Antithrombin III and its interaction with heparin. Comparison of the human, bovine, and porcine proteins by 1H NMR spectroscopy

1H NMR has been used to characterize and compare the structures of antithrombin III from human, bovine, and porcine plasma as well as to investigate the interactions of each of these proteins with heparin fragments of defined length. The amino acid compositions of the three proteins are very similar, which is reflected in the gross features of their 1H NMR spectra. In addition, aromatic and methyl proton resonances in upfield-shifted positions appear to be common to all three proteins and suggest similar tertiary structures. Human antithrombin III has five histidine residues, bovine has six, and porcine has five. The C(2) proton from each of these residues gives a narrow resonance and titrates with pH; the pKa's are in the range 5.15-7.25. It is concluded that all histidines in each protein are surface residues with considerable independent mobility. The carbohydrate chains in each protein also give sharp resonances consistent with a surface location and motional flexibility. The 1H spectra are sensitive to heparin binding. Although heparin resonances obscure protein resonances in the region 3.2-6.0 ppm, difference spectra between antithrombin III with and without heparin show clear perturbation of a small number of aromatic and aliphatic protein protons. These resonances include those of histidine C(2) and C(4) protons, of 10-20 other aromatic protons, of a methyl group, and also of protons with chemical shifts similar to those of lysine and/or arginine side chains. For human antithrombin III, it was shown that heparin fragments 8, 10, and 16 sugar residues in length result in almost identical perturbations to the protein. In contrast, tetrasaccharide results in fewer perturbations.(ABSTRACT TRUNCATED AT 250 WORDS)     

Proton nuclear magnetic resonance studies on bovine lutropin, its subunits, and on the alpha subunit of pregnant mare serum gonadotropin. Assignment of histidine resonances in the alpha subunit.

The pK values of the 3 histidine residues in the common alpha subunits of bovine and equine glycoprotein hormones have been determined from titration curves generated from their C-2 proton nuclear magnetic resonances at different pH values. Assignment of resonances to specific histidines is based on a comparison between the two species, which have 1 histidine residue in different positions in their sequences, and of the bovine alpha subunit after removal of its histidine 94 by treatment with carboxypeptidases. In both species, those histidines closest to the COOH terminus titrate with near normal pK values of 6.2. The histidine residue found in the bovine subunit at position 87 titrates with an approximate pK value of 5.4. Histidine 83, adjacent to an oligosaccharide moiety in both species, does not titrate over a pH range of 4.0 to 8.0 and thus appears inaccessible to solvent. Similarly, in bovine lutropin-beta, 1 of 3 histidine residues does not titrate between pH 5.0 and 7.0. In the intact hormone, 2 "nontitratable" histidine residues are found. Changes in the characteristics of the signals, however, preclude unambiguous assignment of these two resonances to the nontitrating histidines in the isolated subunits. It appears that changes in the environment of at least some histidines occur when the subunits combine to yield intact hormone.     

Examination, by 1H-n.m.r. spectroscopy, of the binding of a synthetic, high-affinity heparin pentasaccharide to human antithrombin III

Binding of a synthetic, high-affinity heparin pentasaccharide and of intact heparin to both native and elastase-modified human antithrombin III have been examined by 1H-n.m.r. spectroscopy. The pentasaccharide perturbs many protein resonances in the same way as does intact heparin. There are, however, differences that seem to arise both from fewer contacts in the heparin binding-site when the pentasaccharide binds and from dissimilar conformational changes in the protein. The resonance of the H-2 atom of the histidine, considered to be the N-terminal residue and to be located in the heparin binding-site, is strongly perturbed by heparin binding both to native and modified antithrombin. The pentasaccharide has little effect on this histidine in either protein. Resonances from two of the remaining four histidine units are sensitive to longer-range conformational changes, and show differences between binding of the two heparin species both in native and modified ATIII. It is concluded that the pentasaccharide only partly fills the heparin binding-site and does not produce a conformational change identical to that caused by intact heparin. This is particularly significant as regards the mechanism of action of heparin, because the synthetic pentasaccharide activates ATIII towards Factor Xa, but not towards thrombin.     

Properties of thrombin- and elastase-modified human antithrombin III

Proteolytically modified forms of human antithrombin III have been prepared by reaction of native antithrombin with thrombin, human neutrophil elastase, or porcine pancreatic elastase. These forms have two chains disulfide linked and are of the same molecular weight as native antithrombin III. 1H NMR spectroscopy has been used to characterize these proteins and to compare them to one another and to native antithrombin III. The three modified proteins have very similar NMR spectra and histidine residues with identical pH titration parameters, and they undergo the same spectral changes upon binding heparin. They differ from native antithrombin III in all of these respects. In addition, the proteins are much more stable than native antithrombin III. The three modified proteins behave identically as a function of temperature; at 372 K, 44 K above the unfolding temperature for native antithrombin III, the proteins are still folded and possess approximately 70 unexchanged amide protons even after several hours. The unfolding of the heparin binding domain at low concentrations of deuteriated guanidine hydrochloride seen in native thrombin III is absent in the modified forms. It is concluded that the thrombin- and elastase-modified forms of antithrombin have identical structures when allowance is made for the slightly different sites of cleavage by the two types of elastase and by thrombin. This structure is very different from that of native antithrombin III.(ABSTRACT TRUNCATED AT 250 WORDS)     

Refolding properties of antithrombin III. Mechanism of binding to heparin

The presence of two unfolding domains in antithrombin III during its denaturation in guanidinium chloride has previously been reported (Villanueva, G. B., and Allen, N. (1983) J. Biol. Chem. 258, 11010-11013). In the present work, we report the results of refolding studies on antithrombin III. Circular dichroism and intrinsic fluorescence studies have demonstrated that the first unfolding domain of low stability (midpoint at 0.7 M guanidinium chloride) is irreversible upon renaturation, whereas the second unfolding domain (midpoint at 2.3 M guanidinium chloride) is reversible. The intermediate form of antithrombin III, termed AT-IIIR, which has lost the structural features of the first domain was investigated. Clotting assays and electrophoretic analyses showed that AT-IIIR had lost 60% of heparin cofactor activity but was still capable of forming sodium dodecyl sulfate-stable complexes with thrombin. Although certain regions of this molecule do not refold to the conformation of native antithrombin III, the tryptophan residues refold to a conformation identical with the native state. This was demonstrated by fluorescence quenching, solvent perturbation, and chemical modification studies. However, the tryptophan-ascribed fluorescence enhancement and absorption difference spectrum which occur when heparin binds to antithrombin III are reduced by 70%. On the basis of these data, the binding of heparin to antithrombin III is interpreted in terms of a two-step mechanism. The primary binding occurs in the region without tryptophan and is followed by a secondary conformational rearrangement which affects the tryptophan environment. The mechanism of the binding of heparin and antithrombin III has been previously studied by kinetic methods, and the data also support a two-step mechanism. The agreement of these two studies employing entirely different approaches to the same problem lends support to the validity of this postulated mechanism.     

Molecular weight analysis of antithrombin III-heparin and antithrombin III-thrombin-heparin complexes

The molecular interactions between components of the heparin-catalyzed antithrombin III/thrombin reaction were investigated by light scattering. When heparin was added to antithrombin III, the molecular weight increased to a maximum and then decreased to that of a 1:1 (antithrombin III X heparin) complex. The initial molecular weights at low heparin to antithrombin III ratios were consistent with the formation of a 2:1 (antithrombin III X heparin) complex in which only one antithrombin III molecule had undergone the conformational change measured by protein fluorescence enhancement. The peak molecular weight never reached that of a complete 2:1 complex. This behavior was observed for bovine and human antithrombin III in the presence of both unfractionated heparin and high molecular weight-high affinity heparin. Pentosane polysulfate also caused some multiple associations. Bovine antithrombin III and thrombin formed a 1:1 complex that underwent further aggregation within minutes, while the human proteins did not aggregate on this time scale after forming the 1:1 complex. In the presence of stoichiometric amounts of heparin, the bovine proteins formed an initial complex of Mr = 230,000 (corresponding to a dimer of heparin-antithrombin III-thrombin) which underwent further aggregation. The human proteins, however, formed a 1:1 (antithrombin III X thrombin) initial complex in the presence of heparin, followed by aggregation. These interactions of thrombin and antithrombin with heparin suggest complex interactions that could relate to heparin function.     

Proton magnetic resonance spectra of adrenodoxin: features of the aromatic region

This paper presents the first 1H-NMR spectra of the aromatic region of adrenodoxin, a mammalian mitochondrial 2Fe-2S non-heme iron ferredoxin. One-dimensional proton NMR spectra of both reduced and oxidized adrenodoxin were recorded as a function of pH. Resonances due to two of the three histidines of adrenodoxin gave sharp signals in the one-dimensional proton NMR spectra. The pKa values of the resolved histidine resonances in the oxidized protein were 6.64 +/- 0.03 and 6.12 +/- 0.06. These values were unchanged when adrenodoxin was reduced by the addition of sodium dithionite. In addition, the oxidized protein showed a broadened histidine C-2H resonance with a pKa value of approx. 7. This resonance was not apparent in the spectra of the reduced protein. The resonances due to the single tyrosine in adrenodoxin were identified using convolution difference spectroscopy. In addition, a two-dimensional Fourier-transform double quantum filtered (proton, proton) chemical shift correlated (DQF-COSY) spectrum of oxidized adrenodoxin was obtained. The cross peaks of the resonances due to the tyrosine, the four phenylalanines, and two of the three histidines of adrenodoxin were resolved in the DQF-COSY spectrum. Reduction of the protein caused several changes in the aromatic region of the NMR spectra. The resonances assigned to the C2 proton of the histidine with a pKa of 6.6 shifted upfield approx. 0.15 ppm. In addition, when the protein was reduced one of the resonances assigned to a phenylalanine residue with a chemical shift of 7.50 ppm appeared to move downfield to 7.82 ppm.(ABSTRACT TRUNCATED AT 250 WORDS)     

1H NMR spectroscopic studies on the interactions between human plasma antithrombin III and defined low molecular weight heparin fragments.

The effects of length and composition upon the antithrombin-binding properties of heparin have been investigated for two series of structurally related heparin oligosaccharides. Each series consists of a tetrasaccharide, hexasaccharide, and octasaccharide heparin fragment composed of alternating hexuronic acid (either iduronate 2-sulfate or glucuronate) and glucosamine 6,N-disulfate residues. These two series represent dominant structural motifs in intact heparin and differ from each other by the presence of a glucuronic acid in one series in place of an iduronate 2-sulfate residue penultimate to the reducing end of the fragment. Perturbations to the 1H resonances in the NMR spectrum of antithrombin upon binding of the two series of heparin fragments are compared to those generated by intact heparin binding, as well as to the effects of binding of a synthetic high-affinity pentasaccharide. All of the heparin fragments examined appear to bind to antithrombin at the same site. Three of the heparin fragments (hexasaccharide-2, octasaccharide-2, and octasaccharide-1) produce almost identical perturbations in the antithrombin 1H NMR spectrum compared to binding of intact heparin, including perturbations of resonances from tryptophan 49. This indicates that neither the glucuronic acid nor the trisulfated glucosamine residue (structural elements known to be part of the high-affinity heparin motif) are necessary for the majority of the conformational changes induced upon heparin fragment binding to antithrombin. However, the low anticoagulant activity of these fragments indicates that the changes in protein conformation upon fragment binding, as manifested by these 1H resonance perturbations, are not sufficient for catalytic activation of the inhibitor.(ABSTRACT TRUNCATED AT 250 WORDS)     

Comparison of the kinetic behavior of human and bovine proteins in the heparin-catalyzed antithrombin III/thrombin reaction

Significant differences between saturation kinetic properties of heparin-stimulated reactions between thrombin and antithrombin III from human and bovine species were observed. In both systems, the apparent Km for antithrombin III was higher than the KD for antithrombin III-heparin interaction, monitored by intrinsic protein fluorescence change. The Km for thrombin and kcat were much higher for proteins of the human species than the bovine species. The apparent Km for one human protein was dependent on the concentration of the other human protein, indicating interaction of the binding events. The reaction product formed from the bovine proteins was a potent inhibitor of the reaction but the product from the human proteins was a poor inhibitor. The major differences between the two species appeared to be related to interaction of thrombin or thrombin derivatives with heparin or heparin-antithrombin III complexes.     

NMR assignments of the four histidines of staphylococcal nuclease in native and denatured states

NMR signals from all four histidine ring C epsilon protons and three of the four histidine C delta protons in the protein staphylococcal nuclease have been assigned by comparing spectra of the wild-type (Foggi strain) protein to spectra of three variants that each lack a different histidine residue. All proteins studied were cloned and overproduced in Escherichia coli. The NMR spectra of the three mutant proteins (H8R, H46Y, and H124L) used to make these assignments were similar to one another and to those of the wild type, except for signals from the mutated residues. The pKa values of those histidines conserved between the wild type and the mutants remained essentially unchanged. Multiple histidine C epsilon proton resonances due to non-native forms of nuclease were observed in both thermally induced and acid-induced unfolding. Residue-specific assignments of H epsilon protons in the thermally denatured forms of the mutant H46Y were obtained from connectivities to the native state by saturation transfer.     

Nuclear magnetic resonance titration curves of histidine ring protons. A direct assignment of the resonances of the active site histidine residues of ribonuclease.

One of the four titrating histidine ring C-2 proton resonances of bovine pancreatic ribonuclease has been assigned to histidine residue 12. This was accomplished by a direct comparison of the rate of tritium incorporation into position C-2 of histidine 12 of S-peptide (residues 1 to 20) derived from ribonuclease S, with the rates of deuterium exchange of the four histidine C-2 proton resonances of ribonuclease S under the same experimental conditions. The same assignment was obtained by a comparison of the NMR titration curves of ribonuclease S, the noncovalent complex of S-peptide and S-protein (residues 21 to 124) with the results for the recombined complex in which position C-2 of histidine 12 was fully deuterated. The second active site histidine resonance was assigned to histidine residue 119 by consideration of the NMR titration results fro carboxymethylated histidines and 1-carboxymethylhistidine 119 ribonuclease. This assignment is a reversal of that originally reported, and has important implications for the interpretation of NMR titration data of ribonuclease.     

Equilibrium studies of the effect of difference in sequence homology on the mechanism of denaturation of bovine and horse cytochromes-c

We have carried out equilibrium studies of the effect of the amino acid residue difference in the primary structure of bovine cytochrome-c (b-cyt-c) and horse cyt-c (h-cyt-c) on the mechanism of their folding <--> unfolding processes at pH 6.0 and 25 degrees C. It has been observed that guanidinium chloride (GdmCl)-induced denaturation of b-cyt-c follows a two-state mechanism and that of h-cyt-c is not a two-state process. This conclusion is reached from the coincidence and non-coincidence of GdmCl-induced transition curves of bovine and horse proteins, respectively, monitored by measurements of absorbance at 405, 530 and 695 nm and circular dichroism (CD) at 222, 416 and 405 nm. These measurements on h-cyt-c in the presence of GdmCl in the concentration range 0.75-2.0 M also suggest that the protein retains all the native far-UV CD but has slightly perturbed tertiary interaction. The intermediate in the presence of these low denaturant concentrations does not have the structural characteristics of a molten globule as judged by the 8-Anilino-1-napthalene sulfonic acid (ANS) binding and near-UV CD experiments. We have also carried out thermal denaturation studies of bovine and horse cyts-c in the presence of GdmCl monitored by absorbance at 405 nm and far-UV CD at 222 nm. The heat-induced denaturation measurements in the presence of the denaturant show (1) that denaturation of b-cyt-c is a two-state process and that of h-cyt-c does not follow a two-state mechanism, and (2) that the enthalpy change on denaturation of both proteins strongly depends on GdmCl concentration.     

Characterization of human, bovine, and horse antithrombin III.

A comparison of the physical-chemical properties of human, bovine, and horse antithrombin III has been made. These three plasma proteins are strong inhibitors of bovine factor Xa and form a 1:1 molar complex with this coagulation enzyme. Human, bovine, and horse antithrombin III are glycoproteins containing hexose, hexosamine, and neuraminic acid. The total carbohydrate was 9, 12, and 16% for human, bovine, and horse antithrombin III, respectively. These proteins have a similar amino acid composition, although some monor variations were noted. Each antithrombin III is composed of a single polypeptide chain with an amino-terminal histidine residue. Of the first 17 amino-terminal residues, only three differences were noted between the three proteins. These occur in position 2 which is occupied by Gly, Arg, and Trp in human, bovine, and horse, respectively; position 6 which has a deletion in human antithrombin III; and position 8 where Ile in human and horse antithrombin III has been replaced by Val in the bovine preparation. The remainder of the first 17 residues is the same in all three proteins. The molecular weights for the bovine and horse preparation were 56 600 and 52 500, respectively, as determined by sedimentation equilibrium in the presence of guanidine hydrochloride. Some immunological cross-reactivity was also observed between the three different proteins.     

Ethoxyformylation of bovine growth hormone

Reactivity of histidines in bovine growth hormone towards ethoxyformic anhydride was investigated and localization in the molecule of two kinetically distinguishable classes was achieved, a slow class including only histidine residue 169 (k = 0.180 min-1) and a fast one composed of histidines 19 and 21 (k = 0.900 min-1). Total ethoxyformylation of bovine growth hormone brought about a complete loss of its capacity to compete with 125I-labelled hormone for rat-liver binding sites, but modification of approximately half of the fast histidine group was enough to produce an important decrease in this capacity. Circular dichroism studies indicated no significant changes in protein conformation with all three histidine residues modified. Practically full binding capacity was restored when these residues were regenerated by treatment with hydroxylamine. These results suggest that one or both of the fast reacting histidine residues are involved in bovine growth hormone binding to its specific receptors.     

Histidine residues at the N- and C-termini of alpha-helices: perturbed pKas and protein stability.

A single histidine residue has been placed at either the N-terminus or the C-terminus of each of the two alpha-helices of barnase. The pKa of that histidine residue in each of the four mutants has been determined by 1H NMR. The pKas of the two residues at the C-terminus are, on average, 0.5 unit higher, and those of the residues at the N-terminus are 0.8 unit lower, than the pKa of histidines in unfolded barnase at low ionic strength. The conformational stability of the mutant proteins at different values of pH has been measured by urea denaturation. C-Terminal histidine mutants are approximately 0.6 kcal mol-1 more stable when the introduced histidine is protonated, both at low and high ionic strength. N-Terminal mutants with a protonated histidine residue are approximately 1.1 kcal mol-1 less stable at low ionic strength and 0.5 kcal mol-1 less stable at high ionic strength (1 M NaCl). The low-field 1H NMR spectra of the mutant proteins at low pH suggest that the C-terminal histidines form hydrogen bonds with the protein while the N-terminal histidines do not form the same. The perturbations of pKa and stability result from a combination of different electrostatic environments and hydrogen-bonding patterns at either ends of helices. The value of 0.6 kcal mol-1 represents a lower limit to the favorable electrostatic interaction between the alpha-helix dipole and a protonated histidine residue at the C-terminal end of the helix.(ABSTRACT TRUNCATED AT 250 WORDS)     

A dilatometric study of the denaturation of bovine serum albumin by guanidine hydrochloride

The denaturation of bovine serum albumin by guanidine hydrochloride was studied using the dilatometric method. From dilatometric measurements the differences between the partial specific volume of the protein in denaturant solutions and water, respectively, were determined. The differences reflect the extent of unfolding as well as the binding of the denaturant. From the differences and the known partial specific volume of the native protein, the partial specific volumes at individual denaturant concentrations were obtained.     

Structure-function relationships in heparin cofactor II: chemical modification of arginine and tryptophan and demonstration of a two-domain structure

Heparin cofactor II and antithrombin III are plasma proteins functionally similar in their ability to inhibit thrombin at accelerated rates in the presence of heparin. To further characterize the structural and functional properties of human heparin cofactor II as compared to antithrombin III, we studied the possible significance of arginyl and tryptophanyl residues and the changes in protein structure and activity during guanidinium chloride (GdmCl) denaturation. Both antithrombin and heparin cofactor activities of heparin cofactor II are inactivated by the arginine-specific reagent, 2,3-butanedione. Saturation kinetics are observed during modification and suggest formation of a reversible protease inhibitor-butanedione complex. Quantitation of arginyl residues following butanedione modification shows a loss of about four residues for total inactivation, one of which is essential for antithrombin activity. Arginine-modified heparin cofactor II did not bind to heparin-agarose and implies a role for the other modified arginyl residues during heparin cofactor activity. N-Bromosuccinimide oxidation (20 mol of reagent/mol of protein) of heparin cofactor II results in modification of approximately two tryptophanyl residues with no concomitant loss of heparin cofactor activity. Moreover, there is no enhancement of intrinsic protein fluorescence during heparin binding to the native inhibitor. Circular dichroism measurements show that the structural transition of heparin cofactor II during denaturation is distinctly biphasic, yielding midpoints at 0.6 and 2.6 M GdmCl. Functional protease inhibitory activities are affected to the same extent following denaturation-renaturation at various GdmCl concentrations. The results indicate that arginyl residues are critical for both antithrombin and heparin binding activities. In contrast, tryptophanyl residues are apparently not essential for heparin-dependent interactions. The results also suggest that heparin cofactor II contains two structural domains which unfold at different GdmCl concentrations.     

Proton NMR studies of the biologically active 1-34 fragment of bovine parathyroid hormone: examination of a structural model

Proton NMR spectra of the biologically active 1-34 fragment of bovine parathyroid hormone (bPTH) were studied as a function of pH over the range of pH 4 to 10, in buffer and in 6 M guanidine DC1. One of the histidine C-2 peaks titrated normally, with a pKa value of 6.8, but the other two histidines in this peptide had pKa values of 6.3. Denatured PTH showed only one histidine C-2 peak with a pKa of 6.7. An aliphatic peak identified as due to either a methionine or a glutamine residue also shifted with pH, and the pKa for this shift was 6.3. Finally, small but significant upfield shifts in the methyl and methylene resonances were observed as a function of pH, and when compared to the denatured peptide. These results indicate that the N-terminal domain of native PTH has considerable structure in solution, and are consistent with a theoretical model for the folding of this peptide.     

Analysis of folding and unfolding reactions of cytochrome b5

The guanidine hydrochloride- (GuHCl-) induced unfolding and refolding of a recombinant domain of bovine microsomal cytochrome b(5) containing the first 104 amino acid residues has been characterized by both transient and equilibrium spectrophotometric methods. The soluble domain is reversibly unfolded and the equilibrium reaction may be monitored by changes in absorbance and fluorescence that accompany denaturation of the native protein. Both probes reveal a single cooperative transition with a midpoint at 3 M GuHCl and lead to a value for the protein stability (DeltaG(uw)) of 26.5 kJ mol(-1). This stability is much higher than that reported for the corresponding form of the apoprotein (approximately 7 kJ mol(-1)). Transient changes in fluorescence and absorbance during protein unfolding exhibit biphasic profiles. A fast phase occupying approximately 30% of the total amplitude is observed at high denaturant concentrations and becomes the dominant process within the transition region. The rates associated with each process show a linear dependency on GuHCl concentration, and at zero denaturant concentration the unfolding rates (k(uw)) are 4.5 x 10(-5) s(-1) and 5.2 x 10(-6) s(-1) at 25 degrees C. The pattern of unfolding is not correlated with covalent heterogeneity, since a wide range of variants and site-directed mutants exhibit identical profiles, nor is the unfolding correlated with cis-trans Pro isomerization in the native state. In comparison with the apo form of cytochrome b(5), the kinetics of refolding and unfolding are more complex and exhibit very different transition states. The data support a model for unfolding in which heme-protein interactions give rise to two discernible rates of unfolding. From an analysis of the activation parameters associated with each process it is established that two structurally similar transition states differing by less than 5 kJ mol(-1) exist in the unfolding reaction. Protein refolding exhibits monophasic kinetics but with distinct curvature apparent in plots of ln k(obs) versus denaturant concentration. The data are interpreted in terms of alternative routes for protein folding in which a "fast track" leads to the rapid ordering of structure around Trp26 for refolding while a slower route requires additional reorganization around the hydrophobic core.     

Folding of beta-sandwich proteins: three-state transition of a fibronectin type III module

An analysis of the folding of the 94 residue tenth fibronectin type III (fnIII) domain of human fibronectin (FNfn10) is presented. Use of guanidine isothiocyanate as a denaturant allows us to obtain equilibrium and kinetic data across a broad range of denaturant concentrations that are unavailable in guanidine hydrochloride. Equilibrium unfolding experiments show that FNfn10 is significantly more stable than has been reported previously. Comparison of equilibrium and kinetic parameters reveals the presence of an intermediate that accumulates at low denaturant concentrations. This is the first demonstration of three-state folding kinetics for a fnIII domain. We have previously shown that a homologous domain from human tenascin (TNfn3) folds by a two-state mechanism, but this does not necessarily indicate that the two proteins fold by different folding pathways.