Aspartyl beta-hydroxylase: in vitro hydroxylation of a synthetic peptide based on the structure of the first growth factor-like domain of human factor IX.

beta-Hydroxylation of aspartic acid is a post-translational modification that occurs in several vitamin K-dependent coagulation proteins. By use of a synthetic substrate comprised of the first epidermal growth factor-like domain in human factor IX and either mouse L-cell extracts or rat liver microsomes as the source of enzyme, in vitro aspartyl beta-hydroxylation was accomplished. Aspartyl beta-hydroxylase appears to require the same cofactors as known alpha-ketoglutarate-dependent dioxygenases. The hydroxylation reaction proceeds with the same stereospecificity and occurs only at the aspartate corresponding to the position seen in vivo. Further purification and characterization of this enzymatic activity should now be possible.     

Identification and purification to near homogeneity of the vitamin K-dependent carboxylase.

Vitamin K-dependent carboxylase catalyzes the modification of specific glutamic acids to gamma-carboxyglutamic acid in several blood-coagulation proteins. This modification is required for the blood-clotting activity of these proteins and has thus been the subject of intense investigation. We have now identified the bovine vitamin K-dependent carboxylase and purified it to near homogeneity by an affinity procedure that uses the 59-amino acid peptide FIXQ/S (residues -18 to 41 of factor IX with mutations Arg----Gln at residue -4 and Arg----Ser at residue -1). The carboxylase as purified has a molecular weight of 94,000. It is also the major protein that can be cross-linked to iodinated FIXQ/S and is the only protein whose cross-linking is prevented by a synthetic factor IX propeptide. The degree of purification is about 7000-fold with reference to ammonium sulfate-fractionated microsomal protein from liver.     

In vitro and in vivo functional characterization of bovine vitamin K-dependent gamma-carboxylase expressed in Chinese hamster ovary cells.

Coagulation factor IX is a serine protease for which high-level expression of biologically active protein in heterologous cells is limited due to inefficient proteolytic removal of the propeptide as well as vitamin K-dependent carboxylation of multiple amino-terminal glutamic acid residues. We have overexpressed the vitamin K-dependent gamma-carboxylase cDNA and monitored its ability to improve factor IX processing in Chinese hamster ovary (CHO) cells. From amino acid sequence analysis of bovine liver vitamin K-dependent gamma-carboxylase, degenerate oligonucleotides were used to isolate a 3.5-kbp bovine cDNA that encoded a 758-residue open reading frame. Expression of the cDNA in COS-1 and CHO cells yielded 17- and 16-fold increases in the in vitro gamma-carboxylase activity of microsomal preparations, respectively. Anti-serum raised against a predicted peptide sequence reacted with a 94-kDa polypeptide in the partially purified bovine liver preparation as well as in stably transfected CHO cells. The amount of antibody reactivity correlated with the increased ability to carboxylate a peptide substrate in vitro. These results strongly support the conclusion that the cDNA encodes the vitamin K-dependent gamma-carboxylase. Transient transfection of the gamma-carboxylase expression vector into factor IX-expressing CHO cells did not improve the specific procoagulant activity of secreted factor IX. In contrast, transfection of an expression vector encoding the propeptide processing enzyme PACE (paired basic amino acid cleaving enzyme) did improve the specific activity of secreted factor IX by 3-fold. These results demonstrate that the ability of CHO cells to modify glutamic acid residues to gamma-carboxyglutamic acid in secreted factor IX is not limited by the expression of the vitamin K-dependent gamma-carboxylase alone.     

beta-Hydroxyasparagine in domains homologous to the epidermal growth factor precursor in vitamin K-dependent protein S.

Vitamin K-dependent protein S is involved in the regulation of blood coagulation. It is a 75-kDa single chain protein with an NH2-terminal gamma-carboxyglutamic acid-containing domain followed by a thrombin-sensitive region and four domains arranged in tandem, each of which is homologous to the epidermal growth factor (EGF) precursor. The NH2-terminal EGF-like domain contains beta-hydroxyaspartic acid, which has been identified in vitamin K-dependent proteins. The following EGF-like repeat has a very pronounced sequence homology (10 consecutive residues identical) to one of the EGF-like units in the EGF precursor. We now show that, in protein S, this EGF-like repeat has one beta-hydroxyasparagine residue formed by hydroxylation of asparagine. The two COOH-terminal EGF-like repeats also contain beta-hydroxyasparagine, an amino acid not previously found in proteins. Sequence comparisons have enabled us to identify a consensus sequence that seems to be required by the hydroxylase(s).     

Primary structure of bovine vitamin K-dependent protein S.

Protein S is a vitamin K-dependent plasma protein that functions as a cofactor to activated protein C in the inactivation of coagulation factors Va and VIIIa. The nucleotide sequence of a full-length cDNA clone, obtained from a bovine liver library, was determined and the amino acid sequence was deduced. In addition, 95% of the structure was determined by protein sequencing. Protein S consists of 634 amino acids in a single polypeptide chain and has one asparagine-linked carbohydrate side chain. The cDNA sequence showed that the protein has a leader sequence, 41 amino acid residues long. The amino-terminal part of the molecule containing gamma-carboxyglutamic acid is followed by a region, residues 42-75, with two peptide bonds that are very sensitive to cleavage by thrombin. Residues 76-244 have four cysteinerich repeat sequences, each about 40 residues long, that are homologous to the precursor of mouse epidermal growth factor. In contrast to the other vitamin K-dependent plasma proteins, the carboxyl-terminal part of protein S is not homologous to the serine proteases.     

Studies of the vitamin K-dependent carboxylase and vitamin K epoxide reductase in rat liver

Vitamin K is required as a cofactor for a microsomal enzyme that converts glutamyl residues in precursor proteins to gamma-carboxyglutamyl residues in completed proteins. These residues are essential for the biological function of prothrombin, factors VII, IX, and X, protein C, and protein S. Current data suggest that recognition of protein substrates by the carboxylase requires an unidentified protein-protein interaction in addition to the Glu substrate binding site. The primary vitamin K-dependent event has now been shown to be the abstraction of the gamma-hydrogen of the substrate Glu residue with the concurrent formation of vitamin K 2,3-epoxide. Coumarin anticoagulants appear to inhibit the microsomal vitamin K epoxide reductase and one of a number of microsomal quinone reductases. They therefore block vitamin K action by preventing the recycling of vitamin K epoxide to the quinone and to the active cofactor form, the hydroquinone. Excess vitamin K can reverse a coumarin anticoagulant effect as the nonsensitive quinone reductase can continue to furnish the active coenzyme.     

Molecular defects of factor IX Chicago-2 (Arg 145----His) and prothrombin Madrid (Arg 271----cys): arginine mutations that preclude zymogen activation

Factor IX Chicago-2 and prothrombin Madrid were purified from patients with hemophilia B and congenital dysprothrombinemia, respectively. Each protein displays defects in zymogen activation secondary to the failure to cleave one of the sessile bonds whose cleavage is necessary for full coagulant activity. These proteins were isolated by immunoaffinity chromatography using conformation-specific antibodies directed at either factor IX or prothrombin. Factor IX Chicago-2 is cleaved abnormally by factor XIa, yielding a pattern consistent with the failure to cleave the sessile bond between Arg 145 and Ala 146. Prothrombin Madrid is cleaved abnormally by factor Xa, yielding a pattern consistent with the failure to cleave the sessile bond between Arg 271 and Thr 272. Peptide mapping was performed on reduced and alkylated factor IX, factor IX Chicago-2, prothrombin, and prothrombin Madrid, and the hydrolysates were separated by high-performance liquid chromatography. The mutant peptide in factor IX Chicago-2 was identified by automated Edman degradation as residues 143 through 188 of factor IX, and had a histidine substituted for arginine at residue 145. The mutant peptide identified in prothrombin Madrid corresponds to residues 267 through 285 of prothrombin and has the substitution of cysteine for arginine at residue 271. These mutations, each occurring at arginines, are identical to those in factor IX Chapel Hill and prothrombin Barcelona. These results suggest that a limited repertoire of point mutations, many affecting arginine residues, may be responsible for hereditary defects of the vitamin K-dependent proteins in patients with normal antigen levels.     

Functional mapping of anti-factor IX inhibitors developed in patients with severe hemophilia B

Development of inhibitory antibodies is a serious complication of treatment with repeated factor IX infusions in a minority of patients with hemophilia B. Such antibodies detected in 8 patients have been characterized. Typing studies revealed that patients' immune response toward factor IX is highly heterogeneous and involves immunoglobulin G (IgG) antibodies, preferentially IgG1 and IgG4. The preservation of the sequence and the 3-dimensional orientation of the amino acids constituting one epitope are highly important for the assembly of an antibody-antigen complex. To localize the epitopes on the factor IX molecule, an original approach was designed using a set of factor X chimeras carrying regions of factor IX. Results showed that some patients' antibodies were directed against both the domain containing the gamma-carboxy glutamic acid residues (Gla domain) and the protease domain of factor IX. In contrast, no binding was observed to the epidermal growth factor-like domains or to the activation peptide. Functional characterization showed that the purified IgG from patients' serum inhibited the factor VIIIa-dependent activation of factor X. Moreover, patients' IgG directed against the Gla domain inhibited the binding of factor IX to phospholipids as well as the binding of factor VIII light chain to factor IXa. These data demonstrate that inhibitors appearing in patients with severe hemophilia B display specificity against restricted functional domains of factor IX.     

Molecular cloning of matrix Gla protein: implications for substrate recognition by the vitamin K-dependent gamma-carboxylase.

Matrix Gla protein (MGP), a low molecular weight protein found in bone, dentin, and cartilage, contains 5 residues of the vitamin K-dependent amino acid gamma-carboxyglutamic acid (Gla). We have used antibodies raised against MGP and oligonucleotide probes to screen a lambda gt11 cDNA library constructed from the rat osteosarcoma cells (line ROS 17/2) that had been pretreated with 1 alpha,25-dihydroxyvitamin D3. By sequencing several cloned cDNAs, we established a 523-base-pair sequence that predicts an 84-residue mature MGP and a 19-residue hydrophobic signal peptide. The 84-residue mature rat MGP predicted from the cDNA sequence has an additional 5 residues at its C terminus (-Arg-Arg-Gly-Ala-Lys) not seen in the sequence of MGP isolated from bovine bone. The structure of rat MGP provides insight into the mechanisms by which the vitamin K-dependent gamma-carboxylase recognizes substrate. The present studies show that MGP, unlike other vitamin K-dependent proteins, lacks a propeptide. The absence of an MGP propeptide demonstrates that gamma-carboxylation and secretion of vitamin K-dependent proteins need not be linked to the presence of a propeptide or to its proteolytic removal. The propeptides of other vitamin K-dependent proteins are structurally homologous, and there is evidence that this homologous propeptide domain is important to substrate recognition by the gamma-carboxylase. Mature MGP has a sequence segment (residues 15-30) that is homologous to the propeptide of other vitamin K-dependent proteins and probably serves the same role in gamma-carboxylase recognition. Rat MGP also has a second sequence that has recently been identified in all known vitamin K-dependent vertebrate proteins, the invariant unit Glu-Xaa-Xaa-Xaa-Glu-Xaa-Cys (EXXXEXC). Since the glutamic residues in this unit are sites of gamma-carboxylation, it has been suggested that the EXXXEXC unit could allow the gamma-carboxylase to discriminate between substrate and product. The demonstration that two structures common to vitamin K-dependent proteins, the homologous propeptides domain and the invariant EXXXEXC unit, are in mature MGP indicates that des-gamma-carboxy-MGP should be an excellent in vitro gamma-carboxylase substrate for analysis of mechanisms involved in substrate recognition and product dissociation.     

Factor IXHollywood: substitution of Pro55 by Ala in the first epidermal growth factor-like domain

Factor IX is a multidomain protein essential for hemostasis. We describe a mutation in a patient affecting the first epidermal growth factor (EGF)-like domain of the protein. All exons and the promoter region of the gene were amplified by the polymerase chain reaction method, and sequenced. Only a single mutation (C----G), that predicts the substitution of Pro55 by Ala in the first EGF domain was found in the patient's gene. This mutation leads to new restriction sites for four enzymes. One new site (Nsi) was tested in the amplified exon IV fragment and was shown to provide a rapid and reliable marker for carrier detection and prenatal diagnosis in the affected family. The factor IX protein, termed factor IXHollywood (IXHW), was isolated to homogeneity from the patient's plasma. As compared with normal factor IX (IXN), IXHW contained the same amount of gamma-carboxy glutamic acid but twice the amount of beta-OH aspartic acid. Both IXHW and IXN contained no detectable free -SH groups. Further, IXHW could be readily cleaved to yield a factor IXa-like molecule by factor Xla/Ca2+. However, IXaHW (compared with IXaN) activated factor X approximately twofold slower in the presence of Ca2+ and phospholipid (PL), and 8- to 12-fold slower in the presence of Ca2+, PL, and factor VIIIa. Additionally, IXaHW had only approximately 10% of the activity of IXaN in an aPTT assay. In agreement with the nuclear magnetic resonance-derived structure of EGF, the Chou-Fasman algorithm strongly predicted a beta turn involving residues Asn-Pro55-Cys-Leu in IXN. Replacement of Pro55 by Ala gave a fourfold decrease in the beta turn probability for this peptide, suggesting a change(s) in the secondary structure in the EGF domain of IXHW. Since this domain of IXN is thought to have one high-affinity Ca2+ binding site and may be involved in PL and/or factor VIIIa binding, the localized secondary structural changes in IXHW could lead to distortion of the binding site(s) for the cofactor(s) and, thus, a dysfunctional molecule.     

First epidermal growth factor-like domain of human blood coagulation factor IX is required for its activation by factor VIIa/tissue factor but not by factor XIa.

Factor IX consists of a gamma-carboxyglutamic acid-rich domain followed by two epidermal growth factor (EGF)-like domains and the C-terminal protease domain. To delineate the function of EGF1 domain in factor IX, we constructed three mutants: an EGF1 domain-deleted mutant (IX delta EGF1), a point mutant (IXQ50P) with a Gln-50-->Pro change, and a replacement mutant (IXPCEGF1) in which the EGF1 domain of factor IX was replaced by that of protein C. These mutants and wild-type (WT) factor IX (IXWT) were expressed in 293 kidney cells by using pRc/CMV vector. The purified proteins had the same gamma-carboxyglutamic acid content as the normal plasma factor IX (IXNP) and were activated normally by factor XIa-Ca2+. In contrast, IX delta EGF1 could not be activated by factor VIIa-tissue factor-Ca2+, and the activation of IXPCEGF1 in this system was markedly slow; however, IXQ50P was activated at a normal rate. In additional studies, both IXWT and IX delta EGF1 were rapidly converted to their respective IX alpha forms by factor Xa-phospholipid-Ca2+. Since this reaction has an absolute requirement for phospholipid, it indicates that the mutants under study are not impaired in their interactions with phospholipid. Relative coagulant activities of factor XIa-activated proteins were IXNP, 100%; IXWT, 75-85%; IX delta EGF1, < or = 1%; IXPCEGF1, < or = 2%; and IXQ50P, 6-10%. We conclude that the EGF1 domain of factor IX is required for its activation by factor VIIa-tissue factor and that the Gln-50 residue is not critical for this activation. Further, the EGF1 domain of factor IX is not essential for phospholipid binding and for its activation by factor XIa. In addition, the low coagulant activities of the activated mutants indicate that the EGF1 domain is also important in factor X activation by factor IXa-factor VIIIa-Ca(2+)-phospholipid complex.     

Substrate specificity of the protein tyrosine phosphatases.

The substrate specificity of a recombinant protein tyrosine phosphatase (PTPase) was probed using synthetic phosphotyrosine-containing peptides corresponding to several of the autophosphorylation sites in epidermal growth factor receptor (EGFR). The peptide corresponding to the autophosphorylation site, EGFR988-998, was chosen for further study due to its favorable kinetic constants. The contribution of individual amino acid side chains to the binding and catalysis was ascertained utilizing a strategy in which each amino acid within the undecapeptide EGFR988-998 (DADEpYLIPQQG) was sequentially substituted by an Ala residue (Ala-scan). The resulting effects due to singular Ala substitution were assessed by kinetic analysis with two widely divergent homogeneous PTPases. A "consensus sequence" for PTPase recognition may be suggested from the Ala-scan data as DADEpYAAPA, and the presence of acidic residues proximate to the NH2-terminal side of phosphorylation is critical for high-affinity binding and catalysis. The Km value for EGFR988-998 decreased as the pH increased, suggesting that phosphate dianion is favored for substrate binding. The results demonstrate that chemical features in the primary structure surrounding the dephosphorylation site contribute to PTPase substrate specificity.     

Characterization of a cDNA coding for human factor VII.

Factor VII is a precursor to a serine protease that is present in mammalian plasma. In its activated form, it participates in blood coagulation by activating factor X and/or factor IX in the presence of tissue factor and calcium. Clones coding for factor VII were obtained from two cDNA libraries prepared from poly(A) RNA from human liver and Hep G2 cells. The amino acid sequence deduced from the cDNAs indicates that factor VII is synthesized with a prepro-leader sequence of 60 or 38 amino acids. The mature protein that circulates in plasma is a single-chain polypeptide composed of 406 amino acids. The amino acid sequence analysis of the protein and the amino acid sequence deduced from the cDNAs indicate that factor VII is converted to factor VIIa by the cleavage of a single internal bond between arginine and isoleucine. This results in the formation of a light chain (152 amino acids) and a heavy chain (254 amino acids) that are held together by a disulfide bond. The light chain contains a gamma-carboxyglutamic acid (Gla) domain and two potential epidermal growth factor domains, while the heavy chain contains the serine protease portion of the molecule. Factor VII shows a high degree of amino acid sequence homology with the other vitamin K-dependent plasma proteins.     

Fatal hemorrhage in mice lacking gamma-glutamyl carboxylase

The carboxylation of glutamic acid residues to gamma-carboxyglutamic acid (Gla) by the vitamin K-dependent gamma-glutamyl carboxylase (gamma-carboxylase) is an essential posttranslational modification required for the biological activity of a number of proteins, including proteins involved in blood coagulation and its regulation. Heterozygous mice carrying a null mutation at the gamma-carboxylase (Ggcx) gene exhibit normal development and survival with no evidence of hemorrhage and normal functional activity of the vitamin K-dependent clotting factors IX, X, and prothrombin. Analysis of a Ggcx(+/-) intercross revealed a partial developmental block with only 50% of expected Ggcx(-/-) offspring surviving to term, with the latter animals dying uniformly at birth of massive intra-abdominal hemorrhage. This phenotype closely resembles the partial midembryonic loss and postnatal hemorrhage previously reported for both prothrombin- and factor V (F5)-deficient mice. These data exclude the existence of a redundant carboxylase pathway and suggest that functionally critical substrates for gamma-carboxylation, at least in the developing embryo and neonate, are primarily restricted to components of the blood coagulation cascade.     

Efficient synthesis of human type alpha transforming growth factor: its physical and biological characterization.

Human transforming growth factor type alpha (TGF-alpha) was synthesized by a stepwise solid-phase method with an overall yield of 26%. Synthetic TGF-alpha, consisting of 50 amino acid residues deduced from a cDNA precursor sequence, was purified in a single HPLC step. The homogeneity and primary structure were confirmed by several criteria including Edman degradation and mass spectrometry. Synthetic TGF-alpha was as active as murine epidermal growth factor in binding to the epidermal growth factor receptor and in stimulation of anchorage-dependent and of anchorage-independent growth of normal indicator cells in culture. Synthetic TGF-alpha stimulated plasminogen activator production in A 431 and HeLa cells; the stimulation was similar to that induced by epidermal growth factor. Furthermore, synthetic human TGF-alpha showed similar immunoreactivity when compared with rat TGF-alpha. Thus, the 50-amino acid TGF-alpha is likely to be the bioactive principle produced and secreted by tumor cell lines.     

Tumor cell autocrine motility factor.

A cell motility-stimulating factor has been isolated, purified, and partially characterized from the serum-free conditioned medium of human A2058 melanoma cells. We term this activity "autocrine motility factor" (AMF). AMF has the properties of a protein with an estimated size of 55 kDa. At concentrations of 10 nM or less, AMF stimulated the random or directed motility of the producer cells. However, AMF is not an attractant for neutrophils. Amino acid analysis of the purified AMF protein revealed a high content of serine, glycine, glutamic acid, and aspartic acid residues. The activity of AMF was not replaced or blocked by known growth factors such as epidermal growth factor or type beta transforming growth factor. Mechanistic studies showed that AMF stimulated the incorporation of [3H]methyl into cell membrane phospholipids after incubation with [methyl-3H]methionine with a sustained increase in the methylation of phosphatidyldimethylethanolamine to phosphatidylcholine. In contrast, AMF did not affect the incorporation of [1,2-14C]choline into phosphatidylcholine. AMF was produced in large amounts by three different clones of ras oncogene-transfected metastatic NIH 3T3 cells but not by the nontransformed parental cells. AMF may play a major role in the local invasive behavior of tumor cells and may also facilitate the concerted invasion by groups of tumor cells.     

Comparison of amino acid sequence of bovine coagulation Factor IX (Christmas Factor) with that of other vitamin K-dependent plasma proteins.

The amino acid sequence of bovine blood coagulation Factor IX (Christmas Factor) is presented and compared with the sequences of other vitamin K-dependent plasma proteins and pancreatic trypsinogen. The 416-residue sequence of Factor IX was determined largely by automated Edman degradation of two large segments, containing 181 and 235 residues, isolated after activating Factor IX with a protease from Russell's viper venom. Subfragments of the two segments were produced by enzymatic digestion and by chemical cleavage of methionyl, tryptophyl, and asparaginyl-glycyl bonds. Comparison of the amino acid sequences of Factor IX, Factor X, and Protein C demonstrates that they are homologous throughout. Their homology with prothrombin, however, is restricted to the amino-terminal region, which is rich in gamma-carboxyglutamic acid, and the carboxyl-terminal region, which represents the catalytic domain of these proteins and corresponds to that of pancreatic serine proteases.     

Isolation and characterization of a cDNA coding for human factor IX.

A cDNA library prepared from human liver has been screened for factor IX (Christmas factor), a clotting factor that participates in the middle phase of blood coagulation. The library was screened with a single-stranded DNA prepared from enriched mRNA for baboon factor IX and a synthetic oligonucleotide mixture. A plasmid was identified that contained a cDNA insert of 1,466 base pairs coding for human factor IX. The insert is flanked by G-C tails of 11 and 18 base pairs at the 5' and 3' ends, respectively. It also included 138 base pairs that code for an amino-terminal leader sequence, 1,248 base pairs that code for the mature protein, a stop codon, and 48 base pairs of noncoding sequence at the 3' end. The leader sequence contains 46 amino acid residues, and it is proposed that this sequence includes both a signal sequence and a pro sequence for the mature protein that circulates in plasma. The 1,248 base pairs code for a polypeptide chain composed of 416 amino acids. The amino-terminal region for this protein contains 12 glutamic acid residues that are converted to gamma-carboxyglutamic acid in the mature protein. These glutamic acid residues are coded for by both GAA and GAG. The arginyl peptide bonds that are cleaved in the conversion of human factor IX to factor IXa by factor XIa were identified as Arg145-Ala146 and Arg180-Val181. The cleavage of these two internal peptide bonds results in the formation of an activation peptide (35 amino acids) and factor IXa, a serine protease composed of a light chain (145 amino acids) and a heavy chain (236 amino acids), and these two chains are held together by a disulfide bond(s). The active site residues including histidine, aspartate, and serine are located in the heavy chain at positions 221, 270, and 366, respectively. These amino acids are homologous with His57, Asp102, and Ser195 in the active site of chymotrypsin. Two potential carbohydrate binding sites (Asn-X-Thr) were identified in the activation peptide, and these were located at Asn157 and Asn167. The homology in the amino acid sequence between human and bovine factor IX was found to be 83%.     

Structure of the glycosaminoglycan domain in the type IX collagen-proteoglycan.

Type IX collagen represents 5-20% of the total collagen in hyaline cartilage. The molecules of this collagen are composed of three genetically distinct polypeptide subunits. One of these subunits, alpha 2(IX), contains covalently bound glycosaminoglycan (chondroitin sulfate or dermatan sulfate). We report here on the structure of the glycosaminoglycan attachment site of type IX collagen-proteoglycan. We show, by a combination of cDNA and peptide sequencing, that the attachment region contains the sequence Gly-Ser-Ala-Asp, located within the noncollagenous domain NC3 of the alpha 2(IX) chain. By comparing the exons encoding the NC3 domain in the alpha 2(IX) and alpha 1(IX) genes, we find that the exon coding for the glycosaminoglycan attachment site in the alpha 2(IX) gene is 48 base pairs long, whereas the homologous alpha 1(IX) exon is 33 base pairs. The NC3 domain is, therefore, five amino acid residues longer in alpha 2(IX) than in alpha 1(IX). The extra sequence in alpha 2(IX), Val-Glu-Gly-Ser-Ala, provides a simple explanation for the kink observed at the NC3 domain of type IX molecules when examined by electron microscopy. The inserted block of amino acid residues also provides the NC3 domain of alpha 2(IX) chains with a serine residue, not present in alpha 1(IX) that serves as attachment site for a glycosaminoglycan side chain. Our data show that the amino acid sequence that surrounds the glycosylated serine residue in type IX collagen-proteoglycan differs from glycosylated sequences in noncollagenous core proteins. The data also provide strong evidence that glycosylation of type IX collagen is not a chance glycosylation of a serine residue in a noncollagenous domain, but is a specific post-translational modification of this unusual collagen molecule.     

Antibody-probed conformational transitions in the protease domain of human factor IX upon calcium binding and zymogen activation: putative high-affinity Ca(2+)-binding site in the protease domain.

The Fab fragment of a monoclonal antibody (mAb) reactive to the N-terminal half (residues 180-310) of the protease domain of human factor IX has been previously shown to inhibit the binding of factor IXa to its cofactor, factor VIIIa. These data suggested that this segment of factor IXa may participate in binding to factor VIIIa. We now report that the binding rate (kon) of the mAb is 3-fold higher in the presence of Ca2+ than in its absence for both factors IX and IXa; the half-maximal effect was observed at approximately 300 microM Ca2+. Furthermore, the off rate (koff) of the mAb is 10-fold higher for factor IXa than for factor IX with or without Ca2+. Moreover, like the kon for mAb binding, the incorporation of dansyl-Glu-Gly-Arg chloromethyl ketone (dEGR-CK) into factor IXa was approximately 3 times faster in the presence of Ca2+ than in its absence. Since steric factors govern the kon and the strength of noncovalent interactions governs the koff, the data indicate that the region of factor IX at residues 180-310 undergoes two separate conformational changes before expression of its biologic activity: one upon Ca2+ binding and the other upon zymogen activation. Furthermore, the dEGF-CK incorporation data suggest that both conformational changes also affect the active site residues. Analyses of the known three-dimensional structures of serine proteases indicate that in human factor IX a high-affinity Ca(2+)-binding site may be formed by the carboxyl groups of glutamates 235 and 245 and by the main chain carbonyl oxygens of residues 237 and 240. In support of this conclusion, a synthetic peptide including residues 231-265 was shown to bind Ca2+ with a Kd of approximately 500 microM. This peptide also bound to the mAb, although with approximately 500-fold reduced affinity. Moreover, like factor IX, the peptide bound to the mAb more strongly (approximately 3-fold) in the presence of Ca2+ than in its absence. Thus, it appears that a part of the epitope for the mAb described above is contained in the proposed Ca(2+)-binding site in the protease domain of human factor IX. This proposed site is analogous to the Ca(2+)-binding site in trypsin and elastase, and it may be involved in binding factor IXa to factor VIIIa.