Antithrombin-III
Created by Audrey Johnston
Human antithrombin-III (PDB ID: 1NQ9) is a protein in the serpin family that functions as an inhibitor of multiple factors within the blood clotting response pathway. Found throughout the blood plasma of Homo sapiens, antithrombin is capable of acting on all coagulation serine proteases, but does so most notably on thrombin and factors Xa and IXa (2,4).Thus, antithrombin serves primarily to prevent clotting action from straying past an isolated injury site. Individuals with an abnormal version of this protein characteristically develop thrombosis, or unnecessary clots in the blood stream (2, 4). In order to effectively carry out its functions, antithrombin assumes specific structural conformations via interactions with itself and other cofactors.
Although its physiological state is typically a single peptide with only one subunit, antithrombin often crystallizes as a dimer that includes both its latent and unactivated states. Thus, its crystal structure has a molecular weight of 98060.27 Da (rather than 58 kDa reported for its single, 432 residue chain monomer) and an isoelectric point of 5.95. The unactivated or native state of antithrombin refers to its conformation sans prior interaction with its cofactors- heparin or heparin sulfate. The secondary structure of this native state is comprised of three beta sheets and nine alpha helices interconnected by random coils, all constructed around a five stranded beta sheet core. This core is referred to as the A-sheet and is an essential component of conformational changing (1).
The native tertiary structure adopted by this protein contains a domain specific for the heparin pentasaccharide ligand. Coordination of the N-terminal region of the protein, along with the N-terminal end of helix A and all of helix D ultimately forms said binding site (2). Antithrombin’s folded structure also includes two significant surface exosites that are essential in determining the protein’s active state, as well as two functional hinge regions that facilitate the serpin’s inhibitory reactions. Three intramolecular disulfide bonds occur between six cysteine residues 8-128, 21-95, and 247-430. These serve to stabilize the native state (4). Also present on the mature molecule’s surface are several distinct locations of glycosylation, specifically N-acetyl-D-glucosamine (NAG) in the dimer crystallization. In addition, antithrombin's physiological state is found to be complexed with several water molecules.
As antithrombin cycles through its occupation as a serine protease inhibitor, it transitions through several allosteric alterations. The first stage in a rather stepwise activation process is considered low affinity binding of heparin pentasaccharide. The characteristic dissociation constant at this stage is about 20 uM (3). Following this weaker intermediate complex is a conformational change that drastically improves further affinity for heparin. An increase in the dissociation constant to around 50 nM demonstrates how accommodating this binding site becomes towards its cofactor (3). Next, high affinity heparin binding induces the expulsion of antithrombin’s functional reactive center loop (RCL), and its serpin activity may be carried out. Antithrombin is now said to be in its active state.
Achieving the active state first involves recognition of the heparin ligand. This occurs between three saccharides of the sugar’s nonreducing end and key residues within antithrombin’s binding site. Several basic residues: Lys-11, Arg-13, Arg-46, Arg-47, Lys-114, Lys-125 and Arg-129 interact ionically and through hydrogen bonds with the negative sulfate and carboxylate groups of the heparin ligand (2). Specifically, Lys-125 and Lys-11 form a critical electrostatic interaction with the 3-O-sulfate of the pentasaccharide (2). Minor structural changes proceed this binding and contribute to the second phase of better-fitting interaction with heparin. Slight allosteric repositioning of residues in the beta core, as well as induced formation of the P-helix, cause further electrostatic rearrangements on the protein’s surface. As a result, previously unfavorable exosite interactions with target proteases Xa and IXa are mitigated.Tyr-253 and Glu-255 exosites are relieved and able to act favorably with these proteases. This freedom thereby increases their interaction and propagates further important conformational changes. Finally, helix D extends, which pushes Lys-125 and Arg-129 of the binding site into a more heparin-friendly position, and the active state is reached (2). Overall, such a mechanism maintains stability of the activated molecule, as well as allowing its function to be specific and regulated.
One of the most important implications of the induced active state of antithrombin is the expulsion of its reactive center loop (RCL), also known as the reactive site domain. This structure projects upward from the beta core and frees Arg-393 in the P helix to fulfill inhibitory contacts with target proteases (4). Without this important structural change, the cleavage mechanism responsible for rendering such clotting factors as Xa and IXa inactive would fail. Thus, the RCL portion of antithrombin is of high structural significance. It is a main component of the ‘suicide pathway’ antithrombin uses for inhibition that results in its own inactivation (2).
The suicide pathway of antithrombin initiates when the P1 arginine (Arg-393) is recognized and bound by a target protease. Once bound, a catalytic serine in the protease “attacks” the P1 arginine, resulting in cleavage of the RCL (2). However, this intermediate cleaved complex remains covalently attached by an ester linkage to the catalytic serine via the P1 arginine residue (1, 2). The cleavage process signals and initiates the major switch-like conformational change that ultimately sequesters coagulation factors in an inactived state. Several critical domains are essential to this function, namely the RCL and two special hinge regions. These two hinge domains are responsible for carrying out this switch mechanism. A proximal hinge is formed by residues 380-384, which are primarily alanines, and residues 402-407 form a distal hinge. A gate loop held together by a disulfide bond between Cys-247 and Cys-430 prevents the hinges from facilitating movement before RCL cleavage (4). Active antithrombin with a cleaved reactive site undergoes a structural change in which the RCL, with the attached acyl intermediate, partially inserts back into the beta core. This triggers gate opening and hinge activation. A massive hinge switch flips the bound protease to the opposite side of the serpin. This final change in conformation irreversibly inactivates the antithrombin-protease complex. Due to this inactivation mechanism, antithrombin protein must be constantly replenished in the body to maintain its function.
Residues along the protruding RCL are not the only important recognition sites for target proteases. Recent research has shown that the specific exosites mentioned in paragraph five are fundamentally important in antithrombin’s anticoagulant function. Once heparin has alleviated constraints on the exosites in the native form, binding affinity for coagulation factors increases upwards of 1000-fold (9). In addition to the conformational changes induced by heparin, exosite interactions too increase the efficiency of antithrombin inhibition. In fact, physiologically required rates of clotting protease inhibition are only achievable through the function of exosites (9).
Along with rate acceleration, exosites also play a key role in determining and controlling substrate specificity (8). For example, antithrombin’s two exosites have an extremely high specificity for the Arg-150 residue of the autolysis loop of factors Xa and IXa. Expectedly, when heparin binding induces allosteric changes that uncover the exosites, the factors and antithrombin readily interact (9). Yet, antithrombin’s other target molecule- thrombin- achieves no such surge in activity. Instead, thrombin inactivation must occur through a bridging technique that is modulated by heparin itself. This is because thrombin lacks the highly conserved and specific Arg-150 residue (2, 3). It cannot be inactivated by the switch mechanism. Thrombin inhibition is therefore only increased two-fold by heparin pentasaccharide association, while Xa and IXa are inhibited at 300-fold increased rate (3). Exosite interaction combined with ligand binding is thus responsible for controlling the majority of antithrombin’s highly specific and regulated function. This model system for serpin exosite specificity can even be manipulated and implemented medically as a way to engineer inhibition of nontarget proteases (9).
Antithrombin’s role in the human coagulation pathway is highly essential. Its typical concentration in the human bloodstream is about 2.3 uM. Though the protein is capable of acting on its targets in its native (unbound) state, heparin binding alone increases affinity for such targets up to 300-fold (1). Therefore, higher heparin binding equates to greater efficiency and activity of antithrombin. Two isoforms of antithrombin-III coexist and have different innate binding affinities for the heparin pentasaccharide ligand. Variance between the two forms is strictly related to protein glycosylation. The more abundant alpha form constitutes about 90% of the total antithrombin protein found in circulation and contains four distinct locations of N-linked glycosylation. Beta antithrombin on the other hand, makes up only 5-10% and is found with only three oligosaccharide side chains (5). The beta form is known to have a substantially higher affinity for heparin despite its rarity. Both chains have identical sequences.
Antithrombin usually bears oligosaccharide chains at asparagine residues 96, 135, 155, and 172. The beta isoform lacks glycosylation at Asn-135. Having an oligosaccharide chain at Asn-135 creates a bulky interference with the heparin binding site (5). Without the presence of this chain, steric constraints are alleviated and more favorable conformational changes ensue. The end result is a beta antithrombin with a three-fold increase in heparin pentasaccharide binding affinity (1). Since association with extra carbohydrate side chains normally functions to preserve the protein’s stability, beta antithrombin is seen to degrade more rapidly in vivo. Consequently, the smaller concentration of beta antithrombin is partially due to the molecule’s reduced half-life. Alpha antithrombin is thought to be converted to the beta form during blood circulation. Lability of the Asn-135 oligosaccharide is the proposed mechanism, and overall the interconversion ability is considered another means of self-regulation (5).
Beta antithrombin is considered to be physiologically more important in thrombin inhibition events. Because it reacts more efficiently with heparin, it can respond to lower concentrations of the ligand. It also has a higher baseline capacity for inhibition without stimulation from heparin. Part of the reason for its comparatively lower concentration may thus be a function of its higher turnover rate in the blood stream (1). Since alpha antithrombin is only sparsely attracted to heparin in its native state (before its high affinity allosteric changes), beta antithrombin’s high affinity that occurs without the multistep activation process is beneficial. It thus exhibits a greater measure of activity. Therefore, beta antithrombin would effectively be used up (converted to its permanent, inactive state) faster, rendering it less abundant.
Antithrombin is classified as a serpin family protein. It is similar in structure and function to another serpin member- antichymotrypsin (PDB ID: 3DLW-A)-also found in Homo sapiens. Protein BLAST and DALI analysis (E=4xe^-142; Z=39.2 and rmsd= 2.1) respectively indicated that the two molecules share homology in both sequence and tertiary structure (significance value: 0.05). PSI BLAST reported a 93% measure of sequence homology. DALI results confirmed the highly comparable three-dimensional folding of the two proteins (10). Antichymotrypsin (ACT) is composed of 12 alpha helices and 18 beta strands, while antithrombin is similarly comprised of 14 helices and 18 beta strands. ACT, like antithrombin, has a tertiary structure built around a large, central beta-sheet core and a functional reactive center loop for protease binding (7). Analogous distal and proximal hinge domains are present in tertiary structure. This is relevant as both molecules engage in the same trapping mechanism for targeting and inhibiting certain proteinase enzymes. However, in its presented form, ACT lacks associated ligands, whereas antithrombin-III is bound to multiple NAG groups as well as heparin.
In particular, this form of ACT is too crystallized in a misfolded, latent phase conformation. Steric strains in native tertiary structure for both proteins allow versatility in assuming different stages and functional conformations, yet make them susceptible to self-association and inhibition as well (such as the antithrombin-III latent-native crystal dimer). Thus both proteins are prone to being found in a nonfunctional, self-induced inactive state. Studying misfolding/self-aggregation pathways for proteins like these may be preventative, as serpins seem to have an innate inclination to form improper, disease-causing conformations (6).
Horse leukocyte elastase inhibitor (PDB ID: 1HLE) is a protein found in Equus callabus that is also related to antithrombin-III in structure and function. PSI BLAST (E= 4e^-128) issued a sequence homology of 75%. While primary structure is only moderately related to antithrombin,their compared tertiary structures through DALI are quite similar (Z= 36.0, rmsd= 2.0) (10). Horse elastase inhibitor’s secondary structure consists of 15 helices and 14 beta sheets, which in turn fold in a remarkably similar way to the general serpin configuration adopted by antithrombin. Unlike antithrombin, this protein consists of two subunits, ones that are nonidentical at that. Antithrombin is primarily a single chain peptide and will occasionally coalesce with itself as a homodimer (but as a native-latent complex). Thus, tertiary similarities are likely the result of similar functions, and not as much due to similar sequence. In all, most data tend to cite alpha-antitrypsin as the preferred model of comparison for both horse elastase inhibitor as well as ACT.
Because antithrombin has multiple critical sites/domains in its tertiary structure, point mutations could easily render the protein useless. Antithrombin deficiency has a reported prevalence of between 1 in 2000 and 1 in 5000 individuals (4). Replacement of essential positively charged amino acid side chains in the heparin binding site can be detrimental to ligand affinity. Indirectly, mutations at Pro-41, Phe-99, Ser-116 and Gln-118 can disrupt heparin binding as well. These specific residues constitute a tightly packed region between the amino end of the protein and the A and D helices that when disrupted alter the binding site domain (4). Creation of an extra glycosylation site can also sterically interrupt heparin binding. Other domains, such as the hinges, are mainly impaired by substitutions with bulky or polar residues. Mutations that impede the reactive site largely interfere with substrate binding, and protease inhibition is expectedly altered. Increased risk of thrombosis, or blood clots, are consequential for individuals with abnormal antithrombin protein. In fact, a complete antithrombin deficiency is lethal in mice embryos (2). Clearly, the role of antithrombin in humans is imperative in nature.