Substrate specificity of 6-O-endosulfatase (Sulf-2) and its

Substrate specificity of 6-O-endosulfatase (Sulf-2) and itsimplications in synthesizing anticoagulant heparan sulfate

Elizabeth H Pempe, Tanya C Burch, Courtney J Law,and Jian Liu1

Division of Chemical Biology and Medicinal Chemistry, Eshelman School ofPharmacy, University of North Carolina, Rm 303, Beard Hall, Chapel Hill,NC 27599, USA

Received on November 16, 2011; revised on May 31, 2012; accepted onJune 1, 2012

Heparan sulfate (HS) 6-O-endosulfatase (Sulf ) catalyzesthe hydrolysis of 6-O-sulfo groups from HS polysacchar-ides. The resultant HS has reduced sulfation levels anddisplays altered biological activities. The Sulfs have beenassociated with several cancers and developmental pro-blems and could function as a tool for editing specificHS structures. Here, we characterize the substrate specifi-city of human Sulf-2 using site-specifically radiolabeledsynthetic polysaccharides. The enzyme was expressedand harvested from the conditioned medium of Chinesehamster ovary cells transfected with Sulf-2 expressionplasmids. The uniquely [35S]sulfated polysaccharides wereprepared using purified recombinant HS biosyntheticenzymes. We found that Sulf-2 is particularly effective inremoving the 6-O-sulfo group residing in the trisulfateddisaccharide repeating unit comprising 2-O-sulfated uronicacid and N-sulfated 6-O-sulfo glucosamine, but can alsohydrolyze sulfo groups from N- and 6-O-sulfated disac-charides. In addition, we found that Sulf-2 treatment sig-nificantly decreases HS’s ability to bind to platelet factor 4(PF4), a chemokine, while binding to antithrombin ismaintained. Because HS–PF4 complexes are the initiatingcause of heparin-induced thrombocytopenia, this findingprovides a promising strategy for developing heparin ther-apies with reduced side effects. Further understanding ofSulf-2 activity will help elucidate HS structure–functionrelationships and provide a valuable tool in tailoring HS-based anticoagulant drugs.

Keywords: heparan sulfate / heparin / substrate specificity /sulfatase

Introduction

Heparan sulfate (HS) is a glycosaminoglycan present on thesurface of all animal tissues as well as within the extracellularmatrix. As the body’s most negatively charged molecule, itinteracts with numerous proteins and has a role in manyphysiological processes, including cell proliferation, coagula-tion and the immune response (Rosenberg et al. 1997; Eskoand Selleck 2002; Taylor and Gallo 2006). A special highlysulfated form of HS, heparin, is isolated from porcine intes-tine and is the world’s most widely used anticoagulant drug.Heparin activates antithrombin (AT), which prevents theaction of thrombin on fibrinogen and the formation of bloodclots (Rosenberg et al. 1997). Heparin is very effective buthas several concerning drawbacks, including contaminationissues like those causing the wide-spread heparin recall in2008 (Guerrini et al. 2009) and potentially deadly side effectslike heparin-induced thrombocytopenia.HS exists as a linear polysaccharide attached to a protein

core by a tetrasaccharide linkage region. The carbohydratechain is made up of repeating disaccharide units of either glu-curonic or iduronic acid (GlcA/IdoA) and glucosamine(GlcN). These disaccharides can be decorated at several posi-tions with sulfo groups, and the GlcN residue can be unsubsti-tuted, acetylated or sulfated. These modifications occur duringHS biosynthesis within the cell. The location of the sulfogroups and the position of IdoA are crucial for the biologicalfunctions of HS.HS biosynthesis occurs primarily in the Golgi apparatus

by series of specialized enzymes (Turnbull et al. 2001). First,a core linkage region consisting of xylose-galactose-galactose-GlcA is attached to the core protein, and HS polymerasesynthesizes a linear polysaccharide chain consisting of alter-nating GlcA and GlcNAc (N-acetyl glucosamine) units (Lindet al. 1998). Next, a series of biosynthetic enzymes modifythe carbohydrate chain. N-deacetylase/N-sulfotransferase (NST)converts GlcNAc to GlcNS, and C5-epimerase (C5-epi) con-verts some GlcA residues to IdoA. 2OST (2-O-sulfotransferase)adds sulfation to the 2-O position of uronic acid residues, and6OST and 3OST sulfate GlcN sugars (Lindahl et al. 1998;Feyerabend et al. 2006). After synthesis, mature HS chains canbe modified by other enzymes such as heparanase and sulfa-tases. HS and heparin differ in their amount of sulfation (0.6 vs2.6 sulfo groups per disaccharide) and their iduronic acidcontent (20 vs 80%) (Peterson et al. 2009).A novel class of sulfatase enzymes known as Sulfs

(6-O-endosulfatases) was found to have endosulfatase activity1To whom correspondence should be addressed: Tel: +1-919-843-6511;Fax: +919-966-0204; e-mail: jian_liu@unc.edu

Glycobiology vol. 22 no. 10 pp. 1353–1362, 2012doi:10.1093/glycob/cws092Advance Access publication on June 12, 2012

© The Author 2012. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 1353

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

and remove 6-O-sulfo groups from HS (Dhoot et al. 2001;Morimoto-Tomita et al. 2002). Two Sulf enzymes (Sulf-1 andSulf-2) have been identified, cloned and expressed in severalcell lines (Dhoot et al. 2001; Morimoto-Tomita et al. 2002;Ohto et al. 2002; Nagamine et al. 2005; Ai et al. 2006). Thetwo Sulfs appear to be functionally redundant and have variedroles in cell signaling, development and cancer. The Sulfspromote some signaling pathways, such as Wnts (Dhoot et al.2001; Ai et al. 2003; Tang and Rosen 2009), bone morpho-genic protein (Viviano et al. 2004) and glial cell-derivedneurotrophic factor (Ai et al. 2007) while inhibiting others,such as fibroblast growth factor-2 (Dai et al. 2005; Lamannaaet al. 2007) and transforming growth factor-β (Yue et al.2008). Gene knockdown (Dhoot et al. 2001) and knock-outstudies (Ai et al. 2007; Holst et al. 2007; Lum et al. 2007;Ratzka et al. 2008) have shown the importance of Sulfs in de-velopment; these mice show aberrant growth, muscle innerv-ation, skeletal tissue and lung development. In cancer, Sulfsare believed to possess both pro-oncogenic (Morimoto-Tomitaet al. 2002; Nawroth et al. 2007; Lai et al. 2008) and tumorsuppressing (Danenberg et al. 2003; Lai et al. 2004; Dai et al.2005) activities. Overexpression of Sulf-2 in particular was re-cently found to promote carcinogenesis in non-small-cell lungcarcinomas, pancreatic cancer and hepatocellular carcinoma(Lemjabbar-Alaoui et al. 2010; Rosen and Lemjabbar-Alaoui2010). Sulf-2 also regulates receptor tyrosine kinase pathwaysand tumor growth in glioblastoma (Phillips et al. 2012),making it an attractive target for cancer therapy.The Sulfs are understood to cleave 6-O-sulfo groups from

trisulfated (N-, 2-O- and 6-O-) disaccharides (Morimoto-Tomita et al. 2002). However, the extent of their substrate spe-cificity and their ability to recognize other disaccharides is notwell understood, largely due to the fact that polysaccharideswith defined sulfation types were not available. In previousstudies, only substrates containing this trisulfated motif havebeen used to test Sulf-2 activity (Morimoto-Tomita et al.2002; Saad and Leary 2003). Recently, we demonstrated thecontrol of the sulfation types in HS polysaccharides using bio-synthetic enzymes (Chen et al. 2007), enabling us to investi-gate the substrate specificity of Sulfs in greater detail. Here,we report the substrate specificity of human Sulf-2 usingenzymatically synthesized 35S-labeled polysaccharides thatallowed us to test Sulf-2′s ability to recognize other disacchar-ide motifs. In addition, for the first time, we have investigatedthe effect of Sulf-2 treatment on the ability of HS to bind toAT and PF4. The interactions between HS and these two pro-teins are important for controlling the anticoagulant activityand side effects of heparin, respectively. Further characteriza-tion of Sulf-2 and its effect on HS–protein interactionswill serve as a valuable tool for modulating HS structures andgenerating specific bioactive polysaccharides for medicalapplications.

ResultsExpression of an active Sulf-2 enzymeThe Sulf-2 enzyme was transiently expressed in chinesehamster ovary (CHO) cells with the goal of obtaining theSulf-2 protein for the substrate specificity study. The

recombinant Sulf-2 enzyme was detected in both the cell lysateas well as in the conditioned medium as measured by westernanalysis (Figure 1D). The recombinant Sulf-2 enzyme was har-vested from the conditioned medium of transfected CHO cells.To test the activity of the crude protein, [35S]HS isolated fromCHO cells that had been metabolically labeled with sodium[35S]sulfate was subjected to Sulf-2 treatment. The resultant HSwas digested with heparin lyases to disaccharides for disacchar-ide compositional analyses. The high-performance liquid chro-matography (HPLC) chromatograms of the disaccharideanalysis of Sulf-2-treated, empty vector-treated and untreated[35S]HS samples (Figure 1) show a clear decrease in the trisul-fated ΔUA2S-GlcNS6S peak in the Sulf-2-treated sample withan increase in the corresponding free [35S]sulfate andΔUA2S-GlcNS peaks. This suggested that the enzyme was infact active, and it could act to desulfate 6-O-sulfo groupslocated on UA2S-GlcNS6S disaccharides as described in previ-ously published reports (Morimoto-Tomita et al. 2002; Saadand Leary 2003).

Preparation of substrates with different sulfation groupsHaving determined that Sulf-2 was active, we sought to deter-mine with greater detail which polysaccharide substrates couldfunction as substrates for the enzyme. To this end, a seriesof polysaccharides carrying different sulfation types with orwithout IdoA were prepared using an enzymatic approach, asdemonstrated in Figure 2. A 35S-label was strategically intro-duced to a specific site by a sulfotransferase, facilitating theidentification of which sulfo groups were removed by Sulf-2.For example, construct 1, IdoA2S-[6-O-35S]GlcNS6S, carriedthe 6-O-[35S]sulfo group at the N,6-O-sulfo glucosamine(GlcNS6S) unit. Thus, a release of [35S]sulfo groups fromconstruct 1 after Sulf-2 treatment unambiguously indicatedthat a 6-O-desulfation reaction occurred. Construct 1 was pre-pared by incubating deacetylated heparosan with unlabeled 3′-phosphoadenosine-5'-phosphosulfate (PAPS) and NST, 2OSTand C5-epi in subsequent steps. This compound was35S-labeled at the 6-O position using [35S]PAPS and 6OST.The rest of the constructs (2–9) were prepared in a similarfashion using recombinant enzymes, as shown in the syntheticscheme (Figure 2). The structure of each substrate was con-firmed by disaccharide analysis as described previously(Peterson and Liu 2010).A disaccharide analysis of the Sulf-2-treated substrate

was also performed for each construct. The HPLC analysis ofuntreated and Sulf-2 treated construct 1 is shown as anexample in Figure 3. The presence of the trisulfated peak,ΔUA2S-GlcNS6S, is nearly completely eliminated by treat-ment with the Sulf-2 enzyme. The reduction in this peak isbalanced by the appearance of a large free sulfate peak in thetreated sample. In addition, there is a partial decrease in theΔUA-GlcNS6S peak that contributed to the free sulfate peak.To quickly quantify the activity of Sulf-2 against the different

substrates, spin column assays were performed. Here, thecolumn separates the analytes based on the size of the molecule.We anticipated that the [35S]sulfate would be trapped in thecolumn while the intact polysaccharides would pass through thecolumn without retardation. Substrates were incubated overnightin the presence of Sulf-2 or water as a control, and the samples

EH Pempe et al.

1354

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

were purified by QuickSpin column to determine how much35S-radioactivity had been converted to smaller [35S]sulfate. Theresults from the spin column assays are shown in Table I. Fromthese assays, it is clear that the best substrate for Sulf-2 is thepolysaccharide containing N-, 2-O- and 6-O-sulfations. Furtheranalysis of the susceptibility of the different sulfated polysac-charides to [35S]sulfate release permitted us to dissect the sub-strate specificities of Sulf-2.

Effect of other sulfo groups in addition to 6-O-sulfationConstructs 2–4 were prepared to examine the effect of othersulfo groups on the cleavage of 6-O-sulfation. IdoA-[6-O-35S]GlcNS6S (construct 2) showed a 24.3 ± 2.7% decrease in35S-labeled groups compared with 8.3 ± 0.1% in the control.Construct 3, GlcA-[6-O-35S]GlcNS6S, lost 43.1 ± 2.4% of itsradiolabeled sulfo groups compared with 6.6 ± 1.6% in the

control sample. These two substrates indicate that Sulf-2 isable to desulfate the 6-O-sulfo group on the GlcNS6S that isflanked by a non-reducing-end GlcA or IdoA residue. Con-struct 4, GlcA-[6-O-35S]GlcNAc6S, which lacked N-sulfation,was not a very effective substrate, showing 28.1 ± 1.3% cleav-age compared with 14.2 ± 3.5% in the control. Despite thefact that Sulf-2 can hydrolyze both constructs 2 and 3 (con-taining -IdoA-GlcNS6S- and GlcA-GlcNS6S-, respectively),Sulf-2 most effectively desulfates the 6-O-sulfation that islocated in trisulfated disaccharide repeating regions in whichN- and 2-O-sulfation are also present.

The reactivity of Sulf-2 toward other sulfation typesTo determine whether Sulf-2 could hydrolyze sulfo groupsfrom positions other than the 6-O position, constructs 5–9were synthesized. Constructs 5 and 6 were 35S-labeled at the

Fig. 1. Disaccharide composition of Sulf-2-treated and untreated [35S]HS. [35S]HS from CHO cells was digested with heparin lyases and analyzed byreverse-phase ion-pairing HPLC. (A) Disaccharide composition of untreated [35S]HS. Peak 2, ΔUA-GlcNS; peak 3, ΔUA-GlcNac6S; peak 4, ΔUA-GlcNS6S;peak 5, ΔUA2S-GlcNS; peak 6, ΔUA2S-GlcNS6S. (B) Disaccharide composition of EV-treated [35S]HS. Peak 2, ΔUA-GlcNS; peak 4, ΔUA-GlcNS6S; peak 5,ΔUA2S-GlcNS; peak 6, ΔUA2S-GlcNS6S. (C) Disaccharide composition of Sulf-2-treated [35S]HS. Peak 1, free sulfate; peak 2, ΔUA-GlcNS; peak 3,ΔUA-GlcNac6S; peak 4, ΔUA-GlcNS6S; peak 5, ΔUA2S-GlcNS; peak 6, ΔUA2S-GlcNS6S. (D) Western blot showing the presence of the Sulf-2 protein. Celllysates (30 µg) and CM were separated on 12% SDS–PAGE, blotted to nitrocellulose and probed with anti-myc antibody. EV, empty vector-tranfected cells. (E)Lyase degradation products of [35S]HS. R1= H/SO3H, R2= H/SO3H/Ac. (F) Lyase degradation products of Sulf-2-treated [35S]HS. R1= H/SO3H, R2= H/SO3H/Ac. Cleavable groups are shown in bold.

Characterization of heparan sulfate 6-O-endosulfatase

1355

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

2-O position and also contained N-sulfation. Unlike 6, con-struct 5 was synthesized using C5-epi and thus containedIdoA. Both of these substrates showed negligible [35S]sulfaterelease when treated with Sulf-2 (12.6 ± 1.3 and 13.7 ± 3.0%compared with 10.4 ± 2.6 and 6.7 ± 1.9% in the control,respectively). Constructs 7 and 8 were 35S-labeled at the 3-Oposition, and construct 9 was labeled at the N position(Table I). These substrates also showed no decrease in35S-labeled groups after Sulf-2 treatment. From these results,we conclude that only 6-O-sulfation can be removed by theSulf-2 enzyme. We also investigated the effect of epimeriza-tion on Sulf-2 recognition. Although GlcA2S-[6-O-35S]GlcNS6S proved difficult to prepare, based on the results ofsubstrates 2 and 3, it appears that Sulf-2 is able to desulfateboth GlcA- and IdoA-containing disaccharides.

Effect of Sulf-2 treatment on the affinity of HS to AT bindingHS interacts with AT to inhibit the activity of thrombin andfactor Xa and regulate the blood coagulation cascade. Thus,

we were interested to know whether Sulf-2 treatment wouldaffect the binding of HS to AT. AT-binding HS was preparedby incubating HS from the bovine kidney with 3OST-1 and[35S]PAPS. A Sulf-2-treated fraction was prepared by incubat-ing this material with Sulf-2. Concanavalin A (ConA)-Sepharose beads were incubated with AT and HS, and theAT-bound HS fraction was eluted using a 1-M NaCl solution(Figure 4). When the treated and untreated fractions wereincubated with ConA-Sepharose and AT, 51.2% of the untreat-ed fraction and 43.6% of the Sulf-2-treated fraction wererecovered, suggesting that Sulf-2 does not remove critical6-O-sulfo groups from the AT-binding pentasaccharide withinthe polysaccharide. [N-35S]HS, which does not considerablybind AT, was used as a negative control.

Effect on Sulf-2 treatment on HS-PF4 bindingIn addition to AT, the binding of heparin to PF4 has signifi-cant clinical relevance. PF4-HS-immunoglobulin G com-plexes initiate an immune response known as heparin-induced

Fig. 2. Synthetic scheme for the 35S-labeled HS constructs. Heparosan or HS were treated with HS biosynthetic enzymes and PAPS to achieve the desiredsulfation groups and epimerization of the uronic acid. In HS, both GlcA and IdoA are present. 35S-labeled groups are shown in bold.

EH Pempe et al.

1356

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

thrombocytopenia, a dangerous and deadly side effect of heparin(Arepally and Ortel 2010). Thus, we examined whether Sulf-2could decrease the binding of HS to PF4. For these studies, weused [35S]HS from CHO cells as our untreated material due to itslarge number of radioactive sulfo groups per unit of HS.A dot blot membrane binding assay was used to compare

the PF4 binding capabilities of Sulf-2-treated and untreatedHS (Figure 4A). The wells contained 6000 cpm of Sulf-2-treated or untreated [35S]HS with increasing amounts of PF4.[35S]HS bound to PF4 was captured by the nitrocellulosemembrane. The untreated samples reached a maximumbinding of 66.3% with 152 nM PF4, but the Sulf-2-treatedsamples bound only up to 3.1% with 608 nM PF4. From thetwo binding curves, it is apparent that Sulf-2 treatment canreduce the binding of HS to PF4 by over 10-fold.For a more quantitative assessment of the effect of Sulf-2

treatment on the binding affinities of HS to AT and PF4, affin-ity co-electrophoresis was performed. This is an establishedmethod for determining the affinity constant of radiolabeledpolysaccharide ligands with proteins (Lee and Lander 1991).The mobility of each lane is determined, and the protein con-centration in each lane can be used to determine a Kd value

based on the Scatchard equation. For AT, gels containingserial dilutions from 0–3.2 µM AT were prepared and 50,000cpm of the AT-binding fraction of 3OST-1-treated [35S]HSfrom CHO cells was added. The gel was run for 2 h, driedovernight, imaged and analyzed using ImageQuant TL soft-ware. For the untreated and treated samples, a plot of R/[AT]vs R gave linear slopes of y = −0.0949 x + 0.048 (R2 = 0.87)and y = −0.094 x + 0.44 (R2 = 0.95), respectively. These corres-pond to Kd values of 10.53 and 10.59 nM, indicating that thebinding affinity of HS to AT is unaffected by treatment withSulf-2.For PF4, serial dilutions of 0–1.41 µM PF4 were used, and

50,000 cpm of 3OST-1-treated [35S]HS from CHO cells wasadded to each gel. The gels were run for 2.5 h and analyzedas above. Linear regression slopes of y = −0.1627 x + 0.1238(R2 = 0.77) and y = −0.0085 x + 0.0037 (R2 = 0.74) were calcu-lated for the untreated and treated substrates, respectively.These correspond to Kd values of 5.47 and 117.6 nM,showing that the binding affinity of HS for PF4 is decreased�20-fold by treatment with Sulf-2. Taken together, our datasuggest that Sulf-2 treatment decreases the binding affinity ofHS to PF4, while the affinity to AT remains intact.

Fig. 3. Disaccharide analysis of construct 1 (IdoUA2S-[6-O-35S]GlcNS6S), the best Sulf-2 substrate, before and after Sulf-2 treatment. (A) Disaccharidecomposition of construct 1. (B) Disaccharide composition of Sulf-2-treated construct 1; disaccharides without 35S-labeled groups are not detected. (C) Lyasedegradation products of construct 1. (D) Lyase degradation products of Sulf-2-treated construct 1. Cleavable groups are shown in bold.

Characterization of heparan sulfate 6-O-endosulfatase

1357

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

Disaccharide analysis of AT-binding HS with and withoutSulf-2 treatmentWe compared the disaccharide composition of 3-O-[35S]sul-fated HS with and without Sulf-2 treatment (Figure 5). The3-O-sulfated glucosamine residue is known to be present atthe center of the AT-binding site in HS (Liu et al. 1996). Thedisaccharide analysis revealed the presence of three3-O-sulfated disaccharides: Glc-AnMan3S, IdoA2S-AnMan3Sand GlcA-AnMan3S6S. We observed that the level of the di-saccharide GlcA-AnMan3S6S in the Sulf-2-treated samplewas decreased to 38.3% from 51.4%, suggesting that Sulf-2removed a 6-O-sulfo group from the 3-O-sulfated glucosa-mine residue. However, the removal of this 6-O-sulfo groupfrom the 3-O-sulfated glucosamine residue has no impact on

the binding affinity to AT. This observation is consistent withthe fact that AT-binding HS is composed of a disaccharideunit of -GlcA-GlcNS3S- (without a 6-O-sulfo group in the di-saccharide unit) and -GlcA-GlcNS3S6S- (with a 6-O-sulfogroup in the disaccharide unit) (Colliec-Jouault et al. 1994).

Discussion

The Sulfs are interesting enzymes due to their involvement indevelopment and cancer and their potential role in preparingHS structures. Although some initial studies of Sulf-2 activityhave been carried out, the full substrate specificity of Sulf-2had not been elucidated. To address this issue, we synthesized

Table I. Substrate specificity of Sulf-2

Constructname

Repeating disaccharide structure [35S]sulfation level(nmol/μg polysaccharide)

Control [35S]sulfate release (%)

Sulf-2-treated [35S]sulfate release (%)

1 [-IdoA2S-[6-O-35S]GlcNS6S-]n 1.7 2.7 ± 1.0 97.5 ± 0.82 [-IdoA-[6-O-35S]GlcNS6S-]n 1.9 8.3 ± 0.1 24.3 ± 2.73 [-GlcA-[6-O-35S]GlcNS6S-]n 1.4 6.6 ± 1.6 43.1 ± 2.44 [-GlcA-[6-O-35S]GlcNAc6S-]n 0.1 14.2 ± 3.5 28.1 ± 1.35 [-[2-O-35S]IdoA2S-GlcNS-]n 1.1 10.4 ± 2.6 12.6 ± 1.36 [-[-2-O-35S]GlcA2S-GlcNS-]n 1.3 6.7 ± 1.9 13.7 ± 3.07 3-O-[35S]HS (consisting of -GlcA-[3-O-35S]GlcNS3S ± 6S-

domain)1.1 13.6 12.2

8 3-O-[35S]HS (consisting of -IdoA2S-[3-O-35S]GlcNS3S ± 6S-and -GlcA-[3-O-35S]GlcNS3S ± 6S- domains)

0.8 10.1 13.9

9 N-[35S]sulfated HS 1.6 11.7 13.2

Synthetic substrates were incubated with or without the Sulf-2 enzyme and isolated from released sulfate groups by QuickSpin column. The remaining [35S]sulfate on the polysaccharide was quantified with a scintillation counter.

Fig. 4. Effect of Sulf-2 treatment on PF4 and AT binding. (A) [35S]HS was incubated with 0.1 mg/mL AT, and the complex of AT and HS was captured usingConA-Sepharose beads. The AT-binding [35S]HS was prepared by incubating HS from bovine kidney with the 3OST-1 enzyme and [35S]PAPS. (B) PF4-bindingassay comparing Sulf-2-treated and untreated radiolabeled HS. [35S]HS obtained from CHO cells was incubated with varying amounts of PF4, and the solutionswere directly applied to a nitrocellulose membrane. The membrane was then washed and subjected to radioactivity analysis using a scintillation counter. Filledcircles, untreated HS; open circles, Sulf-2-treated HS. (C) Binding affinities (Kd values) of Sulf-2-treated and untreated HS to PF4 and AT as determined byaffinity co-electrophoresis.

EH Pempe et al.

1358

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

nine HS polysaccharides with unique sulfation patterns and35S-labeled groups at a specific position. Using these polysac-charides, we were able to determine which constructs acted assubstrates for Sulf-2 based on their decrease in radioactivityafter Sulf-2 treatment. We found that, as previously under-stood, trisulfated IdoA2S-[6-O-35S]GlcNS6S regions servedas an excellent substrate for Sulf-2, showing �90% removalof the radiolabeled group. In addition, this study found for thefirst time that disulfated UA-[6-O-35S]GlcNS6S units experi-enced the hydrolysis of 6-O-sulfo groups. Sulfation at posi-tions other than the 6-O position was unable to be removedby the enzyme. These results were determined by spincolumn assay and by HPLC analysis.We also examined how Sulf-2 treatment affected binding

to AT and PF4. A major application of synthetic HS is thatit can be used to make heparin-like drugs. For heparin to

have anticoagulant activity, it is essential that it interact withAT, which occurs through a specific AT-binding pentasacchar-ide: GlcNAc/NS6S-GlcA-GlcNS3S6S-IdoA2S-GlcNS6S. Itappears from our results that Sulf-2 removes 6-O-sulfo groupsthat are not within the critical AT-binding region, allowing HSto maintain its interaction with AT. It should be noted that theimportance of 6-O-sulfation for AT-binding is well documen-ted. Atha et al. (1985), using a series of synthetic pentasac-charides, demonstrated that a particular 6-O-sulfo group onthe non-reducing end side of the 3-O-sulfated glucosamineresidue contributed the AT binding energy nearly equal to thatof the 3-O-sulfo group. At the polysaccharide level,6-O-sulfation is also important for AT-binding (Zhang et al.2001; Chen et al. 2007). The unique substrate specificity ofSulf-2 allows the removal of 6-O-sulfo groups that are non-essential for AT binding.Reducing dangerous side effects is a major goal for new

heparin therapies. Currently, heparin-induced thrombocyto-penia is one of the most clinically important drug complica-tions, affecting �3% of patients receiving unfractionatedheparin (Arepally and Ortel 2010). Heparin–PF4 complexesare recognized by IgG molecules, which initiate an immunereaction that ultimately leads to platelet degradation and un-controlled bleeding. For this reason, we were interested to de-termine whether Sulf-2 treatment could affect the binding ofHS to PF4. Based on a ligand binding assay and affinityco-electrophoresis, we found that Sulf-2 can reduce binding toPF4 on the order of 10–20-fold. Heparin exhibits its anti-coagulant activity by binding to AT, and the complex inhibitsthe activity of factors Xa and IIa. We anticipate that Sulf-2treatment should not affect the anti-Xa activity becauseanti-Xa activity is directly correlated with AT-binding affinity.However, we do not know whether Sulf-2 treatment affects itsanti-IIa activity as the structural requirements for anti-IIa ac-tivity are unknown at the present time.Given that binding to AT is maintained with Sulf-2 treat-

ment, the decrease in PF4 binding could have exciting impli-cations in the preparation of a new generation of heparindrugs that have reduced side effects. In addition, an extendedunderstanding of the substrate specificity of Sulf-2 could bebeneficial for the treatment of cancers in which this enzyme isup-regulated.

Materials and methodsPreparation of 35S-labeled polysaccharidesRadiolabeled polysaccharide substrates 1–6 were preparedusing �1 µg heparosan, a capsular polysaccharide isolatedfrom the Escherichia coli K5 strain, as a starting material(Vann et al. 1981). Substrates 7–9 were prepared from bovinekidney HS. The starting materials were modified with 5–10µg of C5-epi, NST, 6OST-1/3, 2OST, 3OST-1 and 3OST-5 (asindicated in Figure 2) in sequential 200-µL reactions contain-ing approximately 1 × 106 cpm [35S]PAPS and 10 nmol un-labeled PAPS in 50 mM 2-(N-morpholino)ethanesulfonic acid(MES) (pH 7) and 0.5% Triton X-100. The enzymatic reac-tions were incubated at 37°C for 60 min, heat inactivated andpurified using a diethylaminoethyl (DEAE) column (Vannet al. 1981).

Fig. 5. Reverse-phase ion-pairing HPLC chromatograms of the disaccharideanalysis of 3-O-sulfated HS with and without Sulf-2 treatment. The 3-O-[35S]sulfated HS was prepared by incubating HS with purified 3-OST-1 enzymeand [35S]PAPS. A portion of the 3-O-[35S]sulfated HS was treated withSulf-2. Both Sulf-2-treated and untreated HS were subjected to deacetylationwith hydrazine followed by nitrous acid degradation at pH 4.5 and 1.5. Theresultant disaccharides were analyzed by HPLC. (A) The chromatogram ofthe analysis of untreated 3-O-[35S]sulfated HS. (B) The chromatogram of theanalysis of Sulf-2-treated 3-O-[35S]sulfated HS. The elution positions wereidentified by eluting with standards: 1, GlcA-AnMan3S; 2,IdoA2S-AnMan3S; 3, GlcA-AnMan3S6S. Asterisk indicates the unidentifiedcomponents.

Characterization of heparan sulfate 6-O-endosulfatase

1359

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

Expression of recombinant Sulf-2 in CHO cellsA plasmid consisting of a full-length cDNA encoding humanSulf-2 was purchased from Open Biosystem (Clone ID#7969293). The gene was cloned into a pcDNA3.1A mamma-lian expression vector (from Invitrogen, Carlsbad, CA) usingXhoI/HindIII sites to obtain the Sulf-2 expression plasmiddesignated as pcDNA3.1A-myc/His-HSulf-2. Wild-type CHOcells were transiently transfected with the expression plasmidpcDNA3.1A-myc/His-HSulf-2 or an empty pcDNA3.1A vectoraccording to a standard protocol. Briefly, CHO cells wereseeded in 6-well plates in F-12 media supplemented with 10%fetal bovine serum (FBS) and were maintained in a 5% CO2 hu-midified incubator at 37°C. When the cells reached 90–95%confluence, they were transfected using Lipofectamine 2000reagent (Invitrogen) and Opti-MEM Reduced Serum Media(Invitrogen) according to the manufacturer’s protocol. After4–6 h, the medium was replaced with F-12 media containing10% FBS. The conditioned medium was collected after 48–72h of incubation and centrifuged at 3300 × g for 15 min toremove cellular debris.

Western blottingSulf-2- and EV-transfected cells were collected using trypsinand washed with phosphate-buffered saline. Cells were lysedwith 1.5 M sucrose, 1% Triton X-100 and 1 mM phenyl-methylsulfonyl fluoride. Sulf-2 and EV conditioned medium(CM) and lysates were separated by 12% sodium dodecylsulfate polyacrylamide gel electrophoresis and visualized withan anti-myc primary antibody (Cell Signaling Technology) andthe SuperSignal detection system (Thermo Scientific), revealinga band of �132 kDa in the Sulf-2 samples (Morimoto-Tomitaet al. 2002).

Sulf-2 enzymatic assayA 100-µL reaction, containing 35S-labeled substrate, 50 mMMES (pH 6.5), 10 mM CaCl2, 0.1% Triton X-100 and 50 µLof Sulf-2 enzyme, was incubated overnight at 37°C.Sulf-2-treated and untreated polysaccharide substrates werepurified using Quick Spin Columns for radiolabeled DNApurification (Roche, Indianapolis, IN). Columns were centri-fuged at 1100 × g for 2 min to remove the column buffer, thenplaced into a fresh collection tube and loaded with 50 µL ofreaction mixture. After centrifugation for 4 min at 1100 × g,the eluate was collected and the amount of 35S-labeled poly-saccharide was quantified using a scintillation counter.

Analysis of synthetic polysaccharidesTo determine the structural compositions of the radiolabeledsubstrates, disaccharide analyses were performed usingheparin lyase digestion. Polysaccharide substrates were incu-bated overnight with a mixture of heparin lyases I, II and III(0.1 mg/mL each) in 200 µL of 50 mM sodium phosphate(pH 7) at 37°C. The reaction was terminated by boiling at100°C for 5 min and was loaded onto a Bio-Gel P-2 column(Bio-Rad, Hercules, CA) to isolate disaccharides. These disac-charides were analyzed using reverse-phase ion-pairing HPLCas described previously (Duncan et al. 2006).

Preparation of 35S HS from CHO cellsAT-75 flask of wild-type CHO cells was grown to confluencein F-12 media supplemented with 10% FBS. The cells werethen incubated with 1 mL of 1.0 mCi/mL sodium [35S]sulfate(Perkin Elmer, Boston, MA) for 6 h at 37°C with 5% CO2.Two hundred microliters of a pronase stock solution contain-ing 1 mg/mL pronase (Sigma, St. Louis, MO), 240 mMNaAcO (pH 6.5) and 1.92 M NaCl was added to the cells,and the flask was incubated overnight at 37°C. The pronase-digested sample was centrifuged at 10,000 rpm for 15 minand filtered using a 0.45-µm filter, then purified using aDEAE-Sepharcel column (Sigma), which was equilibratedusing a buffer containing 20 mM NaAcO (pH 5) and150 mM NaCl. The [35S]HS was eluted from the column with1 M NaCl in 20 mM NaAcO and was dialyzed overnightagainst 50 mM ammonium bicarbonate using a 14,000MWCO membrane, then dried. The sample was reconstitutedin 1 mL of water, and 10 µL of a solution containing 10 NNaOH and 0.89 M sodium borohydride was added to breakthe linkage between the core protein and HS. It was incubatedat 46°C for 16 h. The sample was also treated with 20 U/mLchondroitinase ABC to remove chondroitin sulfate before use.

Preparation of 3OST-1-treated [35S]HSTo 3-O-sulfate [35S]HS using unlabeled PAPS for affinityco-electrophoresis, �100,000 cpm of [35S]HS was added to a50-µL reaction containing 0.05 ng of 3OST-1, 10 mM MnCl2,5 mM MgCl2, 75 µg/mL protamine chloride, 0.4 mg/mLchondroitin sulfate, 0.12 mg/mL bovine serum albumin, 1%Triton X-100 and 0.5 mM PAPS. The reaction was incubatedat 37°C for 20 min, inactivated at 80°C for 10 min, thendiluted with 60 µL of H2O and centrifuged at 2000 × g for 10min. Substrates used for AT-binding studies were then isolatedusing a ConA-Sepharose column (Sigma).

Preparation of mPF4PF4 and tobacco etch virus protease expression plasmid. Thefull-length murine PF4 cDNA was obtained from OpenBiosystems (Clone ID: 582960). The heparin-binding domain ofPF4 (Ala33-Ser105) was cloned into a PET32 vector (Novagen)using NcoI and HindIII sites to give a plasmid named asPF4-PET32/tobacco etch virus protease (TEV). In this plasmid,a TEV cleavage hexapeptide sequence, EQLYFQG, wasconstructed between thioredoxin and PF4. The design permittedcleavage of the thioredoxin–PF4 fusion protein to release PF4.The resultant recombinant PF4 protein has five extra amino acidresidues (GSRHG) at the N terminus. A bacterial strain, BL21(DE3)-RIL/pRK793, expressing TEV protease was a generousgift from Dr Lars Pedersen (National Institute of EnvironmentalHealth Sciences).The PF4-PET32/TEV plasmid was introduced into BL21

cells, and the cells were grown in LB medium containing50 µg/mL of carbenicillin and induced with isopropylβ-D-1-thiogalactopyranoside. The cells were pelleted, lysed in25 mM Tris, 500 mM NaCl and 300 mM imidazole (pH 7.5)and purified using a nickel column. The fractions containingprotein were collected and incubated overnight with TEV pro-tease (1:25 w/w ratio of TEV:PF4). After TEV cleavage, thePF4 was dialyzed against 20 mM Tris and 250 mM NaCl

EH Pempe et al.

1360

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

(pH 8) and was purified with a heparin column, which waseluted with 250–2000 mM NaCl gradient in 20 mM Tris (pH8). The purity of PF4 was greater than 90% as determined byNuPAGE 12% Bis–Tris Gel (Invitrogen). PF4 was eluted as aprotein with an MW of 30 kDa on a Sephadex G75 column,suggesting that PF4 is present in a tetrameric form that wasconsistent with a previous report (Mikhailov et al. 1999). Theconcentration of PF4 was determined by amino acid compos-itional analysis.

Dot blot assay for PF4 bindingA dot blot assay was used to determine the binding affinitiesof Sulf-2-treated and untreated HS to PF4. A 300-µL reactioncontaining 130 mM NaCl, 50 mM Tris (pH 7) buffer wasincubated with [35S]HS from CHO cells and 0–608 nM PF4for 30 min at 37°C before being blotted on a nitrocellulosemembrane (GE Healthcare). The membrane was washed usingthe sample buffer, and the spot containing bound HS was cutout to quantify using a scintillation counter.

Affinity co-electrophoresisAffinity co-electrophoresis gels were prepared using a stand-ard protocol (Lee and Lander 1991) with lanes containing0–1.41 µM PF4 or 0–3.2 µM AT and �50,000 cpm3OST-1-labeled [35S]HS per gel. Gels were run for 2 (AT) or2.5 (PF4) h and the bands were imaged using a Storm 860phosphorimager (Molecular Dynamics) and ImageQuant TLv2005 software (GE Healthcare Life Sciences). Kd valueswere determined by plotting R/[protein] vs R, where R = (Mo−M)/Mo; Mo is the mobility of free HS, and M is the mobil-ity of HS at each protein concentration. The slope is equal to−1/Kd.

AT-binding assayTo quantify the binding of Sulf-2-treated and untreated HS toAT, 3OST-1-labeled substrates were incubated with 0.1 mg/mL AT, and the complex of AT and HS was captured usingConA-Sepharose beads (Sigma). For this experiment, theAT-binding [35S]HS was prepared by incubating HS from thebovine kidney with the 3OST-1 enzyme and [35S]PAPS.

Disaccharide analysis of 3-O-[35S]sulfated HS. 3-O-[35S]sulfated HS, with or without Sulf-2 treatment, were subjectedto the nitrous acid degradation to yield 35S-labeleddisaccharides. The HS was deacetylated in a hydrazinesolution contain 12.5 mg/mL hydrazine sulfate followed bythe degradation with nitrous acid at pH 4.5 and 1.5 as well assodium borohydride reductions described previously (Shivelyand Conrad 1976). The resultant disaccharides were desaltedon a BioGel P-2 column and resolved on a C18-column(Vydac) under the reverse phase ion-pairing HPLC conditions(Liu et al. 1999).

Funding

This work was supported by the National Institutes of Health(R01HL094463 and R01AI050050 to J.L., F31AG040927-01to E.H.P.).

Acknowledgements

We also thank Dr Lars Pedersen for providing us the plasmidexpressing recombinant TEV protease.

Conflict of interest

None declared.

Abbreviations

AT, antithrombin; C5-epi, C5-epimerase; ConA, concanavalinA; CHO, chinese hamster ovary; CM, conditioned medium;DEAE, diethylaminoethyl; FBS, fetal bovine serum; GlcNS6S,N,6-O-sulfo glucosamine; HPLC, high-performance liquidchromatography; HS, heparan sulfate; MES, 2-(N-morpholino)ethanesulfonic acid; NST, N-sulfotransferase; PAPS, 3′-phos-phoadenosine-5′-phosphosulfate; PF4, platelet factor 4; 2OST,2-O-sulfotransferase; 3OST-1, 3-O-sulfotransferase isoform 1;3OST-5, 3-O-sulfotransferase isoform 5; 6OST-1/3,6-O-sulfotransferase isoform 1/isoform 3; SDS-PAGE, sodiumdodecyl sulfate polyacrylamide gel electrophoresis; Sulf,6-O-endosulfatase.

References

Ai X, Do AT, Lozynska O, Kusche-Gullberg M, Lindahl U, Emerson CP, Jr.2003. QSulf1 remodels the 6-O sulfation states of cell surface heparansulfate proteoglycans to promote Wnt signaling. J Cell Biol. 162:341–351.

Ai X, Do AT, Kusche-Gullberg M, Lindahl U, Lu K, Emerson CP, Jr. 2006.Substrate specificity and domain functions of extracellular heparan sulfate6-O-endosulfatases, QSulf1 and QSulf2. J Biol Chem. 281:4969–4976.

Ai X, Kitazawa T, Do AT, Kusche-Gullberg M, Labosky PA, Emerson CP, Jr.2007. SULF1 and SULF2 regulate heparan sulfate-mediated GDNF signal-ing for esophageal innervation. Development. 134:3327–3338.

Arepally GM, Ortel TL. 2010. Heparin-induced thrombocytopenia. Annu RevMed. 61:77–90.

Atha DH, Lormeau JC, Petitou M, Rosenberg RD, Choay J. 1985.Contribution of monosaccharide residues in heparin binding to antithrom-bin III. Biochemistry. 24:6723–6729.

Chen J, Jones CL, Liu J. 2007. Using an enzymatic combinatorial approachto identify anticoagulant heparan sulfate structures. Chem Biol.14:986–993.

Colliec-Jouault S, Shworak NW, Liu J, De Agostini AI, Rosenberg RD. 1994.Characterization of a cell mutant specifically defective in the synthesis ofanticoagulantly active heparan sulfate. J Biol Chem. 271:24953–24958.

Dai Y, Yang Y, MacLeod V, Yue X, Rapraeger AC, Shriver Z, VenkataramanG, Sasisekharan R, Sanderson RD. 2005. HSulf-1 and HSulf-2 are potentinhibitors of myeloma tumor growth in vivo. J Biol Chem.280:40066–40073.

Danenberg HD, Szalai AJ, Swaminathan RV, Peng L, Chen Z, Seifert P, FayWP, Simon DI, Edelman ER. 2003. Increased thrombosis after arterialinjury in human C-reactive protein-transgenic mice. Circulation.108:512–515.

Dhoot GK, Gustafsson MK, Ai X, Sun W, Standiford DM, Emerson CP, Jr.2001. Regulation of Wnt signaling and embryo patterning by an extracellu-lar sulfatase. Science. 293:1663–1666.

Duncan MB, Liu M, Fox C, Liu J. 2006. Characterization of theN-deacetylase domain from the heparan sulfate N-deacetylase/N-sulfotransferase 2. Biochem Biophys Res Commun. 339:1232–1237.

Esko JD, Selleck SB. 2002. Order out of chaos: Assembly of ligand bindingsites in heparan sulfate. Annu Rev Biochem. 71:435–471.

Feyerabend TB, Li JP, Lindahl U, Rodewald HR. 2006. Heparan sulfateC5-epimerase is essential for heparin biosynthesis in mast cells. Nat ChemBiol. 2:195–196.

Guerrini M, Shriver Z, Bisio A, Naggi A, Casu B, Sasisekharan R, Torri G.2009. The tainted heparin story: An update. Thromb Haemost.102:907–911.

Characterization of heparan sulfate 6-O-endosulfatase

1361

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022

Holst CR, Bou-Reslan H, Gore BB, Wong K, Grant D, Chalasani S, CaranoRA, Frantz GD, Tessier-Lavigne M, Bolon B. 2007. Secreted sulfatasesSulf1 and Sulf2 have overlapping yet essential roles in mouse neonatal sur-vival. PLoS One. 2:e575.

Lai JP, Chien JR, Moser DR, Staub JK, Aderca I, Montoya DP, Matthews TA,Nagorney DM, Cunningham JM, Smith DI. 2004. hSulf1 Sulfatase pro-motes apoptosis of hepatocellular cancer cells by decreasing heparin-binding growth factor signaling. Gastroenterology. 126:231–248.

Lai JP, Sandhu DS, Yu C, Han T, Moser CD, Jackson KK, Guerrero RB,Aderca I, Isomoto H, Garrity-Park MM. 2008. Sulfatase 2 up-regulates gly-pican 3, promotes fibroblast growth factor signaling, and decreases survivalin hepatocellular carcinoma. Hepatology. 47:1211–1222.

Lamanna WC, Kalus I, Padva M, Baldwin RJ, Merry CL, Dierks T. 2007.The heparanome-the enigma of encoding and decoding heparan sulfate. JBiotechnol. 129:290–307.

Lee MK, Lander AD. 1991. Analysis of affinity and structural selectivityin the binding of proteins to glycosaminoglycans: Development of asensitive electrophoretic approach. Proc Natl Acad Sci USA. 88:2768–2772.

Lemjabbar-Alaoui H, van Zante A, Singer MS, Xue Q, Wang YQ, Tsay D,He B, Jablons DM, Rosen SD. 2010. Sulf-2, a heparan sulfate endosulfa-tase, promotes human lung carcinogenesis. Oncogene. 29:635–646.

Lind T, Tufaro F, McCormick C, Lindahl U, Lidholt K. 1998. The putativetumor suppressors EXT1 and EXT2 are glycosyltransferases required forthe biosynthesis of heparan sulfate. J Biol Chem. 273:26265–26268.

Lindahl U, Kusche-Gullberg M, Kjellen L. 1998. Regulated diversity ofheparan sulfate. J Biol Chem. 273:24979–24982.

Liu J, Shworak NW, Fritze LMS, Edelberg JM, Rosenberg RD. 1996.Purification of heparan sulfate D-glucosaminyl 3-O-sulfotransferase. J BiolChem. 271:27072–27082.

Liu J, Shworak NW, Sinaÿ P, Schwartz JJ, Zhang L, Fritze LMS, RosenbergRD. 1999. Expression of heparan sulfate D-glucosaminyl3-O-sulfotransferase isoforms reveals novel substrate specificities. J BiolChem. 274:5185–5192.

Lum DH, Tan J, Rosen SD, Werb Z. 2007. Gene trap disruption of the mouseheparan sulfate 6-O-endosulfatase gene, Sulf2. Mol Cell Biol. 27:678–688.

Mikhailov D, Young HC, Linhardt RJ, Mayo KH. 1999. Heparin dodecasac-charide binding to platelet factor-4 and growth-related protein-alpha. J BiolChem. 274:25317–25329.

Morimoto-Tomita M, Uchimura K, Werb Z, Hemmerich S, Rosen SD. 2002.Cloning and characterization of two extracellular heparin-degrading endo-sulfatases in mice and humans. J Biol Chem. 277:49175–49185.

Nagamine S, Koike S, Keino-Masu K, Masu M. 2005. Expression of aheparan sulfate remodeling enzyme, heparan sulfate 6-O-endosulfatase sul-fatase FP2, in the rat nervous system. Brain Res Dev Brain Res.159:135–143.

Nawroth R, van Zante A, Cervantes S, McManus M, Hebrok M, Rosen SD.2007. Extracellular sulfatases, elements of the Wnt signaling pathway,

positively regulate growth and tumorigenicity of human pancreatic cancercells. PLoS One. 2:e392.

Ohto T, Uchida H, Yamazaki H, Keino-Masu K, Matsui A, Masu M. 2002.Identification of a novel nonlysosomal sulphatase expressed in the floorplate, choroid plexus and cartilage. Genes Cells. 7:173–185.

Peterson S, Frick A, Liu J. 2009. Designing of biologically active heparan sulfateand heparin using an enzyme-based approach. Nat Prod Rep. 26:61–627.

Peterson SB, Liu J. 2010. Unraveling the specificity of heparanase utilizingsynthetic substrates. J Biol Chem. 285:14504–14513.

Phillips JJ, Huillard E, Robinson AE, Ward A, Lum DH, Polley MY, RosenSD, Rowitch DH, Werb Z. 2012. Heparan sulfate sulfatase SULF2 regu-lates PDGFRalpha signaling and growth in human and mouse malignantglioma. J Clin Invest. 122:911–922.

Ratzka A, Kalus I, Moser M, Dierks T, Mundlos S, Vortkamp A. 2008.Redundant function of the heparan sulfate 6-O-endosulfatases Sulf1 andSulf2 during skeletal development. Dev Dyn. 237:339–353.

Rosen SD, Lemjabbar-Alaoui H. 2010. Sulf-2: An extracellular modulator ofcell signaling and a cancer target candidate. Expert Opin Ther Targets.14:935–949.

Rosenberg RD, Shworak NW, Liu J, Schwartz JJ, Zhang L. 1997. Heparansulfate proteoglycans of the cardiovascular system. Specific structuresemerge but how is synthesis regulated? J Clin Invest. 100:S67–S75.

Saad OM, Leary JA. 2003. Compositional analysis and quantification fheparin and heparan sulfate by electrospray ionization ion trap mass spec-trometry. Anal Chem. 75:2985–2995.

Shively JE, Conrad HE. 1976. Formation of anhydrosugars in the chemicaldepolymerization of heparin. Biochemistry. 15:3932–3942.

Tang R, Rosen SD. 2009. Functional consequences of the subdomain organ-ization of the sulfs. J Biol Chem. 284:21505–21514.

Taylor KR, Gallo RL. 2006. Glycosaminoglycans and their proteoglycans:Host-associated molecular patterns for initiation and modulation of inflam-mation. FASB J. 20:9–22.

Turnbull J, Powell A, Guimond S. 2001. Heparan sulfate: Decoding adynamic multifunctional cell regulator. Trends Cell Biol. 11:75–82.

Vann WF, Schmidt MA, Jann B, Jann K. 1981. The structure of the capsularpolysaccharide (K5 antigen) of urinary-tract-infective Escherichia coli 010:K5:H4. Eur J Biochem. 116:359–364.

Viviano BL, Paine-Saunders S, Gasiunas N, Gallagher J, Saunders S. 2004.Domain-specific modification of heparan sulfate by Qsulf1 modulates thebinding of the bone morphogenetic protein antagonist Noggin. J BiolChem. 279:5604–5611.

Yue X, Li X, Nguyen HT, Chin DR, Sullivan DE, Lasky JA. 2008.Transforming growth factor-beta1 induces heparan sulfate6-O-endosulfatase 1 expression in vitro and in vivo. J Biol Chem. 283:20397–20407.

Zhang L, Beeler DL, Lawrence R, Lech M, Liu J, Davis JC, Shriver Z,Sasisekharan R, Rosenberg RD. 2001. 6-O-sulfotransferase-1 represents acritical enzyme in the anticoagulant heparan sulfate biosynthetic pathway. JBiol Chem. 276:42311–42321.

EH Pempe et al.

1362

Dow

nloaded from https://academ

ic.oup.com/glycob/article/22/10/1353/1987884 by guest on 16 April 2022