Multiple functional domains of the heparin molecule

Clinical efficacy of low molecular weight heparin in postoperative thrombosis prophylaxis

In a randomized controlled clinical trial, the efficacy and safety of two low molecular weight heparin ( LMWH ) fractions in the prophylaxis of deep vein thrombosis (DVT) were assessed. One hundred twenty-six patients undergoing major abdominal surgery received alternatively 2,500 APTT units b.i.d. of two LMWH fractions or 5,000 APTT units b.i.d. of an unfractionated sodium mucosal heparin ( UFH ). LMWH 2 differed from LMWH 1 by presenting a lower mean molecular weight and a higher anti-Xa/APTT ratio in vitro. Patients were randomly allocated to the three groups, and the development of DVT was studied with the 125I-fibrinogen uptake test ( RFUT ). The study was interrupted and the code broken prematurely because of otherwise unexplainable bleeding events. While no thrombosis and no severe bleeding were detected in the UFH group, three (7%) RFUT -positive DVT and two (5%) hemorrhagic complications occurred in the LMWH 1 group. No thrombosis and nine (22%) cases of severe bleeding were observed in the LMWH 2 group. Thus, the latter group differed significantly from the control group with regard to subjective and objective criteria for postoperative bleeding. Although these results do not allow general conclusions as to the value of LMWH fractions in the prevention of DVT, they indicate that these preparations just as ordinary heparin have a limited therapeutic range.

Structural studies on heparin. Tetrasaccharides obtained by heparinase degradation

Three tetrasaccharides representing major structural sequences of heparin were isolated in good yield and characterized after degradation of heparin by purified flavobacterial heparinase. N-Desulfation was necessary to achieve good separation of these closely related compounds from each other. One of the tetrasaccharides was shown to be derived from the fully sulfated repeating segments; to contain L-iduronic acid and six sulfate groups, and have the structure delta 4,5- HexpA -(2-SO4)-(1----4)-alpha-D- GlcpN -(N-SO4)-(6-SO4)-(1- ---4)-alpha -L- IdopA -(2-SO4)-(1----4)-D- GlcN -(N-SO4)-(6-SO4). The second contained a D-glucuronic acid unit that was nonsulfated instead of the L-iduronic acid, and the third, obtained in a fairly low yield, contained five sulfate groups, three of which being located on the disaccharide at the nonreducing end, and having the structure delta 4,5- HexpA -(2-SO4)-(1----4)-alpha-D- GlcpN -(N-SO4)-(6-SO4)-(1- ---4)-alpha -L- IdopA -(2-SO4)-(1----4)-D- GlcN -(N-SO4). All tetrasaccharides had a sulfated, unsaturated uronic acid unit at the nonreducing end, confirming that the heparinase requires sulfated L-iduronic acid units for activity.

Anticoagulant activities of heparin oligosaccharides and their neutralization by platelet factor 4

Oligosaccharides of well-defined molecular size were prepared from heparin by nitrous acid depolymerization, affinity chromatography on immobilized antithrombin III (see footnote on Nomenclature) and gel chromatography on Sephadex G-50. High affinity (for antithrombin III) octa-, deca-, dodeca-, tetradeca-, hexadeca- and octadeca-saccharides were prepared, as well as oligosaccharides of larger size than octadecasaccharide. The inhibition of Factor Xa by antithrombin III was greatly accelerated by all of these oligosaccharides, the specific anti-Factor Xa activity being invariably greater than 1300 units/mumol. The anti-Factor Xa activity of the decasaccharide was not significantly decreased in the presence of platelet factor 4, even at high platelet factor 4/oligosaccharide ratios. Measurable but incomplete neutralization of the anti-Factor Xa activities of the tetradeca- and hexadeca-saccharides was observed, and complete neutralization of octadeca- and larger oligo-saccharides was achieved with excess platelet factor 4. The octa-, deca-, dodeca-, tetradeca- and hexadeca-saccharides had negligible effect on the inhibition of thrombin by antithrombin III, whereas specific anti-thrombin activity was expressed by the octadeca-saccharide and by the larger oligosaccharides. An octadecasaccharide is therefore the smallest heparin fragment (prepared by nitrous acid depolymerization) that can accelerate thrombin inhibition by antithrombin III. The anti-thrombin activities of the octadecasaccharide and larger oligosaccharides were more readily neutralized by platelet factor 4 than were their anti-Factor Xa activities. These findings are compatible with two alternative mechanisms for the action of platelet factor 4, both involving the binding of the protein molecule adjacent to the antithrombin III-binding site. Such binding results in either steric interference with the formation of antithrombin III-proteinase complexes or in displacement of the antithrombin III molecule from the heparin chain.

Evaluation of critical groups required for the binding of heparin to antithrombin

We have examined the quantitative importance of various monosaccharide residues of an octasaccharide domain of heparin that are responsible for the binding of this oligosaccharide to antithrombin. Different fragments of the octasaccharide were prepared by enzymatic digestion and the avidities of these oligosaccharides for antithrombin were determined by equilibrium dialysis. The data show that the non-reducing-end and the reducing-end tetrasaccharides contribute equally to the binding energy of the octasaccharide. The O6-sulfate group of the N-acetyl glucosamine moiety within the nonreducing-end tetrasaccharide is responsible for approximately equal to 45% of the binding energy of the octasaccharide. Neither the two non-sulfated uronic acid groups that flank this residue nor the N-sulfated glucosamine residue on the reducing end of this tetrasaccharide sequence that bears the unique O3-sulfate substituent contribute significantly to the binding energy of the octasaccharide. We suggest that the lack of sulfation of the two uronic acid moieties within the nonreducing-end tetrasaccharide may be required to permit the N-acetyl glucosamine O6-sulfate group to interact with a specific region on the antithrombin molecule. However, we cannot exclude the possibility that the O3-sulfate group plays an important role in orienting this O6-sulfate group within the nonreducing-end tetrasaccharide.

Heparin-catalyzed inhibitor/protease reactions: kinetic evidence for a common mechanism of action of heparin

Three different heparin-catalyzed inhibitor/protease reactions were studied: antithrombin III/thrombin, heparin cofactor II/thrombin, antithrombin III/factor Xa. The three reactions were saturable with respect to both inhibitor and protease. The initial reaction velocity, for each reaction, could be described by the general rate equation for a random-order bireactant enzyme-catalyzed reaction. The kinetic parameters for the heparin-catalyzed antithrombin III/thrombin and antithrombin III/factor Xa reactions differed in terms of apparent maximum velocity (Vmax) and apparent heparin-protease dissociation constant values. The apparent heparin-antithrombin III dissociation constant values were the same for both reactions. The kinetic parameters for the heparin-catalyzed antithrombin III/thrombin and heparin cofactor II/thrombin reactions differed in terms of apparent Vmax and apparent heparin-inhibitor dissociation constant values. The apparent heparin-thrombin dissociation constant values were the same for both reactions. The results are consistent with a general mechanism of action of heparin for the three reactions that, in its simplest form, requires only that both protease and inhibitor bind to heparin for catalysis to occur.

Structure-activity relationships of heparin. Independence of heparin charge density and antithrombin-binding domains in thrombin inhibition by antithrombin and heparin cofactor II

To better understand how heparin structure affects its activity the relationships between the functional domains for inhibitor binding and charge density were investigated to determine how these domains affect heparin-mediated thrombin inhibition by two different heparin-dependent protease inhibitors, antithrombin (AT) and heparin cofactor II (HC II). A series of heparins, fractionated systematically by charge density, was further fractionated on antithrombin agarose to isolate more homogeneous subfractions that were either inactive or highly active with respect to thrombin inhibition by AT. With AT, the activities of the AT-active subfractions increased sharply with heparin charge density, while those with little or no affinity for AT were virtually inactive. In contrast, with HC II inhibitor, the activities of the heparins depended only upon their charge densities and were independent of AT affinity. At any given charge density, the heparin before fractionation by AT affinity and the fractions that were highly active and inactive with AT were all equally active with HC II. The two inhibitors also differed in their reactivity with heparan sulfate and dermatan sulfate. A charge-density effect with the subfractions having similar high affinity for AT demonstrates that charge density represents a heparin functional domain that is independent of the AT-binding domain. The behavior of the AT-inactive heparins, being fully active with HC II, demonstrates the functional domain necessary for AT binding is not needed to produce HC II activity.

Properties of antithrombin-thrombin complex formed in the presence and in the absence of heparin

Purification of antithrombin-thrombin complex by ion-exchange chromatography on DEAE-agarose resulted in predominantly monomeric complex, whereas purification on matrix-linked heparin produced large amounts of aggregated complex. Monomeric antithrombin-thrombin complexes formed in the presence and in the absence of heparin had similar conformations and heparin affinities. Moreover, the first-order dissociation rate constants, measured by thrombin release, of these complexes were similar, 2.3 X 10(-6)-3.4 X 10(-6)S-1, regardless of whether newly formed or purified complex was analysed. Similar dissociation rate constants were also obtained for purified complex formed with or without heparin, from analyses by dodecyl sulphate/polyacrylamide-gel electrophoresis of the release of modified antithrombin, cleaved at the reactive-site bond. No dissociation of intact antithrombin from the complex was detected by activity measurements or by gel electrophoresis. Aggregation of the complex was found to be accompanied by a decrease in apparent dissociation rate. The similar properties of antithrombin-thrombin complexes formed with or without heparin support the concept of a catalytic role for the polysaccharide in the antithrombin-thrombin reaction. Furthermore, the results indicate that the reaction between enzyme and inhibitor involves the rapid formation of an irreversible, kinetically stable, complex that dissociates into active thrombin and modified, inactive, antithrombin by a first-order process with a half-life of about 3 days. The inhibition thus resembles a normal proteolytic reaction, one intermediate step of which is very slow.

Binding of platelet factor 4 to heparin oligosaccharides

Heparin fractions of differing Mr (7800-18 800) prepared from commercial heparin by gel filtration and affinity chromatography on immobilized anti-thrombin III had specific activities when determined by anti-Factor Xa and anti-thrombin assays that ranged from 228 to 448 units/mg. The anti-Factor Xa activity of these fractions could be readily and totally neutralized by increasing concentrations of platelet factor 4 (PF4). That these fractions bound to immobilized PF4 was indicated by the complete binding under near physiological conditions of 3H-labelled unfractionated commercial heparin. An anti-thrombin III-binding oligosaccharide preparation (containing predominantly eight to ten saccharide units), prepared by degradation of heparin with HNO2 had high (800 units/mg) anti-Factor Xa, but negligible anti-thrombin, specific activity. The anti-Factor Xa activity of this material could not be readily neutralized by PF4, and the 3H-labelled oligosaccharides did not completely bind to immobilized PF4. A heterogeneous anti-thrombin III-binding preparation containing upwards of 16 saccharides had anti-thrombin specific activity of just less than one-half the anti-Factor Xa specific activity. This material was completely bound to immobilized PF4 and was eluted with similar concentrations of NaCl to those that were required to elute unfractionated heparins from these columns. Furthermore, increasing concentrations of PF4 neutralized the anti-Factor Xa activity of this material in a manner similar to that of unfractionated heparin. It is concluded that heparin oligosaccharides require saccharide units in addition to the anti-thrombin III-binding sequence in order to fully interact with PF4.

Changes in disaccharide composition of heparan sulphate fractions with increasing degrees of sulphation

Heparan sulphate by-products from the commercial manufacture of pig mucosal heparin were freed of chondroitin sulphate and fractionated according to anionic density. The fractions were treated with HNO2 at pH 1.5, and the resulting mixtures of oligosaccharides were reduced with NaB3H4 and analysed for their disaccharide composition by paper chromatography and by high-pressure liquid chromatography. The results show that the molar ratio of 2-O-sulpho-alpha-L-iduronosylanhydromannose to 6-O-sulpho-(2-O-sulpho-alpha-L-iduronosyl)anhydromannose decreased from 2.5 to 0.04 as the degree of sulphation of the fractions increased. In contrast, the molar ratio of 6-O-sulpho-(beta-D-glucuronosyl)anhydromannose to 6-O-sulpho-(alpha-L-iduronosyl)anhydromannose was approx. 2.4 in all heparan sulphate fractions and decreased to only half of this value in the most highly sulphated heparin fractions. These results are consistent with biosynthetic studies, which have shown that the N-sulpho-(2-O-sulpho-alpha-L-iduronosyl)D-glucosamine disaccharide is the metabolic precursor of the NO-disulpho-(2-O-sulpho-alpha-L-iduronosyl)-D-glucosamine disaccharide in heparin biosynthesis. The high-pressure liquid chromatography of the heparan sulphate oligosaccharides also revealed a number of unidentified oligosaccharides in the deamination mixtures.

Circular dichroism spectroscopy of heparin-antithrombin interactions

We have utilized circular dichroism spectroscopy to examine the interaction of antithrombin with heparin-derived oligosaccharides and mucopolysaccharides of various sizes. Our studies demonstrate that the various complexes exhibit two major types of chiral absorption spectra. The first of these patterns is seen when octasaccharide, decasaccharide, dodecasaccharide, or tetradecasaccharide fragments bind to the protease inhibitor. The circular dichroism spectra of these complexes when compared to the spectrum of free antithrombin show several distinguishing characteristics. On the one hand, there is a marked general increase in positive chiral absorption that is maximal at 296 and 288 nm and 290 and 282.5 nm. These observations indicate perturbation of "buried" and "exposed" tryptophan residues. On the other hand, a significant augmentation in circular dichroism that peaks at 269.5 and 263 nm is noted. These findings are probably due to the summed positive and negative contributions arising from tryptophan residue(s), disulfide bridge(s), and phenylalanine residue(s). Given that these heparin fragments are able to accelerate factor Xa-antithrombin interactions but not thrombin-antithrombin interactions, the above spectral transitions must be associated with either the binding of a critical domain of the oligosaccharides to the protease inhibitor or the "activation" of the protease inhibitor with respect to factor Xa neutralization. The second of these patterns is apparent when octadecasaccharide, low molecular weight heparin (6,500), and high molecular weight heparin (22,000) interact with antithrombin. The circular dichroism spectra of these complexes compared to the spectrum of free protease inhibitor are similar to the first pattern except for changes within the 292- to 282-nm and 275- to 255-nm regions. The subtraction of the first pattern from the second pattern reveals a shallow negative band between 300 and 275 nm with potential negative minima at 290 and 283 nm as well as a deep negative band between 275 and 255 nm with possible negative minima at 268 and 262 nm. This chiral absorption profile is most likely to arise from conformational changes of a disulfide bridge(s). However, we cannot completely exclude the possibility that the above circular dichroism difference curve might be explained on the basis of transitions originating from a tryptophan residue(s). Given our method for generating the above data, these spectral alterations must be associated with the binding of a second critical domain of the mucopolysaccharide to antithrombin that is required for rapid complex formation with thrombin or the activation of the protease inhibitor with respect to the neutralization of the latter enzyme.

Mechanism of inactivation of trypsin by antithrombin

General aspects of the mechanism of antithrombin action were elucidated by a comparison of the inactivation of trypsin by antithrombin with the inactivation of coagulation proteinases by the inhibitor. Bovine antithrombin and bovine trypsin were shown to form an inactive equimolar complex. A non-complexed, proteolytically modified form of antithrombin, electrophoretically identical with that formed in the reaction with coagulation proteinases, was also produced in the reaction with trypsin. In the absence of heparin, the inactivation of trypsin by antithrombin was 20 times faster than the inactivation of thrombin; the second-order rate constant was 1.5 x 10(5)m(-1).s(-1) at 25 degrees C and pH 7.4. However, the inhibition of thrombin was accelerated about 30 times more efficiently by small amounts of heparin than was trypsin inhibition. Dissociation of the antithrombin-trypsin complex at pH 7.4 followed first-order kinetics with a half-life for the complex of about 80h at 25 degrees C. The complex was rapidly and quantitatively dissociated at pH 11, resulting in the liberation of a modified two-chain form of the inhibitor, cleaved at the same Arg-Ser bond as in modified antithrombin released from complexes with thrombin, Factor Xa and Factor IXa. This supports the previous proposal that this bond is the active-site bond of antithrombin. Antisera specific for thrombin-modified antithrombin reacted with purified antithrombin-trypsin complex, indicating that the inhibitor was present in the complex in a form immunologically identical with thrombin-modified antithrombin. The results thus suggest a common mechanism, but different kinetics, for the inhibition of trypsin and coagulation proteinases by antithrombin.

Structure and biological activity of finback-whale (Balaenoptera physalus L.) heparin octasaccharide. Chemical, carbon-13 nuclear-magnetic-resonance, enzymic and biological studies

Finback-whale (Balaenoptera physalus L.) heparin was partially digested with a purified heparinase and an octasaccharide with high affinity for antithrombin III was isolated from the digest by gel filtration, followed by affinity chromatography on a column of antithrombin III immobilized on Sepharose 4B. This octasaccharide possessed high inhibitory activity for Factor Xa in the presence of antithrombin III, but was essentially inactive for thrombin-antithrombin III reaction. The anticoagulant activity determined by the activated-partial-thromboplastin-time method was very low (40-70 units/mg), although the initial whale heparin exhibited high activity (252 units/mg). On the basis of the results of chemical analyses, 13C n.m.r. spectrum and enzymic studies with purified heparinase, heparitinases 1 and 2, the predominant structure of the octasaccharide was proposed as follows: delta UA(2S) alpha 1 leads to 4GlcNS alpha 1 leads to 4IdUA alpha 1 leads to 4GlcNAc(6S) alpha 1 leads to 4GlcUA beta 1 leads to 4GlcNS(3S) alpha 1 leads to 4IdUA(2S) alpha 1 leads to 4GlcNS. Comparing this structure with those of the heparin octasaccharides so far reported, the presence of the critical structural elements for binding to antithrombin III was suggested in the pentasaccharide region situated at the reducing end of this octasaccharide. Binding to antithrombin III of the critical structural elements alone would appear to elicit the acceleration of the Factor Xa-antithrombin III reaction. Additional structural elements required for the acceleration of the thrombin-antithrombin III reaction and for the manifestation of high anticoagulant activity are discussed.

Further characterization of the antithrombin-binding sequence in heparin

An octasaccharide with high affinity for antithrombin, isolated after partial deaminative cleavage of heparin and previously found to have the following predominant structure (see formula in text) has been studied further. High-voltage, paper electrophoresis of the 3H-labelled disaccharides obtained by deamination with HNO2 (pH 1.5) followed by reduction with Na[3H]BH4 showed approximately 25% of mono-O-sulfated components, in addition to L-iduronic acid(2-O-SO3)-2,5-anhydro-D-[3H]mannitol (6-O-SO3). The monosulfated disaccharides were identified by high pressure, ion-exchange chromatography as L-iduronic acid(2-O-SO3)-2,5-anhydro-D-[3H]mannitol, L-Iduronic acid-2,5-anhydro-D-[3H]mannitol(6-O-SO3). and D-glucuronic acid-2,5-anhydro-D-[3H]-mannitol L, iduronic acid 2,5-anhydro-D-[3H]mannitol(6-O-SO3), and D-glucuronic acid-2,5-anhydro-D-[3H]-mannitol. These components originated from the reducing, terminal disaccharide residue (units 7 and 8), as indicated by selective labelling with Na[3H]-BH4. The structural variability within this region suggests that it is not part of the antithrombin-binding sequence. Neither enzymic removal of the non-sulfated L-iduronic acid unit 1 nor N-deacetylation (by hydrazinolysis) at unit 2 had any significant effect on the affinity of the octasaccharide for antithrombin. However, removal of the disaccharide corresponding to units 1 and 2, by selective deamination of the N-deacetylated octasaccharide, yielded a low-affinity hexasaccharide. In addition, a high-affinity deamination product was formed, presumably an octasaccharide containing a 6-sulfated 2-deoxy-2-C-formyl-D-pentofuranosyl unit due to ring contraction in unit 2. These results suggest that the 6-sulfate group in unit 2 may be involved in antithrombin binding. It is concluded that the antithrombin binding site in heparin is represented by the pentasaccharide sequence extending from unit 2 to unit 6 of the octasaccharide studied.