The effect of precursor structures on the action of glucosaminyl 3-O-sulfotransferase-1 and the biosynthesis of anticoagulant heparan sulfate.

To understand how 2-O-sulfation of uronic acid residues influences the biosynthesis of anticoagulant heparan sulfate, the cDNA encoding glucosaminyl 3-O-sulfotransferase-1 (3-OST-1) was introduced into wild-type Chinese hamster ovary cells and mutant pgsF-17 cells, which are defective in 2-O-sulfation. 3-OST-1-transduced cells gained the ability to bind to antithrombin. Structural analysis of the heparan sulfate chains showed that 3-OST-1 generates sequences containing GlcUA-GlcN(SO(3))3(SO(3)) and GlcUA-GlcN(SO(3))3(SO(3))6(SO(3)) in both wild-type and mutant cells. In addition, IdoUA-GlcN(SO(3))3(SO(3)) and IdoUA-GlcN(SO(3))3(SO(3))6(SO(3)) accumulate in the mutant chain. These disaccharides were also observed by tagging [6-(3)H]GlcN-labeled pgsF-17 heparan sulfate in vitro with [(35)S]PAPs and purified 3-OST-1. Heparan sulfate derived from the transduced mutant also had approximately 2-fold higher affinity for antithrombin than heparan sulfate derived from the transduced wild-type cells, and it inactivated factor Xa more efficiently. This study demonstrates for the first time that (i) 3-O-sulfation by 3-OST-1 can occur independently of the 2-O-sulfation of uronic acids, (ii) 2-O-sulfation usually occurs before 3-O-sulfation, (iii) 2-O-sulfation blocks the action of 3-OST-1 at glucosamine residues located to the reducing side of IdoUA units, and (iv) that alternative antithrombin-binding structures can be made in the absence of 2-O-sulfation.

Heparan sulfate (HS) 1 is a linear polymer covalently attached to the protein cores of proteoglycans, which are abundant and ubiquitously expressed in almost all animal cells as integral membrane proteins, glycosylphophatidylinositol-linked membrane proteins, and proteins of the extracellular matrix. HS assembles by the action of a large family of enzymes that catalyzes chain polymerization (alternating the addition of GlcNAc and GlcUA residues), GlcNAc N-deacetylation and N-sulfation, glucuronic acid (GlcUA) epimerization to L-iduronic acid (IdoUA), 2-O-sulfation of uronic acid residues, and 3-O-and 6-O-sulfation of glucosaminyl residues. Tissue-specific and developmentally regulated expression of enzyme isoforms (e.g. at least four Ndeacetylation/N-sulfotransferases (1)(2)(3)(4)(5), three 6-O-sulfotransferases (6,7), and five 3-O-sulfotransferases (8,9)) produces HS chains with distinct sequences. 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. 10 -16).
The overall structure of HS act and HS inact differs at the disaccharide level (21). In general, the AT binding oligosaccharides always have a 2-O-sulfated IdoUA adjacent to the 3-Osulfated glucosamine unit. To understand how 2-O-sulfation contributes to HS act biosynthesis and AT binding, we used a 2-O-sulfation-defective Chinese hamster ovary (CHO) mutant designated pgsF-17 (24). This mutant does not express mRNA for 2-OST and therefore synthesizes HS chains without 2-Osulfated IdoUA or GlcUA residues. Interestingly, this mutant has higher amounts of GlcNS residues compared with parental wild-type cells, suggesting that 2-O-sulfation suppresses the action of the GlcN N-sulfotransferases (24). We now report that 3-OST-1 can act in the absence of 2-O-sulfation and still generate AT-binding sequences. One of the functions of 2-O-sulfation is apparently to restrict the action of 3-OST-1 at certain sites along the chain.

EXPERIMENTAL PROCEDURES
Cell Culture-Wild-type CHO cells (CHO-K1) were obtained from the American Type Culture Collection (CCL-61; ATCC, Manassas, VA). Wild-type cells and the 2-O-sulfotransferase-deficient mutant pgsF-17 (24) 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 fetal bovine serum that had been dialyzed 200-fold against phosphate-buffered saline (PBS) (25). All tissue culture media and reagents were purchased from Life Technologies, Inc. unless otherwise indicated.
The PHOENIX ecotropic retroviral packaging cell line (ATCC no. SD 3444) was a generous gift from Dr. Gary Nolan (Stanford University Medical Center). PHOENIX ecotropic cells were maintained as described (26). Wild-type or mutant pgsF-17 cells (1 ϫ 10 6 ) were transfected with 10 g of pcB7-ECOTROPIC (a generous gift from Dr. Harvey Lodish) using calcium phosphate precipitation (27). The plasmid contains MCAT1, an ecotropic retrovirus receptor, and hygromycin resistance genes from separate constitutive promoters. The transfected cells were selected for resistance to 200 g/ml hygromycin B (Life technologies, Inc.). Each stable hygromycin-resistant clone was assayed for its ability to be infected by the reporter virus MSCVPLAP, containing alkaline phosphatase (28). Virion Production-The retrovirus plasmid pMSCVpac was a generous gift of Dr. Robert Hawley (University of Toronto) (29). pCMV3-OST-1 was digested with BglII and XhoI to release the wild-type murine 3-OST-1 cDNA (8). The cDNA fragment (1,623 base pairs) was cloned into the BglII ϩ XhoI sites in pMSCVpac. All plasmid DNA prepared for transfection was made with the Invitrogen SNAP-MIDI kit according to manufacturer directions.
Infectious virions were produced by programming ecotropic PHOE-NIX packaging cells with recombinant provirus plasmids using the calcium phosphate transfection technique as described previously (26). After the precipitation step, the cells were re-fed with 2 ml/well of fresh Dulbecco's modified Eagle's medium and incubated overnight. Viral supernatants were collected, either flash-frozen in liquid nitrogen, and stored at Ϫ80°C or used directly after low speed centrifugation.
Wild-type and mutant cells containing ecotropic receptors were treated with trypsin and then plated at 150,000 cells/well in a 6-well dish. One day later, target cells (Ͻ70% confluent) were incubated overnight with viral supernatants containing 5 g/ml Polybrene surfactant. After 12 h, the virus containing medium was replaced with fresh growth medium. Wild-type and mutant cells were exposed to recombinant retrovirus three times and selected and maintained in 7.5 g/ml puromycin (Sigma). More than 90% of mutant cell populations express 3-OST-1 after three rounds of viral exposure.
Antithrombin and FGF-2 Labeling-The standard reaction mixture for preparing fluorescent AT contained 20 mM NaH 2 PO 4 (pH 7.0), 0.3 mM CaCl 2 , 25 g of PBS-dialyzed AT (GlycoMed), 4 milliunits neuraminidase (Worthington), 4 milliunits galactose oxidase (Worthington), and 125 g/ml fluorescein hydrazide (C-356, Molecular Probes) in a final volume of 280 l. The mixtures were incubated at 37°C for 1 h. PBS (1 ml) and a 50% slurry of heparin-Sepharose in PBS (100 l) was added and mixed end-over-end for 20 min. After centrifugation, the heparin-Sepharose beads were washed four times with PBS (1 ml). Labeled AT was eluted with four 0.25-ml aliquots of 10ϫ concentrated PBS and desalted by centrifugation for 35 min at 14,000 rpm through two Microcon-10 columns (Millipore). The concentrated AT was diluted with 0.5 ml of 10% FBS in PBS containing 2 mM EDTA and used directly for cell labeling studies.
Fluorescent FGF-2 was prepared by mixing 50 l of 1 M sodium bicarbonate to 0.5 ml of PBS containing 2 mg/ml bovine serum albumin and 3 g of FGF-2. The mixture was then transferred to a vial of reactive dye (Alexa 594, Molecular Probes) and stirred at room temperature for 1 h. The isolation of the labeled FGF-2 was identical to that described above for labeled AT.
Cell Sorting-Nearly confluent monolayers of cells were detached by adding 10 ml of 2 mM EDTA in PBS containing 10% FBS and centrifuged. The cell pellets were placed on ice, and 50 l each of fluorescein-AT and Alexa 594-FGF-2 was added. After 30 min, the cells were washed once and resuspended in 1 ml of 10% FBS 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. Double-positive cells were subsequently single-cell-sorted into a 96-well plate. The single cell clones were expanded and frozen for further analysis.
HS Preparation and Analysis-Cell monolayers were labeled overnight with 100 Ci/ml sodium [ 35 S]sulfate (carrier free, ICN) in sulfatedeficient Dulbecco's modified Eagle's medium, supplemented with penicillin G (100 units/ml), and 10% (v/v) dialyzed FBS. Metabolic labeling with [6-3 H]glucosamine was done by incubating cells overnight in Dulbecco's modified Eagle's medium containing 1 mM glucose, 10% dialyzed FBS, and 100 Ci/ml D-[6-3 H]glucosamine (40 Ci/mmol, ICN). The proteoglycan fraction was isolated by DEAE-Sepharose chromatography (30) and beta-eliminated in 0.5 M NaBH 4 in 0.4 M NaOH at 4°C overnight. The samples were neutralized with 5 M acetic acid until bubble formation ceased, and the released chains were purified by another round of DEAE-Sepharose chromatography followed by ethanol precipitation. The pellet from centrifugation was washed with 75% ethanol and resuspended in water. The glycosaminoglycans were digested with 20 milliunits of chondroitinase ABC (Seikagaku, Inc.) in buffer containing 50 mM Tris-HCl and 50 mM sodium acetate (pH 8.0). Complete digestion of chondroitin sulfate by chondroitinase ABC was assured by monitoring the extent of conversion of the carrier to disaccharides (100 g ϭ 1.14 absorbance units at 232 nm). HS was purified from chondroitinase-degraded products by phenol/chloroform (1:1 (v/v)) extraction and ethanol precipitation. After washing the pellets with 0.5 ml of 75% ethanol, the HS was dissolved in water for further analysis.
HS chains were analyzed by gel filtration HPLC (TSK G3000SW, 60 cm ϫ 7.5 mm inner diameter, TosoHass, Inc.). The column was equilibrated in 100 mM KH 2 PO 4 buffer (pH 6.0) containing 0.2% CHAPS and 0.5 M NaCl and run at a flow rate of 0.5 ml/min. Blue dextran (Sigma) and [ 35 S]sulfate were used to determine the V o and V t of the column, respectively. Radioactivity in the effluent was determined by in-line liquid scintillation spectrometry (Packard) with a 12-s sampling rate, and the data were averaged over 1-min intervals.
The absolute levels of HS were determined by a modified Alcian blue staining assay. A dye stock containing 0.5% Alcian blue, 0.018 M H 2 SO 4 , and 0.4 M guanidine HCl was centrifuged and filtered (0.2 m). The working dye solution contained 0.25% Triton X-100 and 0.018 M H 2 SO 4 ; the 5% dye stock was similarly clarified. Ten l of a solution containing 0.027 M H 2 SO 4 , 0.375% Triton X-100, and 4 M guanidine HCl was added to samples or standards in water (10 l) along with 100 l of working FIG. 5. IPRP-HPLC of iduronidasedigested 3 H-labeled HS disaccharides. An in vitro 3-OST-1 and cold PAPS modified mutant [ 3 H]HS sample was depolymerized completely by hydrazinolysis and nitrous acid digestion, and the products were purified by gel-filtration chromatography. The disaccharides were either directly analyzed by IPRP-HPLC (solid line) or digested with iduronidase and then analyzed by IPRP-HPLC (dotted line) (see "Experimental Procedures"). Peak 1, aMan R and GlcUA-aMan R ; peak 2, aMan R 3S and aMan R 6S; peak 3, IdoUA-aMan R 3S; peak 4, GlcUA-aMan R 3S; peak 5, GlcUA-aMan R 6S; peak 6, IdoUA-aMan R 6S; peak 7, aMan R 3S6S; peak 8, IdoUA-aMan R 3S6S; peak 9, GlcUA-aMan R 3S6S. The broken line indicates the gradient of acetonitrile. dye solution. After centrifugation for 10 min at 10,000 ϫ g in a microcentrifuge the supernatant was removed, and the pellet was dissolved in 8 M guanidine HCl. The absorbance at 600 nm was measured, and the HS concentration was determined from a standard curve.
Separation of HS act and HS inact by AT Affinity Chromatography-AT-HS complexes were created by mixing 3-O-sulfated HS in 500 l of HB buffer (150 mM NaCl, 10 mM Tris-Cl (pH 7.4)) with 2.5 mM AT, 100 g of chondroitin sulfate, 0.002% Triton X-100, and 1 mM each of CaCl 2 , MgCl 2 , and MnCl 2 (32). HB containing ϳ50% slurry of Concanavalin A-Sepharose 4B (60 l) was then added. AT complexes were bound to Concanavalin A by way of the Asn-linked oligosaccharides. After 1 h at room temperature, the beads were sedimented by centrifugation at 10,000 ϫ g. The supernatant was collected, and the beads were washed three times with 1.25 ml of HB containing 0.0004% Triton X-100. The supernatant and washing solutions contained HS inact . HS act was eluted with three successive washes with 100 l of HB containing 1 M NaCl and 0.0004% Triton X-100. After adding 100 g of chondroitin sulfate as carrier to HS act , the sample was extracted with an equal volume of phenol/chloroform followed by chromatography on DEAE-Sepharose and ethanol precipitation. The pellets were washed with 75% ethanol, dried briefly under a vacuum, and dissolved in water.
Disaccharide Analysis of HS-Radiolabeled HS samples were mixed with 10 g of bovine kidney HS (ICN) and depolymerized by hydrazinolysis and nitrous acid degradation (22). Disaccharides were purified by Bio-Gel P2 chromatography and resolved by ion-pairing reversephase HPLC with appropriate disaccharide standards. Bio-Gel P2 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, collecting 0.5-ml fractions. The desired fractions were either pooled or dried individually under a vacuum to remove ammonium bicarbonate. HS was digested at 37°C for 4 h in 100 l of 40 mM ammonium acetate (pH 7.0) containing 1 mM CaCl 2 and 1 milliunit of heparitinase I (EC 4.2.2.8, Seikagaku Corp.) and analyzed on a P6 column (0.75 ϫ 200 cm) as described above. For iduronidase digestion, 10% of the total radiolabeled disaccharides collected from Bio-Gel P2 chromatography was dissolved in 400 l of 1 ϫ incubation buffer (glycosaminoglycan-5006, Oxford GlycoSciences) with 0.5 ␣-iduronidase (glycosaminoglycan-5006) at 37°C for 16 h. After adding 400 l of 333 mM tetrabutylammonium dihydrogen phosphate, the digested sample was injected into ion-pairing reverse-phase HPLC.
Affinity co-electrophoresis was done as described (33). HS act from both wild type and mutants was electrophoresed in an agarose gel through zones containing AT at different concentrations. The migration of HS act was retarded by the presence of AT, with the degree of retardation dependent on AT concentration. AT Inactivation of Thrombin and Factor Xa-Human ␣-thrombin (4 mg/ml 50% glycerol, 3,000 units/mg) and factor Xa (10.4 mg/ml 50% glycerol, 820 units/mg) were from Hematologic Technologies. Factor Xa and ␣-thrombin were diluted 1:200 with PBS containing 1 mg of bovine serum albumin (4 units/ml and 15 units/ml, respectively). AT (2.5 mg/ml) was diluted 1:200 to give a 2 ϫ 10 Ϫ7 M stock solution. The chromogenic substrates, S-2765 (factor Xa assays) and S-2238 (thrombin assays), were from Chromogenix. Both substrates were made up at 1 mM with 1 mg/ml Polybrene in water. Heparin (174 international units/mg, Sigma) was used as a standard. HS act was used for factor Xa (Յ25 ng) and thrombin assays (Յ62. The inhibition rate constant (k 2 ) for AT inactivation of thrombin and factor Xa were derived from the equation where [E] o and [E] t are the concentrations of active enzyme at the initial and measured times of the reaction, respectively. The concentration of Xa was calculated based on the activity measured in the absence of inhibitor (34).

3-OST-1 Transduction Imparts Binding to AT but not to FGF-2-CHO wild-type and pgsF-17 cells do not express
3-OST-1 and therefore do not make detectable 3-O-sulfate containing saccharides or bind to AT. The lack of uronyl 2-Osulfotransferase activity in the mutant also prevents binding to FGF-2 (24). To study the effect of 3-OST-1 on ligand binding, mutant and wild-type cells were stably transduced with multiple copies of 3-OST-1 cDNA. As shown by flow cytometry, the expression of 3-OST-1 gave rise to wild-type and mutant cells that bound AT, but the binding of FGF-2 was unchanged (Fig.  1). The wild-type transductant contained three copies of 3-OST-1 (B, broken line), whereas the population of transduced mutant cells (D, broken line) contained one or more copies of the gene based on the observation that 90% of the cells exhibited binding to AT (D, broken line).
The mutant cells that exhibited high AT binding were singlecell-sorted into a 96-well plate, and 12 clones were expanded for further analysis. Genomic DNA from the clones was digested with EcoRI and probed for 3-OST-1 by Southern blot (Fig. 2). All the clones, including wild-type cells and the original mutant, have an endogenous copy of 3-OST-1. No enzyme activity was found in nontransduced wild-type and mutant cells, indicating that the endogenous gene was inactive. The 12 clones have 1-5 inserted copies of the gene. 3-OST-1 activity roughly correlated with the number of extra copies of the gene (coefficient of correlation ϭ 0.948; data not shown).
Heparan Sulfate from Mutant Cells Transduced with 3-OST-1 Accumulate Unusual Disaccharides-We have shown previously that 3-OST-1 generally acts on glucosamine units  (Fig. 3, solid line) consist of GlcUA-aMan R 6S (peak 2) and IdoUA-aMan R 6S (peak 3). In vitro treatment of mutant HS with pure 3-OST-1 and cold PAPS yielded two new disulfated disaccharides (Fig. 3, dotted line). preparations from transduced F17 clones 2, 3, 10, and 12 (containing one, three, five, and four copies of 3-OST-1, respectively) were analyzed and compared with the transduced wildtype and parental lines. [ 3 H]HS chains were purified and depolymerized completely by hydrazinolysis and nitrous acid, and the resultant disaccharides were analyzed by IPRP-HPLC (see "Experimental Procedures"). As shown previously, the predominant monosulfated disaccharides in CHO cells consist of GlcUA-aMan R 6S (peak 4) and IdoUA-aMan R 6S (peak 5), and the major disulfated species consists of IdoUA2S-aMan R 6S (peak 7) (35). A new disaccharide accumulated after transduction of wild-type cells, namely GlcUA-aMan R 3S6S (peak 6). The transduction of F17 cells also yielded two unknown disaccharides, designated X and Y, that were not present in parental lines (Fig. 4A) or in wild-type cells transduced with 3-OST-1 (Fig. 4, B and D). Several lines of evidence demonstrated that X and Y were IdoUA-aMan R 3S and IdoUA-aMan R 3S6S, respectively. First, 3 H-labeled peak Y co-elutes with an IdoUA-[3-O- 35 S]aMan R 6S standard on IPRP-HPLC (36). Second, when the disaccharides were treated with iduronidase, IdoUA residues were removed from IdoUA-aMan R , IdoUA-aMan R 6S, IdoUA-aMan R 3S, and IdoUA-aMan R 3S6S, whereas GlcUA-aMan R , GlcUA-aMan R 3S, GlcUA-aMan R 6S, and GlcUA-aMan R 3S6S were not affected (Fig. 5). Third, a precursor-product relationship was observed in the IPRP-HPLC profiles (Fig. 3)   To examine how the unusual disaccharides might affect the binding of HS to AT, [6-3 H]GlcN-labeled HS chains from F17 cells were 3-O-sulfated in vitro and digested with heparitinase I, which cleaves the chains in regions poor in sulfate, generating oligosaccharides from domains rich in sulfated residues. The products were analyzed by Bio-Gel P6 chromatography (Fig. 6, A and B), and individual fractions representing oligosaccharides of different lengths were analyzed for AT binding. Only oligosaccharides larger than hexamers bound AT regardless of the source of the material (Table I). The AT-binding oligosaccharides in the wild-type oligosaccharides had a 3 H/ 35 S ratio of 1:(0.9 -1.1), whereas that of the mutant had a ratio of 1:(1.7-2.1) depending on the chain length. This result suggested that the mutant had twice as many 3-O-sulfated disaccharides, which was consistent with the disaccharide analysis shown in Fig. 4.
HS act Chains from Mutant and Wild Type Are the Same Size-Previous studies of F17 HS showed that chains were much longer than in wild-type cells because of the lack of heparanase cleavage sites (37). To determine whether chain length affected sulfation by 3-OST-1, we prepared [6-3 H]GlcNH 2 -labeled chains from mutant and wild-type cells and introduced 3-O-sulfate groups in vitro with PAPS and pure 3-OST-1. The chains were affinity-purified and then compared by gel filtration HPLC to chains that were not affinityfractionated (Fig. 7). HS from the wild type consisted of three populations of chains that differed in size as reported (37). However, the affinity-fractionated HS consisted of only relatively large chains of the same overall size in both mutant and wild-type preparations. Similar results were obtained when the high affinity chains were analyzed from cells transduced with 3-OST-1 (data not shown), indicating that the larger chains preferentially contain the precursor sequence recognized by 3-OST-1.
HS from 3-OST-1 Containing Mutant Cells Has Higher Affinity for AT-To compare the AT-binding affinity of HS from 3-OST-1-transduced wild-type and mutant cells, [ 35 S]HS was purified from the transduced cell lines. HS act was purified by AT affinity chromatography and analyzed by affinity co-electrophoresis using agarose gels containing different amounts of AT (Fig. 8). Using the midpoint of the bands to determine relative mobilities, the data for the mutant indicated a K D value of 17 nM, whereas the wild-type HS act gave a value of 32 nM (linear coefficient values ϭ 0.98 and 0.97, respectively). This finding is consistent with earlier studies of synthetic oligosaccharides that showed that pentasaccharides containing two 3-O-sulfate groups (GlcNS6S-GlcUA-GlcNS3S6S-IdoUA2S-GlcNS3S6S-OCH 3 ) had a higher affinity for AT compared with mono-3-O-sulfated pentasaccharide (GlcNS6S-GlcUA-GlcNS3S6S-IdoUA2S-GlcNS6S-OCH 3 ) (38). However, we have not determined whether the additional 3-O-sulfate groups in the transduced mutant are present on adjacent glucosamine units as they are in the synthetic oligosaccharides.
AT Inhibition of Xa and Thrombin-Based on the higher affinity of the HS from the transduced mutant, we predicted that the HS act fraction should display a greater rate of ATmediated factor Xa inactivation (Fig. 8). It was unknown whether the mutant HS act would mediate AT-dependent thrombin inactivation, because thrombin binding to HS might require 2-O-sulfation of the uronic acids. To examine this question, HS was purified from the mutant and wild-type cells and modified by 3-OST-1 and PAPS in vitro. HS act was fractionated by AT affinity assay, quantitated chemically, and tested for its ability to activate AT-dependent inhibition (see "Experimental Procedures"). HS act from the mutant showed a 2.2-fold increased k i value for factor Xa inhibition but only a 1.2-fold increase in thrombin inhibition compared with the values obtained for the HS act derived from wild-type cells transduced with 3-OST-1 (Table II). Thus, the lack of 2-O-sulfate groups does not diminish factor Xa and thrombin inhibition by AT and in fact may accelerate the reactions directly through an enhanced affinity of the HS chains caused by the increase in 3-O-sulfate groups. DISCUSSION The unexpected finding presented in this report is that in the absence of 2-O-sulfation of uronic acid residues, 3-OST-1 will generate disaccharide units containing IdoUA to the nonreducing side of the 3-O-sulfated glucosamine unit. This specificity is different from that originally described for 3-OST-1(8), in which GlcUA was the preferred unit in this position. We cannot tell if the change in substrate specificity is directly the effect of reduced 2-O-sulfation of IdoUA units or if the higher levels of N-sulfated glucosamine residues in the mutant might play a role (24). However, the extent of N-sulfation increases by 50% (ϳ40% GlcNS in the wild type to ϳ60% in the mutant). In contrast, the change in 2-O-sulfation is quantitative and therefore more likely the cause of altered specificity of 3-OST-1. Previous work showed that HS act had a lower content of IdoUA2S-aMan R compared with HS inact , which is consistent with the idea that 2-O-sulfated substrates inhibit 3-O-sulfation (32). Thus, we imagine an exclusion mechanism catalyzed by uronyl 2-O-sulfotransferase, which normally prevents glucosa- mine residues to the reducing side of IdoUA from acting as substrates for 3-OST-1. Because in vitro 3-O-sulfation can impart anticoagulant activity to inactive HS chains, it has been assumed that 3-Osulfation is the final modification step during HS biosynthesis. This study provides direct evidence that 2-O-sulfation preferentially occurs before 3-O-sulfation catalyzed by 3-OST-1. The facts are: 1) the 2-OST transfectant of the 3-OST-1 expression mutant cells does not form the new 3-O containing disaccharides, and 2) 2-O-sulfation results in less sites recognizable by the enzyme (two in wild-type HS versus eight in mutant 3-OST-1-transduced cells and six in wild type versus 14 in the mutant after in vitro sulfation; Fig. 4). The order of reactions for other isoforms of the 3-O-sulfotransferases is unknown, but the presence of IdoUA-GlcNS3S and IdoUA2S-GlcNS3S units in bovine glomerular basement membrane HS suggests that the reaction sequence may differ (39).
Multiple sulfotransferases have been cloned and shown to be expressed in a tissue-specific and developmentally regulated pattern. Furthermore, the individual isozymes seem to differ in substrate specificity (2,6,40). The results shown here suggest another layer of control determined by the level of alternate precursor structures that are available in the cell as determined by the level of the uronyl 2-O-sulfotransferase. Thus, by altering the level of 2-O-sulfation, the pattern and extent of 3-O-sulfation by 3-OST-1 changes, which in turn affects the formation of other possible binding sequences. Another example of this was described for glucosaminyl 6-O-sulfotransferase-1, which only generates the IdoUA-GlcNS6S sequence when chemically desulfated and re-N-sulfated heparin was used as a substrate but makes equal amounts of IdoUA-GlcNS6S and IdoUA2S-GlcNS6S sequences when bovine kidney HS was used as a substrate (6,41). Thus, the precursor structures play a critical role in determining the extent of modification and, by inference, the formation of specific binding sequences for ligands.
Previous work has shown that the size distribution of HS chains in wild type and the F17 mutant line differ because of the lack of heparanase cleavage in chains deficient in 2-Osulfated uronic acid residues (37). In this report, we have shown that HS act consists of only large chains compared with HS inact , which consists of large, medium, and small chains (Fig.  7). Similar differences have also been observed in F9 cell subpopulations of HS. 2 These results suggest that not only the biosynthesis but also the degradation of HS act precursors in CHO are different from the rest of HS chains. Indeed, the relatively lower level of HS act chains in the turnover products suggests that heparanase cleavage may target the precursor sequence for AT.
Retrovirally transduced cell lines can have different copy numbers of the ectopically inserted genes (Fig. 2). The advantage of this arrangement is that other genes responsible for generating specific HS can be sought after chemical mutagenesis. Taking advantage of this approach, a series of CHO mutants defective in the early steps of HS biosynthesis have been generated. 3 These mutants will provide insights into the effect of precursor structures on the action of HS biosynthetic en-zymes and the function of anticoagulant HS biosynthetic pathways.