently that the defect in the mutant cell line was recovered by the transfection of our HS2ST cDNA into the cells (55). However, the ongoing genomic analyses will give the definitive answer.
Studies on sulfotransferases involved in heparin biosynthesis of mastocytoma tissue have suggested that the 2-and 6-Osulfotransferase activities may share a common subunit or reside in a single protein (34). Considering these observations together, it is likely that several different kinds of O-sulfotransferases with different specificities may be involved in the 6-and 2-O-sulfation processes of heparan sulfate and heparin as observed in the cases of N-deacetylation/N-sulfation (32, 33, 38 -40) and 3-O-sulfation (37,42). Further studies will be needed to clarify the roles of HS6ST and HS2ST in the biosynthesis of the fine structures in heparan sulfate involved in regulating activities and functions of various cell growth factors.