Expression and analysis of COOH-terminal deletions of the human thrombospondin molecule

Thrombospondin (TSP) is a homotrimeric extracellular glycoprotein with a subunit molecular mass of 140 kD. The subunits have a modular or domain-like structure and are held together by interchain disulphide bonds. A number of domains have been identified including those for the binding of collagen, fibrinogen, and heparin. Due to the trimeric form of the TSP molecule, the various domains are trivalent in nature and this contributes to the ability of TSP to mediate cell-substrate interactions. Indeed, TSP has recently been shown not only to promote cell adhesion but also to be intimately involved in cell growth and migration. The adhesive function of TSP is attributable to the solid-phase or matrix-bound form of the molecule. There is some evidence that the heparin-binding domain mediates incorporation of soluble TSP into the insoluble matrix form. The heparin-binding domain of TSP is a compact globular amino-terminal moiety that contains two clusters of basic amino acids and a single intrachain disulphide bond. To delineate the role of the heparin-binding domain in matrix assembly and to define further the precise region of interchain disulphide bonding that results in trimer formation, we have expressed deleted forms of the cDNA encoding TSP in SV-40-transformed. African green monkey kidney cells. The proteins synthesized from the various deleted TSP cDNAs were examined for (a) secretion into the culture medium and incorporation into the extracellular matrix; (b) binding to heparin-Sepharose; (c) immunoprecipitability by a conformation-specific monoclonal antibody; and (d) ability to form trimers. This analysis allowed us to draw the following conclusions. (a) A 218 amino acid NH2-terminal protein that preserves the intrachain disulphide bridge of the heparin-binding domain is capable of binding to heparin-Sepharose and incorporating into the extracellular matrix. (b) A shorter 164 amino acid NH2-terminal peptide that does not contain the intrachain disulphide bridge of the heparin-binding domain is neither able to bind to heparin-Sepharose nor able to incorporate into the extracellular matrix. (c) The region of interchain disulphide bridging necessary for trimer assembly resides within a cluster of seven cysteine residues immediately adjacent to the heparin-binding domain.     

Expression of members of the thrombospondin family by human skeletal tissues and cultured cells.

We have previously shown that the multifunctional platelet glycoprotein thrombospondin-1 (TSP-1) promotes resorption in an in vitro resorption assay. However, TSP-1 is one of a family of multifunctional TSP molecules, and the current study was undertaken to investigate whether it is TSP-1 or another TSP family member which may be involved in regulation of resorption in vivo. RT-PCR was performed on cultured human bone cells, cultured human chondrocytes, and three separate samples of human osteoclastoma tissue using primers specific for each TSP family member. mRNA for TSP-2 was detected in almost all samples, and significantly in all osteoclastomas in the above tissues, while TSP-1 was detected less frequently and was only seen in one of three osteoclastomas. TSP-3, -4, and COMP were detected only in a minority of cases. These results indicate that TSP-2 is the most common TSP family member found in skeletal tissues and that TSP-2, rather than TSP-1, may be the molecule responsible for promoting resorption in vivo.     

Relative abundance of thrombospondin 2 and thrombospondin 3 mRNAs in human tissues.

The levels of thrombospondin 2 (TSP2) and thrombospondin 3 (TSP3) mRNAs in a variety of human tissues were determined by analysis of multiple-tissue mRNA dot blots. For TSP2 mRNA, aorta and fetal heart had the greatest relative abundance. High levels were also detected for muscle, fetal, endocrine, immune, and nerve tissues. The pattern of expression of TSP3 mRNA was very different: kidney, pituitary gland, trachea, uterus, and fetal kidney had the greatest abundance. In general, TSP3 mRNA was expressed at high levels in endocrine, muscle, and fetal tissues. In addition to the tissue-specific differences, a more even distribution of TSP3 mRNA among tissues was observed. The high relative abundance of the two mRNAs in a variety of tissues and the tissue-specific differences in expression could be significant for understanding the diverse roles implicated for TSP2 and TSP3.     

Expression and characterisation of the thrombospondin type I repeats of human properdin

Properdin, an upregulator of the alternative complement pathway, is central to deposition of the activated complement fragment C3b on the surfaces of the pathogens, which it achieves by preventing the dissociation of the Bb catalytic subunit from the inherently labile C3bBb complexes. It is also known to bind sulphated glycoconjugates, such as sulphatides. Properdin has an unusual structure formed by oligomerisation of a rod-like monomer into cyclic dimers, trimers and tetramers. The monomer (approximately 53 kDa) contains an N-terminal region of no known homology, followed by six non-identical repeats of 60 amino acids (based on exon/intron boundaries), called 'thrombospondin type I repeats' or TSR modules. We have expressed and purified the N-terminal region and each of the individual TSR repeats in Escherichia coli. Although the individual recombinant TSRs, after a denaturation-renaturation cycle, appeared to be correctly folded modules, as judged by the one-dimensional (1D)- and 2D-nuclear magnetic resonance spectra of TSR3, they did not show binding to either C3b or sulphatide. Polyclonal antibodies were raised against each TSR and were found to be module-specific. The anti-TSR5 polyclonal antibody was found to inhibit binding of native human properdin to solid-phase C3b, or sulphatides. It could also block properdin-dependent haemolysis of rabbit erythrocytes. These results are consistent with the view that the TSR5 contains the major site in properdin which is involved in both C3b and sulphatide binding. It also suggests that a co-operative intramolecular interaction between TSRs, as found in the native molecule, is required for TSR5 to bind either C3b or sulphatides.     

Expression and initial characterization of a recombinant human thrombospondin heparin binding domain

Thrombospondin (TSP) is a trimeric glycoprotein of Mr 420,000. It was originally described as a major component of human platelet alpha granules and is essential for the secondary phase of platelet aggregation. TSP is also synthesized and secreted by a variety of nucleated cells where it functions in processes involving growth and adhesion of cells to the extracellular matrix. Many of these processes are heparin-inhibitable and are mediated by a proteolytic fragment of TSP called the heparin binding domain (HBD). In order to facilitate the analysis of the structure and function(s) of this domain, we have expressed this molecule in Escherichia coli. A fragment of a TSP cDNA that encodes the heparin binding domain was inserted into the prokaryotic expression vector pJBL6. In bacterial cells grown at 42 degrees C, this vector directs the synthesis of a 24,000-Da polypeptide. Milligram quantities of this protein were purified to homogeneity from E. coli lysates. The structure of the recombinant HBD was confirmed by protein sequencing. The protein was further characterized by analysis of its conformation and function. The recombinant HBD binds [3H]heparin with a Kd of 71 nM, almost identical to that of TSP-derived HBD (80 nM). Additionally, the recombinant HBD is able to compete for TSP binding to 11B carcinoma cells. These studies indicate that the recombinant HBD is synthesized and purified in a native configuration and is functionally equivalent to thrombospondin-derived HBD. They further indicate that glycosylation of the thrombospondin HBD is not necessary for its interaction with heparin and that sequences essential to this interaction reside within the first 229 amino acids of secreted thrombospondin.     

Characterization of pea chloroplastic carbonic anhydrase. Expression inEscherichia coli and site-directed mutagenesis

A cDNA encoding the mature, chloroplast-localized carbonic anhydrase in pea has been expressed inE. coli. The enzyme is fully active and yields of up to 20% of the total soluble protein can be obtained from the bacteria. This expression system was used to monitor the effects of site-directed mutagenesis of seven residues found within conserved regions in the pea carbonic anhydrase amino acid sequence. The effects of these modifications are discussed with respect to the potential of various amino acids to act as sites for zinc coordination or intramolecular proton shuttles.     

Reactions of thrombin-serpin complexes with thrombospondin.

Activated platelets release proteins that form stable complexes with thrombin (J. J. Miller, P. C. Browne, and T. C. Detwiler, Biochem. Biophys. Res. Commun. 151, 9-15, 1988). A working model for the reaction (P. C. Browne, J. J. Miller, and T. C. Detwiler, Arch. Biochem. Biophys. 265, 534-538, 1988) includes a dissociable complex of thrombin with released platelet protease nexin, leading to formation of a nondissociable thrombin-nexin complex that then becomes disulfide linked to thrombospondin. This disulfide-linked complex is converted back to the thrombin-nexin complex by reduction of disulfide bonds. Results that allow elaboration on this model are presented. After longer periods of incubation or after incubation with higher concentrations of thrombin, the amount of thrombin complexed with thrombospondin exceeded the amount of thrombin-nexin complex recovered after reduction of disulfide bonds. When the reaction mixture included inhibitors of formation of the thrombin-nexin complex, a slow formation of the thrombin-thrombospondin complex was observed. It was concluded that there is a nexin-independent as well as the faster nexin-dependent disulfide linkage of thrombin to thrombospondin. Addition of thrombin-antithrombin III complexes to the supernatant solution of activated platelets also led to complexes with thrombospondin, demonstrating that serpins other than platelet protease nexin facilitate incorporation of thrombin into complexes with thrombospondin. By heparin affinity chromatography, it was shown that thrombin-nexin complexes dissociably associate with thrombospondin prior to formation of disulfide-linked complexes. These observations are incorporated into a more detailed model of the reaction.     

Expression of human aldose and aldehyde reductases. Site-directed mutagenesis of a critical lysine 262.

Human aldose reductase (EC 1.1.1.21) and aldehyde reductase (EC 1.1.1.2) are implicated in the development of diabetic complications by a variety of mechanisms, and a number of drugs to inhibit these enzymes have been proposed for the therapy and prevention of these complications. To probe the structure and function of these two enzymes, we used site-directed mutagenesis in the cDNAs of both enzymes to replace lysine 262 with methionine. Wild-type and mutant enzymes were overexpressed in Escherichia coli and purified by anion exchange and affinity chromatography. N-terminal sequence analysis, Western blots, and kinetic studies confirmed the identity of the recombinant wild-type enzymes with the native human placental and liver enzymes. Recombinant aldose reductase (hAR) and aldehyde reductase (hGR) have apparent kinetic constants virtually identical to their respective native enzymes. The mutant aldose reductase (hARK262 greater than M) shows a 66-fold increase in Km for NADPH with respect to the wild type (1.9 +/- 0.4 microM versus 125 +/- 14 microM), whereas the Km for DL-glyceraldehyde increased 35-fold (20 +/- 2 versus 693 +/- 41 microM). The same constants for the mutant aldehyde reductase (hGRK262 greater than M) increased 97- and 86-fold, respectively (from 2.0 +/- 0.4 to 194 +/- 16 microM and from 1.6 +/- 0.4 to 137 +/- 3 mM). These results indicate that lysine 262 in aldose reductase and aldehyde reductase is crucial to their catalytic activity by affecting co-factor binding.     

Expression and mutagenesis of ZntA, a zinc-transporting P-type ATPase from Escherichia coli.

Cation-transporting P-type ATPases comprise a major membrane protein family, the members of which are found in eukaryotes, eubacteria, and archaea. A phylogenetically old branch of the P-type ATPase family is involved in the transport of heavy-metal ions such as copper, silver, cadmium, and zinc. In humans, two homologous P-type ATPases transport copper. Mutations in the human proteins cause disorders of copper metabolism known as Wilson and Menkes diseases. E. coli possesses two genes for heavy-metal translocating P-type ATPases. We have constructed an expression system for one of them, ZntA, which encodes a 732 amino acid residue protein capable of transporting Zn(2+). A vanadate-sensitive, Zn(2+)-dependent ATPase activity is present in the membrane fraction of our expression strain. In addition to Zn(2+), the heavy-metal ions Cd(2+), Pb(2+), and Ag(+) activate the ATPase. Incubation of membranes from the expression strain with [gamma-(33)P]ATP in the presence of Zn(2+), Cd(2+), or Pb(2+) brings about phosphorylation of two membrane proteins with molecular masses of approximately 90 and 190 kDa, most likely representing the ZntA monomer and dimer, respectively. Although Cu(2+) can stimulate phosphorylation by [gamma-(33)P]ATP, it does not activate the ATPase. Cu(2+) also prevents the Zn(2+) activation of the ATPase when present in 2-fold excess over Zn(2+). Ag(+) and Cu(+) appear not to promote phosphorylation of the enzyme. To study the effects of Wilson disease mutations, we have constructed two site-directed mutants of ZntA, His475Gln and Glu470Ala, the human counterparts of which cause Wilson disease. Both mutants show a reduced metal ion stimulated ATPase activity (about 30-40% of the wild-type activity) and are phosphorylated much less efficiently by [gamma-(33)P]ATP than the wild type. In comparison to the wild type, the Glu470Ala mutant is phosphorylated more strongly by [(33)P]P(i), whereas the His475Gln mutant is phosphorylated more weakly. These results suggest that the mutation His475Gln affects the reaction with ATP and P(i) and stabilizes the enzyme in a dephosphorylated state. The Glu470Ala mutant seems to favor the E2 state. We conclude that His475 and Glu470 play important roles in the transport cycles of both the Wilson disease ATPase and ZntA.     

Transforming growth factor-beta complexes with thrombospondin.

Thrombospondin (TSP) was demonstrated to inhibit the growth of bovine aortic endothelial cells, an activity that was not neutralized by antibodies to TSP or by other agents that block TSP-cell interactions but that partially was reversed by a neutralizing antibody to transforming growth factor-beta (TGF-beta). Similar to TGF-beta, TSP supported the growth of NRK-49F colonies in soft agar in a dose-dependent manner, which required epidermal growth factor and was neutralized by anti-TGF-beta antibody. Chromatography of a TSP preparation did not separate the TGF-beta-like NRK colony-forming activity from high molecular weight protein. However, when chromatography was performed at pH 11, this activity was dissociated from TSP. These results suggest that at least some growth modulating activities of TSP are due to TGF-beta associated with TSP by strong non-covalent forces. Most of the active TGF-beta released from platelets after degranulation was associated with TSP, as demonstrated by anti-TSP immunoaffinity and gel permeation chromatography. 125I-TGF-beta binds to purified TSP in an interaction that is specific in the sense that bound TGF-beta could be displaced by TGF-depleted TSP but not significantly by native TSP, heparin, decorin, alpha 2-macroglobulin, fibronectin, or albumin. Hence, TGF-beta can bind to TSP, and the complex forms under physiological conditions. Furthermore, TSP-associated TGF-beta is biologically active, and the binding of TGF-beta to TSP may protect TGF-beta from extracellular inactivators.     

Cloning and sequencing of chicken thrombospondin

Thrombospondin is a multifunction adhesive protein with the ability to bind proteoglycans, cell surface receptors, other proteins, and calcium ions. Several sequence motifs for some of these interactions have been identified in human thrombospondin. To evaluate the potential functional significance of these sequences and to begin a study of the evolution of thrombospondin, we have isolated and sequenced thrombospondin cDNA clones from a chicken embryo library. Comparison of the chicken and human sequences reveals that the NH2-terminal heparin-binding domains are only 34% identical. By contrast, the type 3 repeats and the COOH-terminal domains are 80 and 82% identical, respectively, when comparing human and chicken sequences. Potential cell recognition sequences of RGD and VTCG are conserved, with the chicken sequence containing an additional copy of the VTCG sequence. Whereas substitutions occur in the two potential heparin-binding motifs that have human counterparts, the chicken sequence contains a third potential heparin-binding motif. The results indicate that the evolutionary constraints on the various types of cell-binding motifs may be quite different.     

Monocyte synthesis of thrombospondin. The role of platelets

Monocytes produce thrombospondin (TSP), a trimeric glycoprotein whose primary function is not yet clear. Platelet-poor monocytes (less than 1 platelet/50 monocytes) cultured with [35S]methionine produced [35S] TSP barely detectable by immunoprecipitation with either monoclonal or polyclonal antibody to TSP. Platelet-poor monocytes that had not been so thoroughly depleted of platelets (6-12 platelets/50 monocytes) synthesized readily detectable amounts of [35S] TSP. Addition of increasing numbers of washed platelets to platelet-poor monocytes resulted in increasing synthesis of [35S]TSP. This monocyte-platelet interaction was specific for cell type; neither neutrophils nor red cells could substitute for platelets. The induction of synthesis was specific for TSP; monocyte synthesis of three other proteins was not induced upon the addition of platelets. Platelet lysate or thrombin-induced platelet releasate could not substitute for intact platelets. In fact, platelet lysate inhibited [35S]TSP synthesis by monocyte-platelet cultures. This inhibition was not due to endotoxin contamination, interference with immunoprecipitation, or dilution of the [35S]methionine pool. Platelets required contact with monocytes to exert their effect, as culturing the cell populations with a filter between them prevented increased [35S]TSP synthesis. Monocyte-platelet interactions may serve to specifically increase monocyte synthesis of the adhesive protein, TSP.     

Avian thrombocyte thrombospondin

A high molecular weight glycoprotein (450,000) was obtained from thrombin-treated duck thrombocytes by barium citrate adsorption technique followed by heparin-agarose affinity chromatography. Amino acid composition (high number of acidic amino acids and cystine) as well as carbohydrate contents (1.3 per cent hexosamine, 0.9 per cent sialic acid and 1.5 per cent hexose) showed similarity to mammalian platelet thrombospondin.     

Activation of human neutrophils increases thrombospondin receptor expression.

Thrombospondin (TSP), a 450-kDa extracellular matrix protein secreted by platelets may be instrumental in triggering polymorphonuclear leukocyte (PMN) activation and mediating PMN-endothelial cell interactions. TSP alone had no effect on O-2 generation but caused a significant increase in the chemoattractant FMLP-mediated O-2 generation. Purified HBD, but not the 140-kDa COOH-terminal fragment of TSP, retained the priming activity indicating that the priming effect was mediated through the HBD of TSP. The priming of FMLP-mediated O-2 generation by TSP, and our recent studies demonstrating that TSP stimulates PMN adhesion and motility suggested the presence of specific receptors for TSP on PMN. Binding studies on unactivated PMN, using 125I-TSP and competition with excess unlabeled TSP, demonstrated 2.4 x 10(3) binding sites/cell with an apparent Kd of 7 nM. Heparin did not compete for binding as effectively as unlabeled TSP. There were 1.5 x 10(3) heparin-inhibitable binding sites/cell with an apparent Kd of 8 nM that represented approximately 60% of the TSP-specific sites. Therefore, two distinct TSP receptors appeared to exist on unactivated PMN; one interacting with the heparin-binding domain of TSP and one interacting with a different site. Treating PMN with cytochalasin B followed by FMLP caused a 30-fold increase in TSP receptor expression. Binding studies on activated PMNs revealed 7.6 x 10(4) sites/cell; 60% of which were heparin inhibitable. The majority (5.3 x 10(4) sites/cell) of receptors expressed had an affinity of approximately 20 nM. About 50% of these sites were heparin inhibitable. In addition, there were 2.3 x 10(4) higher affinity sites/cell with an apparent Kd of 6 nM. Heparin-inhibitable sites comprised 70% of the higher affinity sites. The existence of a subset of TSP receptors that were heparin-inhibitable on PMN suggests that binding of TSP may trigger functionally independent responses. Increased receptor expression and expression of two high affinity binding sites following PMN activation may modulate PMN-endothelium or PMN-basement membrane interactions localized at the blood vessel wall.     

Specific inhibition of endothelial cell proliferation by thrombospondin.

Angiogenesis is a multi-step event involving endothelial cell migration, attachment, and proliferation. A thrombospondin (TSP)-like protein has recently been described as a naturally-occurring inhibitor of angiogenesis. We now report that human platelet TSP inhibits the in vitro proliferation of endothelial cells from the rabbit corpus luteum, bovine adrenal cortex and pulmonary artery, and human umbilical vein. The antiproliferative effect of TSP was neutralized by monoclonal antibodies against TSP. The growth arrest seen with TSP was specific for endothelial cells since TSP actually stimulated the growth of vascular smooth muscle cells and human foreskin fibroblasts. These results imply that the angiogenesis-inhibiting effect of TSP is mediated through an inhibition of endothelial cell proliferation. Elucidation of the mechanism of action of TSP on endothelial cell proliferation may lead to potential therapeutic approaches for the control of neovascular diseases.     

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.