6-O-sulfotransferase-1 represents a critical enzyme in the anticoagulant heparan sulfate biosynthetic pathway.

Using recombinant retroviral transduction, we have introduced the heparin/heparan sulfate (HS) 3-O-sulfotransferase 1 (3-OST-1) gene into Chinese hamster ovary (CHO) cells. Expression of 3-OST-1 confers upon CHO cells the ability to produce anticoagulantly active HS (HS(act)). To understand how 6-OST and other proteins regulate HS(act) biosynthesis, a CHO cell clone with three copies of 3-OST-1 was chemically mutagenized. Resulting mutants that make HS but are defective in generating HS(act) were single-cell-cloned. One cell mutant makes fewer 6-O-sulfated residues. Modification of HS chains from the mutant with pure 6-OST-1 and 3'-phosphoadenosine 5'-phosphosulfate increased HS(act) from 7% to 51%. Transfection of this mutant with 6-OST-1 created a CHO cell line that makes HS, 50% of which is HS(act). We discovered in this study that (i) 6-OST-1 is a limiting enzyme in the HS(act) biosynthetic pathway in vivo when the limiting nature of 3-OST-1 is removed; (ii) HS chains from the mutant cells serve as an excellent substrate for demonstrating that 6-OST-1 is the limiting factor for HS(act) generation in vitro; (iii) in contradiction to the literature, 6-OST-1 can add 6-O-sulfate to GlcNAc residues, especially the critical 6-O-sulfate in the antithrombin binding motif; (iv) both 3-O- and 6-O-sulfation can be the final step in HS(act) biosynthesis in contrast to prior publications that concluded 3-O-sulfation is the final step in HS(act) biosynthesis; (v), in the presence of HS interacting protein peptide, 3-O-sulfate-containing sugars can be degraded into disaccharides by heparitinase digestion as demonstrated by capillary high performance liquid chromatography coupled with mass spectrometry.

Heparin/heparan sulfate (HS) 1 is a linear polymer covalently attached to the protein cores of proteoglycans, which are abun-dant and ubiquitously expressed in almost all animal cells. HS is assembled by the action of a large family of enzymes that catalyze chain polymerization (alternating addition of GlcNAc and GlcUA residues), GlcNAc N-deacetylation and N-sulfation, glucuronic acid (GlcUA) epimerization to L-iduronic acid (Ido-UA), 2-O-sulfation of uronic acid residues, and 3-O-and 6-Osulfation of glucosaminyl residues. Tissue-specific and developmentally regulated expression of the biosynthetic enzymes and enzyme isoforms produces HS chains with distinct sequences (1)(2)(3). These different sequences enable interactions to occur with a broad array of protein ligands that modulate a wide range of biological functions in development, differentiation, homeostasis, and bacterial/viral entry (reviewed in Refs. 4 -11).
The specificity of any HS and protein interaction is largely dictated by arrangements of sulfates along the chain. For example, the pentasaccharide sequence, GlcNAc/NS6S-GlcUA-GlcNS3S؎6S-IdoUA2S-GlcNS6S, represents the minimum sequence for antithrombin (AT) binding, where the 3S (3-Osulfate) and 6S (6-O-sulfate) groups constitute the most critical elements involved in the interaction (12)(13)(14)(15)(16). To delineate the biosynthetic pathway that regulates anticoagulant heparan sulfate (HS act ) biosynthesis, our laboratory has purified as well as molecularly cloned 3-OST-1 (17,18). We have demonstrated that 3-OST-1, usually existing in limited amounts, acts upon HS act precursor to produce HS act and upon HS inact precursor to produce 3-O-sulfated HS inact (17,19). When 3-OST-1 is no longer limiting, the capacity for HS act generation is determined by the abundance of HS act precursors (20). In this case, the limitation in HS act production is the presence of the critical 6-O-sulfate groups and/or the availability of proper epimerization/sulfation patterns in the precursors.
In most cases, the critical 6S in the AT binding oligosaccharides sequenced and characterized in heparin is attached to GlcNAc residues (a summary table for heparin is presented in Refs. 21,22). We previously reported that the critical 6S in * This work was supported in part by National Institutes of Health Grants 5-P01-HL41484 and 5-R01-HL58479 (to R. D. R.) and GM-50573 (to R. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB006180.
§ Recipient of National Research Service Award Postdoctoral Fellowship.
ʈ Current address: Division of Medicinal Chemistry, School of Pharmacy, University of North Carolina, Chapel Hill, NC 27599.
‡ ‡ To whom correspondence should be addressed: Massachusetts Institute of Technology, Bldg. 68-480, 77 Massachusetts Ave., Cambridge, MA 02139. Tel.: 617-253-8804; Fax: 617-258-6553; E-mail: rdrrosen@ mit.edu. 1 The abbreviations used are: HS, heparan sulfate; HS act , anticoagulantly active heparan sulfate; HS inact , anticoagulantly inactive heparan sulfate; HS total , total heparan sulfate; CHAPS, 3-[(3-cholamidopropy)dimethylammonio]-1-propanesulfonic acid; GAG, glycosaminoglycan; PAPS, 3Ј-phosphoadenosine 5Ј-phosphosulfate; AT, antithrombin; 3-OST-1, HS act from F9 cells is exclusively attached to GlcNAc residues (20). To date, three 6-O-sulfotransferases have been cloned, but reported substrate specificities indicate that none of them can put a 6-O-sulfate group on GlcNAc residues (2). To understand how specific 6-O-sulfotransferases and other factors regulate HS act biosynthesis, we created cell mutants that are defective in the formation of HS act precursors. We chose Chinese hamster ovary (CHO) cells, because a series of HS biosynthetic mutants have been successfully made in this cell line (23)(24)(25)(26)(27)(28). However, CHO wild-type cells do not generate 3-O-sulfated residues and therefore do not make any HS act . Using recombinant retroviral transduction, we have introduced the HS 3-Osulfotransferase-1 (3-OST-1) gene into CHO cells (29). 3-OST-1 expression gives rise to CHO cells with the ability to produce HS act . A cell line was chosen that has three copies of 3-OST-1 as determined by Southern analysis (29). After chemical mutagenesis of this cell line, mutant cells that were positive for FGF-2 binding but negative for AT binding were FACS sorted and cloned. The scheme for making HS act precursor mutants is outlined in Fig. 1. The advantage of having multiple copies of 3-OST-1 is that other upstream genes that are responsible for generating specific HS precursor structures can be sought after chemical mutagenesis without concern for the loss of 3-OST-1. FGF-2 selection is employed to ensure that the mutant cells still make HS. By using this scheme, we obtained a 6-O-sulfatedefective mutant. After correcting the mutant with 6-O-sulfotransferase-1 transfection, we created a cell line that makes HS, 50% of which is HS act . This represents the highest percentage of HS act production by any cell line reported so far.
The approach used in this study includes placing multiple copies of a downstream enzyme in CHO cells, mutagenizing the cells to obtain mutants deficient in upstream enzymes that are part of the pathway, and then characterizing the mutants. This technique may constitute a general approach for defining and obtaining the components of biosynthetic pathways once the terminal biosynthetic enzyme has been obtained and proteins that recognize the final product of the pathway are available.

EXPERIMENTAL PROCEDURES
Cell Culture-Wild-type Chinese hamster ovary cells (CHO-K1) were obtained from the American Type Culture Collection (CCL-61, ATCC, Rockville, MD). Wild-type and mutant cells were maintained in Ham's F-12 medium supplemented with 10% fetal bovine serum (HyClone), penicillin G (100 units/ml), and streptomycin sulfate (100 g/ml) at 37°C under an atmosphere of 5% CO 2 in air and 100% relative humidity. The cells were passaged every 3-4 days with 0.125% (w/v) trypsin and 1 mM EDTA, and after 10 -15 cycles, fresh cells were revived from stocks stored under liquid nitrogen. Low-sulfate medium was composed of Ham's F-12 medium supplemented with penicillin G (100 units/ml) and 10% fetal bovine serum that had been dialyzed 200-fold against phosphate-buffered saline (PBS) (30). Low-glucose Ham's F-12 medium contained 1 mM glucose supplemented with penicillin G (100 units/ml), streptomycin sulfate (100 g/ml), and fetal bovine serum that had been dialyzed 200-fold against PBS (30). All tissue culture media and reagents were purchased from Life Technologies, Inc. (Gaithersburg, MD) unless otherwise indicated.
3-OST-1 Recombinant Retroviral Transduction-The method of 3-OST-1 recombinant retroviral transduction into CHO cells was described in detail in our previous publication (29).
Antithrombin and FGF-2 Labeling-The same procedure has been used as described previously (29).
Cell Sorting-Nearly confluent monolayers of 3-OST-1-transduced CHO-K1 cells were detached by adding 10 ml of 2 mM EDTA in PBS containing 10% fetal bovine serum and centrifuged. The cell pellets were placed on ice, and 50 l each of fluorescein-AT and Alexa 594-FGF-2 were added. After 30 min, the cells were washed once and resuspended in 1 ml of 10% fetal bovine serum in PBS containing 2 mM EDTA. Flow cytometry and cell sorting was performed on FACScan and FACStar instruments (Becton Dickinson) using dual color detection filters. Cells positive for AT and FGF-2 binding were sorted and subsequently single-cell-cloned into a 96-well plate. The single-cell clones were expanded and frozen for further analysis.
Twelve 3-OST-1-transduced CHO-K1 clones were obtained as described above. The number of copies of 3-OST-1 in the individual clones was determined by Southern analysis. Genomic DNA (10 g) was digested with 40 units of EcoRI overnight at 37°C, electrophoresed on a 0.7% (w/v) agarose gel, transferred to GeneScreen Plus (PerkinElmer Life Sciences), and probed with 3-OST-1 cDNA labeled with the Megaprime labeling kit (Amersham Pharmacia Biotech). Blots were hybridized in ExpressHyb solution (CLONTECH) containing 3-OST-1 probe (2 ϫ 10 6 cpm/ml), followed by autoradiography. The cell clone with three copies of 3-OST-1 was expanded and frozen for further studies.
Mutant Screening-Wild-type CHO cells with three copies of 3-OST-1 were mutagenized with ethylmethane sulfonate as described in the literature (31) and frozen under liquid nitrogen. A portion of cells was thawed, propagated for 3 days, and labeled with both Alexa 594-FGF-2 and fluorescein-AT. The labeled cells were sorted, and FGF-2positive and AT-negative cells were collected. Approximately 1 ϫ 10 4 sorted cells were collected into 1 ml of complete F-12 Ham's media then plated in T-75 flasks. Sorted cell populations were maintained in complete F-12 Ham's medium for 1 week, then the cells were labeled and sorted again as described above. After five rounds of sorting, FGF-2positive and AT-negative cells were single-cell-sorted into a 96-well plate. The single-cell clones were expanded and frozen for further analysis. The sorting profiles of CHO-K1 with three copies of 3-OST-1, mutant, and the 6-OST-1 correctant of the mutant were shown by dual-color fluorescence flow cytometric analysis (see Fig. 2) HS Preparation and Analysis-The method was the same as one published previously (29).
cDNA Cloning and Expression of CHO 6-OST-1-Sequences coding for CHO 6-OST-1 were amplified from a CHO-K1/cDNA quick-clone library (CLONTECH). The reaction mixture contained 2 units of Pfu polymerase (Stratagene), 1 ng of cDNA, and 100 pmol of the primers. The sense primer has an added BglII site (5Ј-GCAGATCTGCAGGAC-CATGGTTGAGCGCGCCAGCAAGTTC-3Ј), and the antisense primer has an added XbaI site (5Ј-GCTCTAGACTACCACTTCTCAATGATGT-GGCTC-3Ј). The 6-OST-1 primer sequences are derived from the human 6-OST-1 cDNA sequence (from residues 240 to 264) and to the complement of this sequence (from residues 1147 to 1172) as reported (32). After 30 thermal cycles (1 min of denaturation at 94°C, 2 min of annealing at 55°C, 3 min of extension at 72°C), the amplification products were analyzed in 1% agarose gels and detected by ethidium bromide staining. The amplification products were excised from the gel and cleaned by Gel Extraction kit (Qiagen). The PCR product was treated with BglII and XbaI, ligated into XbaI-and BamHI-digested pInd/Hygro plasmid (CLONTECH), and transformed into Escherichia coli DH5␣-competent cells. Four clones from each of two separate PCR reactions were sequenced and found to be identical. pInd/Hygro 6-OST-1-containing plasmid was transfected into the CHO mutant cells. Cells positive for AT and FGF-2 binding were sorted and subsequently singlecell-cloned into a 96-well plate. The single-cell clones were expanded and frozen for further analysis.  (20). After ethanol precipitation, the pellets were washed with 75% ethanol, dried briefly under vacuum, and dissolved in water for further analysis.

Separation of HS act and HS inact by AT-affinity Chromatography-
The procedure is identical to that described previously (29).
For low pH nitrous acid degradation, radiolabeled HS samples were mixed with 10 g of bovine kidney HS (ICN) and digested (35).
Disaccharides were purified by Bio-Gel P2 chromatography and resolved by ion pairing reverse-phase HPLC with appropriate disaccharide standards (36). Bio-Gel P2 or P6 columns (0.75 ϫ 200 cm) were equilibrated with 100 mM ammonium bicarbonate. Radiolabeled samples (200 l) were mixed with dextran blue (5 g) and phenol red (5 g) and loaded onto the column. The samples were eluted at a flow rate of 4 ml/h with collection of 0.5-ml fractions. The desired fractions were dried under vacuum, individually or pooled, to remove ammonium bicarbonate.
⌬UA-GlcNS3S Disaccharide Structure Determination by Capillary IPRP-HPLC Coupled with Mass Spectrometry-It has been reported that heparin molecules exhibiting a high affinity for a synthetic peptide (CRPKAKAKAKAKDQTK) mimicking a heparin-binding domain of heparin interacting protein (HIP) also show an extremely high affinity for AT (37). We expected that inclusion of this small peptide in the heparitinase digestion solution would protect 3-O-[ 35 S]sulfate-labeled HS from degrading into tetrasaccharide. Theoretically, we should recover HIP peptide-protected, AT-binding HS oligosaccharides. However, in the presence of the HIP peptide, all the 3-O-[ 35 S]sulfate-labeled sugars were degraded into disaccharides instead of oligosaccharides or tetrasaccharides as judged by their elution position on Bio-Gel P2 and their unique elution positions on IPRP-HPLC (the major 3-O-[ 35 S]sulfate-containing disaccharides eluted right before ⌬UA-GlcNS6S disaccharide standard). Because there is no ⌬UA-GlcNS3S standard reported, we decided to prove the structure. We first made stable isotope PAP 34 S. The PAP 34 S (99% isotope purity determined by ESI-MS) was prepared by incubating ATP and stable isotope Na 2 34 SO 4 (Isonics Corp.) with ATP sulfurylase (Sigma Chemical Co.), adenosine 5Ј-phosphosulfate kinase (a generous gift from Dr. Irwin H. Segel), and inorganic pyrophosphatase (Sigma) (38). HS chains from wild-type CHO cells were labeled with pure 3-OST-1 plus PAP 34 S. We then developed a capillary IPRP-HPLC (LC Packings) method for separating HS disaccharides, which is similar to a conventional IPRP-HPLC (29) except that 5 mM dibutylamine is used as an ion-pairing reagent (Sigma), and then coupled it to an electrospray ionization time-of-flight mass spectrometer (Mariner Workstation, PerSeptive Biosystems, Inc.) to detect the mass of each disaccharide eluted. Six HS disaccharide standards from Seikagaku were separated by capillary HPLC and detected by negative polarity ESI-MS. The accuracy of the ESI-MS is Ϯ 0.001 m/z unit after calibration with the molecular standard sets supplied by the manufacture (bis-trifluorobenzoic acid, heptadecafluorononanoic acid, perfluorotetradecanoic acid). 3-O-34 S-labeled HS was digested with a combination of 1 mU of each heparitinase I, heparitinase II, heparitinase IV, and heparinase in the absence or presence of 0.5 mg/ml HIP peptide. 0.5 g of digested HS was injected into capillary HPLC coupled with mass spectrometry (see Fig. 3). UV peak B eluted at the same time as a ⌬UA-GlcNS6S standard, whereas UV peak D eluted at the same time as a ⌬UA2S-GlcNS standard (see Fig. 3A). Three major ions with m/z 247.5, 496.0, and 625.2 were observed in both UV peaks (see Fig. 3, panels B and D), where 496.0 is z1 (Ϫ1)-charged, 247.5 is z2 (Ϫ2)charged, and 625.2 is one dibutylamine-adducted with z1 (Ϫ1)-charged ⌬UA-GlcNS6S or ⌬UA2S-GlcNS disaccharides. However, when the m/z region 494.0 -501.0 from both peaks B and D were expended (panels C and E), a non-natural abundant, z1-charged molecular ion with m/z 498.0 was observed in UV peak B but not in UV peak D. 498.0 versus 496.0 of disaccharide ions should represent ⌬UA-GlcNS3[ 34 S]S and ⌬UA-GlcNS6S, respectively. The mass for ⌬UA-GlcNS3[ 34 S]S was barely detectable in the absence of HIP peptide, which is consistent with the literature that 3-O-sulfate-containing sugars are usually degraded into tetrasaccharides, not disaccharides, by a mixture of heparitinase digestion (20,33). Therefore, HIP peptide is included in heparitinase digestion when we degrade 3-O-containing HS into disaccharides.
Northern Blot Hybridization and RT-PCR-To generate specific Northern blot hybridization probes, PCR primers were designed that bracket unique sequences within human 6-OST-1, 6-OST-2 and 6-OST-3. A 249-bp PCR product that corresponds to a region within the 3Ј-UTR of the 6-OST-1 gene starting at position 1772 and ending at 2021 was used as an isoform specific probe. Similarly, a 299-bp PCR product that corresponds to a region in the 3Ј-UTR of the 6-OST-2 gene starting at position 1831 and ending at 2130, and another product within the 3Ј-UTR of the 6-OST-3 gene starting at 943 and ending at 1378 (444 bp) were used as a probe. PCR was performed with [␣-32 P]dCTP (PerkinElmer Life Sciences), and isoform-specific radiola-beled probes were purified on G-25 Sephadex spin columns (Roche Molecular Biochemicals). Hybridizations were carried out according to the manufacturer's instructions using 2 ϫ 10 6 cpm probe per milliliter of ExpressHyb solution (CLONTECH). After the hybridizations were complete, the blots were washed twice in 2ϫ SSC containing 0.1% SDS and once with 0.1ϫ SSC containing 0.1% SDS, all at room temperature. Blots were then washed with 0.1ϫ SSC containing 0.1% SDS at 50°C. For blots hybridized with the 6-OST-1 probe, this last wash was repeated twice at 65°C. The membranes were then subjected to autoradiography with BioMax imaging film (Kodak) with a BioMax MS intensifying screen (Kodak).
For RT-PCR, poly(A)-purified or DNase I-treated total RNA was used. Primer pairs were designed that bracket isoform-specific regions within the human sequences for both 6-OST-2 and 6-OST-3. . Total RNA was purified from wild-type and mutant CHO-K1 cells using the RNeasy total RNA kit from Qiagen according to the manufacturer's instructions. RNA was quantitated by absorbance at 260 nm, and 100 g of total RNA was reacted with DNase I (Ambion) at 37°C for 45 min, twice extracted with equal volumes of acid phenol/chloroform, precipitated in ethanol, and reconstituted in diethyl pyrocarbonatetreated water. Further selection of poly(A) plus RNA was carried out with the Oligotex mRNA kit (Qiagen). RNA integrity was checked after electrophoresis on a 1% agarose gel, and all RT reactions were run with Moloney murine leukemia virus reverse transcriptase (Ambion) according to manufacturer's instructions. PCR was performed with Super Taq polymerase (Ambion).
Baculovirus Expression and Purification of 6-OST-1-Human 6-OST-1 recombinant baculovirus was prepared using the pFastBas HT donor plasmid modified by the insertion of honeybee mellitin signal peptide (36) and the Bac-to-Bac baculovirus expression system (Life Technologies, Inc.) according to the manufacturer's protocol, except that recombinant bacmid DNA was purified using an endotoxin-free plasmid purification kit (Qiagen, Inc.) and transfection of Sf9 cells was scaled up to employ 3 g of bacmid DNA and 6 ϫ 10 6 exponentially growing cells in a 100-mm dish. At day 3 post-transfection, baculovirus was precipitated from the medium with 10% polyethylene glycol, 0.5 M NaCl at 12,000 ϫ g, re-suspended in 14 ml of medium, and applied to a 100-mm dish seeded with 1.5 ϫ 10 7 Sf9 cells. Medium from the infected cells was harvested after 90 h of growth at 27°C, centrifuged at 400 ϫ g, made to 10 mM in Tris, adjusted to pH 8.0, and centrifuged at 4000 ϫ g. Clarified medium was diluted with an equal volume of cold 10 mM Tris-HCl, pH 8.0, and stirred for 30 min with 0.6 ml (packed volume) of Toyopearl 650M chromatographic media (TosoHaas). The heparin-Sepharose was packed into a column (0.4 ϫ 4.75 cm), washed with 5 ml of TCG 50 (10 mM Tris-HCl, pH 8.0, 2% glycerol, 0.6% CHAPS, 50 mM NaCl), eluted with 1.2 ml of TCG 1000 (as above, but 1 M in NaCl) containing 10 mM imidazole, and concentrated to 0.25 ml in a Microcon YM-10 centrifugal filter (Millipore Corp.).
Histidine-tagged recombinant 6-OST-1 was affinity-purified by mixing the product eluted from heparin-Sepharose for 90 min at 4°C with nickel-nitrilotriacetic acid magnetic agarose beads (Qiagen, Inc.) magnetically sedimented from 60 l of suspension. The beads were washed twice with 0.125 ml of TCG 400 containing 20 mM imidazole and eluted twice with 0.03 ml of TCG 400 containing 250 mM imidazole. The combined elution fractions contained ϳ25% of the sulfotransferase activity present in the starting medium.
Bacterial Expression and Purification of 6-OST-1-Expression vector pET15b was purchased from Novagen (Madison, WI). E. coli strains BL21 and DH5␣ were obtained through the ATCC (Manassas, VA). An AseI restriction site was introduced at 211-216 bp, and a BamHI restriction site was introduced at 1344 -1349 bp of human 6-OST-1 (32) by PCR. The 6-OST-1 gene was then ligated into NdeI-and BamHIdigested pET15b and transformed into competent E. coli strain DH5␣. A BL21 colony containing 6-OST-1 in pET15b with confirmed sequence was used to inoculate 2 liters of LB containing 100 g/ml ampicillin. The cultures were shaken in flasks at 250 rpm at 37°C. When the optical density at 600 nm reached 1.2, 1 mM isopropyl-1-thio-␤-D-galactopyranoside was added to the cultures. The cultures were then agi-tated at 250 rpm overnight at room temperature. The cells were pelleted at 5000 rpm for 15 min. The supernatant was discarded, and the cell pellet was resuspended in 40 ml of 20 mM Tris, 500 mM NaCl, 0.6% CHAPS, 1% glycerol, and 5 mM imidazole, pH 7.9 ("binding buffer"). The cells were homogenized, and the homogenate was centrifuged at 13,000 rpm for 20 min. The supernatant was filtered through 0.2-m filter paper and loaded onto a BioCAD HPLC system (PerSeptive Biosystems, Cambridge, MA) and purified using Ni 2ϩ chelate chromatography. Briefly, the supernatant was loaded onto the column and washed with binding buffer until unbound material was washed off the column. Then, low affinity material was washed off the column using 20 mM Tris, 500 mM NaCl, 0.6% CHAPS, 1% glycerol, and 55 mM imidazole, pH 7.9, and 6-OST-1 was eluted from the column with 20 mM Tris, 500 mM NaCl, 0.6% CHAPS, 1% glycerol, and 500 mM imidazole, pH 7.9. The purity of the recombinant 6-OST-1 was determined using a silverstained protein gel.

Similar Charge Density in GAGs from HS act Defective Mutant and Wild-type Cells-
The interaction of AT with heparin/HS depends on a pentasaccharide sequence consisting of GlcNAc/NS6S-GlcUA-GlcNS3SϮ6S-IdoUA2S-GlcNS6S, where the 3S and 6S groups constitute the most critical elements involved in the interaction. We reasoned that mutant cells, which make HS lacking the critical 6S group or having an improper N-sulfation/epimerization pattern that fails to generate HS act precursor, should not bind to AT. By using the scheme shown in Fig. 1, we obtained cell mutants that are defective in AT-binding (Fig. 2B). We first examined the synthesis of GAGs in the HS act -defective mutant by biosynthetic labeling studies with [6-3 H]GlcN (Fig. 3). HPLC anion-exchange analysis of the [ 3 H]GAG chains from the mutant cells resolved HS (0.31-0.50 M NaCl) from chondroitin sulfate (0.52-0.60 M NaCl) (Fig. 4, solid tracer). The GAG chains from the 3-OST-1-expressing wild-type resolved into a similar profile to that of the mutant (Fig. 4, broken tracer). This finding implies that the HS from the mutant has a similar charge density to that of the wild-type.
HS act -defective Mutant Makes Less 6-O-Sulfate-containing Disaccharides-Because HS from the mutant has similar charge density to that of the wild-type, the decrease in AT binding in the mutant should correlate with a change in the structure of the HS chains.
HS act -defective Mutant and Wild-type CHO Cells Express 6-OST-1 but Not 6-OST-2 and 6-OST-3 mRNA-To explain the reduced levels of 6-O-sulfate-containing disaccharides in the mutant, we first checked how many 6-OST isoforms are expressed in CHO wild-type versus mutant cells. Three murine 6-OST isoforms have been cloned (2). We have cloned and used human 6-OST-1, 6-OST-2, and 6-OST-3 cDNA as probes and conducted both Northern and RT-PCR studies in the mutant and the 3-OST-1-expressing wild-type CHO cells. Northern analysis indicates that the mutant and the 3-OST-1-expressing wild-type have the same level of 6-OST-1 mRNA. No 6-OST-2 and 6-OST-3 mRNA was detected on the same blot (data not shown). These results suggest that CHO cells might only express 6-OST-1. To prove this, one set of PCR primers for 6-OST-1, three sets of PCR primers for 6-OST-2, and two sets of PCR primers for 6-OST-3 were made and RT-PCR reactions were conducted using wild-type and mutant mRNA as templates (see "Experimental Procedures"). The same level of 6-OST-1 PCR products was observed for both wild-type and the mutant, but no amplification was observed for three sets of  (Table I); however, there was no change in either the level of mRNA expression nor the mutation of the coding sequence of 6-OST-1 (data not shown). These results suggest that the mutant might have a defect in 6-OST-1 activities. To test this hypothesis, sulfotransferase activity assays were conducted using HS from wild-type CHO cells, N-, O-desulfated, re-N-sulfated heparin (CDSNS-heparin) and 6-O-desulfated heparin (a generous gift from Dr. Jeffrey D. Esko) as substrates and crude cell homogenates from wild-type and mutant CHO cells as a source of enzyme. In the wild-type, enzyme activity was proportional to time for ϳ2 h and with the amount of cell protein added up to100 g. The reaction products were digested by a combination of heparitinase digestion, followed by Bio-Gel P2 chromatography. The disaccharides collected were subjected to IPRP-HPLC analysis. . A cell line that has three copies of 3-OST-1 was chosen by Southern analysis. After chemical mutagenesis of this cell line, FGF-2 binding-positive and AT binding-negative mutant cells were FACS-sorted and cloned. The advantage of having three copies of 3-OST-1 is that upstream genes that are responsible for generating specific HS precursor structures can be sought after chemical mutagenesis without losing 3-OST-1. FGF-2 selection is employed to make certain that the mutant cells still make HS. type and mutants. 2-O-Sulfotransferase activity was similar in mutant cells (118 Ϯ 3 pmol/min/mg) and the wild-type cells (122 Ϯ 2 pmol/min/mg) when CDSNS-heparin was used as substrate. However, 30 -39% reduction of 6-O-sulfotransferase activities in the mutant was observed with all three substrates (Table II).
In Vivo 6-OST-1-corrected Mutant (Correctant) Makes HS, 50% of Which Is HS act -Decreased 6-OST-1 activity in the mutant might be responsible for its deficiency in AT binding. To test this idea, the CHO 6-OST-1 coding region was successfully amplified and sequenced from the CHO-K1 quick-clone cDNA library by PCR, because only partial 6-OST-1 coding sequence from CHO cells has been reported (32). The CHO 6-OST-1 sequence has been deposited in GenBank (accession number AB006180; the differences in the amino acid residues from the previously reported partial sequence of the same cDNA (32) are commented upon). CHO 6-OST-1 cDNA was then expressed in the mutant. The stable transfectants were labeled with fluorescein-AT and Alexa 594-FGF-2 and subjected to dual-color FACS. The correctants with high AT binding affinity were single-cell-cloned (Fig. 2G). [ 35 S]Sulfate metabolically labeled HS from correctants was isolated. AT-affinity chromatography of [ 35 S]HS indicates that the correctant makes 50% HS act chain. This represents the highest percentage of HS act production by any cell lines reported so far. We then analyzed the disaccharide composition of the correctant cells.
[ 35 S]Sulfate metabolically labeled HS from 3-OST-1 expression wild-type, mutant, and correctant cell clones were isolated and digested with a mixture of heparitinases. The resulting disaccharides were separated on a Bio-Gel P2 column and were then further resolved by IPRP-HPLC (Fig. 4). The correctant exhibited a greatly increased amount of both GlcNAc6S (9% versus 5%) and GlcNS6S (13% versus 4%) residues compared with the mutant. The amount of GlcNAc6S and GlcNS6S residues in the correctant cells exceeded the amount of GlcNAc6S and GlcNS6S residues in the wild-type CHO-K1 cells with three copies of 3-OST-1 (GlcNAc6S (9% versus 7%) and GlcNS6S (13% versus 9%)) ( Table I).
Contribution of 6-OST-1 in Generating HS act Oligosaccharides-To further locate the 6-O-sulfate addition in AT-binding HS act oligosaccharides, cold mutant HS chains were treated with purified baculovirus-expressed 6-OST-1 with [ 35 S]PAPS overnight. After heparitinase I digestion, HS act oligosaccharides were affinity-purified (7% of 6-O-[ 35 S]sulfate-labeled HS total ). The HS act oligosaccharides were then treated with low pH nitrous acid that cleaves N-sulfated residues, and a combination of heparitinases that cleaves 3-O-sulfate-containing sugar into tetrasaccharides and all other sugars into disaccharides. Treated and untreated HS act oligosaccharides were run on Bio-Gel P6 columns (Fig. 7). Di-and tetrasaccharides were collected from enzyme and low pH nitrous-treated samples as indicated. The tetrasaccharides resistant to a combination of heparitinases I and II and heparinase digestion represented the 3-O-sulfatecontaining tetrasaccharides as reported earlier (20,33). The presence of similar amounts of tetrasaccharides from both nitrous and enzyme degradation suggests the 3-O-containing tetrasaccharides have the structures, UAϮ2S-GlcNAc6 35 S-GlcUA-GlcNS3SϮ6 35 S. To prove this, the tetrasaccharides (Fig. 8A) collected from enzyme digestion (Fig. 7C) were further digested into disaccharides (Fig. 8B) with heparitinase I in the presence of HIP peptide (the same method as shown in Fig. 3). IPRP-HPLC profiles of 6-O-sulfate-tagged HS act di-and tetrasaccharides from Fig. 7C were shown in Fig. 8.  (Fig. 8). In HS act oligosaccharides, 6-OST-1 adds 6-O-sulfates not only at GlcUA/IdoUA-GlcNS, GlcUA-Glc-NAc, and IdoUA2S-GlcNS, but also at GlcUA-GlcNS3S. These results show that 6-OST-1 is the enzyme that not only puts the critical 6-O-sulfate group in HS act oligosaccharides but also other 6-O-sulfate groups in HS act oligosaccharides as well. 3-OST-1 and 6-OST-1 are, therefore, the critical enzymes for the generation of HS act . DISCUSSION A new approach for generating HS act biosynthetic mutants and a novel method for characterizing 6-O-sulfotransferase substrate specificity have been developed in this study. This approach includes placing multiple copies of a downstream enzyme into CHO cells and then mutagenizing these cells. Mutants defective in upstream enzymes that are part of the anticoagulant HS biosynthetic pathway were then sorted by FACS. A mutant was then characterized both in vivo and in vitro to delineate enzymes involved in generating anticoagulant HS. The advantage of sorting is that it targets GAG synthesis selectively and very large populations of cells can be screened, which make it possible to detect specific mutations in effecting distinct HS structures. This technique may constitute a general approach for defining, obtaining, and characterizing the components of a biosynthetic pathway once the terminal biosynthetic enzyme is obtained and proteins that recognize the final product of the pathway are available.
It has been reported that a type of size exclusion chromatography coupled with mass spectrometry is effective for compositional analysis of chondroitin sulfate oligosaccharides. Mass spectrometric detection produces far more information than conventional UV or fluorescent detectors and allows the monosaccharide composition of individual components to be determined (39). Introducing the stable isotope PAP 34 S into the 3-O position of HS by pure 3-OST-1 as described in Fig. 3, a 3-Osulfate-containing disaccharide with a unique mass, which has not been reported before, was readily identified by a combination of capillary IPRP-HPLC coupled with mass spectrometry. The method developed in this paper consumes 0.5 g of total HS for separating and detecting different HS disaccharides. It will serve as a practical way of accomplishing HS disaccharide analysis of general HS samples from cells or tissues without radioisotope labeling. Furthermore, we could treat biologically   Three 6-OST genes have been cloned and shown to be expressed in a tissue-specific pattern. Furthermore, the individual isozymes appear to differ in substrate specificity. We do not know how 6-OST-2 and 6-OST-3 contribute in HS act generation in vivo. However, by comparing baculovirus-expressed and purified human 6-OST-1, 6-OST-2, and 6-OST-3, we found tional competence is responsible for these effects. Unfortunately no good antibodies exist that would allow us to monitor the expression and the post-modification of 6-OST-1.
6-O-Sulfation has been shown to be critical not only in AT binding, but also in FGF-FGFR interactions and other biologically important protein-HS interactions (5-11). The contributions of each 6-OST isoform in generating specific biologically active HS sequences are unknown. Because 6-OST-2 and 6-OST-3 are not present in CHO cells, CHO wild-type cells and our 6-OST-1-defective mutant cells are ideal for studying the in vivo and in vitro substrate specificity for 6-OST-2 and 6-OST-3 in terms of AT-binding, FGF-FGFR-activating, and other biologically relevant sequence generations. H]HS act oligosaccharides were digested with a mixture of heparitinases. The resulting diand tetrasaccharides were separated on a Bio-Gel P6 column (Fig. 7C). The tetrasaccharides collected from Fig. 7C were further degraded into disaccharides in the presence of HIP peptide. The disaccharides collected (Fig. 7C), and the disaccharides derived from the tetrasaccharides (Fig. 7C) were resolved by IPRP-HPLC with appropriate internal standards (Fig. 8)