A conserved epitope on several human vitamin K-dependent proteins. Location of the antigenic site and influence of metal ions on antibody binding

A murine monoclonal antibody (designated H-11) produced by injecting mice with purified human protein C was found to bind several human vitamin K-dependent proteins. Using a solid-phase competitive radioimmunoassay with antibody immobilized onto microtiter plates, binding of 125I-labeled protein C to the antibody was inhibited by increasing amounts of protein C, prothrombin, and Factors X and VII over a concentration range of 1 X 10(-8) to 1 X 10(-6) M. Other vitamin K-dependent proteins including Factor IX and protein S did not inhibit or inhibited only at the highest concentration binding of radiolabeled protein C to the immobilized antibody. Chemical treatment of prothrombin with a variety of agents including denaturation by sodium dodecyl sulfate, reduction with mercaptoethanol followed by carboxymethylation with iodoacetic acid, citraconylation of lysine residues, removal of metal ion with EDTA, or heat decarboxylation did not destroy the antigenic site recognized by the antibody as measured by immunoblotting of prothrombin or prothrombin derivative immobilized onto nitrocellulose. Immunoblotting of purified vitamin K-dependent polypeptides with the monoclonal antibody following sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrophoretic transfer to nitrocellulose indicated that the antigenic site was found on the light chains of protein C and Factor X. Chymotrypsin digestion of prothrombin and isolation on QAE-Sephadex of the peptide representing amino-terminal residues 1-44 of prothrombin further localized the antigenic site recognized by the monoclonal antibody to the highly conserved gamma-carboxyglutamic acid-containing domain. The exact location of the antigenic determinant for antibody H-11 was established using synthetic peptides. Antibody H-11 bound specifically to synthetic peptides corresponding to residues 1-12 of Factor VII and 1-22 of protein C. Comparison of protein sequences of bovine and human vitamin K-dependent proteins suggests that the sequence Phe-Leu-Glu-Glu-Xaa-Arg/Lys is required for antibody binding. The glutamic acid residues in this peptide segment are the first 2 gamma-carboxyglutamic acid residues near the amino-terminal end in the native proteins. Increasing concentrations of Ca2+, Mg2+, or Mn2+ partially inhibited binding of 125I-protein C to the antibody in a solid-phase assay system with half-maximal binding observed at divalent metal ion concentrations of 2, 4, and 0.6 mM, respectively.(ABSTRACT TRUNCATED AT 400 WORDS)     

A new vitamin K-dependent protein. A phospholipid-binding zymogen of a serine esterase.

Conclusive evidence is presented that a recently purified (Stenflo, J. (1976) J. Biol. Chem. 251, 355-363) vitamin K-dependent protein (arbitrarily referred to as Protein C) which is not related to prothrombin, Factors IX or X is also unrelated to Factor VII. It therefore appears to be a new, previously unrecognized vitamin K-dependent protein. In contrast to prothrombin, which binds to negatively charged phospholipid only in the presence of Ca2+ ions, Protein C, like the other vitamin K-dependent proteins, is a precursor of a serine esterase, presumably a protease, but it does not seem to be necessary for blood coagulation. Although the lipid-binding properties of Protein C may suggest that it is associated with membrane structures, its biological function remains unknown.     

Preparation of vitamin K-dependent proteins,such as clotting factors II,VII, IX and X and clotting inhibitor protein C

A review is given of preparative methods for the isolation of the vitamin K-dependent clotting factors II, VII, IX, X and clotting inhibitor protein C, all derived from human plasma. Factor II, activated factor VII and activated protein C are also obtained from recombinant animal cells. The methods for their purification are described. The problem of difference in posttranslational modifications between plasma derived and recombinant protein is discussed with regard to therapeutic proteins.     

A procedure for isolation of human protein C and protein S as by-products of the purification of factors VII, IX, X and prothrombin

A DEAE-Sephadex column chromatography step utilized to purify human Factor VII consistently yields a protein peak between the factor VII activity peak and prothrombin, factor X and factor IX activity peak (S.P. Bajaj, S.I. Rapaport, and S.F. Brown: J. Biol. Chem. 251, 253-259, 1981). We now report that this protein peak contains protein C and protein S. Preparative disc polyacrylamide gel electrophoresis of the proteins in this peak permitted a complete separation of protein C from protein S. Protein C at this step usually contained approximately 5-10% of Factor X, which could be removed by a goat anti-human Factor X antibody column. For a typical preparation, starting with 10L of plasma, the yield of Protein C was 5 mg and of protein S was 4 mg. Both proteins revealed apparent homogeneity in sodium dodecyl sulfate gel electrophoretic system. beta-Protein C and beta-protein S were not observed in our preparations starting with plasma collected directly into citrate anticoagulant containing benzamidine and soybean trypsin inhibitor, suggesting that these beta forms of protein C and protein S, isolated by other investigators, are slightly degraded forms of the native proteins. Antisera generated to these proteins were monospecific and could be used to monitor column fractions during purification. When examined by immunoelectrophoresis, the electrophoretic mobility of protein S in plasma was slower than that of isolated protein S. When exposed to plasmin, protein C was activated slightly and then rapidly degraded.     

A Procedure for Isolation of Human Protein C and Protein S as by-Products of the Purification of Factors VII,IX, X and Prothrombin

Abstract

A DEAE-Sephadex column chromatography step utilized to purify human Factor VII consistently yields a protein peak between the factor VII activity peak and prothrombin, factor X and factor IX activity peak (S.P. Bajaj, S.I. Rapaport, and S.F. Brown: J. Biol. Chem. 251., 253-259, 1981). We now report that this protein peak contains protein C and protein S. Preparative disc polyacryla-mide gel electrophoresis of the proteins in this peak 'permitted a complete separation of protein C from protein S. Protein C at this step usually contained approximately 5-10% of Factor X, which could be removed by a goat anti-human Factor X antibody column. For a typical preparation, starting with 10L of plasma, the yield of Protein C was 5 mg and of protein S was 4 mg. Both proteins     

The propeptides of the vitamin K-dependent proteins possess different affinities for the vitamin K-dependent carboxylase.

The vitamin K-dependent gamma-glutamyl carboxylase catalyzes the modification of specific glutamates in a number of proteins required for blood coagulation and associated with bone and calcium homeostasis. All known vitamin K-dependent proteins possess a conserved eighteen-amino acid propeptide sequence that is the primary binding site for the carboxylase. We compared the relative affinities of synthetic propeptides of nine human vitamin K-dependent proteins by determining the inhibition constants (Ki) toward a factor IX propeptide/gamma-carboxyglutamic acid domain substrate. The Ki values for six of the propeptides (factor X, matrix Gla protein, factor VII, factor IX, PRGP1, and protein S) were between 2-35 nM, with the factor X propeptide having the tightest affinity. In contrast, the inhibition constants for the propeptides of prothrombin and protein C are approximately 100-fold weaker than the factor X propeptide. The propeptide of bone Gla protein demonstrates severely impaired carboxylase binding with an inhibition constant of at least 200,000-fold weaker than the factor X propeptide. This study demonstrates that the affinities of the propeptides of the vitamin K-dependent proteins vary over a considerable range; this may have important physiological consequences in the levels of vitamin K-dependent proteins and the biochemical mechanism by which these substrates are modified by the carboxylase.     

Identification and isolation of vitamin K-dependent proteins by HPLC

Six of the seven known vitamin K-dependent proteins found in plasma were chromatographed on a large-pore propylsilane column using aqueous trifluoroacetic acid/acetonitrile gradients. Prothrombin and Factor VII coeluted, the others were readily resolved. The technique has been used to monitor the purification of protein C and protein S using immobilized anti-protein S. Preliminary evidence is presented which is suggestive of the existence of additional vitamin K-dependent proteins in plasma.     

A new vitamin K-dependent protein. Purification from bovine plasma and preliminary characterization.

Four proteins active in blood coagulation have long been known to require vitamin K for their proper biosynthesis: factors II, VII, IX, and X. This paper describes the purification of a hitherto unrecognized vitamin K-dependent glycoprotein from bovine plasma. The biosynthesis of this protein is interfered with by the vitamin K antagonist Dicoumarol. The molecular weight of the protein is approximately 56,000 and, like factor X, it has two polypeptide chains. The light chain binds Ca2+. Its NH2-terminal amino acid sequence is homologous to the NH2-terminal sequences of the other vitamin K-dependent proteins and it contains vitamin K-dependent gamma-carboxyglutamic acid residues. The biological function of this protein is unknown.     

A method for systematic purification from bovine plasma of six vitamin K-dependent coagulation factors: prothrombin, factor X, factor IX, protein S, protein C, and protein Z

A systematic purification scheme is presented for the isolation of six vitamin K-dependent coagulation factors from bovine plasma in a functionally and biochemically pure state. The vitamin K-dependent proteins concentrated by the ordinary barium citrate adsorption were first separated into four fractions, fractions A, B, C, and D, by DEAE-Sephadex A-50 chromatography. From the pooled fraction A, protein S, factor IX, and prothrombin were purified by column chromatography on Blue-Sepharose CL-6B. Heparin-Sepharose chromatography of the pooled fraction B provided mainly pure factor IX, in addition to homogeneous prothrombin. A high degree of resolution of protein C and prothrombin from the pooled fraction C was obtained with a Blue-Sepharose column. This dye-ligand chromatographic procedure was also very effective for the separation of protein Z and factor X contained in the pooled fraction D. Thus, these preparative procedures allowed high recovery of milligram and gram quantities of six vitamin K-dependent proteins from 15 liters of plasma in only two chromatographic steps, except for protein S, which required three (the third step was rechromatography on Blue-Sepharose CL-6B).     

Chromatographic purification and properties of a therapeutic human protein C concentrate

Protein C deficiency (inherited and acquired) has a relatively high incidence rate in the general population worldwide. For many years, protein C deficient patients have been treated with fresh frozen plasma, prothrombin complex concentrates, heparin or oral anticoagulants, which all have clinical drawbacks. We report the production process of a highly purified human protein C concentrate from 1500 l of cryo-poor plasma by a four-step chromatographic procedure. After DEAE-Sephadex adsorption, protein C was separated from clotting factors II, VII and IX by DEAE-Sepharose FF and further purified, using a new strategy, by an on-line chromatographic system combining DMAE-Fractogel and heparin-Sepharose CL-6B. In addition, the product was treated against viral risks by solvent-detergent and nanofiltration on 15-nm membranes. The protein C concentrate was essentially free of other vitamin K-dependent proteins. Proteolytic activity was undetectable. Neither activated protein C, prekallikrein activator, nor activated vitamin K-dependent clotting factors were found resulting in good stability of the protein C activity. In vitro and in vivo animal tests did not reveal any sign of potential thrombogenicity. The final freeze-dried product had a mean protein C concentration of 58 IU/ml and a mean specific activity of 215 IU/mg protein, corresponding to over 12000-fold purification from plasma. Therefore, this concentrate appears to be of potential benefit for the treatment of protein C deficiency.     

Expression, purification, and characterization of recombinant gamma-carboxylated factor IX synthesized in Chinese hamster ovary cells

Factor IX has been expressed to high levels within a recombinant host cell and the biologically active fraction subsequently purified to homogeneity for characterization. The coding sequence for Factor IX was inserted into a mammalian cell expression vector and transfected into dihydrofolate reductase-deficient Chinese hamster ovary cells. The integrated DNA was amplified to a high copy number by selection for increasingly higher expression levels of the marker gene, dihydrofolate reductase, contained within a co-transfected plasmid. Cloned cell lines secreting over 100 micrograms/ml Factor IX antigen and up to 1.5 microgram/ml native Factor IX antigen have been obtained. Expression of biologically active Factor IX was dependent on the presence of vitamin K in the culture media. The gamma-carboxylated Factor IX was isolated from cell culture fluid by immunoaffinity chromatography using antibodies conformation-specific for the metal-stabilized conformer of Factor IX. This conformation is dependent upon metal ions and gamma-carboxyglutamic acid. Purified recombinant Factor IX migrated as a single band on sodium dodecyl sulfate-polyacrylamide gel electrophoresis with an electrophoretic mobility equivalent to plasma-derived Factor IX. The purified recombinant Factor IX demonstrated Factor IX coagulant activity, measured in Factor IX-deficient plasma, of 35-75 units/mg. Amino acid analysis of the alkaline hydrolysate of recombinant Factor IX demonstrated an average of 6-7 mol of gamma-carboxyglutamic acid per mol of Factor IX. NH2-terminal sequence analysis of the first 17 residues revealed equivalent amino acid sequences for both purified recombinant and plasma-derived Factor IX. The results represent the first purification and characterization of a biologically active, gamma-carboxylated vitamin K-dependent protein expressed in a recombinant DNA system.     

Separation of human vitamin K-dependent coagulation proteins using hydrophobic interaction chromatography

A rapid and simple method was developed to separate human vitamin K-dependent plasma proteins from each other, yielding virtually homogeneous pools. The purification technique is based on the single use of hydrophobic interaction chromatography, starting from prothrombin concentrate (PC or DEFIX, also termed factor IX concentrate) as initial material. Phenyl-sepharose HP demonstrated optimal separation by comparing several hydrophobic resins as well as resins used in standard procedures like immobilised heparin and Cibacron blue. Under ideal conditions, factor X could be separated in a single step as well as prothrombin. Factor IX co-eluted with other minor proteins. Focus was given only on these three proteins due to their relative abundance. Complete separation of all proteins present in the starting material was achieved by MonoQ anion-exchange chromatography following the phenyl-sepharose run. The resulting purified material could be demonstrated to be of equal or higher purity than using described methods. This strategy employing hydrophobic interaction chromatography for blood macromolecules could be of immense value for purifying the human vitamin K-dependent proteins and represents a considerable simplification over other purification schemes. It not only involves minimal sample handling but also can be readily up-scaled and is a cost-efficient alternative.     

Gamma-carboxyglutamic acid

Summary Gamma-carboxyglutamic acid is an amino acid with a dicarboxylic acid side chain. This amino acid, with unique metal binding properties, confers metal binding character to the proteins into which it is incorporated. This amino acid has been discovered in blood coagulation proteins (prothrombin, Factor X, Factor IX, and Factor VII), plasma proteins of unknown function (Protein C, Protein S, and Protein Z), and proteins from calcified tissue (osteocalcin and bone-Gla protein). It has also been observed in renal calculi, atherosclerotic plaque, and the egg chorioallantoic membrane, among other tissues. Gamma-carboxyglutamic acid is synthesized by the post-translational modification of glutamic acid residues. This reaction, catalyzed by a hepatic carboxylase, requires reduced vitamin K, oxygen, and carbon dioxide. The function of -carboxyglutamic acid is uncertain. In prothrombin y-carboxyglutamic acid residues bound to metal ions participate as an intramolecular non-covalent bridge to maintain protein conformation. Additionally, these amino acids participate in the calcium-dependent molecular assembly of proteins on membrane surfaces through intermolecular bridges involving y-carboxyglutamic acid and metal ions.     

Human plasma and recombinant factor VII. Characterization of O-glycosylations at serine residues 52 and 60 and effects of site-directed mutagenesis of serine 52 to alanine

Factor VII is a multidomain, vitamin K-dependent plasma glycoprotein that participates in the extrinsic pathway of blood coagulation. Earlier studies demonstrated a novel disaccharide (Xyl-Glc) or trisaccharide (Xyl2-Glc) O-glycosidically linked to serine 52 in human plasma factor VII (Nishimura, H., Kawabata, S., Kisiel, W., Hase, S., Ikenaka, T., Shimonishi, Y., and Iwanaga, S. (1989) J. Biol. Chem. 264, 20320-20325). In the present study, human plasma and recombinant factor VII were isolated and subjected to enzymatic fragmentation. Peptides comprising residues 48-62 of the first epidermal growth factor-like domain of each factor VII preparation were isolated for comparative analysis. Using a combined strategy of amino acid sequencing, carbohydrate and amino acid composition analysis, and mass spectrometry, three different glycan structures consisting of either glucose, glucose-xylose, or glucose-(xylose)2 were detected O-glycosidically linked to serine 52 in plasma and recombinant factor VII. Approximately equal amounts of the three glycan structures were observed in plasma factor VII, whereas in recombinant factor VII the glucose and the glucose-(xylose)2 structures predominated. In addition to the O-linked glycan structures observed at serine 52, a single fucose was found to be covalently linked at serine 60 in both human plasma and recombinant factor VII. Carbohydrate and mass spectrometry analyses indicated that the fucosylation of serine 60 was virtually quantitative. Metabolic labeling studies using [14C]fucose confirmed the presence of O-linked fucose at serine 60. In order to assess whether the carbohydrate moiety at serine 52 contributes to the biological activity of factor VII, we have constructed a site-specific mutant of recombinant factor VII in which serine 52 has been replaced with an alanine residue. Mutant factor VIIa exhibited approximately 60% of the coagulant activity of wild-type factor VIIa in a clotting assay. The amidolytic activity of mutant factor VIIa was indistinguishable from that observed for recombinant wild-type factor VIIa. In addition, the ability of mutant factor VIIa in complex with either purified relipidated tissue factor apoprotein or tissue factor on the surface of a human bladder carcinoma cell line (J82) to activate either factor X or factor IX was virtually identical to that observed for wild-type factor VIIa. These results indicate that the carbohydrate moiety O-glycosidically linked to serine 52 does not appear to be involved either in the interaction of factor VIIa with tissue factor, or the expression of its proteolytic activity toward factor X or factor IX following complex formation with tissue factor.(ABSTRACT TRUNCATED AT 400 WORDS)     

A novel blood coagulation factor IX/factor X-binding protein with anticoagulant activity from the venom of Trimeresurus flavoviridis (Habu snake): isolation and characterization

Using affinity chromatography on a column of factor X-Cellulofine, we have isolated a novel blood coagulation factor X-binding protein with anticoagulant activity from the venom of Trimeresurus flavoviridis (Habu snake). This anticoagulant protein was also purified by chromatography on Sephadex G-75 and S-Sepharose Fast Flow. The yield of the purified protein was approximately 16 mg from 400 mg of crude venom. The purified protein gave a single band on both analytical alkaline disc-gel electrophoresis and SDS-PAGE. This protein had a relative molecular weight (Mr) after SDS-PAGE of 27,000 before reduction of disulfide bonds and 14,000 after reduction of disulfide bonds. The protein prolonged the clotting time induced by kaolin or factor Xa. In the presence of Ca2+, it formed a complex with factor X, the molar ratio being 1 to 1. Similar complex formation was observed with factor Xa and factor IX/factor IXa, but not with other vitamin K-dependent coagulation factors, i.e., prothrombin, factor VII, protein C, protein S, and protein Z. The interaction of this anticoagulant protein with factor IX/factor X was dependent on gamma-carboxyglutamic acid (Gla) domains, since Gla-domainless derivatives of factor X and factor IXa beta' did not interact with this anticoagulant protein.     

Localization of thrombin cleavage sites in the amino-terminal region of bovine protein S

Protein S is a vitamin K-dependent plasma protein. It functions as a cofactor to activated protein C in the inactivation of factors Va and VIIIa by limited proteolysis. Protein S is very sensitive to proteolysis by thrombin which reduces its calcium ion binding and leads to a loss of its cofactor activity. We have now determined the sequence of the 100 amino-terminal amino acid residues and localized the thrombin cleavage sites. Protein S contains 11 gamma-carboxyglutamic acid residues in the amino-terminal region (residues 1-36). This part of protein S is highly homologous to the corresponding parts in the other vitamin K-dependent clotting factors, whereas the region between residues 45 and 75 is not at all homologous to the other clotting factors. Thrombin cleaves two peptide bonds in this part of protein S, first at arginine 70 and then at arginine 52. The peptide containing residues 53-70 is released from protein S after thrombin cleavage. The amino-terminal fragment, residues 1-52, is linked to the large carboxyl-terminal fragment by a disulfide bond, which involves cysteine 47. After residue 78, protein S is again homologous to factors IX and X and to proteins C and Z, but not to prothrombin. Position 95 is occupied by a beta-hydroxyaspartic acid residue.     

The propeptide region of clotting factor IX is a signal for a vitamin K dependent carboxylase: evidence from protein engineering of amino acid -4.

Homologous "propeptide" regions are present in a family of vitamin K-dependent mammalian proteins, including clotting factors II, VII, IX, X, protein C, protein S and bone "gla" proteins. To test the hypothesis that the propeptide is a signal for the correct gamma-carboxylation of the adjacent gamma-carboxy region, we have mutated amino acid -4 of human factor IX from an arginine to a glutamine residue, by M13-directed site-specific mutagenesis of a cDNA clone. After expression of mutant factor IX in dog kidney cells, we find that it is secreted into the medium in a precursor form containing the propeptide, and is inefficiently gamma-carboxylated compared to the control, wild-type, recombinant factor IX. This result supports the hypothesis that the propeptide region is required for efficient gamma-carboxylation of factor IX in dog kidney cells. Furthermore, it confirms previous results that arginine at amino acid -4 is required for correct propeptide processing.     

Purification of human C4b-binding protein and formation of its complex with vitamin K-dependent protein S.

C4b-binding protein was purified from human plasma in high yield by a simple procedure involving barium citrate adsorption and two subsequent chromatographic steps. Approx. 80% of plasma C4b-binding protein was adsorbed on the barium citrate, presumably because of its complex-formation with vitamin K-dependent protein S. The purified C4b-binding protein had a molecular weight of 570 000, as determined by ultracentrifugation, and was composed of about eight subunits (Mr approx. 70 000). Uncomplexed plasma C4b-binding protein was purified from the supernatant after barium citrate adsorption. On sodium dodecyl sulphate/polyacrylamide-gel electrophoresis in non-reducing conditions and on agarose-gel electrophoresis it appeared as a doublet, indicating two forms differing slightly from each other in molecular weight and net charge. The protein band with the higher molecular weight in the doublet corresponded to the C4b-binding protein purified from the barium citrate eluate. Complex-formation between protein S and C4b-binding protein was studied in plasma, and in a system with purified components, by an agarose-gel electrophoresis technique. Protein S was found to form a 1:1 complex with the higher-molecular-weight form of C4b-binding protein, whereas the lower-molecular-weight form of C4b-binding protein did not bind protein S. The KD for the C4b-binding protein-protein S interaction in a system with purified components was approx. 0.9 X 10(-7) M. Rates of association and dissociation at 37 degrees C were low, namely about 1 X 10(3) M-1 . S-1 and 1.8 X 10(-4)-4.5 X 10(-4) S-1 respectively. In human plasma free protein S and free higher-molecular-weight C4b-binding protein were in equilibrium with the C4b-binding protein-protein S complex. Approx. 40% of both proteins existed as free proteins. From equilibrium data in plasma a KD of about 0.7 X 10(-7) M was calculated for the C4b-binding protein-protein S interaction.     

Protein structural requirements and properties of membrane binding by gamma-carboxyglutamic acid-containing plasma proteins and peptides

The membrane-binding characteristics of a number of modified vitamin K-dependent proteins and peptides showed a general pattern of structural requirements. The amino-terminal peptides from human prothrombin (residues 1-41 and 1-44, 60:40) bovine factor X (residues 1-44), and bovine factor IX (residues 1-42), showed a general requirement for a free amino-terminal group, an intact disulfide, and the tyrosine homologous to Tyr44 of factor X for membrane binding. Consequently, the peptide from factor IX did not bind to membranes. Any of several modifications of the amino terminus, except reaction with trinitrobenzenesulfonic acid, abolished membrane binding by the factor X and prothrombin peptides. Calcium, but not magnesium, protected the amino terminus from chemical modification. The requirement for a free amino terminus was also shown to be true for intact prothrombin fragment 1, factor X, and factor IX. Although aggregation of the peptide-vesicle complexes greatly complicated accurate estimation of equilibrium binding constants, results with the factor X peptide indicated an affinity that was not greatly different from that of the parent protein. The most striking difference shown by the peptides was a requirement for about 10 times as much calcium as the parent proteins. In a manner similar to the parent proteins, the prothrombin and factor X peptides showed a large calcium-dependent quenching of tryptophan fluorescence. This fluorescence quenching in the peptides also required about 10 times the calcium needed by the parent proteins. Thus, the 1-45 region of the vitamin K-dependent proteins contained most of the membrane-binding structure but lacked component(s) needed for high affinity calcium binding. Protein S that was modified by thrombin cleavage at Arg52 and Arg70 showed approximately the same behavior as the amino-terminal 45-residue peptides. That is, it bound to membranes with overall affinity that was similar to native protein S but required high calcium concentrations. These results suggested that the second disulfide loop of protein S (Cys47-Cys72) and prothrombin (Cys48-Cys61) were involved in high affinity calcium binding. Since factor X lacks a homologous disulfide loop, an alternative structure must serve a similar function. A striking property of protein S was dissociation from membranes by high calcium. While this property was shared by all the vitamin K-dependent proteins, protein S showed this most dramatically and supported protein-membrane binding by calcium bridging.