The Antithrombin-binding Sequence of Heparin - Semantic Scholar

THEJOURNALOF BIOLOGICAL CHEMISTRY Vol. 256, No. 5. Issue of March 10, pp. 2389-2394. 1981 Printed In U.S.A.

The Antithrombin-binding Sequence of Heparin LOCATION OF ESSENTIAL N-SULFATE GROUPS* (Received for publication, June 4, 1980, and in revised form, October27, 1980)

Johan Riesenfeld, Lennart Thunberg,Magnus Hook$, and Ulf Lindahl From the Departmentof Medical and Physiological Chemistry, Swedish Uniuersityof Agricultural Sciences, The Biomedical Center, S -751 23 Uppsala, Sweden

Products obtainedby partial depolymerization of pig reports (6-10). One approach to this problem has been to isolate oligosaccharides with high affinity for antithrombin mucosal heparin with nitrous acid were fractionated by affinity chromatographyon immobilized antithrom- following partial depolymerization of heparin. The octasacre- charideshown in Fig. 1 represents the smallest such high bin. A high affinity octasaccharide fraction was covered and reducedwith sodium [3H]borohydride, affinity fragment obtainable by partial depolymerization of yieldingterminal 2,5-anhydro-~-[l-~H]mannitol resi- heparin with nitrous acid (9, 10). Studies on the functional dues. Partial N-desulfation of this material by treat- significance of the various components within this octasacchament with aqueous dimethylsulfoxide resulted in the ride sequence have been initiated. The results described in formation of two distinctspecies, with low and medium the present paper demonstrate that both N-sulfate groups (at affinity for antithrombin, respectively, in addition to positions 4 and 6, respectively, in Fig. 1) are required for high components that retained a high affinity for the protein. affinity binding of heparin to antithrombin. The lowand medium affinity fragments were converted into high affinity components by sulfation of free amino MATERIALSANDMETHODS groups; N-acetylation was without effect. The N-subAn antithrombin-binding heparin octasaccharide was isolated from stituents on the 2 glucosamine units at positions 4 and pig intestinal mucosal heparin by a modification (9) of a procedure 6 (the 3H-labeled end group being at position 8) were reported previously (8). Briefly, heparin was partially depolymerized defined by gel chromatography of the 3H-labeled oli- by treatment with nitrous acida t -10°C for 35 min and the products gosaccharides released on selective deaminationof N- were fractionated by affinity chromatographyon antithrombin-Sephsulfated or N-unsubstituted D-glucosamine residues. In arose. Material eluting at9 1 M NaCl was recovered and fractionated the high affinity fraction, both glucosamine residues, at by gel chromatography on Sephadex G-50. The smallest components, [“HIpositions 4 and 6, respectively, were N-sulfated. The corresponding to octasaccharides, were reduced with sodium medium affinity fraction contained an N-sulfated glu- borohydride and the labeled product(specific activity, 1.0X IO5cpm/ cosamine unitin position 4 and anN-unsubstituted unit pg of uronic acid) was purified by affinity chromatography on antiin position 6. The low affinity components wereeither thrombin-Sepharose followed by gel chromatography on Sephadex G-50. Periodate oxidation of the final product followed by alkaline N-unsubstituted in both positions or N-unsubstituted cleavageyieldedalabeledpentasaccharide asthemainproduct, in position4 and N-sulfated in position 6. It is concluded indicating 2 consecutive sulfated L-iduronic acid residues, and hence that high affinity binding of heparin to antithrombin 2 N-sulfated glucosamine units, between the terminal 2,5-anhydrorequires the presence of two consecutive N-sulfated [:‘H]mannitol residue and the periodate-susceptible uronic acid unit glucosamine residues in specific positions of the anti- (see Fig. 1 and Ref. 8). Disaccharidesandtetrasaccharideswiththegeneralstructures thrombin-binding sequence; lossof either one of these N-sulfate groups (with or without subsequent N-acet- hexuronosyl + 2.5-anhydromannose and hexuronosyl+ N-acetylgluylation) results in a distinct, and appreciable, decreasecosaminyl -+ hexuronosyl -+ 2,5-anhydromannose (0-sulfate groups are not indicated), respectively, were isolated after exhaustive deamin binding affinity (and in anticoagulant activity). inative cleavage of heparin with nitrous acid (8, 11). Radioactivitywasdetermined as describedpreviously (12). Gel chromatography was performed using a column (0.8 x 200 cm) of Sephadex G-25, eluted with 1 M NaCl. Affinity chromatography on Heparin prevents the coagulation of blood by binding to antithrombin-Sepharose was carried out according to the procedure antithrombin (also known in the literature as antithrombin described by Laurent et al. (13), adapted to a total gel volume of 3 ml. 111). thereby accelerating the rate at which this protein inacHeparin fragments were N-desulfated by treatment at 5OoC with tivates the proteinases involved inthe coagulation mechanism dimethylsulfoxide containing 5% water (14). To achieve partial N(1,2). Only a fraction of the molecules inheparin preparations desulfation, the reaction was interrupted after 45 or 75 s; additional binds withhighaffinity to antithrombin and this fraction experimental details are given in the legendto Fig. 3. D-Glucosamine residues with unsubstituted amino groups were N-acetylated by reaccounts for most of the anticoagulant activity (3-5). The withaceticanhydride(15)orN-sulfated by reactionwith identification of the polysaccharide structure required for this action trimethylamine-sulfurtrioxidecomplex (16). high affinity bindinghas been the subjectof numerous recent Treatment of heparin-related saccharides with nitrous acid was performed according to two methods that will be referred to (by the * This work was supported by grants from the Swedish Medical pH of the reaction mixtures) as the pH 1.5 (17) and the pH 3.9 (17; Research Council (2309,5197); the Swedish University of Agricultural Reaction B in Ref.18) procedure, respectively.’ Glucosamine residues Sciences; Kabi AB, Stockholm; and the National Swedish Board for with sulfated amino groups are attacked in the pH 1.5 but not in the Technical Development. The costs of publication of this article were defrayed in part by the payment of page charges, This article must ’ T h e varioustypes of deaminationreactionsareactuallynot therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. properly defined solely by reference to the pH of the reaction mixtures *Present address, Diabetes Research and Training Center, De(17). T h e p H 1.5 procedure and the pH 3.9 procedure refer to the partment of Biochemistry, University of Alabama in Birmingham, reactions involving 0.33 M and 3.9 M nitrite, respectively, botha t room Birmingham, AL 35294. temperature, as shown in Table 111 of Ref. 17.

2389

Antithrombinbinding

2390

pH 3.9 reaction, whereas glucosamine residues with unsubstituted amino groups behave inthe reverse manner. Glucosamine units with acetylated amino groupsare resistant to either reaction. Susceptible glucosamine residues are converted into 2,5-anhydromannose units, alongwith cleavage of the correspondingglucosaminidiclinkages. Due to a side reaction (deaminative ring contraction),the cleavage process is not always quantitative(19). RESULTS

Strategy of the Investigation-Removal of N-sulfate groups from heparin leads to loss of anticoagulant activity (20-22) and it therefore seems reasonable to assume that such groups are essential to the binding of heparin to antithrombin. The aim of the present study was to locate such essential N-sulfate groups in the antithrombin-binding sequenceof the polysaccharide. This sequence is contained within an isolated octasaccharide fragment, which thus represents 4 glucosamine units of the intact polymer (Fig. 1). However, only three of these units retain their N-substituents; the glucosamine residue corresponding to position 8 loses its amino group andis converted into a terminal anhydromannose unit during degradation of the polysaccharide with nitrousacid. Furthermore, one of the remainingglucosamine units, at position 2, is preferentially N-acetylated (7,8). The problem of locating the N-sulfate groups that areessential for thebinding of heparin to antithrombinmay therefore be reduced to the questionas towhetherbothor only one of theN-sulfate groups a t positions 4 and 6, respectively, are involved. This question was approached asfollows. Octasaccharides with high affinity for antithrombin and with 'H-labeled 2,5-anhydromannitol end groups were subjected to partial N-desulfation, and the products were fractionated with regard to affinity for antithrombin. Loss of affinity was correlated to loss of N-sulfate groups by structural characterization of the fractions separated by affinity chromatography. The distributionof N-sulfated and N-unsubstitutedglucosamine residues was deduced from the molecular size of the 3H-labeled oligosaccharides released from the octasaccharides by selective deamination with nitrous acid. The procedure is schematically illustrated in Fig. 2. Cleavage was induced either a t N-unsubstituted glucosamine residues(pH 3.9 procedure) or, after acetylblocking of free amino groups, at N-sulfated glucosamine residues (pH 1.5 procedure), and the products were separated by gel chromatography. Only components containing theinitial end group "H-label would register in the chromatograms.As seen from Fig. 2 and Table I, the combined information obtained by the two degradation procedures will allow the discrimination between all possible combinations of N-substituents on a t positions 4 and 6, the 2 glucosamineresidueslocated respectively. Affinity Chromatography of Partially N-Desulfated Antithrombinbinding Heparin Fragment-'H-labeled antithrombin-binding octasaccharide was partially N-desulfated as described in the legend to Fig. 3 and was then subjected to affinity chromatography on antithrombin-Sepharose. While

Sequence of Heparin

w NAc NSOj N q

I

PARTIAL N-DESULFATW

r"-

* 1

ACETYLATION

NAc N q N A c

\

H Q ; Pl 3.9

" NAc NSUj

rtJq;PH1.5

" NAc NAc FIG. 2. Schematic representation of the analytical procedure. A 3H-labeled antithrombin-binding heparin octasaccharideis subjected to partial N-desulfation. A hypothetical product, isolated by affinity chromatography, is degraded with nitrous acid, either at pH 3.9 or, following N-acetylation,at pH 1.5. The size of the labeled oligosaccharides formed is determined by gel chromatography. The symbolsused are: , glucosamineresidue (X denoting the substituent on the amino group; sulfate,acetyl, or none, as indicated); U,hexuronic acid residue; -A, 2,5-anhydromannose residue; -A, 2,5-anhydr0[1-~H]mannitol residue. The particular desulfation product shown will release a labeled disaccharide on deamination at pH 3.9 and a labeledtetrasaccharide at pH 1.5; see Table I for products obtained withother combinations of N-substituents.

x

the intact fragment emerged as a single peak at about 1.4 M NaCl concentration (Fig. 3A),the modified material separated into three fractions (Fig. 3B). Part of the material, with low affinity for antithrombin,was either not absorbed or eluted at low ionic strength. A fraction of medium affinity eluted as a distinct peak a t about 0.6 M NaCl, whereas a high affinity fraction retained the same affinity for antithrombin as the starting material. Prolonging the reaction with dimethylsulfoxide from 45 to 75 s led to a decrease in the high affinity fraction and acorresponding increase in the low affinity fraction; themedium affinity fraction was essentially unchanged. Each of thethreefractionsretainedits elution position on rechromatography. N-Acetylating the modified fragments prior to chromatographydid not significantly affect the elution patterns (not shown). N-Sulfation of the same material restored the initialhighaffinityfor antithrombin (Fig. 3C), demonstrating that thelack of such affinity following dimethylsulfoxide treatment was indeed due to loss of Nsulfate groups. Structural Characterizationof Fractions-Deamination of the high, medium, and low affinity fractions with nitrous acid at pH 3.9 gave the results shown in Fig. 4. The high affinity fraction did not release any low molecular weight labeled components (Fig. 4A), suggesting thatthe 2 glucosamine residues located at positions 4 and 6, respectively, were both N-sulfated (see Table I). In contrast, both the medium and the low affinity fractions were degraded by the nitrous acid 1 2 3 4 5 6 7 8 treatment, yielding labeled disaccharide (Fig. 4B)and a mixture of di- andtetrasaccharide (Fig. 4 0 , respectively, as degradation products. The amount of disaccharide released from thelow affinity fraction increased with increasing extent of N-desulfation. The formation of disaccharides indicates FIG. 1. Tentative structure for the antithrombin-binding oc- that essentially all medium affinity components anda proportasaccharidefragment isolated afterpartialdeaminative tion of the low affinity molecules contained an N-unsubsticleavage of pig mucosal heparin. X = H or SO3-. The (3H-labeled) tuted glucosamine residue at position 6. The fairly large yield 2,5-anhydromannitol residue in position 8 corresponds toa N-sulfated of labeled tetrasaccharide from the low affinity fraction furglucosamine unit in the intact polysaccharide. The structure shown ther suggests that a significant proportion of these molecules represents the majority of antithrombin-bindingsequencesinpig mucosal heparin; however, structural variants compatible with affin- contained an N-unsubstitutedglucosamine residue at position 4, in combination with an N-sulfategroup at position 6. ity for antithrombin cannotbe excluded.

Antithrombinbinding

Sequence of Heparin

2391

TABLE I Labeled oligosaccharides formed on deamination of heparin octasaccharideswith different N-substituents Only the 2 glucosamine residues (in positions 4 and 6) adjacent to the labeled anhydro['H]mannitol end group (position 8) are considered; it is assumed that these residues are both N-sulfated in the starting material (see "Materials and Methods"). For additional information, see the text and Fig. 1. Terminal sequence" positions Labeled with oligosaccharide deamination formed on

+

UA

GlcN

-+

UA

-P

-+

soaUA

-+

GlcN

GlcN

8

pH 1.5b

pH 3.9

"-*

UA

+

[JH]aMan

Disaccharide

>Tetrasaccharide

I

I

+

at 7

6

5

4

3

+

UA

-+

SO'GlcNHs'

-+

UA

-P

["laMan

Tetrasaccharide

Disaccharide

-+

UA

+

GlcN

-+

UA

"-*

[3H]aMan

Disaccharide

Tetrasaccharide

I

sos+

UA

+

GlcNH3'

I SO'GlcNHa'

UA "-* r3HlaMan >Tetrasaccharide Disaccharide " No 0-sulfate groups are indicated. Abbreviations: UA, unspecified hexuronic acid; GlcNSO:]-, N-sulfated D-glucosamine; GlcNH:,', Nunsubstituted D-glucosamine;aMan, 2,5-anhydro-D-mannitol. I, Unsubstituted amino groups were acetylated prior to deamination at pH 1.5. +

UA

GlcNH:%'

-+

UA

-+

-+

-+

1 I1 I

OS

. (6) . l . . . .i . . . :. . . .

l

I 1

i

-

t

4:

. _ ^. . . . . . . . . .

!(&-

20

30

I

FR*CTION N-R

FIG. 3. Affinity chromatography on antithrombin-sepharose of (A) antithrombin-binding heparin octasaccharide; (B) the same fragment after partial N-desulfation; and (C) the same material as in B, after sulfation of free amino groups. For N-desulfation, the material (4 X IO6 cpm) was converted into the pyridinium salt and was then dissolved in 0.1 ml of dimethylsulfoxide containing 5%water. After heating the solution at 50°C for 45 s, the reaction was terminated by the addition of 0.1 ml of 0.2 mM NaOH in 1 M NaCI. The products were recovered by passage through a column of Sephadex G-15, eluted with 10%ethanol and were then separated by affinity chromatography (B). The two peaks of the low affinity fraction differ only slightly with regard to affinity for antithrombin; the more retarded peak can be eliminated by increasing the initial NaCl concentration of the elution buffer from 0.05 to 0.1 M. After N desulfation for 45 s ( B ) the low, medium, and high affinity fractions constituted 56, 8, and 35%, respectively, of the total labeled material. Increasing the reaction period to 75 s did not affect the elution positions of the individual peaks; however, the distribution of material

EFFLUENT VOLUME (mli

FIG. 4. Gel chromatography on Sephadex G-26 of labeled products obtained on deamination at pH 3.9 of the (A) high affinity fraction; (B) medium affinity fraction; and (C) low affinity fraction of partially N-desulfated antithrombin-binding heparin octasaccharide.The arrowsindicate the peak elution positions of the intact octasaccharide ( * ) and of di- and tetrasaccharides, respectively, obtained by exhaustive deamination of heparin. See legend to Fig. 5 regarding the difference in elution position between the low affinity tetrasaccharide (0and the reference tetrasaccharide. The octasaccharide peaks in Panels ( B )and (0probably reflect the extent of deaminative ring contraction, not accompanied by glycosidic bond cleavage (19).

between the low, medium, and high affinity fractions was changed to fractions werepooled as indicated by the horizontal bars, desalted by passage through columns of Sephadex (3-15, eluted with 10% ethanol, and analyzed as described in the text. 83, 10, and 3%, respectively. Thethree

Antithrombinbinding

2392

Sequence of Heparin TABLEI1

Predominant terminal sequences in fractions ofpartially N-desulfated,antithrombin- binding heparin octasaccharide Terminal sequence" with positions

Fraction 3

High affinity

+

UA

5

4 -+

GlcN

+

UA

I

Medium affinity

+

sOn

UA

--*

['HIaMan

-+

UA

"-f

["HlaMan

-+

UA

--f

["HIaMan

-+

UA

--f

[,'H]aMan

1

UA

-+

GlcN

UA

+

GICNH.~'

,

-+

UA

--f

+

UA

--f

GlcNH:g+ GlcN

I

so:'-

Low affinity + I'

-+

so:r-

~

so:$-

"+

GlcN

8

7

6 +

UA

-+

GlcNH.''

+

UA

+

GlcN&+

No 0-sulfate groups areindicated. For abbreviations, see Table I.

EFFLUENT VOLUME

m O

FIG. 5. Gel chromatography on Sephadex G-25 of labeled products obtained on deamination at pH 1.5 of the ( A ) high ( C ) low affinity fraction; (B) mediumaffinityfraction;and affinity fraction of partially N-deeulfated, N-acetylated antithrombin-binding heparin octasaccharide. For explanation of the arrows, see Fig. 4. The majorcomponent derived fromthe medium aftinity fraction ( B ) eluted somewhat ahead of the tetrasaccharide standard. However, this difference was apparently due to the high 0-sulfate content of the material, demonstrable by paper electrophoresis; the elution patternof the completely desulfated medium affinityoligosaccharide was similar tothat of ahyaluronicacid tetrasaccharide (not shown).

fractions from each other (Fig. 5). The labeled component released from the high affinity fraction was essentially disaccharide (Fig.5A), whereas that derivedfrom the medium affinity fraction wastetrasaccharide (Fig. 5B). Thelow affinity fraction yielded two peaks, correspondingto disaccharide and octasaccharide, respectively (Fig. 512).Taken together with the results of deamination at pH 3.9, these cleavage products provide the information required todefine the N-substituents, in each fraction, of the 2 disaccharide units adjacent to the terminal 2,5-anhydro['H]mannitol residue (Table 11). As seen from Table I, an octasaccharide fragment with an N-sulfate group at position 6 should yield a labeled disaccharide on deamination at pH1.5. The release of such disaccharide from the high affinity fraction is thus in agreement with the results of deamination at pH3.9, indicating N-sulfate groups both at position 4 and at position 6. Conversely, the lack of labeled disaccharide following deamination of the medium affinity fraction at pH 1.5reflects the lack of N-substituent at position 6; the tetrasaccharide formed from the same material shows that the N-unsubstitutedglucosamine residue at position 6 is followed by an N-sulfated unit at position 4. The labeled disaccharide liberated from thelow affinity fraction at pH1.5 waspresumablyderivedfrom the same octasaccharide sequences (with N-sulfate groups at position 6 but not a t position 4) that released labeled tetrasaccharide on deamination at pH 3.9. The remaining low affinity sequencesresisted deamination at pH 1.5, indicating that the low affinity molecules lacking N-substituent at position 6 were also N-unsubstituted at position 4. DISCUSSION

Two principally different procedures have been employed in attempts to elucidate the structure of the antithrombinbinding sequence of the heparin molecule. Rosenberg and coworkers separated heparin into fractions of high and low affinity for antithrombin, respectively, and then searchedfor structural differences by comparing the products obtainedon exhaustive deaminative cleavage with nitrous acid (6, 7). A tetrasaccharide, iduronosyl "-f N-acetylglucosaminyl(6-0-sulfate) -+ glucuronosyl-+ anhydromannose (6-0-sulfate) (where the anhydromannose corresponds to an N-sulfated glucosaThese conclusions were confirmedand extended in another mine unit in the intactpolysaccharide chain), was obtained in set of experiments in which the octasaccharides were degraded significant quantity from thehigh affinity fraction. Since this with nitrousacid, under conditions (pH 1.5) resulting in cleav- tetrasaccharide was practically absent from the deamination age at N-sulfated glucosamine residues. T o ensure maximal products of the low affinity fraction, it wasconsidered to specificity, the three fractions (high, medium,and low affinity) represent the critical site responsible for anticoagulant activof partially N-desulfated octasaccharide were N-acetylated ity. In a different approach, Lindahlet al. subjected unfractionprior to deamination. Theresulting gel chromatography patacid and terns were distinctly different from thoseobserved after deam- ated heparin to partial deamination with nitrous then separated the products(which thus contained intact Nination at pH 3.9, and also clearly distinguished the three

Antithrombinbinding

Sequence of Heparin

2393

sulfate groups) by affinity chromatography on immobilized variations in 6-sulfation that is not reflected by any correantithrombin. A highaffinity fraction was isolated; it was sponding variation in affinity for antithrombin (10). Defining the roles of the two essential N-sulfate groups composed of about 12 monosaccharide units and contained the tetrasaccharide sequence implicatedby Rosenberg et al. helps to explain why no high affinity antithrombin-binding as an integral component (8). More recent evidence showed fragments smaller thanoctasaccharide were isolated after that this material could be fractionated by gel chromatogra- partial deaminative cleavage of heparin (9). Formation of a phy intoa series of oligosaccharides, an octasaccharide being hexasaccharide would involve cleavage of the glucosaminidic the smallest species present in significant amounts (9). The bond between positions 6 and 7, along with loss of the Nstructure of this octasaccharide (9, 10) is shown in Fig. 1; the sulfate group a t position 6. Such a hexasaccharide would still sequence (extendingfrom position specific tetrasaccharidesequence occupiesone-half of the contain the tetrasaccharide molecule, with the nonsulfated iduronicacid unit at the non- 1 to position 4) proposedby Rosenbergand Lam (7) to reducing terminus. Itis obvious that any componentessential represent the critical site responsible for high affinitybinding However, thepresentresults to antithrombin binding, located in the region 5 to 8, would of heparintoantithrombin. demonstrate that sequences having N-sulfate at position 4 escape detection by the procedure of Rosenberg et al. ( 8 ) . being Little is known of the functional role of the various struc- only can show at most (all other structural requirements tures in the antithrombin-bindingregion of the heparin mol- fulfilled) medium affinity for antithrombin; indeed, the meecule. Components not essential to the interaction with anti- dium affinity fraction isolated after partial deamination of thrombin may be recognized simply as structural variants(i.e. heparin contained significant amounts of hexasaccharide (9). Heparin with medium affinity for antithrombin, prepared by residues present in some high affinity sequences but not in with "Hothers), such as the 6-sulfategroup at position 4 (10). On the affinity chromatography on antithrombin-Sepharose, other hand, any component that is unique to (and invariably labeled medium affinity fractionasinternalstandard(not present in) the antithrombin-binding region would presum- shown), had an anticoagulant activity of 165 British Pharably be essential to itsfunction; such a component, a 3-sulfate macopoeia units/mg;the activity of high affinity heparin group at position 4 (Fig. l ) ,has indeed been identified recently (prepared in an analogous manner) was 280 British Pharma(10). However, the inherent difficulties in functional evalua- copoeia units/mg. The heparin fraction studied by Rosenberg tions of this kind are illustratedby the following observations. and Lam was selected for by its very high affinity for antiThe constant occurrenceof nonsulfated iduronic acid in non- thrombin ( 7 ) and should thus contain antithrombin-binding reducing terminal position of the antithrombin-binding octa- sequences with N-sulfate groups both at position 4 and at saccharide (9), and the lower concentrations of this sugar in position 6. heparin molecules that lack high affinity for antithrombin (7, Acknowledgments-We are grateful to Mr. B. Ajaxon, AB Vitrum 8), suggested that nonsulfated iduronic acidis somehow essen(Stockholm), for carrying out the assays for anticoagulant activity. tialtoanticoagulant activity.Unexpectedly, theheptasaccharide obtained on digesting the octasaccharide with a-LREFERENCES iduronidase retainedhigh affinity forantithrombin, indicating that thenonsulfated iduronic acid is itself not essential to the 1. Rosenberg, R. I). (1977) Fed. Proc. 36, 10-18 interaction with the protein.' 2. Nordenman, B., Danielsson, A. & Bjork, I. (1978)Eur. J . Biochem. In the present investigation, the role of N-sulfate groups 90, 1-6 3. Lam, L. H., Silbert, d . E. & Rosenberg, €2. D. (1976) Biochem. has been defined on the molecular level. The N-substituents Biophys. Res. Commun. 69, 570-577 of the glucosamine residues in heparin areof two kinds, acetyl 4. Hook, M., Bjork, I., Hopwood, J . & Lindahl, U. (1976)FEBS Lett. and sulfate groups (23, 24). The importance of the N-sulfate 66,90-93 groups to the anticoagulant activity of the polysaccharide has 5. Anderson, L.-O., Barrowcliffe, T. W., Holmer, E., Johnson, E. A. been well established in previous studies (20-22,25). Removal & Sims, G. E. C. (1976) Thromb. Res. 9, 575-583 of N-sulfate groups thus results in a loss of anticoagulant 6. Rosenberg, R. D., Armand, G. & Lam, L. (1978)Proc. Natl. Acad. Sci. U. S. A. 75, 3065-3069 activity that is restored by resulfation but not by acetylation 7. Rosenberg, R. D. & Lam, L. (1979) Proc. Natl. Acad. Sci. U. S. of the exposed free amino groups.:' The present results demA. 76, 1218-1222 onstrate thathigh affinity bindingof heparin to antithrombin requires that2 consecutive glucosamine residues, correspond- 8. Lindahl, U., Backstrom, G., Hook, M., Thunberg, L., Fransson, L.-A. & Linker, A. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, ing to positions 4 and 6 of the antithrombin-binding sequence 3198-3202 (Fig. l), are both N-sulfated. Loss of the N-sulfate group at 9. Thunberg, L., Backstrom, G., Grundberg, H., Hiesenfeld, J . & position 4 resulted in a drastic decrease in binding affinity.NLindahl, U. (1980) FEBS Lett. 117,203-206 Desulfation a t position 6 produced an appreciable, but less 10. Lindahl, U., Backstrom, G., Thunberg, L. & Leder, I. G . (1980) Proc. Natl. Acad. Sei. U. S. A. 77,6551-6555 pronounced, weakening of the interaction. The most plausible 11. Jacobson, I., Hook, M., Pettersson, I., Lindahl, U., Larm, O., interpretation of these findings is that both N-sulfate groups Wiren, E. & von Figura, K. (1979)Biochem. J . 179, 77-87 bind specific ligands on the antithrombinmolecule, The con- 12. Lindahl, U., Jacobson, I., Hook, M., Backstrom, G. & Feingold, tribution of nonspecific charge effects would seem to be of D. S. (1976) Biochem. Biophys. Res. Commun. 70,492-499 little importance, also for the N-sulfate group a t position 6, 13. Laurent, T. C., Tengblad, A., Thunberg, L., Hook, M. & Lindahl, U. (1978) Biochem. J . 175,691-701 since the decreasein binding affinityon N-desulfation was not even partially reversed by N-acetylation of the resulting pos- 14. Inoue, Y.& Nagasawa, K. (1976) Carbohydr. Res. 46,87-95 itively charged amino groups. Moreover, intact antithrombin- 15. Danishefsky, I. & Steiner, H. (1965)Biochim. Biophys. Acta 101, 37-45 binding octasaccharide displays charge heterogeneity due to 16. Levy, L. & Petracek, F. J. (1962) Proc. Soc. Exp. Biol. Med. 109,

' L. Thunberg, G. Backstrom, U. Lindahl, and L. H. Rome, unpublished observation. This is in contrast to the interaction between heparin and lipoprotein lipase, that can be restored (after partial N-desulfation) not only by re-N-sulfation but also by N-acetylation of the polysaccharide (26).

901-905 17. Shively, J. E. & Conrad, H. E. (1976)Biochemistry 15,3932-3942 18. Lindahl, U., Backstrom, G., Jansson, L. & Hallen, A. (1973) J. Biol. Chem. 248, 7234-7241 19. Shively, J. E. & Conrad, H. E. (1976)Biochemistry 15,3943-3950 20. Cifonelli, J. A. (1975) Adu. Exp. Med. Biol. 52, 95-103 21. Nagasawa, K., Tokuyasu, T. & Inoue, Y. (1977) J. Biochem.

2394

Antithrombin- binding Sequence of Heparin

(Tokyo) 81,989-993 22. Riesenfeld, J., Hook, M., Bjork, I., Lindahl, U. & Ajaxon, B. (1977) Fed. Proc. 36,39-43 23. Cifonelli, J. A. & King, J. (1972) Carbohydr. Res. 21, 173-186 24. Lindahl, U. (1976) in MTP International Review of Science:

Organic Chemistry Two-Carbohydrate Series Chemistry (Aspinall, G . O., ed) Vol. 7, p.283, Butterworths, London 25. Cifonelli, J. A. (1974) Carbohydr. Res. 37, 145-154 26. Bengtsson, G., Olivecrona, T., Hook, M., Riesenfeld, J. & Lindahl, U. (1980) Biochem. J . 189,625-633