Human thrombomodulin gene is intron depleted: nucleic acid sequences of the cDNA and gene predict protein structure and suggest sites of regulatory control

An ultrastructural study of thrombomodulin endocytosis: internalization occurs via clathrin-coated and non-coated pits

The regulation of thrombomodulin (TM) expression has been reported to occur by several mechanisms. We have examined constitutive internalization of TM using immunofluorescent and electron microscopic (EM) methods. A cell model was developed to study this process by introducing TM DNA into COS-7 cells for expression. The recombinant TM was determined to behave similarly to native TM from human umbilical vein endothelial cells (HUVEC) with respect to M(r) and cell surface functional activity. The transfected cells expressed 8-100-fold more functional TM per cell than HUVEC. Immunofluorescent studies on these cells indicated that anti-TM antibody-TM complex was internalized in a time- and temperature-dependent manner, with internalization detectable within 10 minutes. When the cells were incubated at 4 degrees C with gold-labelled anti-TM antibody, most of the gold particles were surface bound and detected by EM as individual particles or clusters of 2 or 3 particles. Initiation of endocytosis for 10 to 60 minutes resulted in a redistribution of gold particles into small clusters predominantly in non-coated pits and rarely in clathrin-coated pits, subsequently in early endosomes, multivesicular bodies, and lysosomes. Similar studies were performed with gold-conjugated thrombin, demonstrating a similar route of intracellular processing. These studies provide ultrastructural evidence that the process of endocytosis of TM involves the participation of both clathrin-coated and non-coated pits and vesicles, but that the latter process predominates. Further structure/function studies are indicated using our cell model, since defects in the endocytic pathway of this important anticoagulant receptor may contribute to the development of thromboembolic disease.

Efficient approach to synthesis of two-chain asymmetric cysteine analogs of receptor-binding region of transforming growth factor-alpha

The putative receptor-binding region of human transforming growth factor-alpha (TGF alpha) has been shown to be contributed by two fragments: an A-chain (residue 12-18) and a 17-residue carboxyl fragment (residue 34-50) that includes a disulfide-containing C-loop (residue 34-43). An approach to the synthesis of two-chain analogs containing an intermolecular disulfide linked A-chain and the 17-residue carboxyl fragment (C-fragment) possessing receptor-binding activity is described. The synthesis was achieved by the solid-phase method using the Boc-benzyl protecting group strategy. The single Cys of the A-chain was activated as a mixed disulfide with 2-thiopyridine to form the intermolecular disulfide bond with Cys41 or Cys46 of the C-fragment on the resin support. Prior to this reaction, the acetamido (Acm) protecting group of Cys41 or Cys46 was removed by Hg(OAc)2 on the resin support. The peptide and side chain protecting groups including the S-methylbenzyl moiety of the Cys34 and Cys43 were concomitantly cleaved by high HF. The intramolecular disulfide with two unprotected Cys was formed in the presence of an intermolecular disulfide. This intramolecular disulfide bond formation was usually not feasible under the traditionally-held scheme at basic pH since disulfide interchange would occur faster than intramolecular oxidation. To prevent the disulfide interchange, a new method was devised. The intramolecular disulfide bond oxidation was mediated by dimethylsulfoxide at an acidic pH, at which the disulfide interchange reaction was suppressed. The desired product was obtained with a 60-70% yield.(ABSTRACT TRUNCATED AT 250 WORDS)

Partial characterization of vertebrate prothrombin cDNAs: amplification and sequence analysis of the B chain of thrombin from nine different species

The cDNA sequence of the B chain of thrombin (EC 3.4.21.5) has been determined from nine vertebrate species (rat, mouse, rabbit, chicken, gecko, newt, rainbow trout, sturgeon, and hagfish). The amino acid sequence identities vary from 96.5% (rat vs. mouse) to 62.6% (newt vs. hagfish). Of the 240 amino acids spanned in all the species compared, there is identity at 110 (45.8%) positions. When conservative changes are included, the amino acid similarity increases to 75%. The most conserved portions of the B chain are the active-site residues and adjacent amino acids, the B loop, and the primary substrate-binding region. In addition, the Arg-Gly-Asp motif is conserved in 9 of the 11 species compared, and the chemotactic/growth factor domain is well conserved in all of the 11 species compared. The least conserved regions of the B chain correspond to surface loops, including the putative thrombomodulin-binding sites and one of the hirudin-binding regions. The extent of the amino acid sequence similarity and the conservation of many of the functional/structural motifs suggests that, in addition to their role in blood coagulation, vertebrate thrombins may also play an important role in the general mechanisms of wound repair.

Thrombomodulin gene regulation by cAMP and retinoic acid in F9 embryonal carcinoma cells

Thrombomodulin (TM) expression was investigated during differentiation of F9 embryonal carcinoma cells into primitive or parietal endoderm. Exposure of F9 cells to retinoic acid (RA) triggers differentiation into primitive endoderm and induces the appearance of barely detectable amounts of TM mRNA, whereas treatment with dibutyryl cAMP plus theophylline (CT) augments the levels of TM mRNA to a 4-fold greater extent than RA. Exposure of F9 cells to RA plus CT initiates differentiation into parietal endoderm and synergistically increases the levels of TM mRNA by 10- to 12-fold compared with CT. The time-dependent establishment of cooperativity between RA and CT appears to be secondary to RA-induced differentiation to primitive endoderm. The above alterations in TM mRNA levels occur by a transcriptional mechanism as judged by nuclear run-on experiments. Transient gene expression experiments show that the human TM promoter is transactivated by coexpression of the human RA receptor beta. Thus, the mechanism of induction of TM expression in F9 cells undergoing differentiation to parietal endoderm appears to be similar, but not identical, to that noted for other late response genes.

Low nucleotide diversity in man

The nucleotide diversity (pi) in humans is studied by using published cDNA and genomic sequences that have been carefully checked for sequencing accuracy. This measure of genetic variability is defined as the number of nucleotide differences per site between two randomly chosen sequences from a population. A total of more than 75,000 base pairs from 49 loci are compared. The DNA regions studied are the 5' and 3' untranslated regions and the amino acid coding regions. The coding regions are divided into nondegenerate sites (i.e., sites at which all possible changes are nonsynonymous), twofold degenerate sites (i.e., sites at each of which one of the three possible changes is synonymous) and fourfold degenerate sites (i.e., sites at which all three possible changes are synonymous). The pi values estimated are, respectively, 0.03 and 0.04% for the 5' and 3' UT regions, and 0.03, 0.06 and 0.11% for nondegenerate, twofold degenerate and fourfold degenerate sites. Since the highest pi value is only 0.11%, which is about one order of magnitude lower than those in Drosophila populations, the nucleotide diversity in humans is very low. The low diversity is probably due to a relatively small long-term effective population size rather than any severe bottleneck during human evolution.

Mapping the receptor-recognition site of human transforming growth factor-alpha

The receptor-recognition site human transforming growth factor-alpha (TGF alpha), a 50-residue tricyclic peptide with three disulfide bonds, was mapped by a set of 46 peptide analogs consisting of linear, monocyclic, bicyclic, and tricyclic structures representing individual and overlapping subdomains of human TGF alpha. Linear overlapping fragments ranging from 7 to 18 residues and spanning the entire length of TGF alpha as well as monocyclic analogs with one disulfide linkage were found to be inactive in both receptor-binding and mitogenic assays. Bicyclic analogs with two disulfide linkage and representing either the amino or carboxyl two-thirds of TGF alpha showed low activity at 0.1-0.9 mM concentrations. Tricyclic analogs containing all three disulfide linkages but lacking either the amino or carboxyl terminal heptapeptide was, respectively, 3% and 0.1% as active as TGF alpha. These results show that determinants for the receptor binding cannot be represented by a short continuous fragment or a single subdomain, but are located on a discontinuous surface on a folded structure with disulfide restraints. Furthermore, these results when combined with our previous results which shows that the middle subdomain (second disulfide loop) is not involved in the receptor binding suggest that the receptor-binding residues are constituted of three fragments located at the first and third subdomains as well as the external carboxyl peptide.

Thrombomodulin activity on human saphenous vein grafts prepared for coronary artery bypass

Thrombomodulin, a membrane glycoprotein present on normal vascular endothelium, binds circulating thrombin and is important in protein C activation. These functions contribute to the nonthrombogenic nature of endothelium. Damage during harvest and ex vivo storage of vein grafts may result in dysfunction of this endothelial anticoagulant barrier and possibly contribute to early graft thrombosis. We studied the functional activity and antigenic expression of thrombomodulin on saphenous veins before (initial) and after (harvested) harvest and storage for coronary artery bypass grafting in 15 patients. Also, fresh saphenous vein was studied after mechanical endothelial stripping. After storage for 2.7 +/- 0.6 hours at room temperature in heparinized saline, thrombomodulin functional activity in harvested vein segments was 28% less than initial segments (p = 0.08). Endothelial stripping resulted in a 79% reduction in thrombomodulin activity compared with initial segments (p = 0.04). Immunohistochemical staining confirmed thrombomodulin antigen on vein grafts after harvest and storage, but not on segments stripped of endothelium. Thrombomodulin functional activity and antigenic expression on human saphenous vein grafts is not significantly changed by harvest and relatively short periods of storage at room temperature in heparinized saline.

Relationship between post-translational glycosylation and anticoagulant function of secretable recombinant mutants of human thrombomodulin

Two glycoforms of a soluble mutant of recombinant human thrombomodulin (rec.TM) were used to identify critical N- and O-linked glycans of the endothelial cell thrombin receptor. While N-linked glycans were not found to be involved in any function of rec.TM, an acidic chondroitin sulphate-like glycosaminoglycan (CSGAG) was found to be critical for all the direct anticoagulant functions of rec.TM, including inhibition of thrombin-mediated platelet aggregation. A glycoform of rec.TM lacking CSGAG had very poor anticoagulant activity. Furthermore, the glycoform of rec.TM possessing CSGAG showed strong inhibition by and had high affinity for poly-cationic basic proteins, whereas the CSGAG-deficient rec.TM did not. Monoclonal antibody binding as well as lectin mapping of rec.TM with agglutinins identified sialic acid containing O-linked glycans in both glycoforms additional to the CSGAG in high molecular weight rec.TM These findings define important molecular interactions modulating the anticoagulant function of TM, which appear to be critically regulated by CSGAG, and also showed that the overall post-translational glycosylation pattern of the two glycoforms was very similar except for the presence of CSGAG. The possibility exists that differently expressed glycoforms of TM may be crucial for the expression of endothelial cell-related anticoagulant potential in different vascular beds.

Different glycoforms of human thrombomodulin. Their glycosaminoglycan-dependent modulatory effects on thrombin inactivation by heparin cofactor II and antithrombin III

The relationship between thrombomodulin-associated O-linked glycosammoglycans (GAGs) and the exogenous GAGs heparin or dermatan sulfate was studied in the inhibition of thrombin by antithrombin III (AT III) or heparin cofactor II (HC II). Both rabbit thrombomodulin (TM) and two glycoforms (a high-Mr form containing GAGs and a low-Mr form lacking the majority of O-linked GAGs) of a recombinant human TM deletion mutant (rec-TM) were used. The rapid inactivation of thrombin by HC II in the presence of dermatan sulfate was prevented by both the high-Mr rec-TM and the rabbit TM. In contrast, both rabbit TM treated with chondroitin ABC lyase to remove O-linked GAGs and the low-Mr form of rec-TM had only weak protecting effects. In the absence of exogeneous dermatan sulfate, thrombin inhibition by a high concentration of HC II was slightly accelerated by the high-Mr form of rec-TM but protected by rabbit TM. When thrombin inhibition by AT III in the presence of heparin was studied, both high-Mr rec-TM and rabbit TM again invoked a similar reduction of inactivation rates, whereas in the absence of exogenous heparin, both high-Mr forms accelerated thrombin inhibition by AT III. The diverse reactivities of various forms of TM towards HC II and AT III were also observed during protein C activation by the thrombin-TM complex. These results suggest that thrombin activity at the vessel wall or in fluid phase may undergo major kinetic modulations depending on the type of protease inhibitor, the presence or absence of exogenous GAGs and the glycosylation phenotype of TM. The dependence of TM anticoagulant function on the presence of an intrinsic GAG moiety suggests that variant glycoforms of this endothelial cell cofactor may be expressed differently in a species-, organ-, or tissue-specific manner as a means to regulate TM function in diverse vasculatures.

Initiation of the protein C pathway

The protein C activation system provides an interesting model for the control of coagulation. Expression of the critical receptor appears to be under the control of inflammatory mediators. Open questions of considerable importance relate to the physiological significance of these observations. Preliminary evidence is emerging that thrombomodulin is down-regulated in patients and animals with inflammation, but it remains to be determined if the loss of thrombomodulin causes the thrombotic complications or occurs in response to these complications.

Structure and chromosomal localization of the gene for the oligodendrocyte-myelin glycoprotein

Utilizing a cDNA clone encoding the oligodendrocyte-myelin glycoprotein (OMgp) to screen a human genomic DNA library, we have obtained a clone that contains the OMgp gene. The genomic clone was restriction mapped and the OMgp gene and its 5' and 3' flanking regions were sequenced. A single intron is found in the 5' untranslated region of the gene, while the coding region is uninterrupted by an intron. This placement of a single intron in the OMgp gene is identical to that of the gene for the alpha-chain of platelet glycoprotein Ib, which, along with OMgp, belongs to a family of proteins sharing two distinct structural domains: an NH2-terminal cysteine-rich domain and an adjacent domain of tandem leucine-rich repeats. Hence, it is possible that this family of proteins is not only related in terms of primary structure, but also through similar gene structure. Sequence comparison of the 5' and 3' flanking regions did not reveal striking similarities to other DNA sequences, and no obvious promoter elements were noted. By hybridization of the genomic clone to metaphase cells, we have localized the human OMgp gene to chromosome 17 bands q11-12, a region to which the neurofibromatosis type 1 gene has been previously mapped.

Equilibrium binding of thrombin to recombinant human thrombomodulin: effect of hirudin, fibrinogen, factor Va, and peptide analogues

Thrombomodulin is an endothelial cell surface receptor for thrombin that acts as a physiological anticoagulant. The properties of recombinant human thrombomodulin were studied in COS-7, CHO, CV-1, and K562 cell lines. Thrombomodulin was expressed on the cell surface as shown by the acquisition of thrombin-dependent protein C activation. Like native thrombomodulin, recombinant thrombomodulin contained N-linked oligosaccharides, had Mr approximately 100,000, and was inhibited or immunoprecipitated by anti-thrombomodulin antibodies. Binding studies demonstrated that nonrecombinant thrombomodulin expressed by A549 carcinoma cells and recombinant thrombomodulin expressed by CV-1 and K562 cells had similar Kd's for thrombin of 1.3 nM, 3.3 nM, and 4.7 nM, respectively. The Kd for DIP-thrombin binding to recombinant thrombomodulin on CV-1(18A) cells was identical with that of thrombin. Increasing concentrations of hirudin or fibrinogen progressively inhibited the binding of 125I-DIP-thrombin, while factor Va did not inhibit binding. Three synthetic peptides were tested for ability to inhibit DIP-thrombin binding. Both the hirudin peptide Hir53-64 and the thrombomodulin fifth-EGF-domain peptide Tm426-444 displaced DIP-thrombin from thrombomodulin, but the factor V peptide FacV30-43 which is similar in composition and charge to Hir53-64 showed no binding inhibition. The data exclude the significant formation of a ternary complex consisting of thrombin, thrombomodulin, and hirudin. These studies are consistent with a model in which thrombomodulin, hirudin, and fibrinogen compete for binding to DIP-thrombin at the same site.

Structure of the human sialophorin (CD43) gene. Identification of features atypical of genes encoding integral membrane proteins

A human sialophorin (CD43) specific genomic clone was isolated, and a 6.5 kb fragment containing the 4.6 kb sialophorin gene was sequenced. The promoter region contains no TATA or CAAT boxes, but is highly enriched in G and C nucleotides and contains short repeat sequences similar to those found in the promoters of 'housekeeping' genes. S1-nuclease protection and primer-extension experiments established that the sialophorin gene has two major transcription initiation sites. There is a single intron of 378 bp that interrupts the sequence specifying the mRNA 5' untranslated region. The gene is therefore unusual in that the discrete extracellular, transmembrane and intracellular regions of the protein, including repeat sequences in the extracellular region, are not encoded by separate exons. Utilization of alternative polyadenylation signals was previously shown to generate two sialophorin mRNAs of 1.9 and 4.3 kb, which differ in the length of their 3' untranslated regions. Sequence analysis of the gene establishes that a single polyadenylation signal 2301 bp downstream of the first major transcription initiation site and five overlapping polyadenylation signals beginning a further 2290 bp downstream define the 3' termini of the 1.9 and 4.3 kb mRNA species respectively. The gene contains potential Z-DNA structures, Aly sequences, and elements that may be involved in regulating mRNA stability.

Protein sequence and gene structure for mouse leukosialin (CD43), a T lymphocyte mucin without introns in the coding sequence

A partial cDNA clone for mouse leukosialin was isolated by use of a rat leukosialin cDNA probe. The mouse cDNA was then used to isolate genomic clones that corresponded to the two mouse genes detected in Southern blots. One gene encoded an open reading frame for the homologue of rat leukosialin and this gene was notable for the absence of introns within the coding sequence. A lack of introns has previously been observed for the human leukosialin gene (Shelley, C. S., Remold-O'Donnell, E., Rosen, F. S. and Whitehead, A. S., Biochem. J., submitted). The other mouse gene was an intronless pseudogene for a leukosialin-related sequence. The presence of only one functional gene that lacked coding-region introns established that molecular heterogeneity in mouse leukosialin could not arise from multiple genes or alternative splicing of exons. The sequence of mouse leukosialin suggested an extracellular segment with a high content of O-linked carbohydrate, as is the case in the rat and human. In addition the mouse molecule had one possible N-linked glycosylation site. The cytoplasmic domain of 124 amino acids was highly conserved between rodent and human leukosialins for the functional genes but not for the pseudogene. This suggests an important functional role for the cytoplasmic domain.

Expression of the human inhibin alpha-subunit gene in preovulatory granulosa-theca cells

Inhibin, a gonadal peptide that suppresses pituitary follicle-stimulating hormone, with lesser or no effect on luteinizing hormone, has recently been purified and the complementary deoxyribonucleic acid sequences cloned. Inhibin contains two subunits, labeled alpha-subunit and beta-subunit. Here we report for the first time the detection of human inhibin alpha-subunit gene expression in preovulatory granulosa-theca cells by Northern analysis. The transcript is the same size as previously reported for human placenta and corpus luteum, suggesting that the same gene is being expressed in all three tissues. These findings are consistent with previously reported Southern analysis of deoxyribonucleic acid, which showed only one copy of the alpha-inhibin gene in the human genome. Thus current data strongly suggest that there is only one copy of the inhibin alpha-subunit gene in the human genome, and this same gene is expressed in granulosa-theca cells, corpus luteum, and placenta.

Identification of the core proteins in proteoglycans synthesized by vascular endothelial cells

Proteoglycans, metabolically labelled with [3H]leucine and 35SO4(2-), were isolated from the spent media and from guanidinium chloride extracts of cultured human umbilical-vein endothelial cells by using isopycnic density-gradient centrifugation, gel filtration and ion-exchange h.p.l.c. The major proteoglycan species were subjected to SDS/polyacrylamide-gel electrophoresis before and after enzymic degradation of the polysaccharide chains. The cell extract contained mainly a heparan sulphate proteoglycan that has a buoyant density of 1.31 g/ml and a protein core with apparent molecular mass 300 kDa. The latter was heterogeneous and migrated as one major and one minor band. After reduction, the apparent molecular mass of the major band increased to approx. 350 kDa, indicating the presence of intrachain disulphide bonds. The proteoglycan binds to octyl-Sepharose and its polysaccharide chains are extensively degraded by heparan sulphate lyase. The proteoglycans of the medium contained 90% of all the incorporated 35SO4(2-). Here the predominant heparan sulphate proteoglycan was similar to that of the cell extract, but was more heterogeneous and contained an additional core protein with apparent molecular mass 210 kDa. Furthermore, two different chondroitin sulphate proteoglycans were found: one 200 kDa species with a high buoyant density (approx. 1.45 g/ml) and one 100 kDa species with low buoyant density (approx. 1.3 g/ml). Both these proteoglycans have a core protein of molecular mass approx. 47 kDa.

Tumor necrosis factor suppresses transcription of the thrombomodulin gene in endothelial cells

Tumor necrosis factor (TNF) dramatically alters the levels of various surface components of the blood vessel wall, such as blood coagulation enzyme receptors, leukocyte-adhesive receptors, and class 1 major histocompatibility complex antigens, which may have relevance to its effects in septic shock, angiogenesis, and tumor growth. However, the precise mechanism by which the cytokine is able to accomplish this remodeling of the endothelial cell surface has not been defined. We have demonstrated that exposure of bovine and human endothelial cells to TNF leads to suppression of the functional cell surface thrombin receptor, thrombomodulin (TM), and TM mRNA of virtually identical magnitude. The cytokine has no significant effect on the stability of TM mRNA or endothelial receptor turnover. Nuclear run-on studies reveal that the treatment of endothelial cells with TNF for short periods reduces TM gene transcription to as little as 3% of control values and that this inhibition does not require new protein synthesis.

Stability of the thrombin-thrombomodulin complex on the surface of endothelial cells from human saphenous vein or from the cell line EA.hy 926

Protein C activation by alpha-thrombin on the surface of endothelial cells depends on an essential membrane-glycoprotein cofactor, thrombomodulin. In the present study we have monitored the activity of thrombin-thrombomodulin complexes on human saphenous-vein endothelial cells (HSVEC) or on the endothelial cell line EA.hy 926. Cell monolayers were exposed for 5 min to 8.5 nM human alpha-thrombin and then washed to remove unbound thrombin. The cells were then incubated at 37 degrees C for 5-180 min. At the end of the respective incubation periods, purified human protein C (120 nM) was added in order to assay the activity of the thrombin-thrombomodulin complexes present on the cell surface. HSVEC pre-exposed to thrombin retained their full capacity to promote protein C activation up to 90 min after free thrombin was removed. This capacity then decreased slowly to reach 56% of control value after 180 min of incubation. Original activity was 3.8 +/- 0.9 pmol of activated protein C formed/min per ml per 10(6) cells (mean +/- S.E.M., n = 5). The capacity of protein C activation of EA.hy 926 cells remained constant for 120 min after free thrombin was removed, then decreased to 76% of control after 180 min. Original activity was 2.0 +/- 0.4 pmol of activated protein C formed/min per ml per 10(6) cells (mean +/- S.E.M., n = 3). Similar results were obtained with cells fixed with 3% paraformaldehyde. However, during the 5-180 min incubation period, non-fixed cells of both types were capable of significantly internalizing fluorescent acetylated low-density lipoprotein. In the experimental protocol used here, an eventual inhibition of thrombin internalization by protein C can be excluded, as protein C is only added at the end of the incubation period. We conclude that there is no evidence of rapid internalization of thrombin-thrombomodulin complexes on HSVEC or the EA.hy 926 cell line, as assessed by the ability of membrane-bound thrombin to activate protein C.

Interference of blood-coagulation vitamin K-dependent proteins in the activation of human protein C. Involvement of the 4-carboxyglutamic acid domain in two distinct interactions with the thrombin-thrombomodulin complex and with phospholipids

Human protein C is the precursor of a serine proteinase in plasma which contains nine 4-carboxyglutamic acid residues and functions as a potent anticoagulant. It is activated by thrombin in the presence of an essential endothelial-cell-membrane glycoprotein cofactor, thrombomodulin. In a purified human system, vitamin K-dependent proteins such as factor X, prothrombin and prothrombin fragment 1 were able to inhibit protein C activation by the thrombin-thrombomodulin complex, using either detergent-solubilized thrombomodulin or thrombomodulin reconstituted into vesicles consisting of phosphatidylcholine and phosphatidylserine (1:1, w/w). Factors VII and IX and protein S were much less efficient. Prothrombin fragment 1 behaved as a non-competitive inhibitor with apparent Ki values of 4 microM in the absence, and of 2-2.5 microM in the presence, of phospholipids. Heat decarboxylation of fragment 1 abolished its ability to interfere in protein C activation, and high phospholipid concentrations could attenuate its inhibitory effect and were responsible for a gradual loss of the non-competitive character. Fragment 1 also inhibited the activation of 4-carboxyglutamic acid-domainless protein C, a proteolytic derivative of protein C lacking the 4-carboxyglutamic acid residues, without any influence from phospholipids. At high thrombin concentrations, with respect to thrombomodulin, the inhibitory effect of fragment 1 was diminished. Fragment 1, at 3.8 microM, inhibited by 50% the activation of protein C (0.1 or 0.3 microM) by thrombin. These results suggest that the 4-carboxyglutamic acid domain of vitamin K-dependent proteins can act as a modulator of the protein C anticoagulant pathway through two distinct types of interaction. The functional 4-carboxyglutamic acid domain would be necessary to allow the enhancement of protein C activation in the presence of anionic phospholipids and it could recognize a phospholipid-independent binding site on the thrombin-thrombomodulin complex.

Tumor necrosis factor suppresses transcription of the thrombomodulin gene in endothelial cells

Tumor necrosis factor (TNF) dramatically alters the levels of various surface components of the blood vessel wall, such as blood coagulation enzyme receptors, leukocyte-adhesive receptors, and class 1 major histocompatibility complex antigens, which may have relevance to its effects in septic shock, angiogenesis, and tumor growth. However, the precise mechanism by which the cytokine is able to accomplish this remodeling of the endothelial cell surface has not been defined. We have demonstrated that exposure of bovine and human endothelial cells to TNF leads to suppression of the functional cell surface thrombin receptor, thrombomodulin (TM), and TM mRNA of virtually identical magnitude. The cytokine has no significant effect on the stability of TM mRNA or endothelial receptor turnover. Nuclear run-on studies reveal that the treatment of endothelial cells with TNF for short periods reduces TM gene transcription to as little as 3% of control values and that this inhibition does not require new protein synthesis.

The organization of the human-plasminogen-activator-inhibitor-1 gene. Implications on the evolution of the serine-protease inhibitor family

Plasminogen activator inhibitor 1 (PAI-1) is a member of the serine protease inhibitor super family (SERPINS) which is thought to play an integral role in the control of plasminogen activation. PAI-1 inhibits both tissue-type plasminogen activator and urokinase-type plasminogen activator and may therefore be implicated in the control of various physiological processes. We have isolated the PAI-1 gene including its 5'-flanking sequence. The gene was characterized by restriction enzyme analysis, Southern blotting and DNA sequencing of all the coding parts as well as the 5'-flanking region. The PAI-1 gene contains nine exons and eight introns distributed over approximately 12.3 kb of DNA. All exon/intron boundaries agree with the 'GT-AG' rule. To characterize the presumptive promoter region, 800 bp of the 5'-flanking region was sequenced and potential binding sites for transacting transcriptional factors were localized. The transcription initiation site was identified by S1 protection experiments and is located 25 base pairs downstream of a TATA consensus sequence. By aligning the gene structure of PAI-1 and four other SERPINS and extrapolating a general tertiary structure to these SERPINS, we find that most introns map between subdomain structures of the proteins. Evidence is presented supporting an intron loss model for the evolution of the SERPIN family.

Topogenic signals in integral membrane proteins

Integral membrane proteins are characterized by long apolar segments that cross the lipid bilayer. Polar domains flanking these apolar segments have a more balanced amino acid composition, typical for soluble proteins. We show that the apolar segments from three different kinds of membrane-assembly signals do not differ significantly in amino acid content, but that the inside/outside location of the polar domains correlates strongly with their content of arginyl and lysyl residues, not only for bacterial inner-membrane proteins, but also for eukaryotic.proteins from the endoplasmic reticulum, the plasma membrane, the inner mitochondrial membrane, and the chloroplast thylakoid membrane. A positive-inside rule thus seems to apply universally to all integral membrane proteins, with apolar regions targeting for membrane integration and charged residues providing the topological information.

The amino-terminal domain of thrombomodulin and pancreatic stone protein are homologous with lectins

Amino acid sequence alignment of the amino-terminal of thrombomodulin and pancreatic stone protein with the hepatic asialoglycoprotein receptor shows that these proteins are homologous. From the known disulfide bridge pattern of other proteins belonging to the same family two disulfide bonds can be predicted. The homology raises the question whether the amino-terminal part of thrombomodulin and the pancreatic protein binds carbohydrate or perhaps like tetranectin have a specific affinity for other proteins.