Regulation of biosynthesis of the basic fibroblast growth factor binding ...

Vol. 268, No. 27, Issue of September 25, pp. 20091-20095, 1993 Printed in U.S.A.

THEJOURNAL OF BIOLXICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc

Regulation of Biosynthesis of the Basic Fibroblast Growth Factor Binding Domains of Heparan Sulfate by Heparan Sulfate-N-Deacetylase/N-Sulfotransferase Expression* (Received for publication, February 26, 1993, and in revised form, April 23, 1993)

Masayuki Ishihara8, Yuchuan Guo, Zheng Wei, Zicheng Yang, and Stuart J. Swiedler From Glycomed Inc., Alameda, California 94501

Ariel Orellana and Carlos B. Hirschberg From the Department of Biochemistry a n d Molecular Biology, University of Massachusetts Medical Center, Worcester, Massachusetts 01655

Heparan sulfate-N-deacetylasefN-sulfotransferase HS possesses a complex polymeric structure in which the catalyzes boththe N-deacetylationand N-sulfationreac- majority of the N-sulfated glucosamine residues and various tions that initiate the modification of the oligosaccha- types of 0-sulfated residues are clustered in a series of short ride backbone of heparan sulfate (HS). The glycosami- domains separatedby relatively long sulfate-poor oligosacchaof these highly N-sulfated noglycan polymer appears to modulate the activity of ride sequences(11, 12). The presence growth factorsby mediating theirinitial binding. To un- domains in HS is the major factor accounting for its participaderstand how the biosynthesisof these binding sites is tion in a diverse range of biological activities; for instance, the of these regulated, a rat liver-derived cDNA encoding the above highly sulfated domains participate in the binding activities was overexpressed in aCOS cell mutant (CM- polymers to bFGF (13-15). In addition, N-sulfated fragments derived from various animal tissues and cells that bind with 15) that has reduced levels of the enzyme and binds poorly to immobilizedbasicfibroblastgrowthfactor highest affinity to bFGF are enriched in 2-0-sulfated iduronate (bFGF). Thisresulted in increasedsynthesis of sulfated residues (13-15). blocks of decasaccharide size or longer. These blocks Heparan sulfate-N-deacetylase/N-sulfotransferase (HSexhibited high affinity binding to bFGF and contained a NdAc/NST) catalyzes boththe N-deacetylation and N-sulfation high content of 2-0-sulfated iduronate andat least five of glucosamine residues inHS (16). These reactions areobligAn increase inthe atory steps for further modifications in the biosynthesisof H S consecutiveN-sulfated disaccharides. synthesis of these high affinity blocks was not seen in in that C5-epimerizationof D-glucuronic to L-iduronic acid a n d COS cells even though they transfectedwild-type various 0-sulfation reactions occur only within or adjacent to showed a 4-fold increase of both enzymeactivities, sug- the N-sulfated residues (17). We have previously reported that gesting that once sufficient levels of highly sulfated a COS cell-derived cell line, CM-15, is partially deficient in blocks of saccharides with high affinity for bFGF are N-sulfotransferase activity and exhibits reduced binding to imattained, no further synthesis of these domains occurs.

mobilized bFGF (18). This cell line is also impaired in the synthesis of large fully N-sulfated, highly 0-sulfated domains (15).Thus, the in HS that display high affinity binding to bFGF Heparin and heparan sulfate (HS)’ have been implicated in level of HS-NdAciNST expression is important in regulating the regulationof cell growth (1-4), although the detailed mechthe formation of these domains in the polymer. In the present anisms underlying their actions are not understood. There is an structures of the highly sulfatedblocks contained within study, emerging consensusthat these polymers may be exerting their the HS producedby CM-15 cells a n d wild-type COS cellsovereffects by modulating the activity of various growth factors; for expressing a recombinantly-derived rat liverHS-NdAc/NST example, both basic fibroblast growth factor (bFGF) and acidic were analyzed biochemically and for their bFGF binding afingrowth factor bind to heparin andHS, a n d the resulting com- ity. The results obtained suggest that there is an upper limit to plexes are known to be resistant to both acid and heat denathe formation of the high affinity domains in HS in COS cells turation and proteolytic degradation (5-7). It has been sugthat is unaltered when the HS-NdAc/NST is overexpressed in gested that heparin and HS induce comformational changes in these cells. bFGF that are prerequisites for binding of the growth factor to EXPERIMENTALPROCEDURES a high affinity transmembrane receptor (8, 9, 10). Zkansfection of COS Cells and Isolation of HS Proteoglycan-Wildtype COS and CM-15 cells (5 x lo5 cells/lO-cm dish) were transfected * This work wassupported by Glycomed Inc., a Fogarty fellowship (to using the DEAE-dextran transfection method (19, 20) with 5 pg of A. 0.1,and National Institutes of Health Grant GM 34396 (to C. B. H.). pCMVSTDNA containing the cDNA encoding the full-length transThe costs of publication of this article were defrayed in part by the membrane HS-NdAdNST(21)and were designated as wild-type (+) and payment of page charges. This article must therefore be hereby marked CM-15(+I cells, respectively. Wild-type and CM-15 cells were alsotrans“advertisement”in accordancewith18U.S.C.Section 1734 solelyto fected withpCMV5 vector DNA alone and were designated as wild-type indicate this fact. (control) and CM-15 (control) cells, respectively. After 48 h of incuba$ To whom correspondence should be addressed: Glycomed Inc., 860 tion, the cellsweremetabolicallylabeledfor 24 h with 50 pCi/ml Atlantic Ave., Alameda, CA 94501. The abbreviations used are: HS, heparan sulfate; bFGF, basic fibro- HZ3%O, or/and 50 pCi/ml[6-3H]glucosamine in Dulbecco’smodified blast growth factor; HS-NdAc/NST, heparan sulfate-N-deacetylaseNV- Eagle’s medium with low glucose (1 gfliter) and lacking sulfate (Life sulfotransferase; IdoA, L-iduronic acid; GlcA, D-glucuronic acid; AManR, Technologies, Inc., formula 89-0053AK) that was supplemented with anhydro-D-mannitol;IdoA(Z-S), L-iduronic acid(2-0-S0,); GlcA(G-S),D- dialyzed fetal bovine serum (lo%,v/v).The labeled cultures were chilled glucuronic acid (6-O-SO3);AManR(6-S),anhydro-~-mannitol(6-O-S0~); on ice, the culture medium was removed, and the cells wererinsed twice PBS, phosphate-buffered saline. with cold phosphate-buffered saline (PBS). The washes were combined

20091

20092

bFGF-binding Domains of HS and HS-NdAcINST

TABLEI Overexpression of HS-NdAcINST inCM-15 and wild-type COS cells CM-15 and wild-type cells were transfected with pCMVST DNA containing thecDNA encoding the full-length transmembraneHS-NdAc/NST (designated as“(+)”) or pCMV5 vector DNA alone (designatedas “(control)”). Extracts were prepared and N-deacetylase, N-sulfotransferase, and 0-sulfotransferase activities in eachcell extract were measureda s described under “Experimental Procedures.” All of the determinations are the average values obtainedfrom a minimum of duplicate cultures. N-Deacetylase activity P H released)

CM-15 (+) cell extract CM-15 (control) cell extract Wild-type (+) cell extract Wild-type (control) cell extract

N-Sulfotransferase activity P5S incorporated)

0-Sulfotransferase activity (”5s incorporated)

cpm

cpm

cpm

29,500 t 1200 2800 * 550 42,400 t 1500 10,600 t 900

49,000 t 1700 6900 t 1000 58,000 t 2100 15,600 -c 1400

6900 -c 300 6700 t 400 7200 -c 400 7150k 510

with the original culture medium to obtain the culture medium pool. cells, showed 3-4 times less N-deacetylase and 2-3 times less The cells were trypsinized from the culture dish, and the dishes were N-sulfotransferase activities as compared to that from wildwashed twice with cold PBS. Thecell suspension and PBS washes weretype (control) cells (Table I). Extractsof mutant and wild-type combined and centrifuged to pellet the cells. The supernatant fluids were then added to the culture medium pool. The labeled glycosami- cells transfected withvector encoding the full-length HS-NdAd noglycan in the culture medium pool was digestedby treatment with 0.1 NST, CM-15 (+) and wild-type (+) cells, showed a 7-8-fold and unit/ml chondroitinaseABC and chondroitinase ACII (Seikagaku Amer- a 4-fold increase in both activities, respectively. The extent of ica) overnightat 37 “C. The HS proteoglycan resistant to this treatment stimulation of the enzyme activities in overexpressing cells was was then purifiedby chromatography on a column of DEAE-cellulose, even greater than thatdescribed above, since it was estimated and 35S0,-labeled or/and 3H-labeled glucosamine incorporation intoHS that only 20-50% of the transfected cells overexpressed HSwas determined by scintillation counting. NdAc/NST as a result of DEAE-mediated transfection utilized N-Deacetylase and N-Sulfotransferase Assays-CM-15 and wild-type COS cells were transfected as describedabove and incubated for 72 h. in the present study. No increase in 0-sulfotransferase activiThe cell were rinsed, scraped off the dishes, and collected by centrifu- ties was detected in extracts from CM-15 (+) and wild-type (+) gation. Cell pellets were resuspended in homogenate buffer (250 mM cells. sucrose, 1 pg/ml leupeptin, 0.4 pg/ml pepstatin, 0.2 mM phenylmethylbFGF Affinity Chromatography of Intact HS from CM-15 (+) sulfonyl fluoride, and 50 mM Tris-HC1, pH 7.5), and disrupted by a and Wild-type (+) Cells-We have recently shown that transDounce homogenizer (20 strokes) to prepare each transfected cell exfection of CM-15 cells with the cDNA encoding the HS-NdAd tract. The N-deacetylase activity was measured by determining the NST restored their ability to bind to immobilized bFGF.2 In release of t3Hlacetate from N-PHIacetylated polysaccharide derived addition, [35S]sulfate incorporation into matrix-associated HS from Escherichia coli K5-derived capsular polysaccharide with a specific activity of 400 cpm/ng (dry weight) after 1 h of incubation as was found to be stimulated in CM-15 (+) cells, while it was described(16).N-Sulfotransferaseand0-sulfotransferaseactivities unaffected in wild-type (+) cells. For the purpose of defining the were measured by determining 35S04incorporation into N-desulfated structural basis for these observations, HS was isolated from heparin and completely desulfated, N-resulfated heparin,respectively, (+), wild-type (+I, and the respective control cells metCM-15 from adenosine 3”phosphate 5’-phosph0[“~S]sulfate (PAP35S)with a specific activity of 2.26 Ci/mmol after a 30-min incubation as described abolically labeled with [35S] sulfate and L3H1 glucosamine fol(18). Twenty-five pg of total cell protein from each transfected cell lowed by examination of their chromatographic behavior on a extract was included in both enzyme reactions. bFGF affinity column. As indicated in Table 11, the [35Sl/[3Hl Isolation of Exclusively N-Sulfated Oligosaccharides-The labeled ratio of HS derived from CM-15 (+) cells wasincreased as HS from transfected CM-15 and wild-type cells was N-deacetylated by compared to that ratio determined for HS derived from CM-15 hydrazinolysis and treated with nitrous acid at pH 4.0 (22). The reac(control) cells, indicating thatsulfation was stimulated relative tion mixtures were then reduced with NaBH4 under alkaline condition (221, and the products were separated according to sizeby gel-filtration to the synthesis of the carbohydrate backbone. This ratio fell chromatography on Bio-Gel P-10 (0.8 cm x 100 cm x 2) equilibrated in short of the value achieved in eitherwild-type transfection; this 0.5 M ammonium acetate at a flow rate of 1.5 ml/h. The resultingpools was anticipated since only 20-50% of the transfectedcells overof labeled oligosaccharides were desalted and recovered by lyophiliza- expressed HS-NdAc/NST. For comparison, chondroitin sulfate tion. isolated from all four groups displayed the same [35Sl/[3HlraAffinity Chromatography-The bFGF-coupled Sepharose beads (15, tio, indicating that the changes observed for HS were specific. 23) were packed into either a large column (2 x 6 cm) or a small column The elution patterns of the HS samples from a bFGF affinity (0.8 x 2 cm) and equilibrated in 10 mM Tris, pH 7.3, with 0.2M NaC1. The column are presented inFig. 1.The profiles of labeled HS from labeled oligosaccharides were dissolved in 100 pl of the equilibration buffer and loaded onto the small column (0.8 x 2 cm), which was then wild-type (+) and wild-type (control) cells were similar; more washed with 3 ml of 0.4 M NaCl in the samebuffer. Fractions (0.5 ml) than 60% of the labeled HS required2 M NaCl for elution from were collected and analyzed for radioactivity by liquidscintillation the column. CM-15 (control) cells exhibited a strikingly differcounting. Larger amounts of 35S0,-labeled oligosaccharides prepared ent behavior; only 5% of the radiolabel eluted at above NaCl from CM-15 (+) cells were fractionated on the large bFGF affinity colconcentration. CM-15 (+) cells showed a 6-fold increase in this umn (2 x 6 cm). After loading, the non-bound oligosaccharides were fraction; again, totalphenotypic conversion was notachieved as collected by washing thecolumn with 100 ml of the equilibrationbuffer, and the remaining bound oligosaccharides were eluted with linear gra- a result of the natureof the transienttransfection protocol. The dients of NaCl(0.2-2 M NaCI) in 10m Tris, pH 7.3. Each fractionated results suggest that overexpression of HS-NdAdNST in muoligosaccharide pool was divided into four subpools differing in affinitytant but notwild-type cells leads to significant increased synfor bFGF as shown by the bars in each panelof Fig. 3. These subpools were then dialyzed against water using M, 1000 cut-off membrane thesis of HS with high affinity binding for bFGF. Analysis of Fully N-Sulfated Oligosaccharides of H S from (Spectrum, Houston, TX) and recovered by lyophilization. CM-15 (+) and Wild-type (+) Cells-We characterized the relaRESULTS tive size distribution of fully N-sulfated oligosaccharide blocks

Expression of HS-NdAcINST cDNA in CM-15 and Wild-type Cells-In order t o verify the overexpression of HS-NdAc/NST in transfected CM-15 and wild-type cells, assays for N-deacetylase and N-sulfotransferase were performed.Extracts of CM-15 cells transfected with vector without insert, CM-15 (control)

from wild-type and CM-15 cells transfected with cDNA encoding HS-NdAdNST. Cells were radiolabeled with [35Slsulfate A. Orellana, C. B. Hirschberg, M. Ishihara, Y. Guo, Z. Wei, S. J. Swiedler, and J. D. Esko, unpublished results.

20093

bFGF-binding Domainsof HS and HS-NdAcINST TABLE I1 [35S]S~lfatel[3Hlglucosamine mtio of double-labeled glycosaminoglycan in CM-15 (+) and wild-type (+) cells CM-15 (+) and wild-type(+) cells andtheir control were metabolically labeledwith 35SO:- and[3Hlglucosamine, the labeledHS was then purified, and the amount of incorporated 35Sand 3Hwas determined as described under “Experimental Procedures.” The results representthe mean of duplicate determinations. 35SPH

CM-15 (+) CM-15 (control)

HS

cs

0.87 f 0.07

1.05 0.09 1.02 i 0.10 1.11 i 0.08 1.10 i 0.06

*

0.68 f 0.04

1.01 f 0.05 1.11 f 0.04

Wild-type (+) Wild-type (control)

+

0

1

AI

In

I

0 d

x o 22

I

CM- 1 5

+

1

a u 0

2

1 1 -

V

m

z 0

0

12

6

18

24

rrart.o,> Nurrlbel

FIG.2. Bio-Gel P-10 fractionation of pH 4 HONO cleavage products of N-deacetylated, [SsS,SH]HSfrom CM-16 (+) and wild-type (+) cells. 35S,3H-Labeled N-deacetylated HS from CM-15 (+) cells (panel A ) , wild-type (+) cells (panel C ) ,CM-15 (control)cells (panel B ) , and wild-type (control)cells (panel D ) were treated withHONO at pH 4 and reduced with alkaline NaBH,. The reaction mixtures were chromatographed on a Bio-Gel P-10 column (1 cm x 100 cm x 2) with 0.5 M NHIOCOCH3 at a flow rate of 1.5 mlh. Fractions (1ml) were collected, and and aliquot of each fraction was measured for 35S(0)and 3H (0). The fractions indicated by the bars were separately pooled for further separation on a bFGF affinity column. The numbers above the bars indicate the number of monosaccharide units in each component.

Wild-type

1

Fracllon Number

30

Fraction Number

FIG.1. I LfFinity chromatoeTaDhs of P S . SH1-labeledHS from CM-16 (+) and wild-type c&li. [35$Sulfaterlabeled HS (0)and [3Hlglucosamine-labeled HS(0) from CM-15 (+) and wild-type (+) cells (+) and their control cells (-) (500,000 cpm each) were sequentially applied to a bFGF-affinity column (0.78 x 2 cm) and eluted with discontinuous gradients of NaCl in 10 mM Tris, pH 7.3. Fractions (0.5 ml) were collected, and 35S and 3H distributions were assessed by liquid

(4

scintillation counting. and [3H]glucosamine, and the labeled HS were subjected to hydrazinolysidpH 4 nitrous acid treatment (22, 24). This destroys the intervening N-acetylatedregions of the polymer while preserving highly sulfated blocks in which all glucosamine residues are N-sulfated and where 0-sulfated residues predominate (15). The products were then separatedon a BioGel P-10 column (Fig. 2). The elution patterns of CM-15 (control) cells showed an approximate%fold reduction in fractions longer than decasaccharides as compared to that of wild-type (control) cells, as previously described (15). The most striking changes by the transfection was about 2-fold increase in the yield of each fractionof decasaccharide sizeor greater from the HS of CM-15 (+) cells as compared to thatfrom CM-15 (control) cells. In addition, modest increases of hexa- and tetrasaccharide sizes were observed in CM-15 (+) cells. In contrast, the elution pattern of fragments from wild-type (+) cells was similar to that of wild-type (control) cells except for modest increases of hexa- and tetrasaccharide sizes and larger oligosac-

charides (more than 16-mer). These results demonstrate that overexpression of HS-NdAdVST results in increased synthesis of fully N-sulfated blocks of oligosaccharides in the size classes of decasaccharide or greater in CM-15 cells. Binding Affinities for bFGF ofFully N-Sulfated Oligosaccharide Blocks of H S from CM-15 (+) andWild-type (+) Cells -Previously we have shown that CM-15 cells are impaired in the synthesis of fully N-sulfated blocks of oligosaccharides of decasaccharide size and longer found in HS (15). To confirm that overexpression of the HS-NdAdNST in CM-15 (+) cells resulted in an increase in synthesis of these domains, sizefractionated, fully N-sulfated oligosaccharide blocks from Fig. 2 were applied separately to bFGF affinity columns and eluted with low (0.4 M), middle (1.0 M), and high (2.0 M) NaCl concentrations. As depicted in Table 111, the oligosaccharides with high binding affinities to bFGF were found exclusively in fractions containing decasaccharides and longer. In addition, the 35SPHratio of these fractions increased with their affinity for bFGF. When compared to CM-15 (control) cells, CM-15 (+) cells showed a 3-fold or greater increase in high affinity subfractions of decasaccharides and longer, while increases of less than 2-fold were observed for fractions with lower bFGF affinity. In contrast, theelution patterns of each oligosaccharide class from wild-type (+) cells were quite similar to that of the control. These results demonstrate that overexpression of HS-NdAd NST in CM-15 cells results in increased synthesis of the long extended N-sulfated and highly 0-sulfated oligosaccharide domains that have the highest affinity for bFGF, while levels of these domains are unchanged in transfected wild-type cells. Distribution of [35SlSulfate Groups in HS-derived Oligosaccharides with Different Affinities for bFGF from CM-15 (+) Cells-We have previously reported that the long, N-sulfated oligosaccharides with highest affinity for bFGF derived from wild-type COS cells are enriched in 2-0-sulfated iduronate residues (15). We therefore determined whether IdoA(2-S)-GlcNSO3 occurred in long, N-sulfated oligosaccharides with high affinity for bFGF from CM-15 (+I cells. Deca- and dodecasaccharides were eluted from a bFGF affinitycolumn with a linear NaCl gradient (Fig. 3). Pooled fractions with increasing affinities for bFGF, numbered 1 4 (Fig. 31, were then analyzed for N-sulfate anddisaccharides as shown in Table W.The percentage of sulfate present as N-sulfate (released asfree sulfate by


20094

TABLE I11 Fractionation of size-defined N-sulfated oligosaccharides into specific subpools with differing affinities for bFGF Each size-defined "S,3H-labeled pool of N-sulfated oligosaccharides (14-,12-, lo-, and 8-mer)from CM-15 (+) and wild-type (+I cells and their controls were applied t o a bFGF affinity column (0.8 x 2 cm) and eluted with0.4 M NaCl(14-L, 12-L, 10-L, and 8-L), 1.0 M NaCl(14-M, 12-M, lo", and 8-M), and 2.0 M NaCl (14-H, 12-H, 10-H, and 8-H) in 10 mM Tris, pH 7.3. Subpool

35s

HS-NdAc/NST-transfected 3H c p m l p g cell protein

c p m lpg cell protein

CM-15 14-H 14" 14-L

14 7.2 2.5

5.4 3.5 1.8

2.6 2.1 1.4

3.8 4.6 1.8

1.5 2.5 1.2

2.5 1.8 1.5

12-H 12" 12-L

12 9.6 4.2

4.8 4.8 3.6

2.5 2.0 1.2

2.9 6.6 2.7

1.2 3.4 1.8

2.4 1.9 1.5

10-H 1.9 10" 10-L

5.2 12.2 10.2

6.0 7.6

2.7 2.0 1.3

1.0 8.2 6.8

0.4 4.0 4.3

2.5 2.1 1.6

8-H 8-M 8-L

0 61 26

0 27 18

2.3 1.4

0 48 19

0 23 14

2.1 1.4

40 11 4.4

15 5.3 2.9

2.7 2.1 1.5

42 8.3 2.6

15 4.1 2.1

2.8 2.0 1.2

12-H 9.8 12" 12-L

30 21 10

12

2.5 2.1 1.4

32 19 7.2

11 9.1 5.8

2.9 2.1 1.2

10-H 10" 9.6 10-L

18 52 14

7.2 25

2.5 2.1 1.5

16 48 12

6.6 22 9.1

2.4 2.2 1.3

0 76 31

0 36 22

-

0 69 26

0 32 18

2.2 1.4

Wild-type 14-H 14" 14-L

8-H 8-M 8-L

7.2

-

2.1 1.4

-

-

wild-type (control) cells (Table 111). The presence of intervening N-acetylated regions does not appear to be sufficient to allow for excess N-sulfation ininstances when HS-NdAc/NST is overexpressed.Indeed, the data suggest that production of the blocks exhibiting high affinity for bFGF is tightly controlled even in thepresence of excess enzyme. Since overexpression of the enzyme did not result in any significant increase in the number of N-unsubstituted glucosamine residues in the polymer (data not shown), the results support the view that Nsulfation is coupled to N-deacetylation in vivo (25) and thata single enzyme carries out both reactions in HS synthesis (16). Another important finding was that the CM-15 cells overexpressing HS-NdAc/NST, compared to control cells, produced proportionally less of the poorly 0-sulfated structures that had lower affinities for bFGF in the 10-14-mer size classes of fully N-sulfated blocks of heparan sulfate. Since the oligosaccharide blocks within each 10-14-mer size class all contained the same number of N-sulfated residues, the result suggests that O-sulfation in the large blocks proceeded to a greater extent in DISCUSSION The major conclusion from this study is that the level of the CM-15 (+) cells in which HS-NdAc/NST activities were higher. HS-NdAc/NST stimulates theproduction of high-affinity bFGF Analysis of the distribution of sulfate groups for the bFGF binding domains in HS produced by COS cells until a defined affinity-fractionated deca- and dodecasaccharide subpools from number of sulfated blocks is formed and further synthesis is CM-15 (+) cells showed that the relative affinity of fully Ninsensitive to overexpression of the enzyme. High levels of the sulfated oligosaccharides for bFGF directly correlated withthe enzyme were unable to significantly alter the overall content of IdoA(2-S)-GlcNS03within each size class (Table IV). N-deacetylation and N-sulfation of HS in wild-type (+I cells. This observation is similar to one made for wild-type derived content of IdoA(X-S)-GlcAlthough modest increases in N-sulfated hexa- and tetrasac- materials (15). Incontrast,the charides and larger oligosaccharides (more than 16-mer size) NSO3(6-S)was relatively constant withchanges in therelative were noted, there were no significant variations in theproduc- bFGF affinity. Approximately one IdoA(2-S)-GlcNSOs(G-S)disbe present for each 10- and tion or bFGF affinities of the sulfated blocks in the octa- to accharide unit is estimated to of CM-15 cells tetradecasaccharide rangebetween the overexpressors and the 12-mer block. Ourinitialcharacterization

the pH 1.5 nitrous acid treatment) decreased within the decaand dodecasaccharides size classes as their bFGF affinity increased. This indicated that these higheraffinity fractions contained a higher percentage of 0-sulfated residues per unit of disaccharide. In agreement with this, fractionswithhigher affinity for bFGF contained higher amounts of radiolabeled sulfate in IdoA(2-S)-AManRin both deca- and dodecasaccharides. It is estimated that about four IdoA(2-S)-AhIanRdisaccharideunitsarepresentinthe highaffinityfraction of dodecasaccharides(12-mer-4) andaboutthreeunitsinthe decasaccharides (10-mer-4). The dominant structural unit of heparin, IdoA(2-S)-AManR(6-S),was found at similar levels in each of the pools from deca- and dodecasaccharides examined. It is estimated that about one IdoA(2-S)-AManR(6-S)disaccharide unit is presentper each deca- and dodecasaccharide. The results support our previous observation (15) that increased levels of HS-NdAciNST promote the synthesis of IdoA(2-S)AManR in the fully N-sulfated blocks.


x1 a u 1

1

FractiorlNumber FIG.3. Affinity chromatographyof 35S-labeledHS-derived oligosaccharides from CM-15 (+) cells. Tetradeca-, dodeca-, deca-, octa-, and hexasaccharides (3,000,000cpm each) preparedby pH 4 HONO cleavage of "sSS-labeledN-deacetylated HS from CM-15 (+) cells and separated by gel chromatography (Fig. 2) were applied to a bFGFaffinity column (2 x 6 cm) and eluted with linear gradientsof NaCl in 10 mM Tris, pH 7.3. Fractions (5 ml) werecollected and analyzedfor ""S distribution. The fractions in the dodeca- and decasaccharides were divided into four subpools (subpools 1 4 )as shown by the bars in the panels for dodeca- and decasaccharides t o analyze their ["5Slsulfate distributions.

TABLEIV I""SISu1fate distribution in HS-derived oligosaccharides from CM-15 (+I cells with differing affinities for bFGF Size-defined,affinity-fractionatedHS-derivedoligosaccharidesubpools were prepared from hydrazinolysis/pH 4.0 HONO treatment as described in Fig.3 and analyzed for ["Slsulfate distribution by pH 1.5 HONO treatment and reverse-phase ion-pairing HPLC. The percent of totalradiolabeledoligosaccharidesineach pool degradedintofree :''SO:- and "S-disaccharides by pH 1.5 HONO treatment ranged from 87 to 92%. 12-mer-1 10-mer-1

12-mer-2 12-mer-3 10-mer-2 10-mer-3 Ir

%

52

47 43

10

21 17

26 25

34 33

20 26

18 22

19 22

16 20

3

2 2

%)

Free 35SOi N-sulfate) (as

%>

54 63 62 50

IdoA(2-S)-AManH

56

11

IdoA(2-S)-AManR(6-S) GlcA-AManH(6-S)or4 GlcA(2-S)-AManR

4 2 2

IdoA-AManR(6-S)

2

2

12-mer-4 10-mer-4

0

2 3 2

1

4 0

~~

~

~

~~

showed 0-sulfotransferase activities were at the same level as wild-type (18).The levels of 0-sulfotransferase activities were not changed by the transfection (Table I). Thus, the [35S]su1fate-distribution analysisof the materials derived from CM-15

(+) cells (Table IV) and wild-type-derived materials (15)suggest that the level of activity of HS-NdAcNST may influence the formation of IdoA(2-S)-GlcNS03 in the large extended N-sulfated blocks. Given the results of our present study, it is possible that the presence of large highly sulfated domains in HS may serve to inhibit further enzymatic processing of N-acetylated glucosamine residues by the HS-NdAcNST once a certain amount of domain formation has occurred. This potential substrate specificity for HS-NdAcNST would explain why sulfated block formation is not significantly altered when the enzyme is overexpressedin the wild-type COS cells. On theotherhand, a concept has been proposed that the N-acetylglucosaminyl-/ glucuronosyltransferase and the N-deacetylaselN-sulfotransferase act in concert during chainelongation of heparin, possibly through theaction of enzyme complexes (26). Petterssonet al. (27, 28) have observed that an N-sulfotransferase,purified from a mast cell tumor thatproduces large amounts of heparin, acquires N-deacetylase activity upon addition of a crude protein such as a histone or a polycationic compound such as Polybrene. The rat liver-derived HS-NdAcNST does not appear to require anycofactors for expression of both enzymatic function (16). Itwill be of interest toexamine the effects of overexpression of the mast cell-derived enzyme in our COS cell system and to document whether it will convert the distribution and sulfate content of HS to that more typical for heparin.

Acknowledgments-We thank Drs. H. E. Conrad, RobertJ. Stack, P. N. Shaklee, and Robert D. Larsen for helpful discussions and suggestions. REFERENCES 1. Ishihara, M., and Conrad, H. E. (1989)J. Cell. Physiol. 138, 4 6 7 4 7 6 2. Fedarko, N. S., Ishihara, M., and Conrad, H. E. (1989) J . Cell. Ph,ysiol. 139, 287-294 3. Ishihara, M., Fedarko, N. E., and Conrad, H. E. (19871 J . Biol. Chem. 262, 47084716 4. Castellot, J. J.,Addonizio, M. L., Rosenherg, R., and Karnovsky, M. J. (1981iJ. Cell Biol. 94, 574-580 5. Burgess, W. H., and Maciag, T. (1989) Annu. Reu. Biochem. 58,575-606 6 . Klagsbrun, M. (19891 in Progress in Growth Factor Research, Vol. 1, pp. 207235, Pergamon Press, Oxford 7. Gospodarwicz, D., Neufeld, G., and Schweigerer, L., (19791 J . Cell. Physiol. 5, 12-26 8. Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991) Cell 64,841-848 9. Rapraeger, A. C., Krufka, A., and Olwin, B. B. (1991)Science 252, 1705-1708 10. Klaesbrun. M., and Baird. A. (1991)Cell 67. 229-231 11. Turnbull, J . E., and Galhaher, J. T. (1991) Eiochem. J. 273. 553-559 12. Turnbull, J. E., and Gallagher, J. T. (1990) Biochem. J. 265, 715-724 13. Turnbull, J. E., Fernig, D. G., Ke, Y., Wilkinson, M. C.. and Gallagher, J. T. (1992) J . Biol. Chem. 267, 10337-10341 14. Habuchi, H., Suzuki, S., Saito, T., Tamura, T., Harada, T., Yoshida, K., and Kimata, K. (1992) Biochem. J. 285, 805-813 15. Ishihara, M., Guo, Y., and Swiedler, S. J. (1993) Glycobiology 3, 83-88 16. Wei. Z., Swiedler, S. J., Ishihara, M., Orellana, A,, and Hirschberg, C. B. (19931 Proc. Natl. Acad. Sei. U. S. A. 90, 3885-3888 17. Bame, K. J., Lidholt, K., Lindahl, U., and Esko,J. D. (1991)J. B i ~ lChem. . 266, 10287-10293 18. Ishihara, M., Kiefer, M. C., Barr,P J., Guo,Y., and Swiedler,S. J. (1992JAnaL Biochem. 206,400-407 19. Seed, B., and ArufFo, A. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 3365-3369 20. Aruffo, A., and Seed, B. (1987)Proc. Natl. Acad. Sci. U. S. A. 84, 85734577 21. Hashimoto, Y., Orellana, A,,Gil, G., and Hirschberg, C. B. (1992)J. Biol. Chem. 267, 15744-15750 22. Guo, Y., and Conrad, H. E. (1989) Anal. Biochem. 176, 96-104 23. Ishihara, M., Tyrrell, D. J., Stauber, G. B., Brown, S., Cousens, L. S., and Stack, R. J. (1993) J . Biol. Chem. 268, 46754683 24. Shively, J. E., and Conrad, H. E. (1976) Biochemistr,y 15, 3932-3942 25. Bame, K. J., Reddy, R. V., and Esko, J. D. (1991) J. Biol. Chem., 266, 1246112468 26. Lidholt, K., and Lindahl, U. (1992) Biochem. J . 287,21-29 27. Pettersson, I., Kusche, M., Unger, E., Wlad, H., Nylund, L., Lindahl, U., and Kjellen, L. (1991) J. Biol. Chem. 266, 8044-8049 28. Kjellen, L., Pettersson, I., Unger, E., and Lindahl, U. (1992) Adu. Exp. Med. B i d 313, 107-111