accelerating ability of synthetic oligosaccharides on antithrombin inhibition of proteinases of the...

© 2004 Schattauer GmbH, Stuttgart

Accelerating ability of synthetic oligosaccharides onantithrombin inhibition of proteinases of the clotting andfibrinolytic systemsComparison with heparin and low-molecular-weight heparin

Steven T. Olson1, Richard Swanson1, Elke Raub-Segall2,Tina Bedsted1, Morvardi Sadri1,Maurice Petitou3, Jean-Pascal Hérault3, Jean-Marc Herbert3, Ingemar Björk2

1Center for Molecular Biology of Oral Diseases, University of Illinois-Chicago, Chicago, Illinois, USA2Department of Molecular Biosciences, Swedish University of Agricultural Sciences, Uppsala Biomedical Center, Uppsala, Sweden3Haemobiology Research Department, Sanofi Recherche,Toulouse, France

Thromb Haemost 2004; 92: 929–39

rate constants of ~106 – 107 M-1·s-1, although thrombin appearsto be the more important target. In contrast, factor IXa inhibi-tion is appreciably less stimulated. The conformational changeof antithrombin induced both by the pentasaccharides andlonger heparins contributes substantially, ~150–500-fold, toaccelerating the inactivation of factors Xa, IXa and VIIa andmoderately, ~50-fold, to that of factor XIIa and tissue plasmino-gen activator inhibition. The bridging effect due to binding ofantithrombin and proteinase to the same, long heparin chain isdominating, ~1000–3000-fold, for thrombin inhibition and isappreciably smaller, although up to ~250-350-fold, for the inactivation of factors IXa and XIa.These results establish theproteinase targets of heparin derivatives currently used in orconsidered for thrombosis therapy and give new insights intothe mechanism of heparin acceleration of antithrombin inhibi-tion of proteinases.

KeywordsHeparin, pentasaccharide, hexadecasaccharide, antithrombin,coagulation proteinases

SummaryThe abilities of three synthetic oligosaccharides to accelerateantithrombin inhibition of ten clotting or fibrinolytic proteinas-es were compared with those of unfractionated, fractionatedhigh-affinity and low-molecular-weight heparins. The resultsshow that the anticoagulant effects of the latter three heparinsunder conditions approximating physiologic are exerted almostexclusively by acceleration of the inactivation of thrombin,factor Xa and factor IXa to near diffusion-controlled rate constants of ~106 – 107 M-1·s-1. All other proteinases are in-hibited with at least 20-fold lower rate constants. The anti-coagulant ability of the synthetic regular (fondaparinux) andhigh-affinity (idraparinux) pentasaccharides is due to a commonmechanism, involving acceleration of only factor Xa inhibitionto rate constants of ~106 M-1·s-1. A synthetic hexadecasaccha-ride, containing both the pentasaccharide sequence and a pro-teinase binding site, exerts its anticoagulant effect by accelerat-ing antithrombin inactivation of both thrombin and factor Xa to

Blood Coagulation, Fibrinolysis and Cellular Haemostasis

929

Correspondence to:Ingemar BjörkDepartment of Molecular BiosciencesSwedish University of Agricultural SciencesUppsala Biomedical CenterBox 575, SE-751 23 Uppsala, SwedenTel.: +46 18 4714191, Fax +46 18 550762E-mail: [email protected]

Received June 18, 2004Accepted after revision September 3, 2004

Financial support:This work was supported by National Institutes of Health Grant HL39888 (S.T.O.),

Swedish Scientific Council–Medicine Grant 4212 (I.B.) and by a Research Grantfrom Sanofi-Synthélabo (S.T.O. and I.B.)

Prepublished online October 5, 2004 DOI: 10.1160/TH04-06-0384

For personal or educational use only. No other uses without permission. All rights reserved.Downloaded from www.thrombosis-online.com on 2013-03-10 | ID: 1001037435 | IP: 152.3.102.242

Olson, et al.: Antithrombin activation by therapeutic heparin derivatives

Introduction

The long-established use of the sulfated polysaccharide,heparin, for the prophylaxis and treatment of venous throm-bosis is based on the ability of heparin to bind to and activate theplasma proteinase inhibitor, antithrombin, and thereby to accel-erate antithrombin inactivation of blood clotting proteinases (1, 2). About a third of all heparin chains, designated high-affin-ity heparin, in commercial heparin preparations is responsiblefor nearly all of this effect by virtue of binding tightly toantithrombin through a specific pentasaccharide sequence (3).The binding of the pentasaccharide allosterically activatesantithrombin by a conformational change, which promotes theinitial recognition of certain clotting proteinases, most impor-tantly factor Xa, and thereby increases the rate of inactivation ofthese enzymes (4-9). In contrast, the conformational changeonly minimally affects the rate of inhibition of other proteinas-es, primarily thrombin. Bridging of antithrombin and thrombinby both binding to the same, at least 18 saccharide units long,pentasaccharide-containing heparin chain instead predominant-ly accounts for the increased rate of reaction with this enzyme(4, 10, 11). Recent work has shown that both the conformation-al change and bridging effects contribute substantially to theheparin-induced increase in the rate of antithrombin inactiva-tion of factor IXa (12, 13).

Because of its highly anionic nature, heparin is involved inmany interactions with blood proteins other than antithrombinand with vessel wall components (14), resulting in an appreci-able individual variability in optimal dosage and an associatedrisk for bleeding complications. Such interactions may also leadto heparin treatment being accompanied by side effects likethrombocytopenia and osteoporosis (15, 16). These problemshave inspired efforts to design safer and more specific therapeu-tic agents that would not be associated with such complications.An initial advantage was provided by low-molecular-weightheparins, which are produced from standard heparin by chemi-cal or enzymatic cleavage, giving chains about one-third thesize (17). This heparin form has an increased bioavailability andis easier to administrate than standard heparin but is not entire-ly devoid of the side effects discussed above.

Progress in complex oligosaccharide chemistry has recentlygiven rise to the synthesis of chemically pure, heparin-relatedoligosaccharides, the structures of which have been optimizedto promote selective interactions with certain proteins andreduce undesired interactions with other biological components(18, 19). Fondaparinux, a synthetic analogue of the antithrom-bin-binding pentasaccharide of heparin (20) was recently shownto be a more effective anticoagulant than standard or low-molecular-weight heparins in phase III clinical trials (21-25)and thus appears to have a therapeutic advantage over thesecompounds. Idraparinux, another synthetic pentasaccharidevariant with an increased affinity for antithrombin, is similarly

potent and has a longer half-life in plasma (26). A heparin-likehexadecasaccharide, containing both the pentasaccharidesequence and a charged proteinase binding site separated by anuncharged domain, shows an antithrombin-dependent ability toinhibit thrombin and factor Xa comparable to that of standardheparin and may prove to lack many side effects of the hetero-geneous polysaccharide (27).

The accelerating ability of the synthetic penta-and hexade-casaccharides of putative therapeutic importance on antithrom-bin inhibition of most clotting and fibrinolytic proteinases hasbeen insufficiently or not at all characterized. Moreover, in spiteof an appreciable body of work on heparin function, the accel-erating abilities of full-length heparin on antithrombin inhibi-tion of several clotting proteinases and related enzymes are stilluncertain or unknown. In the case of many of these proteinases,the relative contributions of conformational activation ofantithrombin and approximation of inhibitor and enzyme to theheparin-induced rate increase are also uncertain. The aim of thiswork was to determine the accelerating effects of fondaparinux,idraparinux and the hexadecasaccharide on antithrombin inhibi-tion of all major clotting and fibrinolytic proteinases. Theseaccelerating effects were compared with those of two high-affinity heparins having lengths of ~26 and ~50 saccharide unitsand with those of unfractionated heparin and low-molecular-weight heparin.

Materials and methods

SaccharidesRegular pentasaccharide (SR90107A; fondaparinux, Arixtra®)(20), high-affinity pentasaccharide (Sanorg34006; idraparinux)(26), hexadecasaccharide (SR123781A) (19,27), unfractionatedheparin (Heparin Na) and low-molecular-weight heparin(Nadroparine Na) were products of Sanofi-Synthélabo Research(Toulouse, France). Full-length heparins with high-affinity forantithrombin and with reduced polydispersity and averagemolecular masses of ~8000 and ~15000 (~26 and ~50 saccha-ride units, respectively) were purified from commercial heparinby gel chromatography and affinity chromatography on matrix-linked antithrombin, as described (28, 29).

Antithrombin and coagulation proteinasesAntithrombin was purified from human plasma by affinity chro-matography on matrix-linked heparin, followed by successiveion-exchange and gel chromatographies (29). Concentrations ofthe protein were determined by absorption measurements at 280 nm from a molar absorption coefficient of 37 700 M-1 cm-1

(30).Human thrombin, factor Xa (predominantly α-form), factor

IXaβ, factor XIa, factor α-XIIa, factor VIIa, plasma kallikrein(β-form) and activated protein C were obtained from EnzymeResearch (South Bend, IN, USA). Human Lys-plasmin was

930For personal or educational use only. No other uses without permission. All rights reserved.

Downloaded from www.thrombosis-online.com on 2013-03-10 | ID: 1001037435 | IP: 152.3.102.242


prepared by activating Glu-plasminogen, glycoform I, withurokinase, followed by purification on soybean trypsininhibitor-agarose (31). Human two-chain tissue plasminogenactivator was obtained from Molecular Innovations (Southfield,MI, USA). Nominal enzyme concentrations were provided bythe manufacturers or, in the case of plasmin, determined byactive-site titration with fluorescein mono-p-guanidinobenzoate(32). Human recombinant soluble tissue factor was a gift fromDr. Yale Nemerson (Mount Sinai School of Medicine, NewYork, NY, USA).

Experimental conditionsAll experiments were done at 25 ± 0.2°C. The buffer in experi-ments in the absence of Ca++ was 0.02 M sodium phosphate,100 µM EDTA, 0.1 % PEG 8000, pH 7.4, with sodium chlorideadded to ionic strengths of 0.1 or 0.15. Analyses in the presenceof Ca++ were done in 0.1 M Hepes, 5 mM CaCl2, 0.1 % PEG8000, pH 7.4, with sodium chloride added to an ionic strengthof 0.15.

Functional concentrations of heparin saccha-rides and affinities of saccharide binding toantithrombinConcentrations of antithrombin-binding species in the prepara-tions of the different heparin saccharides and Kd’s for the bind-ing of saccharides to antithrombin were measured by titrations,monitored by the fluorescence increase induced by the binding,of antithrombin with the saccharides at antithrombin concentra-tions much greater than Kd (0.5-1 µM) or approximating Kd

(0.05-0.2 µM), respectively, in the absence or presence of Ca++

at I 0.1 or 0.15 (29).

SDS–PAGE of proteinases and antithrombin-proteinase complexesThe purity of the proteinases and their ability to form SDS-stable complexes with antithrombin were assessed by SDS-PAGE (33, 34) under nonreducing and reducing conditions.Complexes were formed by incubating 2 µM proteinase with 5 µM antithrombin (except for activated protein C in which case both concentrations were 9-fold higher) in the absence orpresence of Ca++ and without or with 5 µM high-affinity heparinfor reaction times expected to result in trapping of >98% (~80%for activated protein C) of the enzyme in complex with theinhibitor, based on measured rate constants. Unreacted pro-teinase was inactivated with 100-500 µM of appropriate tripep-tide chloromethylketones (obtained from Bachem, Bubendorf,Switzerland or Novabiochem, La Jolla, CA, USA) or 1 mMamidinophenylmethanesulfonyl fluoride prior to SDS-PAGEanalysis. All enzymes were found to be >90% pure by SDS-PAGE. Moreover, the factor α-XIIa preparation contained nodetectable factor β-XIIa and the kallikrein preparation con-tained >90% β-form and <10% α-form. All enzymes formed

SDS-stable complexes with antithrombin in a quantitative ornear-quantitative manner, indicating that the enzyme prepara-tions contained at most small amounts of inactive protein.

Concentrations of active proteinasesActive concentrations of most enzymes were assessed by titra-tions with antithrombin, essentially as in previous work (29, 35). Titrations were done without Ca++ in the absence, andin some cases also in the presence, of heparin, except for factorsIXa and VIIa which required both heparin and Ca++ to ensurecompletion of the reactions. SDS-PAGE showed a minimal sac-charide-induced consumption of antithrombin via the substratepathway (36-38) for those enzymes titrated in the presence ofheparin, indicating that such titrations are reasonably accurate.A series of samples of each enzyme at a constant concentrationwas incubated with increasing amounts of antithrombin inmolar ratios to the enzyme of 0-2. Heparin or soluble tissue factor (when included) were added at a 1-4-fold molar ratio tothe enzyme. After reaction times chosen to result in an essen-tially complete reaction, based on measured rate constants, thesamples were diluted into 100-500 µM solutions of the appro-priate chromogenic substrates (from Chromogenics, Mölndal,Sweden or American Diagnostica, Greenwich, CT, USA.).These were S-2238 for thrombin, Spectrozyme fXa for factorXa, Spectrozyme fIXa for factor IXa, S-2366 for factor XIa,Spectrozyme fXIIa for factor XIIa, Spectrozyme fVIIa for factor VIIa, Spectrozyme Kall or S-2302 for plasma kallikrein,and S-2251 for plasmin. Substrates were dissolved in the stan-dard sodium phosphate buffer in the case of most enzymes butin 0.1 M Hepes, 0.1 M NaCl, 10 mM CaCl2, 33 % (v/v) ethyl-ene glycol, pH 8.0 (39) and 0.1 M Hepes, 0.1 M NaCl, 10 mMCaCl2, pH 7.4 in the analyses of factor IXa and factor VIIa,respectively, to increase assay sensitivity. Polybrene, at a con-centration of 50 µg/ml, was also included in titrations withheparin to neutralize the saccharide. The absorbance increase at405 nm, reflecting the residual enzyme activity, was measuredfor several minutes. Concentrations of proteinase active in bind-ing to antithrombin were obtained from the intercepts on theabscissa of the linear segments of plots of the residual enzymeactivity vs. antithrombin concentration. The active concentra-tion of tissue plasminogen activator was determined by a simi-lar titration of 5-10 nM proteinase with plasminogen activatorinhibitor-1 (Molecular Innovations), after establishing the activeconcentration of the latter by titration with active-site-titrated β-trypsin (40). The concentration of activated protein C wasassumed to be that given by the manufacturer, as this proteinasecould not be titrated in the above manner because of its slowreaction with antithrombin.




Kinetics of proteinase inactivation by freeantithrombin and complexes of antithrombinwith heparin saccharidesSecond-order rate constants for reactions of antithrombin andantithrombin-saccharide complexes with proteinases in theabsence or presence of Ca++ were determined essentially as inprevious work (29, 35, 41-43). The proteinase (0.01-1 µM) wasallowed to react with antithrombin at a concentration (0.1 to 100 µM) at least ten-fold higher than that of the proteinase,ensuring pseudo-first-order conditions, in the absence or pres-ence of heparin saccharide. Reactions were done with a range ofcatalytic saccharide concentrations, i.e. at molar ratios ofantithrombin to saccharide >10, when rate constants exceeded~105 M-1s-1 and otherwise with saccharide concentrations com-parable with the concentrations of the inhibitor, i.e., at molarratios of antithrombin to saccharide from 1.2-2.5. All saccharideconcentrations were those of antithrombin-binding species,determined by fluorescence titrations. Concentrations ofantithrombin-saccharide complexes were calculated from meas-ured dissociation constants for complex formation.

Reactions were initiated by addition of the proteinase toantithrombin without or with saccharide and were terminatedafter different times by a further addition of chromogenic sub-strate as in the analyses of active proteinase concentrations. The

substrate for activated protein C was 100 µM Spectrozyme PCa(American Diagnostica) in the standard sodium phosphatebuffer. Observed pseudo-first-order rate constants (kobs) weregenerally obtained by fitting the decrease in enzyme activitywith time to a single-exponential function with an end-point ofzero activity or, in some cases (for factor IXa, factor VIIa andplasma kallikrein), with a finite end-point, based on the ob-servation of a small fraction (typically <10%) of enzyme lesssusceptible to inhibition and likely due to degraded, althoughstill active, forms of the enzyme (44, 45). Reactions of factorXIIa with antithrombin–full-length heparin complexes requiredfitting by a double-exponential decay function with a zero end-point. Second-order rate constants for antithrombin-proteinasereactions in the absence of heparin saccharides were obtainedby dividing kobs by the inhibitor concentration. Second-orderrate constants for reactions of the antithrombin-saccharide com-plexes with the proteinases were determined by first subtractingthe calculated contribution of the reaction of free antithrombinto kobs and then dividing by the concentration of the antithrom-bin-saccharide complex. In both cases, several values, usuallydetermined at different antithrombin or antithrombin-saccharidecomplex concentrations, were averaged. It should be noted thatbecause of the pseudo first order conditions used, measured rateconstants should be minimally affected by even appreciableerrors in the proteinase concentration.

Results

Antithrombin-saccharide dissociation constantsDissociation equilibrium constants, Kd, for the binding of thedifferent heparin saccharides to antithrombin, necessary for calculation of concentrations of antithrombin-saccharide com-plexes, were measured by fluorescence titrations in the absence,and for some saccharides also in the presence, of Ca++ (Table 1).The Kd values for unfractionated heparin and low-molecular-weight heparin should only be considered to be apparent, average values, as the preparations of the two forms containboth high-affinity heparin species of varying lengths and affini-ties, as well as a large proportion of low-affinity species. Ca++

had a negligible effect on the affinity of both the regular pentasaccharide and high-affinity heparin with ~50 saccharidesand was therefore assumed also not to affect the Kd for unfrac-tionated and low-molecular-weight heparins.

Association rate constants for proteinase inhibition by free antithrombin andantithrombin-saccharide complexesSecond-order association rate constants were measured for theinactivation of proteinases by antithrombin alone and by com-plexes of the inhibitor with the different heparin saccharides(Table 2). The ability of each heparin saccharide to accelerate

932

Table 1: Dissociation equilibrium constants (nM) for the binding of different heparin saccharides to humanantithrombin in the absence and presence of Ca++ at25°C, pH 7.4, I 0.15. Dissociation constants for binding of regular pentasaccharide (fondaparinux) (H5), hexadecasaccharide(H16), full-length heparin with ~26 saccharide units (H26), full-length heparin with ~50 saccharide units (H50), unfractionatedheparin (UFH) and low-molecular-weight heparin (LMWH) toantithrombin were determined by fluorescence titrations, asdescribed in Materials and methods.The buffer was 20 mM sodium phosphate, 100 µM EDTA, pH 7.4, I 0.15 in experimentswithout Ca++ and 0.1 M Hepes, 5 mM CaCl2, pH 7.4, I 0.15 inexperiments with Ca++, in both cases with 0.1 % (w/v) poly(ethylene glycol) 8000. Values are averages ± SEM of three to five measurements. The Kd for high-affinity pentasaccharide (idraparinux) (H5*) in the absence of Ca++ was estimated from a Kd, obtained by similar fluorescence titrations, of 27±4 nM at I 0.5. The slope of the ionic-strength-dependence of Kd for ahighly similar high-affinity pentasaccharide (DEFG’H* in (50))was used in the extrapolation of this value to I 0.15. ND, notdetermined.



proteinase inactivation is expressed as the ratio between the rateconstant for the antithrombin-saccharide complex and that forantithrombin alone (Table 2). The contribution of conformation-al activation of antithrombin to the accelerating effect of eachheparin saccharide is indicated by the rate increase produced by the regular pentasaccharide (fondaparinux) (Table 2). The contribution of approximation or bridging of proteinase andinhibitor by both binding to the same heparin chain to the accelerating effect is indicated by the rate increase given by

each heparin saccharide over that shown by the regular pentasaccharide.

Most data were obtained in the absence of Ca++ ions.However, the regular pentasaccharide-accelerated reaction ofantithrombin with factor IXa and the full-length heparin-accel-erated reactions of antithrombin with factors Xa and IXa wereshown to be stimulated by Ca++ (Table 3), in agreement withprevious work (12, 46, 47). In contrast, the acceleration by thesetwo saccharides of the reactions of antithrombin with all other

933

Table 2: Second-order association rate constants (M-1·s-1)for inhibition of human blood clotting and fibrinolysisproteinases by human antithrombin and its complexeswith different heparin saccharides at 25 °C, pH 7.4, I0.15. Association rate constants for reactions with proteinasesof free antithrombin (kuncat) and of complexes of the inhibitorwith regular pentasaccharide (fondaparinux) (kH5), high-affinitypentasaccharide (idraparinux) (kH5*), hexadecasaccharide (kH16),full-length heparin with ~26 saccharide units (kH26), full-lengthheparin with ~50 saccharide units (kH50), unfractionated heparin(kUFH) and low-molecular-weight heparin (kLMWH) were deter-mined as described in Materials and methods.The buffer was

20 mM sodium phosphate, 100 µM EDTA, pH 7.4, I 0.15 in mostexperiments but 0.1 M Hepes, 5 mM CaCl2, pH 7.4, I 0.15 in oneset of experiments with each of thrombin, factor Xa and factorIXa and in all experiments with factor VIIa without and with tissue factor; both buffers also contained 0.1 % (w/v) poly(ethylene glycol) 8000. Values are averages ± SEM of three toseven measurements, except where noted. Numbers in squarebrackets indicate the ratios between the rate constants for theantithrombin-saccharide complexes and that for antithrombinalone, reflecting saccharide accelerating ability. ND, not deter-mined.TF, tissue factor.



proteinases was either unaffected or somewhat impeded by Ca++ (Table 3). Therefore, more complete data sets for the saccharide-accelerated reactions with factors Xa and IXa, aswell as with thrombin, in the presence of Ca++ were alsoacquired. Moreover, all reactions with factor VIIa, without andwith tissue factor, were analysed in the presence of Ca++, as tissue factor requires Ca++ ions for high-affinity binding to factor VIIa (48).

Regular pentasaccharide (fondaparinux)Fondaparinux transforms antithrombin into a specific and rapidinhibitor of factor Xa. The complex between this pentasaccha-ride and antithrombin thus inhibited factor Xa considerablyfaster than any other enzyme, with a near diffusion-limited rateconstant approaching 106 M-1·s-1, due to a ~300-fold increase ofthe inhibition rate from that of the free inhibitor (Table 2) (4).The rate constant was not detectably affected by Ca++.

The conformational change of antithrombin induced by fondaparinux also leads to a substantial acceleration of the reactions of the inhibitor with several other coagulation andfibrinolysis proteinases. The rate constants for inhibition of factor IXa and of factor VIIa with or without tissue factor wereincreased 150-400-fold, i.e. to an extent approaching or evensomewhat exceeding that for the inhibition of factor Xa (Table2), in general agreement with previous work (12, 13, 49). Therate constant for factor IXa inhibition by the antithrombin-pentasaccharide complex was enhanced ~8-fold by Ca++, result-ing in the highest rate constant by which this complex inhibitedother proteinases than factor Xa (Table 2). Antithrombin inacti-vation of factor XIIa and tissue plasminogen activator was alsoaccelerated by an appreciable ~50-70-fold by fondaparinux,whereas an intermediate accelerating effect of 12-24-fold wasobserved for the reactions with kallikrein, activated protein Cand plasmin (Table 2). Only antithrombin inhibition of throm-bin and factor XIa was minimally, ~2-3-fold, enhanced by thispentasaccharide, as shown earlier for thrombin (4). The effectsof conformational activation of antithrombin on increasing theinhibition rates are thus markedly variable, depending on thetarget proteinase.

Although fondaparinux thus substantially accelerates thereaction of antithrombin with several proteinases, the rate constants for the inactivation of all other proteinases than factorXa by the antithrombin-pentasaccharide complex are >25-foldlower than that for factor Xa, <~3 × 104 M-1·s-1 (Table 2). Theserate constants imply that the inactivation of proteinases otherthan factor Xa contributes only little to the pharmacologiceffects of fondaparinux. The low rates attained in spite of someappreciable accelerations are due to antithrombin inhibition ofmost of these enzymes in the absence of any saccharide beingconsiderably slower than that of factor Xa. Only thrombin, fac-tor XIa, kallikrein and plasmin are inhibited by free antithrom-bin with rate constants approaching or exceeding that of factorXa, but fondaparinux has little effect on these reactions.

High-affinity pentasaccharide (idraparinux)The complex between idraparinux and antithrombin inhibitedall clotting and fibrinolysis proteinases studied with rate con-stants experimentally indistinguishable from those of the com-plex with fondaparinux (Table 2). When bound to antithrombin,the high-affinity and regular pentasaccharides thus activateantithrombin in the same manner. Like fondaparinux, idraparin-

934

Table 3: Second-order association rate constants (M-1·s-1) for inhibition of human blood clotting and fibri-nolysis proteinases by complexes of human antithrom-bin with regular pentasaccharide (fondaparinux) or full-length heparin with ~50 saccharide units in the presenceof 5 mM Ca++ at 25°C, pH 7.4, I 0.15. Association rate con-stants for reactions with proteinases of complexes of antithrom-bin with regular pentasaccharide (fondaparinux) (kH5) and full-length heparin with ~50 saccharide units (kH50) were determinedas described in Materials and methods.The buffer was 0.1 MHepes, 5 mM CaCl2, pH 7.4, with NaCl added to I 0.15 and with 0.1 % (w/v) poly(ethylene glycol) 8000.Values are averages± SEM of three to eight measurements, except where noted.Numbers in square brackets indicate the ratios between therate constants in the presence of Ca++ and those in the absenceof Ca++, taken from Table 2. Such ratios could not be derived forfactor VIIa without and with tissue factor, as the rate constantsfor the inhibition of this proteinase in the absence of calciumcould not be measured. ND, not determined.TF, tissue factor.



ux thus also converts antithrombin to an essentially specificinhibitor of factor Xa. The only difference between the two sac-charides is that idraparinux binds 35-250-fold more tightly toantithrombin (Table 1) (26), presumably mainly because of alower dissociation rate constant (50).

HexadecasaccharideIn contrast to the two pentasaccharides transforming antithrom-bin into a selective inhibitor of factor Xa, the synthetic hexade-casaccharide converts antithrombin into a rapid inhibitor of both factor Xa and thrombin. The rate constants for the reactions ofantithrombin with the two proteinases were thus both increasedby the hexadecasaccharide to the diffusion-limited, physiologi-cally and pharmacologically important range of 106-107 M-1·s-1

(Table 2). These high rate constants, not previously quantified,account for the reported ability of the hexadecasaccharide topromote antithrombin inhibition of the two enzymes (27). Ca++ had no effect on the hexadecasaccharide-catalyzed throm-bin inhibition rate but increased the rate of inhibition of factorXa by ~3-fold (Table 2). The high thrombin inactivation ratewas due to the hexadecasaccharide increasing the rate of reac-tion of the enzyme with antithrombin by ~4000-fold, i.e. ~2000-fold better than the regular pentasaccharide (fondaparinux)(Table 2). This effect arises from the hexadecasaccharide con-taining an additional sequence of four highly negatively chargedsaccharides on the nonreducing side of the specific pentasac-charide region, which allows thrombin to bind in an optimalmanner adjacent to antithrombin in a ternary bridging complex(19, 27). The hexadecasaccharide-accelerated rate constant forantithrombin inhibition of thrombin was comparable with thoseproduced by the ~50-saccharide high-affinity and unfractionat-ed heparins (Table 2). However, it was ~6-fold and ~2-foldhigher than those given by low-molecular-weight heparin andthe ~26-saccharide high-affinity heparin, respectively (Table 2),despite the average chain-length of the latter heparins beingcomparable with and even larger than that of the hexadecasac-charide (17). This difference is probably due to an appreciablefraction of the high-affinity chains of low-molecular-weightheparin and a smaller fraction of the ~26-saccharide high-affin-ity heparin chains having an insufficient extension of the chaintowards the nonreducing end from the pentasaccharide region toprovide a thrombin binding site. In contrast to thrombin inhibi-tion, the hexadecasaccharide increased the factor Xa inactiva-tion rate only slightly better than the regular pentasaccharide,~2-fold in the absence and ~6-fold in the presence of Ca++,reflecting only small contributions of bridging to the enhancedrate of inactivation of this proteinase (Table 2).

Similar to its effect on antithrombin inhibition of factor Xa,the hexadecasaccharide increased the rate of antithrombin inac-tivation of the remaining enzymes studied only slightly, in mostcases <~3-fold, better than the regular pentasaccharide (Table2). Most of these enzymes thus differ from thrombin in that the

additional, negatively charged saccharide sequence of the hexa-decasaccharide does not appear to be of sufficient length to form functionally relevant ternary complexes. However, the hexade-casaccharide increased the rate of inhibition of factor IXa in thepresence of Ca++ and of factor XIa 7-9-fold more efficientlythan the regular pentasaccharide, indicative of bridging ofinhibitor and enzyme contributing to some extent to these rateenhancements. The accelerating effect on factor IXa inhibitionin the presence of Ca++ led to a rate constant of ~2 × 105 M-1·s-1, the third highest rate constant of all enzymes studied (Table2). All other enzymes than thrombin, factor Xa and factor IXain the presence of Ca++ were inactivated by the hexadecasac-charide-antithrombin complex with appreciably lower rate constants, <2 × 104 M-1·s-1 (Table 2), which are unlikely to bepharmacologically relevant. The anticoagulant effect of thehexadecasaccharide (27) therefore must be mainly due to thissaccharide enhancing antithrombin inhibition of thrombin andfactor Xa, although inhibition of factor IXa may also contributeto a limited extent.

High-affinity and unfractionated heparinsUnfractionated heparin and the ~50 saccharide high-affinityheparin, the latter representing the average chain-length inunfractionated heparin, had in general comparable abilities toaccelerate antithrombin inhibition of the proteinases studied(Table 2). These accelerations were typically greater than thoseof the ~26-saccharide high-affinity heparin (Table 2), reflectinga bridging effect that increases with heparin chain length(12,47,51). In the absence of Ca++, the bridging effect made apredominant contribution to the acceleration of thrombin inhi-bition by all three full-length heparins but only a small contri-bution to the inhibition of factor Xa, whereas it was of interme-diate importance for factor IXa inhibition (Table 2), in agree-ment with previous work (4,11,12,47,51). The total acceleratingeffects of the three full-length heparins were only slightly dif-ferent for thrombin and factor Xa, ~2000- – ~6000-fold and~600- – ~2500-fold, respectively. However, the three heparinforms accelerated antithrombin inhibition of factor IXa to alarger extent than that of any other enzyme, ~8000- – ~40000-fold, as a result of both a large conformational change effect, asdiscussed above, and appreciable bridging contributions, ~20- –~100-fold (Table 2) (12). The inactivation rate constants were inthe diffusion-controlled range for thrombin (~2 × 107 – ~5 × 107

M-1·s-1) and factor Xa (~2 × 106 – ~7 × 106 M-1·s-1) but some-what lower for factor IXa (~8 × 104 – ~4 × 105 M-1·s-1) (Table 2).

The effect of Ca++ on the full-length heparin-acceleratedrate of antithrombin inactivation of thrombin, factor Xa and fac-tor IXa was studied in most detail for high-affinity heparin with~50 saccharide units. Ca++ had no effect on the rate by whichthis heparin form accelerated thrombin inhibition but increasedthe rate of inhibition of factor Xa by ~5-fold (Table 2), due to anincreased bridging contribution, in general agreement with




previous work (46,47). Ca++ increased the high-affinity heparin-accelerated rate of factor IXa inactivation even more, by ~50-fold, as a result of both a higher rate constant due to con-formational activation and an increased bridging effect (Table 2)(12). As a consequence of these Ca++-induced rate increases, thecomplex between the high-affinity heparin form and antithrom-bin inhibited thrombin and factor Xa with comparable rate con-stants, ~2 × 107 M-1·s-1, at physiological Ca++ concentrations,and the rate constant for factor IXa inhibition was only ~2-foldlower, ~1 × 107 M-1·s-1 (Table 2). These rate constants are all inthe diffusion-limited, physiologically and pharmacologicallyrelevant range.

Of all other proteinases than thrombin and factors Xa andIXa, the high-affinity and unfractionated heparins only acceler-ated antithrombin inhibition of factors XIa and XIIa appreciablybetter than the regular pentasaccharide (Table 2), due to a bridg-ing effect. The contribution of this effect to the rate enhance-ment of factor XIIa inactivation was associated with a biphasicinactivation of the enzyme, 30-40 % of the latter being inhibit-ed ~20-fold more slowly than the main fraction. Since there wasno factor β-XIIa in the preparation used, the slow phase mayreflect factor XIIa bound in a less than optimal manner in theternary heparin-antithrombin-proteinase complex. The bridgingeffect showed a marked dependence on heparin chain length, inparticular in the case of factor XIa, for which it increased from~15 fold for high affinity heparin with ~26 saccharide units to~250 fold for the form with ~50 units (Table 2). A very longheparin chain therefore appears to be required to optimally bindfactor XIa, in agreement with the heparin binding site beingmore extended in this proteinase than in thrombin (52,53). Therate enhancements of factor XIa and XIIa inactivation caused bythe full-length heparins (Table 2) were appreciably higher thanthose of maximally 10-50-fold reported previously(45,52,54,55). This discrepancy is most likely due to the largeexcess of heparin present in the earlier studies decreasing thebridging effect by binding to the proteinase. In spite of the sub-stantial rate enhancements observed in this work, the resultingrate constants for inactivation of factors XIa and XIIa were only~105 M-1 · s-1 (Table 2). The high-affinity and unfractionatedheparins accelerated antithrombin inactivation of the remainingenzymes at most slightly better than the regular pentasaccharide(Table 2), due to minimal bridging effects, and the rate constants were even lower, ≤ 3 × 104 M-1·s-1 (Table 2). The physiologicaland pharmacological effects of the full-length heparin formstherefore presumably are primarily exerted through their abilityto increase the rate of antithrombin inhibition of thrombin, fac-tor Xa and factor IXa. However, the action of plasma cofactorsof some of the other enzymes, such as the potentiation by high-molecular-weight kininogen of the heparin-dependentantithrombin inhibition of kallikrein (44), could result in furtherenhancements of heparin-accelerated rate constants to the physiological and pharmacological relevant range.

Low-molecular-weight heparinIn general agreement with the results for the high-affinity andunfractionated heparins, low-molecular-weight heparin trans-forms antithrombin into a rapid inhibitor of thrombin and factorXa and also into a reasonably fast inhibitor of factor IXa in thepresence of Ca++. However, the accelerating effects and theinactivation rate constants for the reactions with these enzymes(Table 2) were lower than those caused by the full-lengthheparins, due to lower bridging contributions, in agreement with the shorter chain-lengths of low-molecular-weight heparin. Theinactivation rate constants produced by low-molecular-weightheparin in the absence of Ca++ were ~5 × 106 and ~1 × 106

M-1 · s-1 for the inhibition of thrombin and factor Xa, respective-ly, but only ~3 × 104 for factor IXa inhibition (Table 2). Ca++

would be expected to negligibly affect the rate constant forthrombin inhibition but to appreciably increase that for factorXa inhibition, in analogy with its effect on the rate constantscaused by the full-length heparins, although these values in thepresence of Ca++ were not determined for low-molecular-weightheparin. However, Ca++ was experimentally demonstrated toincrease the low-molecular-weight heparin-accelerated rateconstant for factor IXa inactivation ~15-fold, up to a value of ~5 × 105 M-1·s-1 (Table 2). The rate constants produced by low-molecular-weight heparin for antithrombin inhibition ofthrombin and factor Xa under physiological conditions are thusin the physiologically and pharmacologically relevant range,whereas that for factor IXa inhibition is somewhat lower,although close to this range.

Low-molecular-weight heparin also accelerated antithrom-bin inhibition of all the other enzymes studied appreciably lessthan did the full-length heparins, due to small or virtually absentbridging contributions (Table 2). A decreased approximationeffect was especially pronounced for factors XIa and XIIa, con-sistent with only long heparin chains being able to bridgeantithrombin and these enzymes, as discussed above. FactorXIa and plasmin were inactivated by the complex between low-molecular-weight heparin and antithrombin with the highestrate constants, 1–2 × 104 M-1·s-1 (Table 2), of all enzymes otherthan thrombin and factors Xa and IXa, but also these values arepresumably of marginal pharmacologic significance. Theantithrombotic effect of low-molecular-weight heparin underphysiological conditions must therefore be due to its ability toaccelerate antihrombin inhibition of thrombin, factor Xa andmost likely also to some extent factor IXa.

Discussion

This work reports the first comprehensive studies of the accel-erating effects of a number of natural and synthetic heparin saccharides on antithrombin inhibition of all major clotting andfibrinolytic proteinases. The results show that the highest rateconstants that can be attained for such inhibition under physio-




937

logical conditions are 2-5 × 107 M-1·s-1, observed for the inacti-vation of thrombin and factor Xa by complexes of antithrombinwith full-length heparins. The corresponding rate constants for factor IXa inhibition are almost as high, 4 × 106 – 1 × 107

M-1·s-1, whereas those for the inactivation of all other enzymesare appreciably lower. Since antithrombin–full-length heparincomplexes are formed with rate constants ≥ 2 × 107 M-1·s-1

under the same conditions (4, 42), the association of antithrom-bin with heparin is unlikely to limit the rate at which coagula-tion proteinases are inactivated. Heparin acceleration ofantithrombin-proteinase reactions under physiological condi-tions is highest, up to ~200000-fold, for the effect of full-lengthheparins on factor IXa inhibition, but also substantial, ~2000- –6000-fold, for the corresponding effect on the inhibition ofthrombin, factor Xa and factor XIIa,. In contrast, acceleration ofthe inactivation of all other enzymes is at most moderate. Onlyprocoagulant enzymes are thus the principal targets ofantithrombin-heparin complexes, whereas anticoagulant andprofibrinolytic enzymes are not inhibited to any appreciableextent, in keeping with the putative physiological role ofheparin-like polysaccharides as anticoagulants (1, 2).

The conformational change induced in antithrombin bybinding of the specific heparin pentasaccharide region con-tributes substantially, ~150-500-fold, to increasing the rate ofinhibition of factors Xa, IXa and VIIa and moderately, ~50-fold‚ to acceleration of the inactivation of factor XIIa and tissueplasminogen activator. Antithrombin recognition of theseenzymes must therefore be appreciably facilitated by the alteredelectrostatic surface around the antithrombin reactive-site loopresulting from the conformational change (8, 9, 56). In contrast,this change is of less importance for the rate of inhibition of allother enzymes and insignificant for that of thrombin and factorXIa. The bridging effect due to binding of antithrombin and proteinase to the same, long heparin chain is dominating,~1000-3000-fold, for acceleration of thrombin inhibition, isappreciably smaller, although up to ~250-350-fold, for the inhibition of factors IXa and XIa, and contributes modestly ornegligibly to the inactivation of all other enzymes. The bridgingeffect thus well reflects the relative affinities of the enzymes forheparin (53).

The results also show that the heparin derivatives studied,which are currently being used or evaluated for treatment ofvenous thrombosis, exert their anticoagulant effects throughpartially different mechanisms. The effects of unfractionatedand low-molecular-weight heparins under physiological condi-tions are both due to acceleration of antithrombin inhibition ofthrombin, factor Xa and factor IXa. However, acceleration offactor IXa inhibition appears to be less important for low-molecular-weight heparin than for unfractionated heparin.Moreover, the relative importance of the inactivation of thethree enzymes may vary with the low-molecular-weight heparinpreparations, due to different average chain-lengths and content

of antithrombin-binding pentasaccharide sequences of suchpreparations (17, 57). The maximal rate constants for the inactivation of the three enzymes attained by the preparation oflow-molecular-weight heparin characterized are 5-10-foldlower than those effected by unfractionated heparin, due tolower bridging contributions. This pharmacological disad-vantage is presumably compensated by the lower nonspecificbinding of the shorter heparin chains in low-molecular-weightheparin to other plasma proteins than antithrombin and thelonger half-life of these chains in blood (17).

The anticoagulant effects of the regular and high-affinitypentasaccharides (fondaparinux and idraparinux) under physio-logical conditions are also due to a common mechanism, which,in contrast to that for longer heparins, only entails accelerationof antithrombin inhibition of factor Xa. The remarkableantithrombotic effects of fondaparinux seen in clinical trials(21-25) must be related to this unique ability to greatly andspecifically increase the rate of factor Xa inactivation. Asexpected from their structures, fondaparinux and idraparinuxproduce identical rate constants for the inhibition of factor Xa,as well as of the other enzymes, which, however, are somewhatlower than those given by full-length heparins. Nevertheless, the specific factor Xa inhibition and the better metabolic and phar-macokinetics properties, in particular the prolonged half-life inplasma, should make the two pentasaccharides better antithrom-botic agents, as demonstrated for fondaparinux.

Similar to unfractionated and low-molecular-weightheparins, the hexadecasaccharide exerts its anticoagulant effectunder physiological conditions by accelerating antithrombininhibition of both thrombin and factor Xa. However, the effecton factor IXa inhibition appears to be of appreciably less impor-tance for the hexadecasaccharide than for the two heparins. Afurther difference is that the hexadecasaccharide promotesantithrombin inactivation of thrombin better than low-molecu-lar-weight heparin and the inactivation of factor Xa not as wellas unfractionated heparin. The anticoagulant and antithrombot-ic effects of the hexadecasaccharide (27) is therefore presum-ably exerted more through thrombin inhibition and less throughfactor Xa inhibition than the corresponding effects of theunfractionated and low-molecular-weight heparins.

The finding that a 2.5-mg dose of fondaparinux injectedonce a day results in a more efficient thrombosis prophylaxisthan either two daily doses of 30 mg or one daily dose of 40 mglow-molecular-weight heparin (21-25) is not easily explainedonly on the basis of the accelerating effects on antithrombininhibition of proteinases reported in this work. This clinicalobservation demonstrates that specific inhibition of factor Xa isa highly powerful means to achieve an antithrombotic effect butalso shows that the in vitro anticoagulant potency of a drug isonly one determinant of this effect. A critical role is also playedby the distribution, metabolism and pharmacokinetic propertiesof the drug, in turn governed by the structure of the latter. In the



938

case of fondaparinux, a combination of the metabolism andpharmacokinetics profile and the enzyme inhibitory characteris-tics results in a significantly improved antithrombotic potency.Clinical trials with idraparinux and the hexadecasaccharideshould reveal potential therapeutic advantages of these newagents.

AbbreviationsH5, regular pentasaccharide (fondaparinux); H5*, high-affinity pentasac-charide (idraparinux); H16, hexadecasaccharide; H26 and H50, full-lengthheparin with high affinity for antithrombin and containing ~26 and ~50saccharide units, respectively; Kd, dissociation equilibrium constant;LMWH, low-molecular-weight heparin; UFH, unfractionated heparin.

References1. Olson ST, Björk I. Regulation of thrombin

activity by antithrombin and heparin. SeminThromb Hemost 1994; 20: 373-409.

2. Björk I, Olson ST. Antithrombin. A bloodyimportant serpin. In: Church FC, CunninghamDD, Ginsburg D, Hoffman M, Stone SR,Tollefsen DM, eds. Chemistry and Biology ofSerpins. New York, Plenum Press 1997; 17-33.

3. Petitou M, Casu B, Lindahl U. 1976-83, a critical period in the history of heparin: thediscovery of the antithrombin binding site.Biochimie 2003; 85: 83-9.

4. Olson ST, Björk I, Sheffer R, et al. Role of theantithrombin-binding pentasaccharide inheparin acceleration of antithrombin-pro-teinase reactions. Resolution of the antithrom-bin conformational change contribution toheparin rate enhancement. J Biol Chem 1992;267: 12528-38.

5. Huntington JA, Olson ST, Fan BQ, et al.Mechanism of heparin activation ofantithrombin. Evidence for reactive centerloop preinsertion with expulsion upon heparinbinding. Biochemistry 1996; 35: 8495-503.

6. Jin L, Abrahams JP, Skinner R, et al. The anti-coagulant activation of antithrombin byheparin. Proc Natl Acad Sci U S A 1997; 94:14683-8.

7. Huntington JA, McCoy A, Belzar KJ, et al.The conformational activation of antithrom-bin – A 2.85-Å structure of a fluorescein deriv-ative reveals an electrostatic link between thehinge and heparin binding regions. J BiolChem 2000; 275: 15377-83.

8. Chuang YJ, Swanson R, Raja SM, et al.Heparin enhances the specificity of antithrom-bin for thrombin and factor Xa independent ofthe reactive center loop sequence – Evidencefor an exosite determinant of factor Xa speci-ficity in heparin-activated antithrombin. J BiolChem 2001; 276: 14961-71.

9. Chuang YJ, Swanson R, Raja SM, et al. Theantithrombin P1 residue is important for targetproteinase specificity but not for heparin acti-vation of the serpin. Characterization of P1antithrombin variants with altered proteinasespecificity but normal heparin activation.Biochemistry 2001; 40: 6670-9.

10. Danielsson Å, Raub E, Lindahl U, et al. Roleof ternary complexes, in which heparin bindsboth antithrombin and proteinase, in the accel-eration of the reactions between antithrombinand thrombin or factor Xa. J Biol Chem 1986;261: 15467-73.

11. Olson ST, Björk I. Predominant contributionof surface approximation to the mechanism ofheparin acceleration of the antithrombin-thrombin reaction. Elucidation from salt concentration effects. J Biol Chem 1991; 266:6353-64.

12. Bedsted T, Swanson R, Chuang Y-J, et al.Heparin and calcium ions dramaticallyenhance antithrombin reactivity with factorIXa by generating new interaction exosites.Biochemistry 2003; 42: 8143-52.

13. Wiebe EM, Stafford AR, Fredenburgh JC, etal. Mechanism of catalysis of inhibition offactor IXa by antithrombin in the presence ofheparin or pentasaccharide. J Biol Chem 2003;278: 35767-74.

14. Conrad HE. Heparin-binding proteins. SanDiego, CA, Academic Press; 1998.

15. Warkentin TE, Chong BH, Greinacher A.Heparin-induced thrombocytopenia: towardsconsensus. Thromb Haemost 1998; 79: 1-7.

16. Devlin VJ, Einhorn TA, Gordon SL, et al.Total hip arthroplasty after renal transplanta-tion. Long-term follow-up study and assess-ment of metabolic bone status. J Arthroplasty1988; 3: 205-13.

17. Barrowcliffe TW. Low molecular weightheparin(s). Br J Haematol 1995; 90: 1-7.

18. van Boeckel CAA, Petitou M. The uniqueantithrombin III binding domain of heparin: Alead to new synthetic antithrombotics. AngewChem 1993; 32: 1671-90.

19. Petitou M, Hérault JP, Bernat A, et al.Synthesis of thrombin-inhibiting heparinmimetics without side effects. Nature 1999;398: 417-22.

20. Choay J, Petitou M, Lormeau JC, et al.Structure-activity relationship in heparin: asynthetic pentasaccharide with high affinityfor antithrombin III and eliciting high anti-factor Xa activity. Biochem Biophys ResCommun 1983; 116: 492-9.

21. Turpie AGG, Gallus AS, Hoek JA. A synthet-ic pentasaccharide for the prevention of deep-vein thrombosis after total hip replacement. NEngl J Med 2001; 344: 619-25.

22. Eriksson BI, Bauer KA, Lassen MR, et al.Fondaparinux compared with enoxaparin forthe prevention of venous thromboembolismafter hip-fracture surgery. N Engl J Med 2001;345: 1298-304.

23. Bauer KA, Eriksson BI, Lassen MR, et al.Fondaparinux compared with enoxaparin forthe prevention of venous thromboembolism

after elective major knee surgery. N Engl JMed 2001; 345: 1305-10.

24. Turpie AGG, Bauer KA, Eriksson BI, et al.Fondaparinux versus enoxaparin for the prevention of venous thromboembolism inmajor orthopedic surgery. A meta-analysis of 4randomized double-blind studies. Arch InternMed 2002; 162: 1833-40.

25. Eriksson BI, Bauer KA, Lassen MR, et al.Influence of the duration of fondaparinux(Arixtra®) prophylaxis in preventing venousthromboembolism following major orthopedicsurgery. J Thromb Haemost 2003; 1: 383-4.

26. Herbert JM, Hérault JP, Bernat A, et al.Biochemical and pharmacological propertiesof SANORG 34006, a potent and long-actingsynthetic pentasaccharide. Blood 1998; 91:4197-205.

27. Herbert JM, Hérault JP, Bernat A, et al.SR123781A, a synthetic heparin mimetic.Thromb Haemost 2001; 85: 852-60.

28. Danielsson A, Björk I. Binding to anti-thrombin of heparin fractions with differentmolecular weights. Biochem J 1981; 193: 427-33.

29. Olson ST, Björk I, Shore JD. Kinetic charac-terization of heparin-catalyzed and un-catalyzed inhibition of blood coagulation pro-teinases by antithrombin. Methods Enzymol1993; 222: 525-60.

30. Nordenman B, Nyström C, Björk I. The sizeand shape of human and bovine antithrombinIII. Eur J Biochem 1977; 78: 195-203.

31. Bock PE, Day DE, Verhamme IM, et al.Analogs of human plasminogen that arelabeled with fluorescence probes at the catalytic site of the zymogen. Preparation,characterization, and interaction with strep-tokinase. J Biol Chem 1996; 271: 1072-80.

32. Melhado LL, Peltz SW, Leytus SP, et al. p-Guanidinobenzoic acid esters of fluoresceinas active-site titrants of serine proteases. J AmChem Soc 1982; 104: 7299-306.

33. Schägger H, von Jagow G. Tricine-sodiumdodecyl sulfate-polyacrylamide gel elec-trophoresis for the separation of proteins in therange from 1 to 100 kDa. Anal Biochem 1987;166: 368-79.

34. Laemmli UK. Cleavage of structural proteinsduring the assembly of the head of bacterio-phage T4. Nature 1970; 227: 680-5.

35. Arocas V, Bock SC, Olson ST, et al. The roleof Arg46 and Arg47 of antithrombin in heparinbinding. Biochemistry 1999; 38: 10196-204.


36. Björk I, Jackson CM, Jörnvall H, et al. Theactive site of antithrombin. Release of thesame proteolytically cleaved form of theinhibitor from complexes with factor IXa, factor Xa, and thrombin. J Biol Chem 1982;257: 2406-11.

37. Björk I, Fish WW. Production in vitro andproperties of a modified form of bovineantithrombin, cleaved at the active site bythrombin. J Biol Chem 1982; 257: 9487-93.

38. Olson ST. Heparin and ionic strength-depen-dent conversion of antithrombin III from aninhibitor to a substrate of alpha-thrombin. JBiol Chem 1985; 260: 10153-60.

39. Stürzebecher J, Kopetzki E, Bode W, et al.Dramatic enhancement of the catalytic activ-ity of coagulation factor IXa by alcohols.FEBS Lett 1997; 412: 295-300.

40. Olson ST, Swanson R, Day D, et al.Resolution of Michaelis complex, acylation,and conformational change steps in the reac-tions of the serpin, plasminogen activatorinhibitor-1, with tissue plasminogen activa.tor and trypsin. Biochemistry 2001; 40:11742-56.

41. Björk I, Ylinenjärvi K, Olson ST, et al.Decreased affinity of recombinant antithrom-bin for heparin due to increased glycosylation.Biochem J 1992; 286: 793-800.

42. Turk B, Brieditis I, Bock SC, et al. Theoligosaccharide side chain on Asn-135 of �-antithrombin, absent in �-antithrombin,decreases the heparin affinity of the inhibitorby affecting the heparin-induced confor-mational change. Biochemistry 1997; 36:6682-91.

43. Desai U, Swanson R, Bock SC, et al. Role ofarginine 129 in heparin binding and activationof antithrombin. J Biol Chem 2000; 275:18976-84.

44. Olson ST, Sheffer R, Francis AM. Highmolecular weight kininogen potentiates theheparin-accelerated inhibition of plasmakallikrein by antithrombin: Role forantithrombin in the regulation of kallikrein.Biochemistry 1993; 32: 12136-47.

45. Pixley RA, Schapira M, Colman RW. Effect ofheparin on the inactivation rate of human acti-vated factor XII by antithrombin III. Blood1985; 66: 198-203.

46. Rezaie AR. Calcium enhances heparin cataly-sis of the antithrombin factor Xa reaction by atemplate mechanism - Evidence that calciumalleviates Gla domain antagonism of heparinbinding to factor Xa. J Biol Chem 1998; 273:16824-7.

47. Rezaie AR, Olson ST. Calcium enhancesheparin catalysis of the antithrombin-factorXa reaction by promoting the assembly of anintermediate heparin-antithrombin-factor Xabridging complex. Demonstration by rapidkinetics studies. Biochemistry 2000; 39:12083-90.

48. Ruf W, Kalnik MW, Lund-Hansen T, et al.Characterization of factor VII association withtissue factor in solution. High and low affinitycalcium binding sites in factor VII contributeto functionally distinct interactions. J BiolChem 1991; 266: 15719-25.

49. Lormeau JC, Herault JP, Herbert JM.Antithrombin-mediated inhibition of factorVIIa-tissue factor complex by the syntheticpentasaccharide representing the heparin bind-ing site to antithrombin. Thromb Haemost1996; 76: 5-8.

50. Desai UR, Petitou M, Björk I, et al.Mechanism of heparin activation of anti-thrombin - Role of individual residues of thepentasaccharide activating sequence in therecognition of native and activated states

of antithrombin. J Biol Chem 1998; 273:7478-87.

51. Holmer E, Kurachi K, Söderstrom G. Themolecular-weight dependence of the rate-enhancing effect of heparin on the inhibitionof thrombin, factor Xa, factor IXa, factor XIa,factor XIIa and kallikrein by antithrombin.Biochem J 1981; 193: 395-400.

52. Zhao MM, Abdel-Razek T, Sun MF, et al.Characterization of a heparin binding site onthe heavy chain of factor XI. J Biol Chem1998; 273: 31153-9.

53. Badellino KO, Walsh PN. Localization of aheparin binding site in the catalytic domain offactor XIa. Biochemistry 2001; 40: 7569-80.

54. Scott CF, Colman RW. Factors influencing theacceleration of human factor XIa inactivationby antithrombin III. Blood 1989; 73: 1873-9.

55. Wuillemin WA, Eldering E, Citarella F, et al.Modulation of contact system proteases byglycosaminoglycans - Selective enhancementof the inhibition of factor XIa. J Biol Chem1996; 271: 12913-8.

56. Izaguirre G, Zhang W, Swanson R, et al.Localization of an antithrombin exosite thatpromotes rapid inhibition of factors Xa andIXa dependent on heparin activation of theserpin. J Biol Chem 2003; 278: 51433-40.

57. Schoen P, Lindhout T, Hemker HC. Ratios ofanti-factor Xa to antithrombin activities ofheparins as determined in recalcified humanplasma. Br J Haematol 1992; 81: 255-62.

58. Streusand VJ, Björk I, Gettins PGW, et al.Mechanism of acceleration of antithrombin-proteinase reactions by low affinity heparin.Role of the antithrombin binding pentasaccha-ride in heparin rate enhancement. J Biol Chem1995; 270: 9043-51.




accelerating ability of synthetic oligosaccharides on antithrombin inhibition of proteinases of the...

Documents