Circular dichroism spectroscopy ofheparin-antithrombin interactions · Bio-Gel P-10, Bio-Gel P-100, and hydroxyl-apatite (Bio-GelHTP) wereobtained from Bio-Rad. Whatman...

Proc. NatL Acad. Sci. USAVol. 79, pp. 7190-7194, December 1982Biochemistry

Circular dichroism spectroscopy of heparin-antithrombininteractions

(oligosaccharides/mucopolysaccharides/conformational change)

AUDREY L. STONE*, DAVID BEELERt*, GARY OOSTAt*, AND ROBERT D. ROSENBERGt#*National Institute of Mental Health, National Institutes of Health, Bethesda, Maryland 20205; tSidney Farber Cancer Institute, Harvard Medical School, Boston,Massachusetts 02115; and WMassachusetts Institute of Technology, Cambridge, Massachusetts 02139

Communicated by Elkan R. Blout, September ), 1982

ABSTRACT We have utilized circular dichroism spectroscopyto examine the interaction of antithrombin with heparin-derivedoligosaccharides and mucopolysaccharides of various sizes. Ourstudies demonstrate that the various complexes exhibit two majortypes of chiral absorption spectra. The first of these patterns isseen when octasaccharide, decasaccharide, dodecasaccharide, ortetradecasaccharide fragments bind to the protease inhibitor. Thecircular dichroism spectra of these complexes when compared tothe spectrum of free antithrombin show several distinguishingcharacteristics. On the one hand, there is a marked general in-crease in positive chiral absorption that is maximal at 296 and 288nm and 290 and 282.5 nm. These observations indicate pertur-bation of "buried" and "exposed" tryptophan residues. On theother hand, a significant augmentation in circular dichroism thatpeaks at 269.5 and 263 nm is noted. These findings are probablydue to the summed positive and negative contributions arisingfrom tryptophan residue(s), disulfide bridge(s), and phenylalanineresidue(s). Given that these heparin fragments are able to accel-erate factor Xa-antithrombin interactions but not thrombin-an-tithrombin interactions, the above spectral transitions must beassociated with either the binding of a critical domain of the oli-gosaccharides to the protease inhibitor or the "activation" of theprotease inhibitor with respect to factor Xa neutralization. Thesecond ofthese patterns is apparent when octadecasaccharide, lowmolecular weight heparin (6,500), and high molecular weight hep-arin (22,000) interact with antithrombin. The circular dichroismspectra of these complexes compared to the spectrum of free pro-tease inhibitor are similar to the first pattern except for changeswithin the 292- to 282-nm and 275- to 255-nm regions. The sub-traction ofthe first pattern from the second pattern reveals a shal-low negative band between 300 and 275 nm with potential negativeminima at 290 and 283 nm as well as a deep negative band between275 and 255 nm with possible negative minima at 268 and 262 nm.This chiral absorption profile is most likely to arise from confor-mational changes of a disulfide bridge(s). However, we cannotcompletely exclude the possibility that the above circular dichro-ism difference curve might be explained on the basis of transitionsoriginating from a tryptophan residue(s). Given our method forgenerating the above data, these spectral alterations must be as-sociated with the binding of a second critical domain of the mu-copolysaccharide to antithrombin that is required for rapid com-plex formation with thrombin or the activation of the proteaseinhibitor with respect to the neutralization of the latter enzyme.

A small fraction of all heparin preparations binds tightly to an-tithrombin and is responsible for the anticoagulant activity ofthe mucopolysaccharide (1, 2). A unique tetrasaccharide se-quence that contains two nonsulfated uronic acid residues anda N-acetylglucosamine group (3, 4) as well as a 3,6-disulfatedglucosamine moiety (5) is found within molecules of heparin

that have anticoagulant activity. It is assumed that this portionof the mucopolysaccharide represents a major portion of thebinding site recognized by antithrombin and is also sufficientto accelerate factor Xa-antithrombin interactions (4-6). A sec-ond region of heparin with an unknown structure is locatedsome distance from the above tetrasaccharide and appears tobe required for augmenting the rate of neutralization of throm-bin, factor IXa, and factor XIa by antithrombin (6). On the basisof these data, we have proposed that heparin contains multiplefunctional domains that are able to differentially modulate theanticoagulant properties of the protease inhibitor (6). In thiscommunication, we show by near ultraviolet circular dichroismspectroscopy that these two major regions of the mucopolysac-charide interact with separate areas of antithrombin and prob-ably induce different conformations of the protease inhibitor.

MATERIALS AND METHODSColumn Chromatographic Matrices. Sepharose 4B, Seph-

adex G-100, and concanavalin A-Sepharose were purchasedfrom Pharmacia. Bio-Gel P-10, Bio-Gel P-100, and hydroxyl-apatite (Bio-GelHTP) were obtained from Bio-Rad. WhatmanDE-52 DEAE-cellulose was provided by Reeve-Angel (Clifton,NJ). Heparin-Sepharose 4B was prepared by coupling a purifiedform ofthe mucopolysaccharide to cyanogen bromide-activatedSepharose 4B as described (7).

Chemicals. All chemicals were reagent grade or better.Proteins. Human thrombin, human factor Xa, human an-

tithrombin, and bovine antithrombin were prepared in physi-cally homogeneous form by methods previously reported (8).

Measurement of Protein or Mucopolysaccharide Concen-tration. The protein concentrations ofhuman thrombin, humanfactor Xa, human antithrombin, and bovine antithrombin weredetermined by absorbance measurements at 280 nm, assumingextinction coefficients of 16.2, 11.6, 6.5, and 6.7 M ' cm-',respectively (8, 9). Mucopolysaccharide concentrations of hep-arin fractions were estimated colorimetrically by assay ofuronicacid according to the method of Bitter and Muir (10).

Biologic Activity of Heparin Fragments. The potencies ofoligosaccharide or mucopolysaccharide samples with respect tothrombin or factor Xa inhibition were determined by two-stageassay procedures (6).

Highly Active Heparin Preparations. Crude porcine heparinofintestinal origin (stage 14, Wilson Laboratories, Chicago) waspurified by cetylpyridinium chloride precipitation and Sepha-dex G-100 gel filtration (11). Components of molecular weight:22,000 (HMW pool) or -6,500 (LMW pool) were individuallycombined and then separately fractionated by affinity with lim-

Abbreviations: HMW and LMW heparins, high and low molecularweight heparins.

7190

The publication costs ofthis article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-ment" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

16, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982) 7191

iting amounts of bovine antithrombin as outlined (12). The re-sultant products, termed HMW heparin and LMW heparin,exhibited anticoagulant potencies of 731 and 350 units/mg, re-spectively, as quantitated by the thrombin inhibition assay aswell as 491 and 259 units/mg, respectively, as determined bythe factor Xa inhibition assay.

Heparin oligosaccharides were prepared by cleaving 4 g ofporcine mucopolysaccharide with limiting amounts of nitrousacid as described (6). These products were chromatographedat 14 ml/hr on a column of Bio-Gel P-10 (1. 8 X 200 cm) equil-ibrated with 0.5 M NH4HCO3, pH 7.6 (buffer A). The resultantoctasaccharides, decasaccharides, dodecasaccharides, tetra-decasaccharides, and octadecasaccharides were individuallypooled and separately fractionated by affinity with bovine an-tithrombin as outlined above (12). These antithrombin-bindingoligosaccharides were rechromatographed at 2 ml/hr on a col-umn of Bio-Gel P-10 (0.6 x 200 cm) equilibrated with bufferA so that species of larger or smaller size could be excluded fromthe final preparations. In particular, mucopolysaccharide frag-ments present within the early fractions of the tetradecasac-charide pool were eliminated from the final product becausethey were able to accelerate the inhibition of thrombin by an-tithrombin. The specific factor Xa-inhibiting activities of theanticoagulant octasaccharide, decasaccharide, dodecasacchar-ide, tetradecasaccharide, and octadecasaccharide were 182,196, 176, 225, and 110 units/mg, respectively. The specificthrombin-inhibiting activities of the anticoagulant octasacchar-ide, decasaccharide, dodecasaccharide, as well as tetradecasac-charide were essentially zero, whereas the anticoagulant octa-decasaccharide exhibited a specific thrombin-inhibiting activityof 80 units/mg.

+15

+10h.0xe -1j

Circular Dichroism. Circular dichroism measurements wereperformed at 24°C on a Jasco series 500J spectropolarimeter.Cylindrical cells of high optical quality (Optical Cell, Woodbine,MD) with a path length of 2 cm were utilized to obtain spectraldata in the 254- to 303-nm region at 24 + 1°C with a sensitivityof 1 millidegree/cm. The environmental conditions utilizedwere 0.15 M NaCl in 0.01 M Tris-HCl, pH 7.5. The final con-centration of antithrombin employed in the above studies was27.8 ,uM. The final level of mucopolysaccharide utilized wascalculated to attain at least 98% saturation of the protease in-hibitor. In the case of HMW heparin, sufficient amounts ofreactants were used so that 85% of the antithrombin-muco-polysaccharide complexes were tertiary (two antithrombin mole-cules) and 15% were binary. The binding constants employedin the above calculations were established by fluorescence spec-troscopy and were either reported in prior communications (8,11, 12) or will be described in a subsequent publication.

RESULTSFig. 1 Left shows the circular dichroism spectra of three sep-arate aliquots of human antithrombin in the near ultravioletregion. The experimental data exhibit excellent reproducibility,with sample-to-sample variations at the various wavelengthsthat are usually no greater than 1,000 molar ellipticity units(degrees cm2/dmol). The chiral absorption pattern of humanantithrombin in the region 250-310 nm should arise from thesummed contributions ofconformationally inherent or acquiredasymmetries of 4 tryptophan residues, 8 tyrosine residues, 23phenylalanine residues, and 3 disulfide bridges (13-15). Byutilizing established criteria, the major features of the abovecircular dichroism spectra can be tentatively associated with

-4 ~E

§,T W-

290 300 o 260Wavelength, nm

FIG. 1. Near ultraviolet circular dichroism spectra of human antithrombin and oligosaccharide or mucopolysaccharide-protease inhibitor com-plexes. (Left) The various oligosaccharides were added to antithrombin to form complexes with: *, octasaccharide; *, decasaccharide; o, dodecasaccha-ride; and v, tetradecasaccharide. The final levels of heparin fragments employed were sufficient to attain at least 98% saturation of antithrombinwith the final concentration of protease inhibitor set at 27.8 ,tM. The spectra of antithrombin were obtained with three aliquots of the proteaseinhibitor (A, a, v). Small positive variations from the solvent baseline were observed in control scans with free glycosaminoglycans. Control valuesfor these deviations are expressed in terms of their effects on the calculated molar ellipticities ([6]) of antithrombin within the various complexes.These apparent molar ellipticities of free HMW heparin, free decasaccharide, and free dodecasaccharide are depicted at various wavelengths as thefirst, second, and third bars in the sets along the zero line. The levels of heparin fragments and the experimental conditions were identical to thosefor experimental spectra in Left and Right. (Right) The various fractions of glycosaminoglycans were admixed with antithrombin to generate oc-tadecasaccharide complex (A), LMW heparin complex (o), and HMW heparin complex (o). The final levels of mucopolysaccharide employed weresufficient to attain at least 98% saturation of antithrombin with the final concentration of protease inhibitor set at 27.8 ,uM. The spectrum of an-tithrombin was obtained by averaging results depicted in Left (A). The bottom of Right shows the circular dichroism difference, A[ ],c between thecorrected spectra of HMW heparin complex and dodecasaccharide complex (W).

Biochemistry: Stone et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

16, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982)

various transitions among the four kinds ofamino acid moieties(14, 15).

The spectral region above 289 nm contains four types ofbands that arise from two classes oftryptophan residues (14, 15).The first band is noted as a minimum at 302 nm, which cor-responds to the O-OLa transition ofa tryptophan residue(s) thatexhibits strong interactions with, or is hydrogen bonded to,neighboring polar amino acid moieties. The second band ap-pears as a poorly resolved shoulder at 297 nm, which representseither the 0-11La transition of the above group(s) or the O-O'Latransition ofa second tryptophan residue(s) that is more reactivewith the polar solvent. The third band is observed as one of apair ofmaxima at 293.5 and 286 nm that are characteristic oftheO-OLb and the 0-1'Lb transitions of "buried" tryptophan res-idue(s). The fourth band is seen as one of a pair of minima at290 and 282.5 nm that are typical of tryptophan residue(s) thatare "exposed" to the aqueous solvent.The spectral region from 289 to 270 nm exhibits a series of

multiple maxima that are generated by complex contributionsfrom tryptophan residue(s), tyrosine residue(s), and disulfidebridge(s) (14, 15). Although it is not possible at the present timeto make unequivocal assignments of these bands, the maximaaround 276.5 nm with the accompanying positive shoulderaround 281 nm most probably represent the O-OLb and 0-1'Lbtransitions of solvent-exposed tyrosine residue(s).The spectral region from 270 to 255 nm depicts transitions

originating from phenylalanine residue(s), disulfide bridge(s),and tryptophan residue(s) (14, 15). The well-defined pairedmaxima at 267 and 260-261 nm probably correspond to the 0-O'Lb and 0-1'Lb transitions of phenylalanine moieties. Thesharp minimum at 263 nm is consistent with contributions fromother phenylalanine residue(s). The disulfide bridge(s) and the1La transition of tryptophan residue(s) are likely to play a sig-nificant role in generating the overall ellipticity within this re-gion because the chiral absorption of these groups is maximalaround 268 nm (14, 15).We initially attempted to characterize the mucopolysaccha-

ride-induced conformational transitions within the protease in-hibitor that are responsible for the acceleration of factor Xa-antithrombin but not thrombin-antithrombin interaction.§ Tothis end, sufficient amounts of anticoagulant octasaccharide,decasaccharide, dodecasaccharide, or tetradecasaccharide wereindividually mixed with human antithrombin to achieve satu-ration of the protease inhibitor that was in excess of 98%. Itshould be emphasized that each of these heparin fragments iscapable of augmenting the velocity of factor Xa neutralizationby antithrombin but is not able to influence the kinetics of in-activation of thrombin by antithrombin.The circular dichroism spectra of the four oligosaccharide-

antithrombin complexes in the near ultraviolet region are de-picted in Fig. 1 Left. The average baseline variations obtainedwith the oligosaccharides alone expressed in molar ellipticityunits of antithrombin are shown as bars along the zero line.Formation of the various oligosaccharide-antithrombin com-plexes is associated with a dramatic increase in chiral absorptionacross the whole near ultraviolet region. The enhancements atwavelengths above 290 nm are maximal for decasaccharide,dodecasaccharide, or tetradecasaccharide complexes. The changes

§ The electronic transitions associated with the chromophoric groupsof heparin could couple with those of the protein and might result insignificant contributions to the circular dichroism spectra. However,the spectral transitions of the mucopolysaccharide occur in the farultraviolet region. Given that the heparin-induced spectral alterationsare quite large in the near ultraviolet transitions ofthe aromatic groupsofantithrombin, it is unlikely that the above phenomenon contributessignificantly to the chiral absorption changes outlined in this paper.

obtained with octasaccharide interaction product are about 80%of the change observed with the above species. All four oligo-saccharide-protease inhibitor complexes possess very similarmolar ellipticities between 290 and 255 nm. Furthermore, thefine structures of the various peaks within the entire near ul-traviolet region as compared to those of antithrombin revealedeither no shift in the resolved wavelength maxima or minimaor only slight red shifts of <1 nm.We subsequently tried to identify the heparin-induced con-

formational alterations within the protease inhibitor that arerelated to the acceleration of thrombin-antithrombin interac-tions as well as those of factor Xa and antithrombin. For thispurpose, the required levels of anticoagulantly active octadeca-saccharide, LMW heparin, and HMW heparin were individ-ually added to aliquots of human antithrombin to attain satu-rations of the protease inhibitor that were in excess of 98%. Itshould be pointed out that the three heparin species are ableto dramatically enhance the velocities of thrombin and factorXa neutralization by antithrombin.The circular dichroism spectra of the three mucopolysac-

charide-antithrombin complexes and that of free protease in-hibitor in the near ultraviolet region are shown in Fig. 1 Right.The average baseline change obtained with HMW heparinalone expressed in molar ellipticity units ofantithrombin is alsoprovided (Fig. 1 Left). Generation ofthe three interaction prod-ucts resulted in a similar significant augmentation of chiral ab-sorption across the entire near ultraviolet spectrum. In addi-tion, the fine structure of the various peaks as compared to thatof the free protease inhibitor showed either no shift in the re-solved wavelength maxima or minima or slight red shifts ofsl nm.The circular dichroism spectra of the four oligosaccharide

complexes are similar in most respects to the chiral absorptionpatterns of the three mucopolysaccharide interaction products.Both types of complexes are characterized by large increasesin positive ellipticity that begin at wavelengths exclusively as-sociated with the 'La transitions of tryptophan residue(s). Inaddition, both sets of interaction products exhibit considerableaugmentation in chiral absorption within the spectral regionsthat contain transitions of tryptophan, tyrosine, phenylalanine,and cystine residues. Furthermore, the two classes of com-plexes also show significant change in the putative negativebands at 290, 282.5, 269.5, and 263 nm when compared to freeprotease inhibitor. These latter alterations may, in part, be sec-ondary to small shifts in peak position.The circular dichroism spectra of the four oligosaccharide

complexes are also significantly different from the chiral ab-sorption pattern of the three mucopolysaccharide interactionproducts. Indeed, the former complexes show increased molarellipticities throughout the 300-255 nm region when comparedto the latter interaction products.To more easily define the conformational alterations that oc-

cur when oligosaccharides or mucopolysaccharides bind to an-tithrombin, we generated the circular dichroism differencespectra of the various complexes by subtracting the averagechiral absorption curve ofthe protease inhibitor from that oftheindividual interaction products (Fig. 2).

It is apparent that all of the complexes exhibit to a greateror lesser extent several distinct difference bands. First, thereis positive difference peak at 296 nm that probably representsthe 0-1'La transition of buried tryptophan residue(s). The 0-0'La transition at 301 nm could be obscured by the large positiveenvelope, but the shoulder at 288-289 nm may constitute the0-3 transition ofthis series (16). Second, a prominent differencepeak is observed at 290 nm with an accompanying maximumat 282.5 nm that suggests perturbations of an exposed trypto-

719Z Biochemistry: Stone et aL

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

16, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982) 7193

0

x

10

xc)

270 280 290

Wavelength, nm

FIG. 2. Difference circular dichroism spectra. The difference chiralabsorption for each complex relative to free antithrombin, A[6I, wascalculated from data provided in Fig. 1. A, Octasaccharide complex;o, decasaccharide complex; o, dodecasaccharide complex; *, tetrade-casaccharide complex; *, octadecasaccharide complex; x, LMW hep-arin complex; and m, HMW heparin complex. (Lower) Pattern, A[6l1,generated by subtracting the average of a total of five oligosaccharidecomplexes of the kind depicted in Fig. 1 Left (octasaccharide excluded)from the average of a total of five mucopolysaccharide complexes ofthe type shown in Fig. 1 Right.

phan group(s) (16). Third, two sharp difference maxima arenoted at 269.5 and 263 nm that result from the summed positiveand negative contributions of the 'La transition of tryptophanresidue(s), the spectral alterations originating from disulfidebridge(s), and the changes in chiral absorption arising fromphenylalanine residue(s) (16).

It is also clear that heparin fragment-antithrombin com-

plexes, which are able to dramatically accelerate inactivation offactor Xa, exhibit difference bands of greater ellipticity in boththe 296- to 282-nm and the 270- to 255-nm regions comparedto the mucopolysaccharide-protease inhibitor interaction prod-ucts, which are capable of rapidly neutralizing thrombin as wellas factor Xa.

To readily appreciate the spectral characteristics that distin-guish between these two functional types of oligosaccharide or

mucopolysaccharide-antithrombin complexes, we have sub-tracted the average oligosaccharide circular dichroism curve

computed by employing data in Fig. 1 Left from the averagemucopolysaccharide chiral absorption pattern calculated byutilizing information in Fig. 1 Right (see Fig. 2 Lower). Similarmanipulations have also been conducted with the individualspectra obtained from HMW heparin and dodecasaccharide in-teraction products (see bottom of Fig. 1 Right). These profilesofchiral absorption versus wavelength should depict mucopoly-saccharide-induced conformational alterations in antithrombin

that are associated with the ability of the protease inhibitor torapidly neutralize thrombin. These spectra exhibit a shallownegative band between 300 and 275 nm that may be resolvedthrough a relatively large scatter of data into possible negativeminima at 290 and 283 nm as well as a rather deep negative bandbetween 275 and 255 nm with superimposed possible negativeminima at 268 and 262 nm. These alterations may arise fromconformational changes ofdisulfide bridge(s) or negative 'La and'Lb transitions of tryptophan residue(s) (14-16). Furthermore,minor contribution to the circular dichroism curves below 270nm may originate from phenylalanine residue(s) (14-16).

DISCUSSIONWe have randomly cleaved heparin preparations by appropriatechemical techniques and have isolated various oligosaccharidesby gel filtration chromatography. These molecular species werefurther purified by affinity fractionation utilizing antithrombinto yield octasaccharide, decasaccharide, dodecasaccharide, tet-radecasaccharide, and octadecasaccharide fragments. Oligosac-charides from octasaccharide to tetradecasaccharide were ca-pable of dramatically accelerating factor Xa-antithrombininteractions but not thrombin-antithrombin interactions. Oc-tadecasaccharide was able to enhance the velocity of thrombinas well as factor Xa neutralization by the protease inhibitor.These results are similar to those previously reported and havebeen rationalized on the basis of multiple functional domainswithin the mucopolysaccharide (6).The first of these domains contains a unique tetrasaccharide

sequence of two nonsulfated uronic acid residues and a N-ace-tylglucosamine group as well as a 3,6-disulfated glucosaminemoiety that allows the heparin molecule to bind tightly to an-tithrombin (3-5). As shown in a prior communication (6), thisarea of the mucopolysaccharide is closely linked to the regionof heparin that is responsible for accelerating factor Xa-anti-thrombin interactions. The second of these domains is presentas an oligosaccharide segment extending from the nonreducingend of the tetrasaccharide sequence and is able to directly "ac-tivate" antithrombin with respect to thrombin, factor IXa, orfactor Ma neutralization as well as to "approximate" the freeenzymes with the protease inhibitor (6). These different func-tional aspects of the second domain of the mucopolysaccharidecan be dissected by utilizing recently developed kinetic tech-niques (6). The anticoagulant hexadecasaccharide, which pos-sesses only the "activation" portion of the second domain, is notincluded in the present study because this molecular specieswas significantly contaminated with tetradecasaccharide. How-ever, we have also prepared LMW heparin and HMW heparinthat have not been exposed to the chemical cleavage procedureemployed above (12). The LMW heparin contains a single bind-ing site for antithrombin, whereas the HMW heparin possessestwo independent binding sites for the protease inhibitor. Bothof these mucopolysaccharides are able to accelerate neutrali-zation of thrombin as well as factor Xa by antithrombin (12).The overall structure of antithrombin and its interactions

with polydisperse forms of heparin have previously been ex-amined by a variety of physical techniques such as circular di-chroism spectroscopy, ultraviolet difference spectroscopy, andfluorescence spectroscopy (11, 17-20). These detailed studieshave revealed that half of the two tryptophan residues and fourtyrosine residues are located on the surface of antithrombin,whereas the remaining half of each of these two types of groupsare buried within the protease inhibitor (17, 19, 20). Duringformation of the mucopolysaccharide-antithrombin complex,an exposed tryptophan residue in the heparin-recognition siteis brought into close proximity with the highly charged ligand

Biochemistry: Stone et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

16, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982)

(18, 21) and the conformation of buried tryptophan residue(s)is altered as part ofa putative structural transition (19, 20). Priorto our study, it was unclear whether the above molecular eventswere related to the binding of mucopolysaccharide with anti-thrombin or to the activation of the protease inhibitor with re-spect to the different enzymes of the coagulation cascade.

Utilizing the family of complex carbohydrates describedabove and- circular dichroism spectroscopy, we have attemptedto show that interaction ofone or both ofthe functional domainsof the mucopolysaccharide with antithrombin results in differ-ent conformational states of the protease inhibitor.

Our studies demonstrate that the various oligosaccharide-or mucopolysaccharide-antithrombin complexes exhibit twotypes of circular dichroism spectra. The first of these patternsis observed when octasaccharide, decasaccharide, dodecasac-charide, or tetradecasaccharide interacts with the protease in-hibitor. The chiral absorption patterns of the complexes, com-pared to free antithrombin, indicate perturbations of bothexposed and buried tryptophan residue(s) as evidenced by pos-itive difference bands at 296 and 288 nm as well as 290 and 282.5nm (Fig. 2). In addition, our preliminary (unpublished) inves-tigation of the above interaction products compared to free an-tithrombin shows a positive achiral absorption peak at 298 nmas well as negative achiral absorption troughs at 291 and 280 nm(ultraviolet difference spectroscopy) and a 40% augmentationof intrinsic, fluorescence at 340 nm (fluorescence differencespectroscopy). These observations strengthen. our conclusionthat the above heparin fragments are able to induce changes inboth exposed and buried tryptophan moieties of the proteaseinhibitor (14, 15). Furthermore, the chiral absorption spectraofoligosaccharide-antithrombin.complexes as compared to freeprotease inhibitor exhibit positive difference bands at 269 and263 nm that probably represent contributions from the. 'La tran-sition of tryptophan residue(s), the spectral alterations origi-nating from disulfide bridge(s), and the changes in chiral ab-sorption arising from phenylalanine residue(s) (14, 15, 22).

It is apparent from the above description of the circular di-chroism pattern and preliminary data obtained, by ultravioletspectroscopy as well.as fluorescence spectroscopy that most ofthe previously documented heparin-induced alterations of an-tithrombin are observed with octasaccharide, decasaccharide,dodecasaccharide, and tetradecasaccharide. Therefore, theseconformational changes must be associated with the binding ofthe first domain. of the mucopolysaccharide to antithrombin orthe activation of the protease inhibitor with respect to neutral-ization of factor Xa but not thrombin.,The second ofthese circular dichroism patterns isnoted when

octadecasaccharide, LMW heparin, or HMW heparin interactswith antithrombin. The chiral absorption spectra of these com-plexes compared to free protease inhibitor is similar to the firstpattern outlined above except for changes within the 292- to282-nm and 275- to 255-nm regions. The subtraction of the av-eraged first type.of circular dichroism spectra from the averagedsecond type ofcircular dichroism spectra indicates that the sec-ond domain of the mucopolysaccharide induces a characteristicpattern of chiral absorption (bottom of Fig. 1 Right and Fig. 2Lower). These spectral profiles probably arise from conforma-tional changes of a disulfide bridge(s) (14, 15). This hypothesisis supported by our preliminary results obtained with ultravioletdifference spectroscopy and fluorescence spectroscopy, whichreveal no significant distinctions between the two major typesof mucopolysaccharide-antithrombin interaction products.However, we cannot completely exclude the possibility that theabove circular dichroism curve might be explainedon the basis

of transitions originating from a second surface tryptophanresidue.

It should be emphasized that, independent of the origins ofthe spectral changes outlined above, these transitions are.strongly correlated with the binding of the second domain ofthe mucopolysaccharide to antithrombin. Indeed, the additionof a tetrasaccharide to the tetradecasaccharide endows the re-sultant oligosaccharide with the ability to accelerate thrombin-antithrombin interactions as well as the capacity to induce thesecond type of spectral transition.. Furthermore, heparin spe-cies of various molecular sizes that possess.both.classes offunc-tional domains exhibit similar circular dichroism spectra. Theseobservations provide direct experimental evidence for our hy-pothesis that heparin exerts its multiple anticoagulant actionsby binding to different regions ofantithrombin and thereby in-ducing different structural alterations of the protease inhibitor.The potential ability of this system to differentially inactivatethe various enzymes of the coagulation cascade suggests that itmay represent a critical control point for modulating the he-mostatic mechanism.

We are pleased to acknowledge our gratitude to Prof. Elkan Blout(Harvard Medical School) for his interest and generosity in providingus with a stimulating environment and the use ofhis laboratory facilitiesduring part of this investigation. This work was supported by NationalInstitutes of Health Grants HL-28625 and GM26625.

1. Lam, L. H., Silbert, J. E. & Rosenberg, R. D. (1976) Biochem.Biophys. Res. Commun. 69, 570-577.

2. Ho6k, M., Bjbrk, I., Hopwood, J. & Lindahl, U. (1976) FEBSLett. 66, 90-93.

3. Rosenberg, R. D., Armand, G. & Lam, L. H. (1978) Proc. NatlAcad; Sci. USA 75, 3065-3069.

4. Rosenberg, R. D. & Lam, L. H. (1979) Proc. Natl Acad. Sci. USA76, 1218-1222.

5. Lindahl, U., B6ckstrom, G., Thunberg, L. & Leder, I. G. (1980)Proc. NatW Acad. Sci. USA 77, 6551-6555.

6. Oosta, G. M., Gardner, W. T., Beeler, -D. L. & Rosenberg, R.D. (1981) Proc. NatW Acad. Sci. USA 78, 829-833.

7. Rosenberg, R. D. & Damus, P. S. (1973) J. Biol Chem. 248,6490-6505.

8. Jordan, R. E., Oosta, G. M., Gardner,.W. T. & Rosenberg, R.D. (1980) J. Biol Chem. 255, 10073-10080.

9. Nordenman, B., Nystrom, C. & Bjork, I. (1977) Eur.J. Biochem.78, 195-203.

10. Bitter, T. & Muir, H. M. (1962) Anal. Biochem. 4, 330-334.11. Jordan, R. E., Beeler, D. & Rosenberg, R. D. (1979) J. Biol

Chem. 254, 2902-2914.12. Jordan, R. E., Favreau, L. V., Braswell, E. H. & Rosenberg, R.

D. (1982) J. BioL Chem. 257, 400-406.13. Petersen, T. E., Dudek-Wojciechowska, G., Sottrup-Jensen, L.

& Magnusson, S. (1979) in The Physiological Inhibitions of Co-agulation and Fibrinolysis, eds. Collen, D., Wilman, B. & Ver-straete, M. (Elsevier/North-Holland, Amsterdam), pp. 43-54.

14. Strickland, E. H. (1974) Crit. Rev. Biochem. 2, 113-175.15. Colman, D. L. & Blout, E. R. (1967) in Conformations of Bio-

polymers, ed. Ramachandran, G. N. (Academic, London), Vol.1, p. 23.

16. Strickland, E. H., Horowitz, J. & Billups, C. (1969) Biochemistry8, 3205-3213.

17. Villaneuva, G. B. & Danishefsky, I. (1977) Biochem. Biophys.Res. Commun. 74, 803-809.

18. Nordenman, B. & Bjork, I. (1978) Biochemistry 17, 3339-3344.19. Olson, S. T. & Shore, J. D. (1981) J. Biol. Chem. 256, 11065-

11072.20. Olson, S. T., Srinivasan, K. R., Bj6rk, I. & Shore, J. D. (1981)

J. Biol Chem. 256, 11073-11079.21. Blackburn, M. & Sibley, C. S. (1980) J. Biol Chem. 255, 824-

826.22. Longas, M. O., Ferguson, W. S. & Finley, T. H. (1980) J. Biol.

Chem. 255, 3436-3441.

7194 Biochemistry: Stone et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

16, 2

020