Thrombospondin fragmentation by alpha-thrombin and resistance to gamma-thrombin.

In the presence of 2mM-Ca2+, alpha-thrombin slowly cleaved thrombospondin (Mr 180 000) into 150 000-Mr and 30 000-Mr fragments. In the absence of Ca2+, the platelet glycoprotein was progressively and completely hydrolysed by 3 units of the enzyme/ml to 130 000-Mr, 95 000-Mr and 65 000-Mr fragments. In contrast, the nonclotting enzyme form, gamma-thrombin, did not hydrolyse the platelet protein either in the presence or in the absence of Ca2+, even at 10-fold higher concentrations of enzyme. Protein-interacting regions removed from the catalytic site, like those required for fibrinogen recognition, are necessary for thrombin proteolysis of thrombospondin.     

Interdomain cleavage of plasma fibronectin by zinc-metalloproteinase from Serratia marcescens

Limited proteolysis of porcine plasma fibronectin by the 56 kDa proteinase (56K proteinase) (EC 3.4.24.4) from Serratia marcescens released six polypeptides: a 27 kDa peptide, the heparin-binding domain which comprises the NH2-terminal end; a 50 kDa peptide, a mid-molecule that mediates binding to gelatin or collagen; a 160 kDa peptide, that contained the heparin-binding domain with cell-spreading activity; and a 140 and a 20 kDa peptide which released from the 160 kDa peptide. Each fragment was purified and characterized by its chemical and biological properties, and it was found that they were respectively different domains. Both the 160 and the 140 kDa peptide contained one cysteine per mole of peptide. The 160 kDa peptides were connected by a 6 kDa peptide, which was present at the COOH-terminal end of the molecule and was biologically inactive. Only 6 kDa peptide contained a disulfide bond and produced 3 kDa peptide after reduction, whereas other fragments did not change with or without reduction on SDS-polyacrylamide gel electrophoresis. NH2-terminal sequence analyses of the released peptides showed that the 56K proteinase cleaved the fibronectin between the Arg-Thr (located at two different sites), Leu-Ser and Gln-Glu bonds. Out of 118 Arg residues, there are nine sequences containing Arg-Thr, and two of them near or at an interdomain location (at Arg 259 and 2239) were cleaved. Out of 124 Leu residues, there are 11 Leu-Ser sequences and only one, at 687, was cleaved. The above fragments with functional domain activity could be aligned according to the previously reported amino-acid sequence of human or bovine plasma fibronectin. The treatment of fibroblast cells by the 56K proteinase resulted in loss of morphological integrity and extracellular matrix.     

Platelet factor XIII is activated by calpain

The action of calpain (EC 3.4.22.17; Ca2+-dependent cysteine proteinase) on platelet factor XIII has been studied. Calpain I activated platelet factor XIII up to 76% of the maximum level observed with thrombin. Activation was accompanied by the limited proteolysis of the a subunit of platelet factor XIII to produce a 76 kDa fragment which was comparable to the proteolytic product by thrombin. Activation of platelet factor XIII by calpain was inhibited by EDTA, leupeptin, and endogenous calpain-specific inhibitor calpastatin. These findings suggest that calpain is responsible for the intracellular activation of platelet factor XIII.     

Characterization of two murine monoclonal antibodies (P10, P12) directed against different determinants on human blood platelet thrombospondin

Thrombospondin, a 450-kDa glycoprotein composed of three disulphide linked chains, is located in human blood platelet alpha-granules and is released from platelets upon stimulation. This glycoprotein is thought to play a major role in platelet aggregation. The aim of this study was to characterize two monoclonal antibodies (P10 and P12) directed against human blood platelet thrombospondin. When the released material obtained after stimulation of platelets with thrombin in the presence of 2 mM calcium was immediately treated with EDTA, labelled with 125I and incubated with monoclonal antibodies P10 and P12, both immunoprecipitated a major labelled protein band with a molecular mass of 160 kDa and a weaker band at 146 kDa, as analysed on reduced dodecyl sulphate/polyacrylamide gels. The major band corresponds in molecular mass to the thrombospondin subunits. If, however, the released material was left in the presence of Ca2+ for 48 h, then the main band was at 130 kDa and in addition one minor protein band (75 kDa) was immunoprecipitated by P10 whereas P12 recognized two minor protein bands (75 and 60 kDa). When P10 and P12 were incubated with 125I-labelled platelet releasates treated for 48 h at 4 degrees C with 10mM EDTA, three major protein bands (160, 146 and 130 kDa) were immunoprecipitated in addition to the minor bands mentioned above. These results indicate that thrombospondin is probably degraded by the endogenous platelet calcium-dependent protease. Investigation of tryptic peptide fragments of thrombospondin isolated by fast protein liquid chromatography showed that 125I-labelled antibody P10 bound to 400-kDa and 120-kDa fragments whereas 125I-labelled P12 only recognized a 400-kDa fragment. Competition studies involving solid-phase antibody binding and double antibody sandwich assays showed that P10 and P12 were directed against different determinants of thrombospondin. Purified thrombospondin, isolated in the presence of calcium, either directly or after treatment with EDTA, haemagglutinated trypsinized, formaldehyde-fixed sheep erythrocytes identically. The haemagglutination activity of EDTA-treated thrombospondin was inhibited by P10 and enhanced by P12. On the other hand, P10 and P12, despite their binding to calcium-treated thrombospondin, had no effect on its haemagglutination activity. Monoclonal antibodies P10 and P12 could be useful tools to investigate the role of thrombospondin in platelet aggregation.     

Isolation of the fibrinogen-binding region of platelet thrombospondin

Purified platelet thrombospondin binds to immobilized fibrinogen if both Ca++ and Mg++ are present. Digestion of the purified molecule with thermolysin results in a limited number of discrete proteolytic fragments. When such digests are subjected to affinity chromatography on immobilized fibrinogen, only the fragments with Mr of 120,000 and 140,000 are specifically bound and subsequently eluted by the addition of EDTA to the column buffer. Examination by SDS-PAGE under both reducing and nonreducing conditions reveals that the fibrinogen-binding domain is derived from the region of the thrombospondin molecule containing the interchain disulfide bonds. The requirement for Ca++ and Mg++ for optimal binding to fibrinogen is also manifest by the Mr 120,000/140,000 thermolytic fragments.     

Conversion of low-affinity platelet factor 4 to β-thromboglobulin by plasmin and trypsin

Low-affinity platelet factor 4 and β-thromboglobulin are platelet-secreted proteins that bind with low affinity to heparin. They show extensive immunological cross-reactivity and appear to differ in amino acid sequence only by an amino-terminal peptide unique to low-affinity platelet factor 4. The possibility that β-thromboglobulin is derived from low-affinity platelet factor 4 by proteolysis was investigated by exposing this protein to the action of plasmin, thrombin and trypsin. While thrombin had no effect, plasmin and trypsin converted low-affinity platelet factor 4 to a species with the same electrophoretic mobility and isoelectric point as β-thromboglobulin. We conclude that β-thromboglobulin is a breakdown product of low-affinity platelet factor 4.     

Conversion of low-affinity platelet factor 4 to beta-thromboglobulin by plasmin and trypsin

Low-affinity platelet factor 4 and beta-thromboglobulin are platelet-secreted proteins that bind with low affinity to heparin. They show extensive immunological cross-reactivity and appear to differ in amino acid sequence only by an amino-terminal peptide unique to low-affinity platelet factor 4. The possibility that beta-thromboglobulin is derived from low-affinity platelet factor 4 by proteolysis was investigated by exposing this protein to the action of plasmin, thrombin and trypsin. While thrombin had no effect, plasmin and trypsin converted low-affinity platelet factor 4 to a species with the same electrophoretic mobility and isoelectric point as beta-thromboglobulin. We conclude that beta-thromboglobulin is a breakdown product of low-affinity platelet factor 4.     

Multiple domains are involved in the interaction of endothelial cell thrombospondin with fibronectin

Thrombospondin is a large multifunctional glycoprotein synthesized, secreted and incorporated into the extracellular matrix by several cell types in culture. It is also present in the blood platelet and is secreted following platelet activation. We have previously shown that thrombospondin co-distributes with fibronectin in the extracellular matrix and that it can bind directly to purified fibronectin. In order to elucidate the chemical aspects of thrombospondin incorporation into the extracellular matrix, we studied the interaction of endothelial cell thrombospondin and fibronectin. We find that endothelial cell thrombospondin has two distinct binding domains for fibronectin. One domain is on the 70-kDa core fragment, probably similar to that of platelet thrombospondin. The other domain is on the 27-kDa N-terminal fragment and is unique to endothelial cell thrombospondin. The dissociation constant of the intact endothelial-cell-derived molecule is 0.7 +/- 0.2 x 10(-7) M. Following fragmentation, the separate domains bind with somewhat lower affinity: the core domain binds with a Kd of 3.4 +/- 1.5 x 10(-7) M and the N-terminal domain binds with a Kd of 8.8 +/- 1.8 x 10(-7) M. Binding of the intact molecule is Ca2+-independent. By contrast, following tryptic fragmentation, binding of the 70-kDa fragment is practically lost. It can be restored, however, by removal of Ca2+, indicating that the binding site on this domain is either sequestered or becomes so following fragmentation. Heparin, which also binds to both fragments, competed with fibronectin binding to the 27-kDa fragment but not to the 70-kDa domain. The fact that heparin also competitively inhibits fibronectin binding of the intact molecule further supports sequestration of the fibronectin-binding domain on the 70-kDa core fragment. Our data suggest that endothelial-cell thrombospondin possesses two distinct binding sites for fibronectin, a low-affinity constitutively available one and a high-affinity one, possibly sequestered on the intact unbound molecule.     

The binding of divalent metal ions to platelet factor XIII modulates its proteolysis by trypsin and thrombin

We investigated the effect of divalent metal ions on the proteolytic cleavage and activation of platelet Factor XIII by thrombin and trypsin. In the absence of metal ions (5 mM EDTA), trypsin and thrombin rapidly degraded platelet Factor XIII (80 kDa) to low-molecular-mass peptides (50-19 kDa) with simultaneous loss of transglutaminase activity. Divalent metal ions protected Factor XIII from proteolytic inactivation with an order of efficacy of Ca2+ greater than Zn2+ greater than Mg2+ greater than Mn2+. Calcium (2 mM) increased by 10- to 1000-fold the trypsin and thrombin concentrations required to degrade Factor XIII to a 19-kDa peptide. Factor XIIIa formed by thrombin in the presence of 5 mM EDTA had one-half the specific activity of Factor XIIIa formed in the presence of calcium. Factor XIII was cleaved by trypsin in the presence of 5 mM Ca2+ to a 51 +/- 3-kDa fragment that had 60% of the original Factor XIIIa activity. A similar tryptic peptide formed in the presence of 5 mM EDTA did not have transglutaminase activity. In the presence of 5 mM Mg2+, thrombin cleaved Factor XIII to a major 51 +/- 3-kDa fragment that had 60% of the Factor XIIIa activity. Mn2+ (0.1-5 mM) limited trypsin and thrombin proteolysis. The resulting digest containing a population of Factor XIII fragments (50-14 kDa) expressed 50-60% transglutaminase activity of Factor XIIIa. Factor XIII was fully activated by both trypsin and thrombin in the presence of 5 mM Zn2+, resulting in two fragments of 76 and 72 kDa. We conclude that the binding of divalent metal ions to platelet Factor XIII induces conformational changes in the protein that alter its susceptibility to proteolysis and influence the expression of transglutaminase activity.     

Platelet thrombospondin mediates attachment and spreading of human melanoma cells

Human platelet thrombospondin adsorbed on plastic promotes attachment and spreading of human G361 melanoma cells. Attachment is rapid, and spreading is maximal by 90 min with 60-90% of the attached cells spread. In contrast, thrombospondin promotes attachment but not spreading of human C32 melanoma cells, which attach and spread only on laminin substrates. The specificity of these interactions and the regions of the thrombospondin molecule involved in attachment and spreading were examined using proteolytic fragments of thrombospondin and by inhibition studies. The sulfated fucan, fucoidan, and monoclonal antibody A2.5, which is directed against the heparin-binding domain of thrombospondin, selectively inhibit spreading but only weakly inhibit attachment. Monoclonal antibodies against some other domains of thrombospondin, however, are potent inhibitors of attachment. The amino-terminal heparin-binding domain of thrombospondin does not promote attachment. Large fragments lacking the heparin-binding domain support attachment but not spreading of G361 cells. Attachment activity is lost following removal of the 18-kD carboxyl-terminal domain. These results suggest that at least two melanoma ligands are involved in cell attachment and spreading on thrombospondin. The carboxyl-terminal region and perhaps other regions of the molecule bind to receptor(s) on the melanoma surface that promote initial attachment but not cell spreading. Interaction of the heparin-binding domain with sulfated glycoconjugates on melanoma surface proteoglycans and/or sulfated glycolipids mediates spreading. Monoclonal antibodies A2.5 and C6.7 also reverse spreading of G361 cells growing on glass culture substrates, suggesting that binding to thrombospondin mediates attachment of these melanoma cells in culture.     

Characterization of an anti-thrombospondin monoclonal antibody (P8) that inhibits human blood platelet functions. Normal binding of P8 to thrombin-activated Glanzmann thrombasthenic platelets

Stimulated human blood platelets release thrombospondin, an alpha-granule glycoprotein of 450 kDa. The aim of this work was to characterize an anti-thrombospondin monoclonal antibody (P8) in order to study the role of thrombospondin in platelet functions. The presence of thrombospondin receptor sites on resting and thrombin-stimulated platelets of three Glanzmann's thrombasthenia patients and normal donors was investigated using the P8 monoclonal antibody. Monoclonal antibody P8 was extensively characterized using ELISA, immunoprecipitation, immunoadsorbent affinity chromatography combined with tryptic peptide map analysis and crossed immunoelectrophoretic techniques. Labelled P8 bound strongly to thrombin-stimulated normal platelets (n = 14917 +/- 420, mean +/- SD) (Kd = 9.2 +/- 3.0 nM) and poorly to resting platelets (n = 2697 +/- 1278) (Kd = 24.8 +/- 18.6 nM). Moreover, the number of binding sites for P8 on thrombin-stimulated platelets from three Glanzmann's thrombasthenia patients, lacking the IIb-IIIa glycoprotein complex, were found similar to normal samples. F(ab')2 fragments of P8 inhibited aggregation of, and reduced secretion from, washed platelets stimulated by low concentrations of thrombin (0.05-0.06 U/ml) and collagen (0.5-0.6 microgram/ml). F(ab')2 fragments of P8 inhibited thrombin-induced platelet aggregation, but did not reduce fibrinogen binding (n) nor affect its dissociation constant (Kd). Inhibition of platelet aggregation by P8 suggests that thrombospondin plays an active role in promoting platelet aggregation, at low concentrations of thrombin and collagen. Normal binding of P8 to thrombin-stimulated Glanzmann thrombasthenic platelets indicates the presence of a thrombospondin receptor on the platelet surface distinct from the GPIIb-IIIa complex.     

Evidence that platelet and skeletal sarcoplasmic reticulum Ca2+-ATPase are structurally distinct

Proteolytic digestion and indirect immunostaining were used to compare the platelet and sarcoplasmic reticulum Ca2+-ATPase proteins. When the platelet and sarcoplasmic reticulum Ca2+-ATPase proteins were digested in the native state with trypsin, the platelet Ca2+-ATPase, which had an apparent undigested molecular mass of 103 kDa, yielded 78-kDa and 25-kDa fragments. Calcium transport activity depended on the integrity of the 103-kDa protein, while the digested protein had residual ATPase activity. Tryptic digestion of the sarcoplasmic reticulum pump protein, which also had an undigested molecular mass of 103 kDa, yielded products with apparent molecular masses of 55 kDa, 36 kDa, and 26 kDa. Distinct patterns were also observed when the platelet and sarcoplasmic reticulum calcium pump proteins were digested with chymotrypsin and Staphylococcus aureus protease in the presence of sodium dodecyl sulfate. Chymotrypsin digestion of the platelet protein resulted in the appearance of products with apparent molecular masses of 70 kDa, 39 kDa, and 31 kDa, while a similar digestion of the sarcoplasmic reticulum calcium pump protein yielded 54-kDa, 52.5-kDa, 46-kDa, 41-kDa, and 36-kDa fragments. Exposure of the sarcoplasmic reticulum and platelet Ca2+-ATPase proteins to S. aureus protease also yielded dissimilar fragmentation patterns. These results indicate that the Ca2+-ATPases from platelets and sarcoplasmic reticulum are distinct proteins.     

Alternative model for the internal structure of laminin

A monoclonal antibody to laminin, LMN-1, was generated by immunizing rats with laminin from the EHS tumor and fusing the rat spleen cells with mouse NS-1 myeloma cells. Laminin fragments were generated by proteolytic digestion with thrombin, thermolysin, and chymotrypsin. Monoclonal antibody binding fragments were identified by immunoblotting. Fragments which bound monoclonal antibody LMN-1 included a 440-kilodalton (kDa) chymotrypsin fragment and thermolysin fragments of 440 and 110 kDa. These fragments could also be generated from within a 600-kDa thrombin fragment. Digestion of the 440-kDa chymotrypsin fragment with thermolysin generated the 110-kDa antibody binding fragment and a 330-kDa nonbinding fragment. Immunoblotting was performed on extracts of PYS-2 cells and EHS cells using polyclonal and monoclonal antibodies to laminin. Polyclonal antibodies stained the intact 850-kDa complex and the 200- and 400-kDa subunits, while monoclonal LMN-1 stained only the 400-kDa subunit and the complete molecule. Rotary shadowing of monoclonal LMN-1 bound to laminin molecules indicated that the binding site was within the long arm of laminin. Changes in the model of the internal organization of the laminin molecule are proposed, based on the binding of LMN-1 to the 400-kDa subunit and specific proteolytic fragments. The locations of the major thrombin and chymotrypsin fragments in the model are rotated 180 degrees relative to the previously described model [Ott, U., Odermatt, E., Engel, J., Furthmayr, H., & Timpl, R. (1982) Eur. J. Biochem. 123, 63-72] to include part of the 400-kDa subunit of laminin.     

Localization of two binding domains for thrombospondin within fibronectin.

Thrombospondin is a major glycoprotein of the platelet alpha-granule and is secreted during platelet activation. Several protease-resistant domains of thrombospondin mediate its interactions with components of the extracellular matrix including fibronectin, collagen, heparin, laminin, and fibrinogen. Thrombospondin, as well as fibronectin, is composed of several discretely located biologically active domains. We have characterized the thrombospondin binding domains of plasma fibronectin and determined the binding affinities of the purified domains; fibronectin has at least two binding sites for thrombospondin. Thrombospondin bound specifically to the 29-kDa amino-terminal heparin binding domain of fibronectin as well as to the 31-kDa non-heparin binding domain located within the larger 40-kDa carboxy-terminal fibronectin domain generated by chymotrypsin proteolysis. Platelet thrombospondin interacted with plasma fibronectin in a specific and saturable manner in blot binding as well as solid-phase binding assays. These interactions were independent of divalent cations. Thrombospondin bound to the 29-kDa fibronectin heparin binding domain with a Kd of 1.35 x 10(-9) M. The Kd for the 31-kDa domain of fibronectin was 2.28 x 10(-8) M. The 40-kDa carboxy-terminal fragment bound with a Kd of 1.65 x 10(-8) M. Heparin, which binds to both proteins, inhibited thrombospondin binding to the amino-terminal domain of fibronectin by more than 70%. The heparin effect was less pronounced with the non-heparin binding carboxy-terminal domain of fibronectin. By contrast, the binding affinity of the thrombospondin 150-kDa domain, which itself lacked heparin binding, was not affected by the presence of heparin. Based on these data, we conclude that thrombospondin binds with different affinities to two distinct domains in the fibronectin molecule.     

Domain mapping of chicken gizzard caldesmon

Limited proteolysis, affinity chromatography, and immunoblotting have been used to define the domains of chicken gizzard caldesmon, caldesmon120, that interact with calmodulin, F-actin, and a monoclonal antibody prepared using human platelet caldesmon. Treatment of caldesmon120 with chymotrypsin produces groups of fragments near 100, 80, 60, 38, and 20 kDa. Further digestion produces peptides between 40 and 50 kDa. The 100- and 80-kDa peptides cross-react with the monoclonal antibody; the smaller polypeptides do not. The kinetics of cleavage and the antibody studies indicate that the 38- and 80-kDa fragments are the two major pieces of the 120-kDa protein. The 38-kDa fragment, purified by high performance liquid chromatography, and several of its subfragments at 21 and 25 kDa sediment with F-actin, bind to calmodulin-Sepharose in the presence of Ca2+, and are displaced from F-actin by Ca2+-calmodulin. The 80-kDa fragments did not interact with F-actin or calmodulin. We have tentatively placed the 38-kDa fragment at the C-terminal using polyclonal antibodies selected against a beta-galactosidase-caldesmon120 fusion protein produced by a lambda gt11 lysogen. The 38-, 25-, and 21-kDa fragments cross-react with these antibodies; the 80- and 60-kDa fragments do not. Caldesmon77 from human platelets also cross-reacts with these selected antibodies. The results suggest that interacting calmodulin and F-actin binding sites are localized on a 38-kDa C-terminal fragment of caldesmon. The smallest subfragment of this peptide that binds to both F-actin and calmodulin-Sepharose is about 21 kDa. The monoclonal antibody epitope is tentatively localized near the N-terminal of caldesmon77 and must be within 50 kDa of the N-terminal on caldesmon120.     

Structural analysis of p36, a Ca2+/lipid-binding protein of the annexin family, by proteolysis and chemical fragmentation

Limited proteolysis of the core domain of the 36-kDa protein p36 by trypsin gives a first insight into the structural organization of the four annexin repeats. Trypsin opens only a single peptide bond, situated between residues 204 and 205. The two fragments (of 20 kDa and 15 kDa), each containing two annexin repeats, remain as a tight complex (nicked core), which binds phospholipids in a Ca2(+)-dependent manner. After denaturation by 9 M urea, the nicked core is again formed upon renaturation provided both fragments are present. If the fragments are separated by chromatography in urea prior to renaturation, they show different behaviour. The 15-kDa C-terminal repeats aggregate, while the 20-kDa N-terminal repeats stay in solution. In comparison to p36, fragments with two (20-kDa fragment) or one (N-terminal CNBr fragment) annexin repeats show a conformational alteration in CD spectroscopy and hydrodynamics and display an increased susceptibility to proteases. In line with these differences, their Ca2(+)-dependent affinity to phospholipids is more than 10-20-fold decreased. Thus the four annexin repeats form together an integrated domain with multiple contacts between the repeats. Although stable derivatives with less than four repeats can be obtained, their Ca2+/phospholipid binding affinities are noticeably reduced.     

Identification of a 15 kilodalton actin binding region on gizzard caldesmon probed by chemical cross-linking

Fluorescent labeling, limited proteolysis, amino acid sequence determinations, affinity chromatography and specific chemical crosslinking were used to determine the smallest fragment of gizzard caldesmon that interacts with actin. The time course of cleavage with thrombin or submaxillaris arginase-C protease indicates that 90kDa and 35kDa fragments are the two major pieces of the 120kDa native protein. Amino acid sequence determination indicates that the 90kDa fragment is the N-terminal portion of the molecule. Further degradation gave rise to a 15kDa product whose N-terminal amino acid sequence was determined within the first 28 amino acids. Carbodiimide crosslinking with actin revealed that the 15kDa part of the molecule is probably not involved in the actin binding process but may participate in a twisting of the F-actin filament and be responsible of the caldesmon regulatory function during smooth muscle contraction.     

The protein S-binding site localized to the central core of C4b-binding protein

Human C4b-binding protein (C4BP) is a regulator of the classical pathway of the complement system. It appears in two forms in plasma, as free protein and in a noncovalent complex with the vitamin K-dependent coagulation protein, protein S. In the electron microscope C4BP has a spider-like structure with a central core and seven extended tentacles, each of which has a binding site for C4b, although the protein S-binding site has not been unequivocally pinpointed. C4BP was subjected to chymotrypsin digestion which yielded two major fragments, one of 160 kDa representing the central core, and one of 48 kDa representing the cleaved-off tentacles. We have now localized the protein S-binding site to the 160-kDa central core fragment. Using immunoblotting with a panel of polyclonal antisera, the isolated central core was shown to be completely devoid of 48-kDa fragments. The protein S-binding site was susceptible to proteolysis by chymotrypsin, but was protected by a molar excess of protein S included during the proteolysis. The 160-kDa central core fragment consisted of identical, disulfide-linked 25-kDa peptides and a proper disulfide bond arrangement was crucial to protein S binding. Using a direct binding assay it was shown that the isolated central core had the same affinity for protein S as intact C4BP.     

Structure of thrombospondin

The NH2-terminal amino acid sequences of thrombospondin and of a 30,000-Da heparin-binding peptide derived from thrombospondin by treatment with plasmin are identical. The heparin-binding peptide is homogeneous in size but slightly heterogeneous in charge with the predominant isoelectric points being 6.1 and 5.7. Electron microscopy of tungsten replicas of thrombospondin reveals a tripartite structure resembling a "bola" which is about 60 nm across when fully extended. Each part of the molecule terminates in a globular node or head which disappears upon limited plasmin digestion, suggesting that the heparin-binding peptide is located in the head region. In addition to the heparin-binding peptide, a 20,000-Da peptide also apparently associated with the head region is liberated during proteolysis. The electron micrographs indicate that the legs of the bola-like structure must be folded into an extended, flexible, tertiary structure. These legs, each of about 65,000 Da, appear to be attached near the ends opposite the heads, probably by disulfide bonds. Each leg possesses a tab or protein (approximately 20,000 Da) which juts out from this attachment point.     

Interaction of platelet factor 4 with human platelets

Human washed resting platelets bound 125I-labeled platelet factor 4 in a reaction which was saturable and approached equilibrium within 15-30 min. Scatchard plot analysis of the binding isotherms suggested a single class of specific binding sites. Excess of unlabeled protein and low- and high-affinity heparin competed for platelet factor 4 binding sites on the platelet surface and caused a partial displacement of this molecule. Anti-platelet factor 4 Fab fragments caused inhibition of binding of 125I-platelet factor 4 to platelets. Most of the labeled platelet factor 4 which was bound to intact platelets was recovered in the Triton X-100-insoluble cytoskeletal fraction prepared from the same platelets after their stimulation by thrombin. The association with the cytoskeleton was inhibited by anti-platelet factor 4 Fab fragments and by low-affinity heparin. Anti-platelet factor 4 125I-labeled Fab fragments bound to resting platelets, and this binding was greatly increased following platelet stimulation with thrombin. This suggested that endogenously secreted platelet factor 4 also binds to the platelet surface. No significant binding to platelets of 125I-labeled beta-thromboglobulin and 125I-labeled anti-beta-thromboglobulin Fab fragments was observed. Fab fragments of monospecific anti-human platelet factor 4 antibody raised in rabbits inhibited platelet aggregation and secretion induced by low concentrations of thrombin. Fab fragments of anti-beta-thromboglobulin antibody had no inhibitory effect. We suggest that the binding of alpha-granule-derived platelet factor 4 to the specific sites on the surface of platelets may modulate platelet aggregation and secretion induced by low levels of platelet agonists.