• allostery;
  • heparin;
  • serpin;
  • thrombosis


  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

Summary.  Serpins are the predominant protease inhibitors in the higher organisms and are responsible, in humans, for the control of many highly regulated processes including blood coagulation and fibrinolysis. The serpin inhibitory mechanism has recently been revealed by the solution of a crystallographic structure of the final serpin–protease complex. The serpin mechanism, in contrast to the classical lock-and-key mechanism, involves dramatic conformational change in both the inhibitor and the inhibited protein. The final result is a stable covalent complex in which the properties of each component are altered so as to allow clearance from the circulation. Several serpins are involved in hemostasis: antithrombin (AT) inhibits many coagulation proteases, most importantly factor Xa and thrombin; heparin cofactor II (HCII) inhibits thrombin; protein C inhibitor (PCI) inhibits activated protein C and thrombin bound to thrombomodulin; plasminogen activator inhibitor 1 inhibits tissue plasminogen activator; and α2-antiplasmin inhibits plasmin. Nearly all of these reactions are accelerated through interactions with glycosaminoglycans (GAGs) such as heparin or heparan sulfate. Recent structures of AT, HCII and PCI have revealed how in each case the serpin mechanism has been fine-tuned by evolution to bring about high levels of regulatory control, and how seemingly disparate mechanisms of GAG binding and activation can share critical elements. By considering the serpins involved in hemostasis together it is possible to develop a deeper understanding of their complex individual roles.

The serpin mechanism

  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

The serpins (acronym of serine protease inhibitors) were identified as a protein family by Hunt and Dayhoff who reported on the similarities between antithrombin (AT), α1-antitrypsin (α1PI), and ovalbumin in 1980 [1]. Since then, over 500 serpins have been identified in the genomes of prokaryotic and eukaryotic organisms, and are involved in diverse physiologic processes [2,3]. Although most serpins are inhibitors of serine proteases, as the acronym suggests, many have been demonstrated to be incapable of protease inhibition, while others appear to have functions in addition to protease inhibition (e.g. cell signaling, hormone carriers) [4]. Serpins are also capable of inhibiting certain cysteine proteases and have been identified intracellularly, where they contribute to the control of the apoptotic pathways [5]. To date, 35 serpins have been identified in the human genome, which makes it the most abundant family of protease inhibitors in humans. In the blood, serpins also predominate and account for roughly 10% of the plasma proteins [4].

Although sequence identity for serpins is often very weak, their structures are highly conserved. Figure 1(a) shows the structure of the prototypical serpin α1PI [6], which also happens to be the most highly concentrated serpin in blood plasma. The serpin architecture consists of three β-sheets (A, B, and C) and nine α-helices (A–I), which are organized into an upper β-barrel domain and a lower helical domain. These two domains are bridged by the main structural feature of the serpins, the five-stranded β-sheet A. At the top of native α1PI is the reactive center loop (RCL) that, as the name suggests, is the region that interacts with the protease, and in most cases contains the sole determinants of protease specificity. In contrast to most other proteins, the native fold of the serpin is not the most stable, and serpins will readily undergo a rearrangement in topology with the insertion of the RCL as the fourth strand in the now six-stranded β-sheet A. Thus, serpin ‘loop insertion’ results in a molecule that is hyperstable. In the absence of cleavage of the RCL, the transition to the so-called latent state occurs with varying ease and is believed to play a role in serpin senescence (Fig. 1b) [7–9]. The same conformational transition occurs upon proteolytic attack anywhere within the RCL. In fact, the first serpin structure solved by X-ray crystallography was that of RCL-cleaved α1PI (Fig. 1c) [10]. Since then, many structures of native, latent and cleaved serpins have been solved, and together with the recently solved structures of the initial [11] and final [12] serpin/protease complexes, give a complete picture of the extraordinary serpin mechanism.


Figure 1. Topological changes of the serpin fold. Native serpins are metastable and can thus be thought of as kinetically trapped folding intermediates. (a) The prototypical native serpin, α1PI, is composed of a lower helical domain and an upper β-barrel domain, which are bridged by its main A β-sheet (red). The reactive center loop (RCL, yellow) interacts with target proteases predominantly by recognition of the sequence at the reactive center (P1, ball-and-stick) and the flanking region (green). The stability of the serpin is typically doubled by the incorporation of the RCL as the 4th strand of the now 6-stranded β-sheet A, either in (b) the absence, or (c) the presence of proteolytic nicking of the RCL.

Download figure to PowerPoint

The mechanism of inhibition that has emerged is complex compared with that of the other inhibitor families such as the Kunitz class, which include bovine pancreatic trypsin inhibitor (BPTI, also known as aprotinin) and tissue factor pathway inhibitor. The contrasting mechanisms are illustrated in Fig. 2. BPTI and α1PI both interact with proteases via the reactive center loop, but for BPTI the RCL is shorter and held rigidly in place by disulfide bonds and other non-covalent interactions with the core structure. The rigidity of the RCL of BPTI prevents the progression into catalysis, and inhibitory Kis are thus independent of the presence of catalytic residues (such as Ser195 in the protease) [13,14]. The serpin RCL is much longer (typically 22 amino acids) and is extremely flexible. In most crystallographic structures of native serpins, only a fraction of the intact RCL can be traced, due to the flexible nature of the loop. The length and flexibility of the serpin RCL renders it highly proteolytically susceptible, and in contrast to the Kunitz class of protease inhibitors, serpin inhibitory activity is absolutely dependent on the activity of the protease. Serpins have evolved to be very poor substrates, with the very last step in the proteolytic cycle of serine proteases, deacylation, slowed by several orders of magnitude. Deacylation involves the nucleophilic attack by a water molecule on the carbonyl carbon of the ester bond between the P1 residue of the substrate or serpin, and the active site Oγ of Ser195 of the protease. After nucleophilic attack the reaction proceeds through a tetrahedral intermediate during which a negative charge exists on the carbonyl oxygen.


Figure 2. The serpin mechanism of protease inhibition. The inhibitory domain of the serpin is about eight times larger than that of the other protein families of protease inhibitors. (a) A typical Kunitz type inhibitor, BPTI, inhibits trypsin (magenta) by a lock-and-key mechanism due to the complementary composition and conformation of the RCL (colored as in Fig. 1), which is held rigid by disulfide bonds and other interactions. This reversible complex does not depend on the activity of the protease. (b) The serpin mechanism proceeds through a similar Michaelis complex, but continues to the acyl-enzyme intermediate step where full incorporation of the RCL prevents deacylation. Thus the serpin mechanism is thermodynamically irreversible. The incorporation of the RCL results in a hyperstable serpin and a partially denatured protease (broken coil). The acyl-enzyme intermediate is stabilized by the distortion of the catalytic architecture illustrated in (c). An active serine protease (green) is composed of the catalytic triad (Asp102, His57 and Ser195) and the oxyanion hole (main chain amide hydrogens of Ser195 and Gly193, indicated by arrows). In complex (yellow) with serpins (magenta), Ser195 is pulled out of the active site cleft by 3.5 Å, due to its ester bond with the P1 carbonyl carbon, and the oxyanion hole is destroyed.

Download figure to PowerPoint

As Fig. 2(b) illustrates, the serpin has undergone full loop insertion to form the final complex with the protease, and as the protease is still covalently attached it is translocated some 70 Å from the ‘top’ to the ‘bottom’ of the serpin as a consequence of loop insertion (see video of mechanism at The loop insertion is triggered by the acylation step, and progresses at a rate sufficient to compete with that of the deacylation step [15,16]. Upon final loop insertion the serpin is in a hyperstable state, and the protease is ‘crushed’ up against the serpin through the pulling force exerted on the covalently linked active site loop of the protease. The effect of the pulling force is a stretching out of the catalytic loop of the protease that contains the active site serine and the residues that form the oxyanion hole (Fig. 2c). It is now understood in molecular detail how serpins inhibit serine proteases by slowing down deacylation.

In addition to the conformational change in the serpin, the protease also experiences a significant alteration in its structure. Approximately half of the protease cannot be resolved in the crystallographic structure (Fig. 2b). This is due to an unfolding of the protease caused by the elongation of the catalytic loop and through contacts with the now hyperstable serpin. As illustrated in Fig. 2(c), the pulling stress placed on the active site serine has elongated the active site loop. One of the consequences is the removal of Asp194 from its contacts with the N-terminal amino group of Ile16. This contact is formed upon cleavage activation of the zymogen form and results in the ordering of a region known as the zymogen activation domain [17]. The breaking of this salt-bridge through the extension of the catalytic loop results in a reversal of zymogen activation and the consequent disordering of a portion of the protease structure. In addition, several main chain clashes exist when a correctly folded protease is superimposed on the folded portion of the protease seen in the crystallographic structure of the final complex. These two factors conspire to destabilize the protease, which results in the release of bound cofactors [18,19] and the acquisition of a profound proteolytic susceptibility [20–22]. The first effect will aid in receptor-based clearance mechanisms, and the second allows for the in situ destruction of the protease.

The advantages of the serpin mechanism are manifold. The first and most obvious advantage is stoichiometric, and essentially irreversible, protease inhibition. This is of particular importance when a proteolytic cascade is involved, as very low concentrations of proteases early in the cascade can result in very high local levels of proteases at the end of the cascade. The second advantage is that a conformational change in both the inhibitor and the inhibited proteins allows for rapid clearance through release from cofactors and improved recognition by receptors. These conformational changes are also capable of signaling through cell surface receptors (e.g. the inflammatory response [23,24]). A mechanism dependent on conformational change also introduces the possibility of allosteric regulation. In this paper I will illustrate the regulatory advantage of the serpin mechanism in controlling the proteolytic pathways of hemostasis, in particular the glycosaminoglycan (GAG)-mediated mechanisms of activation of AT towards factor (F) Xa and thrombin, and of activation of heparin cofactor II (HCII) towards thrombin. Recent structures of AT and HCII combined with a careful analysis of the available biochemical data reveal an unexpected relationship between the two mechanisms. Finally, a recent structure of protein C inhibitor illustrates how diverse mechanisms can be more fully understood by considering the subgroup of GAG-activatable serpins together.

Thrombin and FXa

  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

In many ways the serpin mechanism resembles a mousetrap [25]. A mousetrap has stored energy that it uses to crush and kill the mouse; it requires bait; and it depends on the activity of the mouse to spring the trap. The story of thrombin inhibition by serpins would thus be incomplete without a description of thrombin's properties, in particular because the serpins described in this article exploit the unique structural aspects of thrombin in different ways. In hemostasis thrombin plays a major part in its own up- and down-regulation. This apparently paradoxical ability is the subject of great interest and has led to the development of thrombin variants with anticoagulant but no procoagulant activity [26]. Thrombin is a chymotrypsin-like serine protease with a typical catalytic mechanism; however, it is about 50% larger than chymotrypsin [27]. There are several insertions within the sequence and so insertion loop residues are lettered. Two of the insertion loops, the 60-insertion loop and the γ-loop are found on either side of the active site cleft and result in a constriction of the active site of thrombin relative to other proteases. The exquisite specificity of thrombin is due in part to the restrictions placed on active site access due to the presence of these bulky loops.

The structure of human thrombin in various forms has been solved over 100 times, and its major features have been described in great detail [27]. Of particular importance to thrombin inhibition by serpins is the existence of two anion-binding exosites on either side of the active site cleft. Exosite I is a basic patch on the P′ side of the active site cleft and is a secondary binding site for fibrinogen, and exosite II, the most basic site, has been identified as the heparin-binding site. It has been demonstrated through mutagenesis that exosite I is necessary for the interaction with HCII [28], and that exosite II is critical for the heparin-accelerated inhibition of thrombin by AT [29]. Although no structure of thrombin bound to heparin has been solved, several of the residues involved in binding have been identified through mutagenesis [29,30], and several others are suggested by their contribution to the electropositive patch [31].

In contrast to thrombin, FXa has a wide-open active site. The surface representations of the active site clefts of thrombin and FXa are shown in Fig. 6(c). FXa has recently been shown to interact with heparin in the presence of calcium or with its acidic Gla domain removed, and upon heparin binding, its inhibition by AT approaches the rate of heparin-enhanced thrombin inhibition [32,33]. The residues involved in heparin binding have been identified through mutagenesis [34,35].


  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

Antithrombin is the most important serpin in hemostasis. It circulates at a concentration of 2.3 µmol L−1 and is capable of inhibiting several of the proteases in the coagulation cascade, but, based on rates of inhibition, its primary targets are FXa and thrombin. The importance of AT is demonstrated by the high association of deficiency with venous thrombosis [36], by the embryonic lethal phenotype in the mouse knockout model [37], and by the success of heparin therapy. The anticoagulant effect of natural heparin and the new synthetic heparins is mediated predominantly through the activation of AT. In this section, I review the mechanism of heparin binding to and activation of AT towards its two main targets, FXa and thrombin.

Heparin is not a physiologic activator of AT. Instead it is the closely related heparan sulfate that lines the vascular wall, and that, through interaction with a fraction of the circulating AT, ensures the fluidity of the microvasculature [38]. Heparin is more highly sulfated and possesses a larger fraction of the highly flexible iduronic acid, and both factors are key in determining affinity for AT. Heparan sulfate is thought to have patches containing high sulfation and iduronic acid levels spaced between unsulfated regions along the length of the proteoglycan, which would provide the sites of interaction with AT. While some 30% of heparin chains bind AT with high affinity, only a small fraction of heparan sulfate chains possess the high-affinity site. Over the years the determinants of high-affinity binding have been identified, and can be reduced to a five-saccharide unit fragment [39,40], commonly referred to as the ‘natural pentasaccharide’. A schematic for the pentasaccharide and its interactions with AT is given in Fig. 3(a). The interaction is exquisitely specific and small alterations in either the pentasaccharide or in AT can significantly affect affinity. However, of the eight sulfates on the natural pentasaccharide, only four interact with AT, and thus the paucity of highly sulfated sites on heparan sulfate is not inconsistent with the presence of high-affinity binding sites. Furthermore, heparan sulfate synthesis depends on cell type, and one would expect a higher degree of sulfation on the endothelial cells lining the microvasculature.


Figure 3. Heparin binding to antithrombin (AT). AT interacts with heparin specifically via a pentasaccharide fragment present in ∼30% of heparin chains. (a) A schematic of the interactions (solid lines are ionic and dashed are hydrogen bonds) predicted for the natural pentasaccharide based on the crystallographic structure of a high-affinity derivative. (b) Binding is a two step process, with the formation of an initial weak complex followed by a conformational change to a tight binding state. The conformational change in the heparin-binding region brings about the expulsion of the preinserted portion of the RCL, which improves its flexibility and allows the reorientation of the P1 arginine side chain (green, ball-and-stick). (c) The heparin-binding site of native AT with helix A in green and helix D in cyan, with important residues indicated (ball-and-stick, labels included for orientation). The local conformational changes for (d) activated and (e) latent ATs are shown in yellow. The only change in secondary structure that is unique to the five-stranded, high-affinity, activated state is the C-terminal extension of helix D.

Download figure to PowerPoint

The binding of heparin by AT has been extensively studied, and is still an area of intense interest. It is not within the remit of this review to detail all of the developments that have led to the current understanding, and only selected results are presented here. The heparin-binding site of AT was first identified through sequence alignment by Huber and Carrell in 1989 [41], and has since been refined through mutagenic [42–45] and crystallographic studies [46]. The heparin-binding residues lie predominantly on helix D, with some contacts on helix A and with the N-terminal loop. Stopped-flow studies have found that binding occurs through an induced-fit mechanism [47], with a small portion of antithrombin in a non-induced high-affinity conformation [48]. Antithrombin is thus known to undergo a significant conformational change upon interaction with heparin, and to be in equilibrium between the two forms even in the absence of heparin. The crystallographic structure of native AT was published in 1994 [49,50], and was followed by that of a pentasaccharide-complexed AT in 1997 [46]. The heparin-binding mechanism of AT is summarized in Fig. 3(b). Antithrombin interacts weakly with the natural pentasaccharide, with an initial dissociation constant of ∼25 µmol L−1 followed by a conformational change to a high-affinity conformation, resulting in an overall dissociation constant of ∼50 nmol L−1. Both local and global conformational changes take place. In the heparin-binding region, helix D is extended towards its C-terminus, a new helix P extends from its N-terminus, and helix A appears to extend towards its N-terminus (Fig. 3c,d). The global conformational change involves the expulsion of the ‘preinserted’ portion of the RCL and the closing of β-sheet A to resemble the native conformation of α1PI. Just how local changes in the heparin-binding region result in the expulsion of the RCL and the closure of sheet A is unknown. Recent work suggests a link between the C-terminal elongation of helix D and loop expulsion [51,52]; however, the most important contact for stabilizing the activated conformation involves Lys114, which is on helix P [43,54]. Because it is an induced-fit binding mechanism, one would expect that the local conformation responsible for tight binding would only exist in the activated conformer, and not be accessible in six-stranded forms, such as the latent form. The crystal structure of activated and latent AT bound to the pentasaccharide reveal identical positions and contacts for Lys114, and helix P exists in both conformations (Fig. 3d,e). The C-terminal extension of helix D is the only conformational change that is unique to the activated conformation, and is thus likely to be the local conformational change that triggers the global structural rearrangement.

Heparin activation is kinetic. The rates of inhibition of FXa and thrombin in the absence of heparin are 2000 and 7000 M−1 s−1, respectively, and the binding of high-affinity heparin results in a ∼10 000-fold acceleration [33,54]. Antithrombin has two distinct mechanisms of activation for the two protease targets. Since the pentasaccharide alone accelerates the inhibition of FXa by 300-fold and of thrombin by only 2-fold, heparin enhancement of FXa inhibition is said to be through an ‘allosteric mechanism’, whereas the enhancement of thrombin inhibition requires longer heparin chains, and has been labeled the ‘bridging mechanism’[55] (Fig. 4a). Thrombin has a well-defined heparin-binding site within its anion-binding exosite II, but until recently FXa was considered incapable of interacting with heparin with reasonable affinity. It has recently been demonstrated that in the presence of physiologic levels of Ca2+ FXa inhibition is further accelerated through the addition of a bridging mechanism to the allosteric mechanism [33]. Although both FXa and thrombin are now known to be capable of bridging, only FXa inhibition is significantly accelerated by the pentasaccharide, and therefore it is still useful to make the distinction between the allosteric and bridging mechanisms.


Figure 4. Schematic representations of the mechanisms by which antithrombin (AT) and heparin cofactor II (HCII) are activated by heparin. (a) AT binds to a specific pentasaccharide region (shaded) of heparin and undergoes a conformational change that results in tight binding to heparin and improved recognition by factor Xa (top), and is thus known as the ‘allosteric’ mechanism. The conformational change does not lead to improved recognition by thrombin (bottom), and acceleration of the reaction is effected through a template effect known as the ‘bridging’ mechanism when thrombin binds to the same heparin chain as AT via its anion-binding exosite II. (b) Glycosaminoglycan (GAG) activation of HCII by heparin (Hep) or dermatan sulfate (DS) requires the displacement of an acidic N-terminal extension from its interactions with the body of the protein. This schematic assumes that the tail is in contact with the basic heparin-binding region and that its release is caused through direct competitive displacement by the GAG. The tail, once released, recruits thrombin by interacting with anion-binding exosite I, which approximates thrombin to the RCL of HCII that contains the unfavorable Leu at the P1 position (arrow). In addition to this allosteric mechanism, heparin can also bind thrombin to form a bridged complex, but dermatan sulfate does not interact with thrombin.

Download figure to PowerPoint

At the last ISTH meeting in Paris, Chuck Esmon suggested to me an AT paradox: how can one inhibitor use the same effector molecule to inhibit efficiently two very different proteases? Until the crystal structures of the encounter complexes are solved, there will always be some uncertainty in any response to this question. The long-standing explanation was that the conformational change that improved recognition by FXa was merely the release of the RCL, and that in its more flexible state the RCL would be preferentially recognized by FXa over thrombin, since the sequence of the reactive center, IAGR, is closer to the consensus sequence of FXa than of thrombin [4]. Recently, however, several groups have approached the problem using mutagenesis strategies, and have developed the hypothesis that exosites are exposed on AT in response to pentasaccharide binding, and are then recognized by FXa but not by thrombin. Our approach has been to attempt the crystallization of the complexes, but to date without success. However, in studying the related serpin, heparin cofactor II (HCII), we have obtained a valuable new insight into the nature of the serpin-protease Michaelis complex, which may help answer the ‘AT paradox’.

Heparin cofactor II

  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

HCII circulates at a concentration comparable to that of AT, and when activated by glycosaminoglycans, heparin, heparan sulfate or dermatan sulfate, is a potent inhibitor of thrombin [56], but its physiologic roles have yet to be established. Although HCII deficiency has been found in patients with thrombotic disorders, the frequency of deficiencies is similar to that found in the general population [57,58]. It is thus unclear whether HCII plays a role in the prevention of thrombosis. The high circulating concentration of 1.2 µmol L−1 and an ability to bind to, and be activated by, the heparan sulfate lining the vasculature indicate that it could add to the antithrombotic protection afforded by AT, although HCII cannot naturally substitute for AT in the case of AT deficiency [59]. HCII can be activated in vivo by the introduction of dermatan sulfate, indicating that dermatan sulfate may provide a potential alternative to heparin for the prevention and treatment of thrombosis [60,61]. This is particularly true for the inhibition of clot-bound thrombin, which appears to be uniquely vulnerable to dermatan sulfate-activated HCII [62]. The fact that HCII binds to dermatan sulfate suggests that it may function extravascularly and provide protection after tissue damage. Dermatan sulfate proteoglycans are found on fibroblasts, smooth muscle cells and other tissues, and when isolated will activate HCII [63,64]. Consistent with such a function is the observation that proteolytic cleavage of the N-terminus of HCII results in a chemotactic peptide corresponding to residues 49–60 [65]. HCII may thus have a small protective function against venous thrombosis, and extravascularly, the dual role of protector from spread of tissue damage and signal of tissue damage to the immune system.

The reactive center P1 residue of HCII is a leucine, which makes it a more effective inhibitor of chymotrypsin and cathepsin G than of thrombin [66–68]. Its specificity for thrombin is conferred through an interaction between thrombin exosite I and an 80-residue, highly acidic N-terminal extension of HCII that is sequestered in the native state and exposed after glycosaminoglycan binding [69]. The acidic domain contains two hirudin-like repeats, PEGEEDDDY and IFSEDDDYIDI, the first of which is required for glycosaminoglycan-mediated thrombin inhibition [69]. Heparin and dermatan sulfate are both cofactors for HCII and effect the same level of acceleration of thrombin inhibition (∼10 000-fold) [70,71]. HCII has been reported to bind to dermatan sulfate with a higher affinity and an apparent specificity for a hexasaccharide containing six sulfates [72], although full acceleration towards thrombin requires significantly longer chains. Its relatively low affinity for heparin explains why the antithrombotic effect of therapeutic heparin is almost wholly mediated through activation of AT.

HCII has a heparin-binding sequence on helix D almost identical to that of AT. The heparin and dermatan sulfate-binding regions are both primarily on helix D, as defined by sequence alignment with AT and mutagenesis studies, and appear to be distinct but overlapping, with heparin binding N-terminally to dermatan sulfate [73–75]. Helix A residues have also recently been demonstrated to play a role in GAG binding to both AT [42] and HCII [76]. Although the activated levels are identical using optimal amounts of heparin and dermatan sulfate (1 × 107 M−1 s−1), two distinct glycosaminoglycan mechanisms of activation have been proposed [69] (Fig. 4b). Heparin, which binds to thrombin via exosite II and is known to bridge AT and thrombin, was believed to accelerate the HCII–thrombin reaction primarily through an analogous bridging mechanism. Since dermatan sulfate does not interact with thrombin, dermatan sulfate activation of HCII was thought to be solely allosteric, with the direct displacement of the acidic tail from the GAG-binding site and its subsequent interaction with exosite I of thrombin. However, recent studies have demonstrated that an allosteric mechanism common to heparin and dermatan sulfate accounts for the majority of activation (Fig. 4b) [69,77].

The overall structure of native HCII is nearly identical to that of native AT (Fig. 5a). This finding was a surprise for two reasons: the model of GAG displacement of the tail as the basis of activation would not require an AT-like allosteric mechanism; and the inactivity of native HCII towards thrombin is already assured through the combination of the sequestered tail and the Leu at the reactive center. Thus, the possibility was raised of a truly allosteric mechanism where GAG binding causes loop expulsion, and that a global conformational change, not direct displacement, is the cause of the release of the acidic tail from its contacts with the body of HCII [78]. However the tail is released for interaction with thrombin, the mechanistic link between HCII and AT is now closer than was previously suspected.


Figure 5. Crystallographic structures of heparin cofactor II (HCII). (a) Native HCII (colored as before) was found to be in a conformation nearly identical to that of native AT, with a partially inserted RCL. The position of the acidic tail (magenta), however, could not be resolved due to crystal contacts. The crystallographic dimer was formed by the β-linkage of strand 1C (orange), which may have displaced the tail from the modeled position shown. (b) The structure of the Michaelis complex between HCII and thrombin demonstrates the basis of specificity. Loop expulsion is presumably caused by heparin binding, as occurs with antithrombin, and is required for the simultaneous interactions at the reactive center and on the ‘top/left’ of HCII where thrombin is docked. The role of the acidic tail can be understood in terms of creating an exosite at the interaction interface. The interaction surfaces of (c) thrombin, (d) HCII, and HCII with the acidic tail are colored according to electrostatic potential (red, negative; blue, positive). In the absence of the tail the complex would result in the apposition of the basic exosite I of thrombin (oval) with a basic surface of HCII (oval). The acidic tail reverses the electrostatic potential of the interface, resulting in improved binding.

Download figure to PowerPoint

The structure of HCII in its Michaelis complex with thrombin further supports the mechanistic link with AT [78] (Fig. 5b). The inserted portion of the RCL is expelled from β-sheet A as predicted from the native structure through analogy with AT activation, and HCII has adopted the five-stranded, active conformation. The RCL is not only expelled, but stretched to the fullest extent possible in order to make contacts with thrombin, which is docked to the rear/left of HCII (in the normal orientation; Fig. 5b). The contacts between the body of HCII and thrombin are extensive, and are partially mediated by the ‘sandwiching’ of the acidic tail between HCII and thrombin. This Michaelis complex is predicted to be of great stability due primarily to these exosite interactions, but the interactions between the active site of thrombin and the reactive center loop of HCII are also favorable [78]. The crucial importance of the exosite interactions at the HCII–thrombin interface has been demonstrated by the poor reactivity of mutant HCII where the tail has been deleted [79].

Although the mechanism probably involves the recruitment of thrombin through an interaction between the acidic tail and exosite I of thrombin prior to the docking of thrombin seen in the Michaelis complex, the role the tail plays in stabilizing the Michaelis complex can be understood in terms of providing an improved interaction interface through its intramolecular interactions. Thus, the docking surface of HCII has properties unfavorable for interaction with thrombin in the absence of the acidic tail. This is illustrated in Fig. 5(c,d,e) where the tail is removed from activated HCII for comparison, and the resulting surface is colored according to electrostatic properties. The tail reverses the electrostatics of the interaction interface so that in place of a repulsive, positive-to-positive interface, a favorable negative-to-positive interface exists (Fig. 5c,d,e). Once the electrostatics are generally favorable the hydrophobic surfaces can come into contact, and it is expected that the majority of the binding energy is provided from the burying of the hydrophobic patches.

Models for the AT Michaelis complexes

  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

Until recently very little was known about the nature of the serpin/protease Michaelis complex. Clearly it requires the substrate-like binding of the RCL in the active site cleft of the protease, but what of the state of the RCL, and what other sites help determine specificity? Studies on RCL mutants indicate that for most serpin/protease interactions, specificity is wholly determined by the sequence of the RCL [4]. In addition, it has been suggested that the RCL undergoes a preinsertion event upon initial interaction with the protease [80]. Solution studies on a α1PI with anhydrotrypsin [81], and the recent structure of the Manduca sexta serpin K with S195A trypsin [82], argue against any loop insertion preceding cleavage of the RCL and agree with earlier findings suggesting no stabilizing contacts outside the RCL. The structure of the HCII–thrombin Michaelis complex appears to be the exception, with extensive contacts between the bodies of the two molecules. There are thus currently two modes by which serpins and proteases can productively interact: intimately, where specificity is determined significantly through extensive exosite interactions; or independently, with the RCL behaving as an isolated peptide substrate loop. Any attempt to build a molecular picture of the Michaelis complex between AT and its targets, FXa and thrombin, must proceed from these two distinct models. Fortunately, AT and HCII have been demonstrated to possess a high degree of structural and mechanistic similarity, and they both inhibit thrombin. Therefore, the starting place for modeling the AT Michaelis complexes must be the HCII–thrombin Michaelis complex.

As mentioned above, recent studies strongly suggest that recognition of AT by FXa depends on exosites that are made available after the pentasaccharide binds to AT. There are two ways of interpreting this observation: (i) that exosites are ‘exposed’ by the conformational rearrangement that accompanies heparin binding to AT; or (ii) that loop expulsion is required for simultaneous interactions with the reactive center loop and the pre-existing exosites. In order to distinguish between these possibilities, surface representations of the ‘top’ of AT (with the RCL removed) were generated for the native and activated forms (Fig. 6a). It is clear that neither shape, charge nor hydrophobic properties are significantly altered in response to heparin binding in the region of AT likely to interact with FXa in the initial complex. This is perhaps not surprising considering the designation of this region as a rigid domain by Whisstock et al.[83]. In contrast, the Michaelis complex between HCII and thrombin does involve the creation of a novel exosite due to the ‘docking’ of the acidic tail on to the top of HCII (Fig. 5d,e). Although the exosite could thus be thought of as being ‘exposed’, it is not available for simultaneous interaction with the reactive center without the expulsion of the RCL from β-sheet A. Similarly, the exosite on AT responsible for FXa recognition may be preformed in native AT, but the simultaneous interactions within the exosite and the RCL can only be formed after the loop has been released from the constraints imposed by its partial insertion into β-sheet A. Thus, a model of FXa docked to pentasaccharide-activated AT has been made on the template of the thrombin–HCII Michaelis complex (Fig. 6b). The interface between the two molecules has been analyzed and was found to be free of steric clashes, and with several favorable ionic interactions. Of particular importance is the predicted interaction between Arg150 of FXa with Glu237 of AT, and between the acidic 30-loop of FXa (Glu36, Glu37 and Glu39) with a basic patch on AT composed of Arg262, Lys287, Lys403 and Arg406 (Fig. 6b,c). The contribution of Arg150 to the interaction of FXa with AT has recently been demonstrated by Manithoday et al., who found a 10-fold reduction in the rate of FXa inhibition by pentasaccharide-activated AT when Arg150 is altered [84], and Quinsey et al. have shown a 5-fold effect of the E36A mutant [85]. To date, no mutations within the basic patch of AT have been characterized, but the model would predict that alanine mutations in this region would significantly reduce the contribution of allostery towards FXa inhibition by AT. Additional support for this model comes from the observation that FXa and AT can be bridged by heparin in a manner similar to the bridging of thrombin and AT. Heparin binding by FXa requires Ca+2 or the removal of the Gla domain, and the residues important for heparin binding have been identified. The position of FXa in the model places the heparin-binding site around the back of AT, and thus, over 30 saccharide units would be predicted to be required to bridge FXa and AT (Fig. 6d). Again, the model is supported by a recent study by Rezaie and Olson, which demonstrated that a 26-unit heparin chain is insufficient to bridge AT to FXa [33]. This is in striking contrast to the minimum heparin length requirement of 15 units for the bridging of AT to thrombin [86], and helps explain the basis of the extra length required by FXa.


Figure 6. Models of the antithrombin (AT) /factor Xa (FXa) Michaelis complex. Unlike heparin cofactor II (HCII), AT undergoes no conformational change at the anticipated interaction interface. (a) Surface representations of the ‘top’ of AT with the RCL removed demonstrate that neither electrostatic (colored as before) nor hydrophobic (green) properties change in response to heparin activation. It is thus likely that the properties required for exosite interactions are preformed, but not accessible to the protease without the expulsion of the RCL. (b) A stereo representation of the model of the AT/FXa (cyan) Michaelis complex built on the structure of HCII with S195A thrombin reveals likely candidates for the favorable exosite interactions. In FXa, Arg150 (blue ball) would interact with Glu237 on AT (red ball), and the 30-loop (red balls on FXa) would interact with a basic patch on AT (blue balls). The surface properties corresponding to these residues is shown in (c). The heparin-binding residues of FXa are shown as blue balls in panel (d), and allow the modeling of the bridged heparin complex. The position of the heparin-binding site in the rear of FXa, which is itself docked to the rear of AT, explains why 26-mer heparin chains are insufficient to bridge AT and FXa.

The HCII–thrombin Michaelis complex, however, does not seem to be a good model for the AT–thrombin Michaelis complex. The superposition of active AT on HCII in complex with thrombin creates several steric clashes between the 60- and γ-insertion loops on thrombin and the body of AT. To relieve these clashes it is necessary to increase the distance between thrombin and AT, which is made possible by the additional three amino acids on the P′ side of the RCL in AT relative to HCII (Fig. 7a). Although it is possible to relieve the steric clashes by increasing the distance between the two surfaces, the electrostatic properties at the interface would suggest repulsion, rather than attraction (Fig. 6c). Thus, in contrast to the intimate FXa–AT interface, the thrombin–AT interface would be predicted to be one in which the distance between the two molecules would be maximized in order to avoid contact. Also inconsistent with an AT–thrombin model based on the HCII–thrombin Michaelis complex is that over 30 saccharide units would be required to bridge the complex, whereas it is known that 15 units is sufficient. Thus the AT–thrombin Michaelis complex would probably resemble that of the Manduca sexta serpin with trypsin, and that heparin bridging would be possible with short chains (15-mer) through the tilting of thrombin in the direction of the heparin-binding site of AT (Fig. 7a). Support for a AT–thrombin Michaelis complex with no significant interactions outside the active site is provided by exosite I variants of thrombin, which have no effect of the rate of thrombin inhibition by AT in the absence or presence of heparin [28]. In addition, the contention that the interface is in fact repulsive in nature has been demonstrated by the truncation of the P′ insertion loop by one or two residues (P7′ and P8′) [87]. These mutants show improved basal rates of FXa inhibition and reduced rates of thrombin inhibition.


Figure 7. Models for the bridged complexes between AT and thrombin, and PCI and APC. To avoid steric clashes between thrombin and AT, and because a 15-mer heparin is known to be sufficient to bridge, thrombin must be pulled away from the body of AT and toward the heparin-binding region as shown in (a). Such a movement is made possible by a three-residue insertion in the P′ side of the RCL (orange). The position of thrombin in this Michaelis complex suggests that in the absence of heparin binding, the RCL would be in a native-like conformation with its partial insertion into β-sheet A. The recent structure of PCI has allowed the postulation of a novel mechanism to explain its heparin activation toward APC, which uses elements of the HCII and AT mechanisms. (b) The surface representations reveal charge repulsion at the predicted interaction interface where the basic heparin-binding regions (oval) come into contact in the model. The role of heparin would thus be to alleviate the unfavorable electrostatic repulsion by interposing between the two molecules. (c) The model for the mechanism is that one-dimensional diffusion along a heparin chain will eventually result in a productive encounter.

Download figure to PowerPoint

One implication of the thrombin–AT model is that loop expulsion is not required for the basal rate of thrombin inhibition by AT. Thus, the flexibility evident in the hinge region in the proposed Michaelis complex would allow the partial insertion of the hinge region as is seen in the native structure of AT. The predicted Michaelis complex between thrombin and AT in the absence of heparin would have a partially inserted RCL and an elongated P′ region, and the slow basal rate would be due to the energetic cost of elongating the P′ side of the RCL and the poor complementarity of the RCL for the active site of thrombin. In contrast, the basal rate of FXa inhibition by AT would be dependent on the expulsion of the loop, and would either suggest a rapid equilibrium between the native and activated forms with FXa interacting only with the activated conformation, or an induced-fit mechanism where loop expulsion is brought about by the initial weak interaction between FXa and native AT.

A proposed mechanism for the PCI, APC, heparin complex

  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References

AT and HCII are the only two serpins involved in hemostasis for which we have a detailed structural understanding. It has become clear from many of the structures of these serpins that the serpin family members share many unexpected mechanistic features that have allowed us to combine our knowledge of the individual mechanisms into a larger and more complete framework. In this paper I have described how clues to the outstanding questions about AT and HCII have been provided by analogy, and how this has led to new hypotheses as to how proteases are specifically inhibited by the serpins controlling coagulation. As more structural data becomes available for the other serpins involved in the maintenance of hemostasis, a similar approach can be used to push forward the boundaries of knowledge. A recent example is provided by the structure of proteolytically modified PCI [88]. PCI inhibits APC in a heparin-dependent manner, believed to be analogous to the bridging of AT and thrombin by heparin [89]. The heparin-binding site of APC has been well established through mutagenic studies [90] and that of PCI has been shown to involve helix H [91]. The structure of PCI has allowed closer examination of its surface electrostatic properties and predicts a more extended heparin-binding site near the top of the serpin. When the PCI–APC complex was built on the template of the HCII–thrombin complex the heparin-binding sites of PCI and APC were in direct apposition (Fig. 7b). Thus we proposed that heparin is sandwiched in between PCI and APC in the Michaelis complex (Fig. 7c), and that instead of bridging, heparin provides a new exosite for improved recognition in a manner analogous to the tail of HCII altering its properties at the interaction interface. Support for this model is provided by heparin-binding variants with improved basal rates of inhibition [90] and a heparin length dependency, but a minimal length of only 7 residues [92].


  1. Top of page
  2. Abstract
  3. The serpin mechanism
  4. Thrombin and FXa
  5. Antithrombin
  6. Heparin cofactor II
  7. Models for the AT Michaelis complexes
  8. A proposed mechanism for the PCI, APC, heparin complex
  9. Acknowledgements
  10. References
  • 1
    Hunt LT, Dayhoff MO. A surprising new protein superfamily containing ovalbumin, antithrombin- III, and alpha 1-proteinase inhibitor. Biochem Biophys Res Commun 1980; 95 (2): 86471.
  • 2
    Silverman GA, Bird PI, Carrell RW, Church FC, Coughlin PB, Gettins PGW, Irving JA, Lomas DA, Luke CJ, Moyer RW, Pemberton PA, Remold-O'Donnell E, Salvesen GS, Travis J, Whisstock JC. The serpins are an expanding superfamily of structurally similar but functionally diverse proteins. Evolution, mechanism of inhibition, novel functions, and a revised nomenclature. J Biol Chem 2001; 276: 332936.
  • 3
    Irving JA, Steenbakkers PJ, Lesk AM, Op den Camp HJ, Pike RN, Whisstock JC. Serpins in prokaryotes. Mol Biol Evol 2002; 19: 188190.
  • 4
    Gettins PGW, Patston PA, Olson ST. Serpins: Structure, function and biology. Austin, TX: R.G Landes Co, 1996.
  • 5
    Bird PI. Serpins and regulation of cell death. Results Probl Cell Differ 1998; 24: 6389.
  • 6
    Elliott PR, Abrahams JP, Lomas DA. Wild-type alpha 1-antitrypsin is in the canonical inhibitory conformation. J Mol Biol 1998; 275: 41925.
  • 7
    Carrell RW, Evans DL, Stein PE. Mobile reactive centre of serpins and the control of thrombosis. Nature 1991; 353 (6344): 5768.
  • 8
    Mottonen J, Strand A, Symersky J, Sweet RM, Danley DE, Geoghegan KF, Gerard RD, Goldsmith EJ. Structural basis of latency in plasminogen activator inhibitor-1. Nature 1992; 355 (6357): 2703.
  • 9
    Zhou A, Huntington JA, Carrell RW. Formation of the antithrombin heterodimer in vivo and the onset of thrombosis. Blood 1999; 94: 338896.
  • 10
    Loebermann H, Tokuoka R, Deisenhofer J, Huber R. Human alpha 1-proteinase inhibitor. Crystal structure analysis of two crystal modifications, molecular model and preliminary analysis of the implications for function. J Mol Biol 1984; 177: 53157.
  • 11
    Li J, Wang Z, Canagarajah B, Jiang H, Kanost M, Goldsmith EJ. The structure of active serpin 1K from Manduca sexta. Structure Fold Des 1999; 7: 1039.
  • 12
    Huntington JA, Read RJ, Carrell RW. Structure of a serpin-protease complex shows inhibition by deformation. Nature 2000; 407 (6806): 9236.
  • 13
    Olson ST, Bock PE, Kvassman J, Shore JD, Lawrence DA, Ginsburg D, Bjork I. Role of the catalytic serine in the interactions of serine proteinases with protein inhibitors of the serpin family. Contribution of a covalent interaction to the binding energy of serpin-proteinase complexes. J Biol Chem 1995; 270: 3000717.
  • 14
    Stone SR, Le Bonniec BF. Inhibitory mechanism of serpins. Identification of steps involving the active-site serine residue of the protease. J Mol Biol 1997; 265: 34462.
  • 15
    Lawrence DA, Olson ST, Muhammad S, Day DE, Kvassman JO, Ginsburg D, Shore JD. Partitioning of serpin-proteinase reactions between stable inhibition and substrate cleavage is regulated by the rate of serpin reactive center loop insertion into beta-sheet A. J Biol Chem 2000; 275: 583944.
  • 16
    Olson ST, Swanson R, Day D, Verhamme I, Kvassman J, Shore JD. Resolution of Michaelis complex, acylation, and conformational change steps in the reactions of the serpin, plasminogen activator inhibitor-1, with tissue plasminogen activator and trypsin. Biochemistry 2001; 40: 1174256.
  • 17
    Huber R, Bode W. Structural basis of the activation and action of trypsin. Acc Chem Res 1978; 11: 11422.
  • 18
    Bock PE, Olson ST, Bjork I. Inactivation of thrombin by antithrombin is accompanied by inactivation of regulatory exosite I. J Biol Chem 1997; 272: 1983745.
  • 19
    Fredenburgh JC, Stafford AR, Weitz JI. Conformational changes in thrombin when complexed by serpins. J Biol Chem 2001; 276: 4482834.
  • 20
    Kaslik G, Patthy A, Balint M, Graf L. Trypsin complexed with alpha 1-proteinase inhibitor has an increased structural flexibility. FEBS Lett 1995; 370: 17983.
  • 21
    Stavridi ES, O'Malley K, Lukacs CM, Moore WT, Lambris JD, Christianson DW, Rubin H, Cooperman BS. Structural change in alpha-chymotrypsin induced by complexation with alpha 1-antichymotrypsin as seen by enhanced sensitivity to proteolysis. Biochemistry 1996; 35: 1060815.
  • 22
    Egelund R, Petersen TE, Andreasen PA. A serpin-induced extensive proteolytic susceptibility of urokinase-type plasminogen activator implicates distortion of the proteinase substrate-binding pocket and oxyanion hole in the serpin inhibitory mechanism. Eur J Biochem 2001; 268: 67385.
  • 23
    Hoffman M, Pratt CW, Brown RL, Church FC. Heparin cofactor II-proteinase reaction products exhibit neutrophil chemoattractant activity. Blood 1989; 73: 16825.
  • 24
    Parmar JS, Mahadeva R, Reed BJ, Farahi N, Cadwallader KA, Keogan MT, Bilton D, Chilvers ER, Lomas DA. Polymers of alpha (1) – antitrypsin are chemotactic for human neutrophils: a new paradigm for the pathogenesis of emphysema. Am J Respir Cell Mol Biol 2002; 26: 72330.
  • 25
    Huntington JA, Carrell RW. The serpins: nature's molecular mousetraps. Sci Prog 1901; 84: 12536.
  • 26
    Gruber A, Cantwell AM, Di Cera E, Hanson SR. The thrombin mutant W215A/E217A shows safe and potent anticoagulant and antithrombotic effects in vivo. J Biol Chem 2002; 277: 275814.
  • 27
    Stubbs MT, Bode W. A player of many parts: the spotlight falls on thrombin's structure. Thromb Res 1993; 69: 158.
  • 28
    Myles T, Church FC, Whinna HC, Monard D, Stone SR. Role of thrombin anion-binding exosite-I in the formation of thrombin- serpin complexes. J Biol Chem 1998; 273: 312038.
  • 29
    Sheehan JP, Sadler JE. Molecular mapping of the heparin-binding exosite of thrombin. Proc Natl Acad Sci USA 1994; 91: 551822.
  • 30
    Tsiang M, Jain AK, Gibbs CS. Functional requirements for inhibition of thrombin by antithrombin III in the presence and absence of heparin. J Biol Chem 1997; 272: 120249.
  • 31
    Huntington JA, Esmon CT. The molecular basis of thrombin allostery revealed by a 1.8 Å structure of the "Slow" form. Structure 2003; 11: 46979.
  • 32
    Rezaie AR. Calcium enhances heparin catalysis of the antithrombin-factor Xa reaction by a template mechanism. Evidence that calcium alleviates Gla domain antagonism of heparin binding to factor Xa. J Biol Chem 1998; 273: 168247.
  • 33
    Rezaie AR, Olson ST. Calcium enhances heparin catalysis of the antithrombin-factor Xa reaction by promoting the assembly of an intermediate heparin-antithrombin-factor Xa bridging complex. Demonstration by rapid kinetics studies. Biochemistry 2000; 39: 1208390.
  • 34
    Rezaie AR. Identification of basic residues in the heparin-binding exosite of factor Xa critical for heparin and factor Va binding. J Biol Chem 2000; 275: 33207.
  • 35
    Rezaie AR. Heparin-binding exosite of factor Xa. Trends Cardiovasc Med 2000; 10: 3338.
  • 36
    Hirsh J. Blood tests for the diagnosis of venous and arterial thrombosis. Blood 1981; 57: 18.
  • 37
    Kojima T. Targeted gene disruption of natural anticoagulant proteins in mice. Int J Hematol 2002; 76 (Suppl. 2): 369.
  • 38
    Casu B, Lindahl U. Structure and biological interactions of heparin and heparan sulfate. Adv Carbohydr Chem Biochem 2001; 57: 159206.
  • 39
    Lindahl U, Thunberg L, Backstrom G, Riesenfeld J. The antithrombin-binding sequence of heparin. Biochem Soc Trans 1981; 9: 49951.
  • 40
    Casu B, Oreste P, Torri G, Zoppetti G, Choay J, Lormeau JC, Petitou M, Sinay P. The structure of heparin oligosaccharide fragments with high anti-(factor Xa) activity containing the minimal antithrombin III-binding sequence. Chemical and 13C nuclear-magnetic-resonance studies. Biochem J 1981; 197: 599609.
  • 41
    Huber R, Carrell RW. Implications of the three-dimensional structure of alpha 1-antitrypsin for structure and function of serpins. Biochemistry 1989; 28: 895166.
  • 42
    Arocas V, Bock SC, Olson ST, Bjork I. The role of Arg46 and Arg47 of antithrombin in heparin binding. Biochemistry 1999; 38: 10196204.
  • 43
    Arocas V, Bock SC, Raja S, Olson ST, Bjork I. Lysine 114 of antithrombin is of crucial importance for the affinity and kinetics of heparin pentasaccharide binding. J Biol Chem 2001; 276: 4380917.
  • 44
    Desai U, Swanson R, Bock SC, Bjork I, Olson ST. Role of arginine 129 in heparin binding and activation of antithrombin. J Biol Chem 2000; 275: 1897684.
  • 45
    Schnedin-Weiss S, Desai UR, Bock SC, Gettins PGW, Olson ST, Bjork I. Importance of lysine 125 for heparin binding and activation of antithrombin. Biochemistry 2002; 41: 477988.
  • 46
    Jin L, Abrahams JP, Skinner R, Petitou M, Pike RN, Carrell RW. The anticoagulant activation of antithrombin by heparin. Proc Natl Acad Sci USA 1997; 94: 146838.
  • 47
    Olson ST, Srinivasan KR, Bjork I, Shore JD. Binding of high affinity heparin to antithrombin III. Stopped flow kinetic studies of the binding interaction. J Biol Chem 1981; 256: 110739.
  • 48
    Desai UR, Petitou M, Bjork I, Olson ST. Mechanism of heparin activation of antithrombin. Role of individual residues of the pentasaccharide activating sequence in the recognition of native and activated states of antithrombin. J Biol Chem 1998; 273: 747887.
  • 49
    Carrell RW, Stein PE, Fermi G, Wardell MR. Biological implications of a 3Å structure of dimeric antithrombin. Structure 1994; 2: 25770.
  • 50
    Schreuder HA, De Boer B, Dijkema R, Mulders J, Theunissen HJ, Grootenhuis PD, Hol WG. The intact and cleaved human antithrombin III complex as a model for serpin–proteinase interactions. Nat Struct Biol 1994; 1: 4854.
  • 51
    Meagher JL, Olson ST, Gettins PG. Critical role of the linker region between helix D and strand 2A in heparin activation of antithrombin. J Biol Chem 2000; 275: 2698704.
  • 52
    Belzar KJ, Zhou A, Carrell RW, Gettins PG, Huntington JA, Helix D elongation and allosteric activation of antithrombin. J Biol Chem 2002; 277: 85518.
  • 53
    Mushunje A, Zhou A, Huntington JA, Conard J, Carrell RW. Antithrombin ‘DREUX’ (Lys 114Glu): a variant with complete loss of heparin affinity. Thromb Haemost 2002; 88: 43643.
  • 54
    Turk B, Brieditis I, Bock SC, Olson ST, Bjork I. The oligosaccharide side chain on Asn-135 of alpha-antithrombin, absent in beta-antithrombin, decreases the heparin affinity of the inhibitor by affecting the heparin-induced conformational change. Biochemistry 1997; 36: 668291.
  • 55
    Olson ST, Bjork I. Role of protein conformational changes, surface approximation and protein cofactors in heparin-accelerated antithrombin-proteinase reactions. Adv Exp Med Biol 1992; 313: 15565.
  • 56
    Tollefsen DM. Heparin cofactor II. Adv Exp Med Biol 1997; 425: 3544.
  • 57
    Andersson TR, Larsen ML, Handeland GF, Abildgaard U. Heparin cofactor II activity in plasma: application of an automated assay method to the study of a normal adult population. Scand J Haematol 1986; 36: 96102.
  • 58
    Bertina RM, Van DL I, Engesser L, Muller HP, Brommer EJ. Hereditary heparin cofactor II deficiency and the risk of development of thrombosis. Thromb Haemost 1987; 57: 196200.
  • 59
    Griffith MJ, Carraway T, White GC, Dombrose FA. Heparin cofactor activities in a family with hereditary antithrombin III deficiency: evidence for a second heparin cofactor in human plasma. Blood 1983; 61: 1118.
  • 60
    Fernandez F, Van Ryn J, Ofosu FA, Hirsh J, Buchanan MR. The haemorrhagic and antithrombotic effects of dermatan sulphate. Br J Haematol 1986; 64: 30917.
  • 61
    Carrie D, Caranobe C, Gabaig AM, Larroche M, Boneu B. Effects of heparin, dermatan sulfate and of their association on the inhibition of venous thrombosis growth in the rabbit. Thromb Haemost 1992; 68: 63741.
  • 62
    Liaw PC, Becker DL, Stafford AR, Fredenburgh JC, Weitz JI. Molecular basis for the susceptibility of fibrin-bound thrombin to inactivation by heparin cofactor ii in the presence of dermatan sulfate but not heparin. J Biol Chem 2001; 276: 2095965.
  • 63
    McGuire EA, Tollefsen DM. Activation of heparin cofactor II by fibroblasts and vascular smooth muscle cells. J Biol Chem 1987; 262: 16975.
  • 64
    Whinna HC, Choi HU, Rosenberg LC, Church FC. Interaction of heparin cofactor II with biglycan and decorin. J Biol Chem 1993; 268: 39204.
  • 65
    Church FC, Pratt CW, Hoffman M. Leukocyte chemoattractant peptides from the serpin heparin cofactor II. J Biol Chem 1991; 266: 7049.
  • 66
    Griffith MJ, Noyes CM, Tyndall JA, Church FC. Structural evidence for leucine at the reactive site of heparin cofactor II. Biochemistry 1985; 24: 677782.
  • 67
    Church FC, Noyes CM, Griffith MJ. Inhibition of chymotrypsin by heparin cofactor II. Proc Natl Acad Sci USA 1985; 82: 64314.
  • 68
    Pratt CW, Tobin RB, Church FC. Interaction of heparin cofactor II with neutrophil elastase and cathepsin G. J Biol Chem 1990; 265: 60927.
  • 69
    Van DV, Tollefsen DM. The N-terminal acidic domain of heparin cofactor II mediates the inhibition of alpha-thrombin in the presence of glycosaminoglycans. J Biol Chem 1991; 266: 2022331.
  • 70
    Tollefsen DM, Pestka CA, Monafo WJ. Activation of heparin cofactor II by dermatan sulfate. J Biol Chem 1983; 258: 67136.
  • 71
    Maimone MM, Tollefsen DM. Activation of heparin cofactor II by heparin oligosaccharides. Biochem Biophys Res Commun 1988; 152: 105661.
  • 72
    Maimone MM, Tollefsen DM. Structure of a dermatan sulfate hexasaccharide that binds to heparin cofactor II with high affinity. J Biol Chem 1991; 266: 14830.
  • 73
    Blinder MA, Andersson TR, Abildgaard U, Tollefsen DM. Heparin cofactor II Oslo. Mutation of Arg-189 to His decreases the affinity for dermatan sulfate. J Biol Chem 1989; 264: 512833.
  • 74
    Whinna HC, Blinder MA, Szewczyk M, Tollefsen DM, Church FC. Role of lysine 173 in heparin binding to heparin cofactor II. J Biol Chem 1991; 266: 812935.
  • 75
    Blinder MA, Tollefsen DM. Site-directed mutagenesis of arginine 103 and lysine 185 in the proposed glycosaminoglycan-binding site of heparin cofactor II. J Biol Chem 1990; 265: 28691.
  • 76
    Hayakawa Y, Hirashima Y, Kurimoto M, Hayashi N, Hamada H, Kuwayama N, Endo S. Contribution of basic residues of the A helix of heparin cofactor II to heparin- or dermatan sulfate-mediated thrombin inhibition. FEBS Lett 2002; 522: 14750.
  • 77
    Sheehan JP, Tollefsen DM, Sadler JE. Heparin cofactor II is regulated allosterically and not primarily by template effects. Studies with mutant thrombins and glycosaminoglycans. J Biol Chem 1994; 269: 3274751.
  • 78
    Baglin TP, Carrell RW, Church FC, Esmon CT, Huntington JA. Crystal structures of native and thrombin-complexed heparin cofactor II reveal a multistep allosteric mechanism. Proc Natl Acad Sci USA 2002; 99: 1107984.
  • 79
    Sheffield WP, Blajchman MA. Deletion mutagenesis of heparin cofactor II. defining the minimum size of a thrombin inhibiting serpin. FEBS Lett 1995; 365: 18992.
  • 80
    Mellet P, Bieth JG. Evidence that translocation of the proteinase precedes its acylation in the serpin inhibition pathway. J Biol Chem 2000; 275: 1078895.
  • 81
    Peterson FC, Gordon NC, Gettins PG. Formation of a noncovalent serpin-proteinase complex involves no conformational change in the serpin. Use of 1H−15N HSQC NMR as a sensitive nonperturbing monitor of conformation. Biochemistry 2000; 39: 1188492.
  • 82
    Ye S, Chech AL, Belmares R, Bergstrom RC, Tong Y, Corey DR, Kanost MR, Goldsmith EJ. The structure of a Michaelis serpin-protease complex. Nat Struct Biol 2001; 8: 97983.
  • 83
    Whisstock JC, Pike RN, Jin L, Skinner R, Pei XY, Carrell RW, Lesk AM. Conformational changes in serpins: II. The mechanism of activation of antithrombin by heparindagger. J Mol Biol 2000; 301: 1287305.
  • 84
    Manithody C, Yang L, Rezaie AR. Role of basic residues of the autolysis loop in the catalytic function of factor Xa. Biochemistry 2002; 41: 67808.
  • 85
    Quinsey NS, Whisstock JC, Le Bonniec B, Louvain V, Bottomley SP, Pike RN. Molecular determinants of the mechanism underlying acceleration of the interaction between antithrombin and factor Xa by heparin pentasaccharide. J Biol Chem 2002; 277: 159718.
  • 86
    Duchaussoy P, Jaurand G, Driguez PA, Lederman I, Ceccato ML, Gourvenec F, Strassel JM, Sizun P, Petitou M, Herbert JM. Assessment through chemical synthesis of the size of the heparin sequence involved in thrombin inhibition. Carbohydr Res 1999; 317: 8599.
  • 87
    Rezaie AR. Partial activation of antithrombin without heparin through deletion of a unique sequence on the reactive site loop of the serpin. J Biol Chem 2002; 277: 12359.
  • 88
    Huntington JA, Kjellberg M, Stenflo J. Crystal structure of protein C inhibitor provides insights into hormone binding and heparin activation. Structure 2003; 11: 20515.
  • 89
    Kazama Y, Niwa M, Yamagishi R, Takahashi K, Sakuragawa N, Koide T. Specificity of sulfated polysaccharides to accelerate the inhibition of activated protein C by protein C inhibitor. Thromb Res 1987; 48: 17985.
  • 90
    Friedrich U, Blom AM, Dahlback B, Villoutreix BO. Structural and energetic characteristics of the heparin-binding site in antithrombotic protein C. J Biol Chem 2001; 276: 241228.
  • 91
    Shirk RA, Elisen MG, Meijers JC, Church FC. Role of the H helix in heparin binding to protein C inhibitor. J Biol Chem 1994; 269: 286905.
  • 92
    Aznar J, Espana F, Estelles A, Royo M. Heparin stimulation of the inhibition of activated protein C and other enzymes by human protein C inhibitor – influence of the molecular weightof heparin and ionic strength. Thromb Haemost 1996; 76: 9838.