the structure and function of mouse thrombomodulin · training grant t32 hlbi 07088 from the...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No. 30, Issue of October 25, PP. 15815-15S22,1988 Printed in U. S. A.

The Structure and Function of Mouse Thrombomodulin PHORBOL MYRISTATE ACETATE STIMULATES DEGRADATION AND SYNTHESIS OF THROMBOMODULIN WITHOUT AFFECTING mRNA LEVELS IN HEMANGIOMA CELLS*

(Received for publication, May 3, 1988)

William A. Dittman, Toshihiko Kumada, J. Evan Sadlert, and Philip W. Majeruss From the Division of the HematologylOncology, Departments of Internal Medicine and Biological Chemistry, Washington University School of Medicine and the $.Howard Hughes Medical Institute, St. Louis, Missouri 63110

Thrombomodulin is an endothelial membrane anticoagulant protein that is a cofactor for protein C activation. We have evaluated the expression of thrombomodulin in cultured mouse hemangioma cells before and after treatment with phorbol myristate acetate (PMA), an agent that stimulates protein kinase C. We also isolated a cDNA encoding 481 amino acids of mouse thrombomodulin and the entire 3“untranslated portion of its mRNA. The deduced amino acid sequence of mouse thrombomodulin is similar to those determined for human and bovine thrombomodulin. An S1 nuclease protection assay was used to measure thrombomodulin mRNA in hemangioma cells. The half-life for thrombomodulin mRNA was 8.9 f 1.8 h (S.D.) in cells treated with actinomycin D. Treatment with PMA had no effect on thrombomodulin mRNA levels. Thrombomodulin turnover was evaluated by immunoprecipitation of [36S]methionine-labeled thrombomodulin. The ts was 19.8 f 3.9 h (S.D.); PMA treatment decreased the tu to 10.9 f 1.1 h (S.D.) while increasing the rate of synthesis to a maximum of 190% of control. Protein C cofactor activity on hemangioma cells was reduced 35 f 4% by treatment with PMA within 30 min. This decrease was associated with a parallel decline in cell surface thrombomodulin antigen and with enhanced phosphorylation of thrombomodulin on serine residues. We conclude that thrombomodulin is phosphorylated in response to treatment of hemangioma cells with PMA which leads to decreased protein C cofactor activity and both increased degradtion and synthesis of thrombomodulin.

Thrombomodulin is an endothelial cell membrane protein that is a cofactor for thrombin-catalyzed activation of protein C (1, 2). Activated protein C is an anticoagulant that inacti- vates coagulation factors Va and VIIIa (3-5). Thrombomod- ulin on the cell surface also removes thrombin from the circulation by binding it; thrombin bound to thrombomodulin stimulates endocytosis and degradation of the thrombin- thrombomodulin complex in lysosomes (6). Endocytosis is

* This research was supported by Grants HLBI 14147 (Specialized Center for Research in Thrombosis), HLBI 16634 (to P. W. M.), Training Grant T32 HLBI 07088 from the National Institutes of Health, and by the Monsanto-Washington University Biomedical Research Grant. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 504060.

§To whom reprint requests should be addressed.

inhibited by protein C but not by activated protein C (7). Thrombomodulin is an important antithrombotic component of endothelium. We have shown that injection of anti-thrombomodulin antibodies in mice decreases the expression of endothelial thrombomodulin cofactor activity and thereby potentiates the toxicity of injected thrombin (8). Increases or decreases in endothelial cell surface thrombomodulin might alter the occurrence of vascular thrombosis; therefore, we have studied the regulation of thrombomodulin expression in cultured endothelial cells. We have used a mouse hemangioma cell line (9) for this purpose since it is stable, expresses thrombomodulin in large amounts, and retains properties of normal endothelium. We prepared mouse thrombomodulin cDNA in order to characterize thrombomodulin transcription in these cells. We also measured the rates of synthesis and degradation of thrombomodulin and have studied the mechanism by which phorbol myristate acetate (PMA)’ decreases cell surface thrombomodulin.

EXPERIMENTAL PROCEDURES

Materials-Radioisotopes were obtained from Amersham Corp., except [cY-~’P]~CTP which was from ICN Radiochemicals (Irvine, CA) and [32P]orthophosphate which was from Du Pont-New England Nuclear. Restriction endonucleases, calf intestine alkaline phospha- tase, T4 polynucleotide kinase, single strand binding protein, and T7 DNA polymerase were from United States Biochemical Corporation (Cleveland, OH). Maloney murine leukemia virus reverse transcriptase, S1 nuclease, RNase H, and DNA and RNA molecular weight markers were from Bethesda Research Laboratories, EcoRI methylase and T4 DNA ligase were from New England BioLabs (Beverly, MA), oligo(dT)12-18, individual nucleotides, EcoRI linkers, and oligo(dT) cellulose were from Pharmacia LKB Biotechnology Inc. X packaging extracts and Bluescript plasmids were from Stratagene Cloning Sys- tems (La Jolla, CA). Actinomycin D and PMA were from Sigma. Other chemical reagents were from Sigma or Fluka Chemical Cor- poration (Ronkonkoma, NY). Human thrombin (lo), human protein C (4), and polyclonal anti-rat thrombomodulin IgG (8) were isolated as described previously. Modified Medium 199 (“199; M199 with basal medium Eagle’s vitamins and amino acids and Earle’s salts), “199 without phosphate or without methionine, and Dulbecco’s phosphate-buffered saline were from the Washington University Medical School Center for Basic Cancer Research. Tissue culture flasks, wells, and dishes were from Becton Dickinson Labware (Lin- coln Park, NJ) or Corning Glass Works (Corning, NY).

Preparation of Mouse Lung cDNA Library and Isolation of Mouse Thrombomodulin cDNA Clones-RNA was isolated from 10 g of fresh mouse lungs by repeated ethanol precipitation of a guanidine hydro- chloride extract as described by MacDonald et al. (11). Polyadenylated RNA was selected on oligo(dT) cellulose (12). First strand cDNA was synthesized by Maloney murine leukemia virus reverse transcriptase treatment of oligo(dT) primed RNA, and the second strand was produced by Escherichia coli DNA polymerase I synthesis on RNase

The abbreviations used are: PMA, 4@-phorbol l2@-myristate 1301- acetate; “199, modified medium 199; SDS, sodium dodecyl sulfate; EGF, epidermal growth factor.

15815

15816 The Structure and Function of Mouse Thrombomodulin H-treated heteroduplexes (13). After methylation, addition of EcoRI linkers and size selection in low melting temperature agarose (3.0- 4.5 kilobase pairs) double-stranded cDNA was ligated into Xgt.10 (14). The X g t l O library was screened (14) by colony hybridization with random hexadeoxyribonucleotide primed 32P-labeled probes (15) generated from a human thrombomodulin cDNA, HTMlO (16). Positive clones pMTM.lb1 and pMTM.2b2 were subcloned into Bluescript plasmids and were sequenced by the dideoxy chain termination method (17) with [Q-~ 'S]~ATP and the T7 DNA polymerase system (181, using double-stranded plasmid as templates and synthetic oligonucleotide primers (16). Areas of compression on gradient, dena- turing sequencing gels were resolved with a combination of dITP and single strand binding protein, used according to the manufacturer's instructions.

S l Nuclease and RNA Analysis-A 187-base fragment including the 128 5' bases of mouse thrombomodulin cDNA and 59 bases from the plasmid was prepared by TaqI and Sac1 restriction endonuclease digestion of pMTM.2b2. The probe was labeled on the 5' end of the TaqI site with [y-32P]ATP and polynucleotide kinase and purified by electrophoresis for use as an S1 probe as described by Favaloro et al. (19). A probe generated by EcoRI and Hind111 digestion of a plasmid containing sequence for murine Pz-microglobulin was labeled at the EcoRI site as described above and was used as control for total amounts of RNA (20, 21). RNA was harvested from cells grown to confluence in 100-mm plates by the method of Chomczynski and Sacchi (23) and measured by absorbance at 260 nm. The mouse thrombomodulin probe was hybridized with RNA samples overnight at 55 "C, whereas the pp-microglobulin probe was hybridized at 50 "C. Fifty or 100 units of S1 nuclease were incubated with the hybrids for 30 min at 37 "C, and the reaction products were analyzed on a gel containing 6% acrylamide and 8.3 M urea. After autoradiography, radioactive bands were cut from the gel, and radioactivity was measured by scintillation counting. Northern blot analysis was performed as described by Lehrach et al. (22). Total RNA was electrophoresed in 1% agarose in the presence of 2.2 M formaldehyde, transferred to nitrocellulose, hybridized with 32P-labeled probes prepared from cDNAs as described above, and washed with 300 mM sodium chloride, 30 mM sodium citrate (pH 7.0), 0.1% sodium dodecyl sulfate (SDS) at 25 "C for 30 min, and then in 15 mM sodium chloride, 1.5 mM sodium citrate, 0.5% SDS at 50 "C for 30 min, and autoradiographed.

Metabolic Labeling-The rate of degradation of thrombomodulin was determined as follows. Hemangioma cells in 35-mm dishes (approximately 4.5 X lo5 cells/dish) were incubated for 48 h in methionine-free "199, 20% fetal calf serum, with [35S]methionine at 400 pCi/ld ml/dish. Cells were washed twice with Dulbecco's phosphate- buffered saline and incubated for the indicated times with complete medium containing PMA (200 ng/ml) or dimethyl sulfoxide (0.01%). PMA was dissolved in dimethyl sulfoxide at a concentration of 2 mg/ ml and stored at -20 "C. Maximal effects on synthesis and degradation of thrombomodulin were found at PMA concentrations of 100 ng/ml.

Thrombomodulin synthesis was measured after cells were incubated for 0-12 h with complete medium containing PMA (200 ng/ ml) or dimethyl sulfoxide (0.01%), washed, and then labeled for 4 h in methionine-free "199 containing [35S]methionine at 200 pCi/ 1.5 ml/dish and PMA or dimethyl sulfoxide.

Thrombomodulin was labeled with [32P]orthophosphate at 2 mCi/ 2 ml/dish in 60-mm dishes (approximately 1.4 X lo6 cells/dish) for 2 or 18 h in phosphate-free "199 supplemented with 10% fetal calf serum and 10% dialyzed fetal calf serum. Labeled cells were treated with PMA (200 ng/ml) or dimethyl sulfoxide (0.01%) for 10 min at 37 "C and then the medium was removed. The cell layers were washed with cold 10 mM Tris-HC1 (pH 7.2) 150 mM sodium chloride, 1 mM EDTA, and then solubilized in cold 10 mM sodium phosphate (pH 7.2) 150 mM sodium chloride, 1% Triton X-100, 1% sodium deoxy- cholate, 0.1% SDS, 2 mM EDTA, 23 pg/ml aprotinin (solubilization buffer) with 2.9 mM diisopropylfluorophosphate added. After centrif- ugation at 30,000 X g for 30 min at 4 "C, the supernatant fraction of the cell extracts and the medium were used for immunoprecipitation. Trichloroacetic acid-precipitable radioactivity was measured in por- tions of the samples.

Thrombomodulin Immunoprecipitation-Immunoprecipitation was performed by a modification of the method of Kessler (24). 35S- Labeled cell extracts and medium were treated twice with 150 pl of Pansorbin (10% in solubilization buffer containing 1 mg ovalbumin/ ml (Calbiochem, Behring Diagnostics, La Jolla, CA)) followed by treatment with 40 pg of pre-immune IgG and 150 p1 of Pansorbin. The supernatant fractions were then incubated at 4 "C for 16 h with

40 pg of anti-rat thrombomodulin IgG. The immune complexes were harvested with 30 pl of Pansorbin or 40 pl of protein A-Sepharose (25%, Sigma) and sequentially washed five times with solubilization buffer, two times with 20 mM Tris-HC1 (pH 7.4) 150 mM sodium chloride, 2 M urea, 100 mM glycine, and two times each with 50 mM Tris-HC1 (pH 8.0) 500 mM sodium chloride, and with water. The proteins were dissolved in 40 p1 of Laemmli's gel sample buffer (25) by incubation for 5 min in boiling water and electrophoresed through a 7.5% polyacrylamide SDS gel (25). Measurement of %-labeled thrombomodulin was performed by fluorography (26) with EN3HANCE solution (Du Pont-New England Nuclear) and by den- sitometry of autoradiograms using a soft laser scanning densitometer (Zeineh, Biomed Instruments, Chicago, IL). Immunoprecipitation of 32P-labeled thrombomodulin was performed as above except that all the solubilization and washing buffers contained 200 p~ sodium vanadate and 10 mM sodium fluoride to inhibit dephosphorylation reactions. Liquid scintillation counting was performed in a Beckman LS 6800 using ScintiVerse I (Fisher).

Assay of Thrombomodulin Cofactor Activity-Assay of thrombomodulin cofactor activity was performed on monolayers of hemangioma cells (1-2 X lo4 cells/well) in 96-well tissue culture plates. The cells were washed three times with "199 and then treated with 40 pl of "199 containing PMA (100 ng/ml) or dimethyl sulfoxide (0.005%) for the indicated times at 37 "C. In some experiments, the medium was removed after a 30 min incubation and complete medium or "199 was added back. After incubation, the treated cells were washed three times with 150 mM NaCl, 20 mM Tris-HC1 (pH 7.4) 1 mM CaClZ, 5 Tg/ml bovine serum albumin, and then incubated in 40 pl of the same buffer with 4 units/ml thrombin, and 0.5 PM protein C for 15 or 30 min at 37 "C. Activated protein C generated in 30 pl of the reaction mixture was determined by measuring the cleavage of the chromogenic substrate S-2238 (Kabi Vitrum, Stockholm, Sweden) in the presence of antithrombin I11 and hirudin as described previously (27). Control inct bations in the absence of cells were performed. Cell number was measured in an automated cell counter (Coulter Electronics, Hialeah, FL).

Binding of Anti-rat Thrombomodulin IgG to Hemangioma Cells- Affinity purified polyclonal anti-rat thrombomodulin IgG (8) was isolated on a human thrombomodulin Affigel-15 column as described previously for affinity-purified anti-human thrombomodulin IgG (16) and radiolabeled with carrier-free NalZ6I using Bolton-Hunter reagent (28) to a specific activity of 20-4500 cpm/ng protein. Labeled antibody was used at 100 nM to measure hemangioma cell surface thrombomodulin as described previously (6, 7). Nonspecific binding was determined in the presence of 10 p~ unlabeled antibody. Protein concentrations were determined by a dye binding assay with bovine serum albumin as the standard (Bio-Rad).

Phosphoamino Acid Analysis-Phosphoamino acid analysis was performed on thrombomodulin isolated from [32P]orthophosphate- labeled cells by immunoprecipitation followed by polyacrylmide gel electrophoresis. After autoradiography, the excised gel slices corre- sponding to the 32P-labeled thrombomodulin were hydrated in 2 ml of 50 mM ammonium bicarbonate, 100 pg/ml L-1-tosylamido-2-phen- ylethylchloromethyl ketone trypsin (Cooper Biomedical, Malvern, PA), 1.5 mM dithiothreitol. After incubation for 24 h at 37 "C, the eluted tryptic peptides were dried in a Speed-Vac (Savant Instru- ments, Hicksville, NY) and hydrolyzed in 6 N hydrochloric acid (Pierce Chemical Co.) for 1 h at 110 "C. Samples were dried and dissolved in 10 ml of water, and phosphoamino acids were isolated on Dowex AG 1-X8 as described by Cooper et al. (29). The eluted samples were electrophoresed for 1.5 h at 700 V at 4 "C on a 20 X 20- cm thin layer cellulose plate (Eastman Kodak) at pH 3.5 in pyri- dine:acetic acidwater (1:10189). Phosphoamino acid standards were localized by ninhydrin, and radioactivity was detected by autoradiography with EN3HANCE spray and intensifying screens (Du Pont Cronex Lightning Plus, E.I. du Pont de Nemours & Co.) and X-Omat film (Eastman Kodak).

RESULTS

Mouse lung was chosen as the source for RNA isolation because of its high thrombomodulin levels. A size-selected cDNA library was prepared in XgtlO and screened with a human thrombomodulin cDNA probe. Two clones were isolated from 2 X IO4 recombinant clones screened. The longer of these, pMTM.2b2, was sequenced on both strands and contains 3212 nucleotides (Fig. 1). An open reading frame of

The Structure and Function of Mouse Thrombomodulin 15817

1443 nucleotides, encoding 481 amino acids, is followed by a thrombomodulin and 86% of the nucleotide sequence includ- 3”untranslated region of 1745 nucleotides and a poly(A) tail. ing all of the 3”untranslated region by comparison with the Apotential polyadenylation signal (AATAAA) begins 23 bases human cDNA. The mouse and human proteins are similar 5’ to the poly(A) tail. The cDNA encodes 83% of mouse with 68% identity of amino acids. The mouse cDNA is also

1

1 0 1

2 0 1

3 0 1

4 0 1

5 0 1

6 0 1

7 0 1

8 0 1

9 0 1

1 0 0 1

1 1 0 1

1 2 0 1

1 3 0 1

1 4 0 1

1 5 0 1

1 6 0 1

1 7 0 1

1 8 0 1

1 9 0 1

2 0 0 1

2 1 0 1

2 2 0 1

2 3 0 1

2 4 0 1

2 5 0 1

2 6 0 1

2 7 0 1

2 8 0 1

2 9 0 1

3 0 0 1

3 1 0 1

3 2 0 1

CCGGTGCATCTCGGGCCCCTGCGCGGCTTCCAGTGGGTTACTGGCGATAACCACACCAGTTACAGCAGGTGGGCGCGGCCCAACGACCAGACGGCTCCAC P V H L G P L R G F Q W V T G D N H T S Y S R W A R P N D Q T A P L

TCTGCGGCCCTCTGTGCGTCACGGTCTCGACAGCAACTGAAGCTGCACCCGGCGAGCCGGCCTGGGAAGAGAAGCCATGCGAGACTGAGACCCAGGGTTT C G P L C V T V S T A T E A A P G E P A W E E K P C E T E T Q G F

CCTCTGTGAGTTTTACTTCACAGCTTCCTGCAGGCCTCTGACGGTGAATACTCGCGATCCTGAGGCTGCCCACATCTCTAGTACCTACAACACCCCGTTC L C E F Y F T A S C R P L T V N T R D P E A A H I S S T Y N T P F

GGGGTCAGTGGTGCGGACTTTCAAACGCTGCCGGTAGGCAGTTCCGCCGCGGTGGAGCCCCTTGGCTTGGAGCTGGTGTGCAGGGCCCCGCCCGGAACTT G V S G A D F Q T L P V G S S A A V E P L G L E L V C R A P P G T S

CAGAGGGACACTGGGCTTGGGAAGCGACAGGAGCCTGGAATTGCAGCGTGGAGATGGTGGCTGTGAGTACTTGTGCAATAGGAGCACGAATGAACCCAG E G H W A W E A T G A W N C S V E N G G C E Y L C N R S T N E P R

ATGCCTCTGCCCCAGAGACATGGACCTGCAGGCCGATGGACGTTCGTGTGCAAGACCTGTGGTTCAATCGTGCAACGAACTCTGCGAGCATTTTTGTGTC C L C P R D M D L Q A D G R S C A R P V V Q S C N E L C E H F C V

AGCAACGCTGAAGTGCCAGGCTCTTACTCCTGTATGTGTGAGACAGGCTACCAGTTGGCTGCAGACGGACACCGGTGTGAGGACGTGGATGACTGTAAGC S N A E V P G S Y S C M C E T G Y Q L A A D G H R C E D V D D C K Q

AGGGGCCCAATCCATGTCCCCAGCTCTGTGTTAACACCAAGGGCGGCTTCGAATGCTTCTGCTATGATGGCTATGAGTTGGTGGATGGAGAGTGCGTGGA G P N . P C P Q L C V N T K G G F E C F C Y D G Y E L V D G E C V E

GCTTCTGGATCCGTGTTTCGGATCTAACTGCGAGTTTCAGTGCCAGCCAGTGAGCCCCACCGACTACCGATGCATCTGCGCTCCAGGCTTCGCACCCAAG L L D P C F G S N C E F Q C Q P V S P T D Y R C I C A P G F A P K

CCGGATGAACCGCACAAGTGCGAAATGTTCTGCAATGACTTCGTGCCCAGCAGACTGTGACCCTAACTCTCCTACTGTTTGTGAATGCCCTGAAGGCT P D E P H K C E M F C N E T S C P A D C D P N S P T V C E C P E G F

TCATCCTGGACGAGGGTTCCGTATGCACGGACATTGATGAGTGCAGTCAAGGCGAATGCTTCACCAGTGAATGTCGAAACTTCCCTGGCTCCTATGAGTG I L D E G S V C T D I D E C S Q G E C F T S E C R N F P G S Y E C

TATCTGCGGGCCTGACACAGCCCTTGCTGGTCAGATTAGCAAAGACTGTGACCCCATCCCTGTTAGGGAAGACACCAAGGAAGAGGAGGGCTCTGGGGAG I C G P D T A L A G Q I S K D C D P I P V R E D T K E E E G S G E

CCTCCAGTCAGCCCTACGCCAGGCTCTCCGACAGGTCCCCCTTCTGCAAGGCCAGTGCACTCTGGCGTGCTCATTGGCATTTCCATTGCCAGCCTGTCCC P P V S P T P G S P T G P P S A R P V H S G V L I G I S I A S L S L

TGGTGGTGGCGCTTTTGGCGCTTCTCTGTCACCTGCGCAAGAAGCAGGGCGCTGCTCGTGCAGAGCTGGAGTACAAGTGCGCATCTTCCGCCAAGGAGGT V V A L L A L L C H L R K K Q G A A R A E L E Y K C A S S A K E V *

AGTGCTGCAGCACGTCAGGACTGATCGGACGCTGCAGAAGTTCTGAGGGATTTGCTCCAGAGACCCAGGTGGCCTTTGTCTTTCCGGGCTCTGTACCTCT . * * .

V L Q H V R T D R T L Q K F

CCTCTCCTCTCTCTCTCTCCAGCCTCCCAGCTGTGTTCTCTGGCAACTT~GCACCCTGGCTGGTATAATAACCAGAGAAGAGCCCATCCTCTCAGGA

CAAGCGAGGGTAGGGAGGACTTGAAGCAGGACAGCCCAGTTTCTTCCAAGTAGATACTGGACAACTGGGCGGAGGTGGCAAATACAGCGGAGATCCCAGA

GTACCCCAGTCCCTCACCTCACCTCCTAGTGCTGCTGATCTGTAGGCTTGAAGGCARACCTTGACCCCATGGGCTGGAGATGACCCAGATATTTATTTTT

TTTAAAGTATTTAGTATTTTCTTCCCTCCAGTTTTCTTCTGCTTGTAAGTCTCCAGCCCCCCACAGCTTTCTCAGTCCCTCCATTCCCCCCCTCCTGTCA

TTCTCCTCCCCAAACCTGATCATAACTTTGCCCTTACCGTTGTTTCCARACTCTTATGTGAAACAGARAAGACACTARAAGCAG~CGTTCTTTTT

CACTGGCTTTGGGTATTTAGTCAGAAATTTCAGGTAACCAAAGCAAAAGAATTTTAACAAAAGCTAAAATATTTCAGCTGAACACTMCTAGTCAATAGT

GCTGGMTGTCACAGAAAATAAACTTAAGGAAGTAGGGTTTTTTTTTTTTTTGAAATCTTTGTTTTTGAAAGGGTGAGCCTGGGTTTTATGATTGTTGCT

GTTGTTTTGAATGGGAATGACAAAAGAGGTCATTATTGTTAAGATTTTTATGCAGGCTCTACAGTGTTATTAATTTTTGACAGTGTTC~TGTGCAGA

GGATCCTTTGTCCAACCCTTTGACATGACAATAGGACATTGCTATCTTGAGACATACTGGGCCACATTCATAGCTTTCCAAGGATGTATGTGGTCCTGCC

*.

TCAACATATCAGAGCCTGACAGATGGAAGCACCTTCCAGTAAAGCATGAGTTGTGTGCTTCGTGCCGAGCTGACTCTCAACTGTGCCTGCCCCTTGTAGT

CCCGAAATACAAGCAATGTGCTGCTGAGGGAAACATGGAAACTTGGGAATGGAGTCTGGGGGTGCCTAGATGGGGCTTTCTTTTMTGAGACTCTTGAAC

AATATCTCGTAATTCAGAGGGATCTTCTAGCCCTGGCCACTGGCCTGTACACAAGAATTGGGACCTCGCTTGGGATCTGGCTAGMTTGCAAAATCCTAG

CCCCCACCCCTGCCCCACCCCAGTGTGCCAGTTCATMGAATCTGCATTTTGACAACATCCACAGGGACATTGTCCAGTCATTTCAGGACAACTGGTCTT

AAGAGTTTCCAACCTTTGTAGAACATTTAAATGTCGGTTAATAATAAGTAGCAGGCCATGTTAAGGCCATTTATTATCAAGAAACTGAGGAATTTTCTCT

GCATAGCTTTGCTTTCTGGATACAATAAAATGAGAAGGTACACACCTCTAGATAGTGCCACACAGAGTCCAGAAGGGTTTTGTTTTAAGTAAGCTAGGAA

TGAGTTCATATGTTAGTGTAAGGAACAAATGTATTATATGTGTATCTTTTGTAAAGAAAGGTTTTTCTTTACGGTTTTGTAAGCTCAGCATATTTGTACA

TATTTATTTATTGGAGTTTCGCTAGAACACACAAGCAAAGCCTTTGCTTATGACGTCACATGTACAAAATAAATAGATGACAGTGTACTG~

AAAAAAAAAAAA 3 2 1 2

FIG. 1. Nucleotide sequence of pMTM.2b2 and predicted amino acid sequence of mouse thrombomodulin. The EGF-like repeats similar to those found in human and bovine thrombomodulin are underlined. The TTATTTAT motif described in the text is boldly underlined. Potential sites of cytoplasmic phosphorylation are starred.

1 0 0

2 0 0

3 0 0

400

5 0 0

600

7 0 0

8 0 0

9 0 0

000

1 0 0

2 0 0

3 0 0

1 4 0 0

1 5 0 0

1 6 0 0

1 7 0 0

1 8 0 0

1 9 0 0

2 0 0 0

2 1 0 0

2 2 0 0

2 3 0 0

2 4 0 0

2 5 0 0

2 6 0 0

2 7 0 0

2 8 0 0

2 9 0 0

3 0 0 0

3 1 0 0

3 2 0 0

15818 The Structure and Function of Mouse Thrombomodulin

similar to a partial bovine cDNA (30) with 67% amino acid identity. The 3”untranslated region of the cDNAs, especially the extreme 3‘ ends, are also highly conserved. The entire 3’- untranslated region is very AT-rich, and a specific AT-rich motif (TTATTTAT) (31) observed in clones for human and bovine thrombomodulin is present in the mouse. We found a single mRNA of 3800 bases by Northern blot analysis of mouse lung RNA.

We investigated the turnover of thrombomodulin mRNA by inhibiting transcription with actinomycin D and determining subsequent changes in mRNA levels by S1 nuclease analysis. The tH of thrombomodulin mRNA determined from five experiments was 8.9 f 1.8 h (S.D.) (Fig. 2). By 8 h, levels were 55 f 20% of control, and by 24 h levels were 22 k 9% of control. fi2-microglobulin mRNA was also measured as control for recovery of total mRNA in the reaction mixtures. Fig. 3 is a Northern blot analysis of mRNA before actinomycin D treatment and after 3 h demonstrating relative stability of the message for thrombomodulin. The same blot was hybridized with a c-myc probe (32) showing rapid disappearance of c- myc mRNA after actinomycin D treatment.

Treatment of hemangioma cells with PMA had no effect on the levels of thrombomodulin mRNA determined as a ratio of thrombomodulin mRNA to fi2-microglobulin mRNA. f i2-

microglobulin mRNA levels were also unchanged. One of four similar experiments with PMA treatment is shown in Fig. 4. There was no difference in the results whether or not serum was present during exposure to PMA.

Thrombomodulin was labeled with [35S]methionine in pulse-chase experiments followed by immunoprecipitation of protein as demonstrated in Fig. 5A. The protein was degraded with a tlA of 19.8 f 3.9 h (S.D.) in untreated cells. Treatment with PMA resulted in a decrease in tlh to 10.9 f 1.1 h (mean of four experiments as shown in Fig. 5B). The rate of synthesis of thrombomodulin was also measured as shown in Fig. 6. Synthesis increased to 1.35 times control when measured over

:h E S l 0’ E

I YWY TWROUBOUOWLlN 5, P lOBL I

0 1 0 20 30 TIME (hr)

FIG. 2. S1 nuclease protection analysis of mRNA levels. Hemangioma cells were grown to confluence in P-100 tissue culture dishes. The medium was changed to M199, containing 20% fetal bovine serum and 10 pg actinomycin D/ml. At the times indicated, total RNA was extracted, and 10 pg was analyzed as described under “Experimental Procedures.” Upper right insert shows control &- microglobulin (j32M) protected fragments (upper panel) and the 128- nucleotide-protected mouse thrombomodulin (MTm) fragment (lower panel). This experiment is representative of five individual experiments. The mouse thrombomodulin probe and expected S1 nuclease- resistant product are shown in the lower left.

Probe MTm cMyc Time(Hr) 0 3 0 3

c

FIG. 3. Northern blot analysis of hemangioma mRNA. Total cellular RNA was isolated as described in Fig. 2, and 10 pg was electrophoresed through 2.2 M formaldehyde, 1% agarose, and transferred to nitrocellulose. On the right, RNA isolated before and 3 h after exposure to 10 pg actinomycin D/ml was hybridized with a probe to c-myc. The membrane was washed and reprobed with labeled mouse thrombomodulin (MTm) cDNA shown on the [eft. The migra- tion of 28 S and 18 S ribosomal RNA subunits is indicated on the right.

Hours 0 I 4 8 PMA + + + +

I

FIG. 4. S1 nuclease analysis of PMA-treated and control hemangioma cells. Cells were grown as described in Fig. 2. At the indicated times, medium was changed to M199 containing 20% fetal bovine serum and 100 ng PMA/ml, in 0.005% dimethyl sulfoxide or dimethyl sulfoxide alone. Cells were harvested, RNA isolated, and 10 pg of total RNA was analyzed by S1 nuclease protection as in Fig. 2. Control &microglobulin (82M) levels are in the upper panel (205 bases), mouse thrombomodulin (MTm) in the lower (128 bases). This experiment is representative of four experiments.

4 h in the presence of PMA. Prior treatment with PMA increased synthesis to 1.9 times control at 2 h and 1.8 times control a t 6 h. Since PMA increases both the synthesis and degradation of thrombomodulin, there should be little net effect on cellular thrombomodulin levels.

Thrombomodulin activity was assayed on intact cells by measuring thrombin-stimulated rates of protein C activation after exposure to PMA as shown in Fig. 7. There was a rapid drop in cofactor activity reaching 65.1 f 3.7% of control levels within 30 min. In another experiment at times between 5 and 120 min (not shown), surface activity reaches a nadir a t 30 min. Thrombomodulin activity reappeared after several hours. The same pattern of thrombomodulin cofactor activity decline and return to control values was observed in separate experiments where PMA was continuously present for 12 h (data not shown).

Since the turnover of thrombomodulin is slow relative to the rapid changes in cofactor activity, we evaluated the pos- sibility that PMA induced endocytosis of thrombomodulin with loss of cell surface thrombomodulin. We measured cell

The St ruc ture and Function of Mouse Thrombomodulin 15819

A TIME(h1 0 6 6 12 12 24 24

PMA - + - + - + -

MTm -.

FIG. 5. Disappearance of “S-labeled thrombomodulin after incubation of hemangioma cells with PMA. Monolayers of hemangioma cells were radiolabeled for 48 h with [3sS]methionine and then chased for the indicated time in complete medium without (X) or with PMA (A, 4, 0, and V) or dimethyl sulfoxide (O,O, A, and V). Radiolabeled thrombomodulin was isolated from cell extracts by immunoprecipitation followed by SDS, 7.5% polyacrylamide gel electrophoresis in 2-mercaptoethanol and fluorography ( A ) . Densitimetry was performed on the autofluorograms. Data derived from each of the four experiments is plotted in B with each experiment represented by a different symbol. MTm, mouse thrombomodulin.

surface thrombomodulin by determining the binding of Iz5I- anti-thrombomodulin IgG to hemangioma cells. These meas- urements proved difficult since 75% of the binding of this antibody to hemangioma cells was “nonspecific,” i.e. not dis- placed by unlabeled IgG. However, PMA did reduce specific binding of antibody to hemangioma cells from 0.10 pg/106 cells (total IgG bound 0.40 pg/106 cells, nonspecific IgG bound 0.30 pg/106 cells) to 0.06 pg/106 cells (total IgG bound 0.28 pg/106 cells, nonspecific IgG bound 0.22 pg/106 cells). In two other experiments, PMA reduced binding similarly.

The PMA-induced decrease in surface antigen and protein C cofactor activity was associated with increased phosphorylation of thrombomodulin as shown in Fig. 8. PMA-treated cells showed a 38 f 14% increase in 32P incorporated into thrombomodulin. Phosphoamino acid analysis of the labeled thrombomodulin showed increased incorporation into serine as shown in Fig. 9.

DISCUSSION

Analysis of the cDNAs for murine, bovine, and human thrombomodulin shows similarity between the three species

TIME (hr 1

FIG. 6. Rate of thrombomodulin synthesis in hemangioma cells treated with PMA with various times. Monolayers of hemangioma cells were pretreated for 0-12 h with PMA (0) or dimethyl sulfoxide (O), washed, and then radiolabeled with [35S] methionine for an additional 4 h. Radiolabeled thrombomodulin content in cell extracts was determined as described in Fig. 5, and the change in radioactivity was expressed as the percentage of the pretreatment control k S.D. of two to six determinations, except for those of dimethyl sulfoxide-treated values a t 6 and 12 h which represent individual dishes.

4ot

TIME (hr) FIG. 7. Protein C activation on hemangioma cells after ad-

dition of PMA. Monolayers of hemangioma cells were treated with serum-free “199 containing PMA (0) or dimethyl sulfoxide (0) for 30 min. The medium was changed and at the indicated times, thrombomodulin cofactor activity was assayed for 30 min. Throm- bomodulin cofactor activity is expressed as the percentage of the pretreatment control &S.D. of three to five determinations from two experiments. Control values were 7.2 f 1.4 pmol of protein Ca formed/ ml/104 cells/30 min (eight experiments).

and certain highly conserved regions which may be important for thrombomodulin function (Fig. 10). Mouse thrombomodulin, like human and bovine, contains a putative amino- terminal ligand-binding domain, a region of epidermal growth factor (EGF)-like repeats, a serine/threonine-rich region, a membrane-spanning domain, and a cytoplasmic tail. EGF- like repeats may preserve the ability of intracellular receptors to effectively release ligand, avoid degradation, and recycle to the cell surface as has been suggested for the low density lipoprotein receptor (33,34), where alteration of these repeats by mutagenesis interfered with low density lipoprotein receptor recycling (35). EGF domains also have been reported to contain sequences that direct /3-hydroxylation of asparagine or aspartate residues (36). The consensus sequence described by Stenflo et al. (36) (CX(N/”)XXXX(,/~)XCXC) is found in EFG repeats 3 and 6. Although not demonstrated for murine thrombomodulin, acid hydrolysates of bovine thrombomodulin have been shown to contain 8-hydroxyaspartic acid (37). Human (16), bovine (30), and murine thrombomodulin all have the consensus sequence in EGF repeats 3 and 6 implying that they contain the amino acid modification. Mouse throm-

15820

PMA The Structure and Function of Mouse Thrombomodulin

- +

MTm -

FIG. 8. Autoradiogram of 32P-labeled thrombomodulin immunoprecipitated from hemangioma cells. Monolayers of hemangioma cells were radiolabeled for 18 h with [32P]orthophosphate and then treated with PMA or dimethyl sulfoxide for 10 min. 32P- Thrombomodulin was immunoprecipitated from cell extracts and analyzed by SDS/7.5% polyacrylamide gel electrophoresis in 2-mercaptoethanol and autoradiography. The proteins that migrate below thrombomodulin are contaminants which are also observed in control precipitates with protein A prior to the addition of antibodies. MTm, mouse thrombomodulin.

PMA - +

-Inorganic Phosphate

c Phosphoserine - Phosphothreonine

-Phosphotyrosine

FIG. 9. Phosphoamino acids of thrombomodulin immunoprecipitated from “P-labeled hemangioma cells. Monolayers of hemangioma cells labeled with [32P]orthophosphate for 18 h were treated with PMA or dimethyl sulfoxide. The experiment was performed as described in the legend to Fig. 8. The phosphorylated thrombomodulin hydrolyzed and phosphoamino acids were separated by high voltage electrophoresis as described under “Experimental Procedures.” The radioactive phosphoamino acids comigrated pre- cisely with authentic phosphoserine internal standards, although the standards migrated slightly differently in the two lanes.

bomodulin has a 39-amino acid serine-threonine-rich domain compared to 34 for human and 42 for bovine, with 7 hydroxy amino acids in mouse and bovine thrombomodulin compared to 8 for human; presumably this is a region for 0-linked glycosylation of thrombomodulin by analogy to the low density lipoprotein receptor (38). The most highly conserved

region of thrombomodulin is the putative membrane-spanning domain with 20 of 23 amino acids identical in all three species; the nonidentical amino acids are conservative substi- tutions. The first 8 residues of the cytoplasmic tail of thrombomodulin and the potential sites for threonine and tyrosine phosphorylation are identical in all three species; serines are present but in different locations.

In the regions where only mouse and human thrombomodulin can be compared, two regions have near identity; 18 of 20 amino acids are identical starting at human amino acid 207, and 22 of 23 starting a t human amino acid 102. The similarity suggests an essential function for these regions.

The conservation of the AT-rich motif in the 3”untranslated region of mRNA from all three species suggests a role for these sequences. Indeed, this motif is associated with short mRNA half-life for many cytokines (39, 40) and destabiliza- tion of hybrid mRNA with the 5’ sequence of globin and 3‘- untranslated region with this motif has been demonstrated (39). We found, however, that despite the presence of the conserved AT motifs the half-life of thrombomodulin is 8.9 h (Figs. 2 and 3), too long to allow for rapid decreases of thrombomodulin activity due to changes in transcription.

PMA had no effect on thrombomodulin mRNA levels (Fig. 4). The genomic sequence for human thrombomodulin (41, 42) does not contain a 9-base pair 5’ motif common to PMA- inducible genes (43), consistent with the observed lack of increase of thrombomodulin mRNA levels in response to PMA.

PMA appears to stimulate endocytosis of thrombomodulin and also increases receptor degradation as indicated by the shortened survival of thrombomodulin (Fig. 5). The increase in thrombomodulin synthesis that follows treatment with PMA may compensate for the increased degradation (Fig. 6) and thereby maintain cellular thrombomodulin levels. The changes in protein degradation in response to PMA cannot explain the observed rapid decrease in protein C activation (Fig. 7). The half-life of the protein remains too long to allow a rapid alteration in activity. Thrombomodulin is subject to endocytosis that is blocked by protein C (6,7) but stimulated in response to PMA as indicated by decreased binding of affinity-purified antibodies directed against thrombomodulin. Endocytosis of other receptors, including the transferrin receptor (44), the tumor necrosis factor receptor (45), and the epidermal growth factor receptor (46, 47), increases in response to PMA. In addition, the transferrin receptor is phosphorylated on serine 24 of the cytoplasmic tail in response to PMA (48,49), although studies involving mutagenesis of this serine suggest that the phosphorylation is not essential for endocytosis in some cell lines (50-52). Whether this is true also for thrombomodulin will require alteration of the cytoplasmic phosphate acceptor. We have been unable to phos- phorylate either human or murine thrombomodulin in uitro with protein kinase C, and therefore, the phosphorylation of thrombomodulin after PMA is probably not a direct effect of protein kinase C.

The relatively slow turnover of thrombomodulin mRNA and protein implies that thrombomodulin levels in mouse hemangioma cells will not fall quickly in response to external stimuli. Presumably, thrombomodulin functions only under conditions where thrombin is generated, and there is no obvious reason to propose a mechanism to decrease thrombomodulin levels quickly. Although thrombomodulin protein levels do not fall quickly, protein C-activating cofactor activity does rapidly decline in response to PMA. Cofactor activity has been shown to fall on the surface of endothelial cells exposed to tumor necrosis factor (53), interleukin 1 (54), and

The Structure and Function of Mouse Thrombomodulin 15821

1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... MLGVLVLGAL ALAGLGFPAP AEPQPGGSQC VEHDCFALYP GPATFLNASQ ICDGLRGHLM TVRSSVAADV ISLLLNGDGG VGRFUUWIGL QLPPGCG

l . . . . . . .................... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . MTm

BTm HTm

MTm HTm BTm

MTm HTm BTm

MTm HTm BTm

MTm HTm BTm

MTm HTm BTm

101 200 PaT;ai?J; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 0 1

NAI PGAP .......................

E G F R l l

AE 30

RCEDVDD KQ PD RCEDVDD I L NQTSY LH RCEDVDD AQ E

0 D H G

EGFRXP EGFRX3 EGFRX4 -

EGFRX5 EE RGTPEDY

EGFRXC

TRANSMEMBRANE

FIG. 10. Alignment of mouse (MTm), human (HTm), and bovine (BTm) thrombomodulin amino acid sequences. Identical amino acids are enclosed in boxes. The epidermal growth factor-like repeats (EGFR) are underlined as is the transmembrane domain.

endotoxin (55) . The mechanism of decline in response to these cytokines has not been determined, but the long survival 17, for thrombomodulin implies that these cytokines, like PMA, 18, may induce the endocytosis of thrombomodulin.

Acknowledgments-We thank Dr. John C. Hoak for providing 19.

hemangioma cells, Drs. Peter Lobe1 and Tim Ley for advice on cloning 20, and S1 analysis, Victoria Masakowski for providing &microglobulin and c-myc probes, Lisa Westfield for synthetic oligonucleotide 21, primers, Roger Inhorn and Dr. Tom Connolly for helpful discussions, 22, and Lois Isenberg for typing this manuscript.

1.

2. 3.

4.

5.

6.

7. 8.

9.

10.

11.

12.

13.

14.

15.

16.

REFERENCES Esmon, C. T., and Owen, W. G. (1981) Proc. Natl. Acad. Sci.

Esmon, C. T. (1987) Science 235, 1348-1352 Kisiel, W., Canfield, W. M., Ericsson, E. H., and Davie, E. W.

Suzuki, K., Stenflo, J., Dahlback, B., and Teodorsson, B. (1983)

Fulcher, C. A., Gardiner, J. E., Griffin, J. H., and Zimmerman,

Maruyama, I., and Majerus, P. W. (1985) J. Biol. Chem. 260,

Maruyama, I., and Majerus, P. W. (1987) Blood 69, 1481-1484 Kumada, T., Dittman, W. A., and Majerus, P. W. (1988) Blood

Fry, G. L., Czervionke, R. L., Hoak, J. C., Smith, J. B., and

Miletich, J. P., Broze, G. J., and Majerus, P. W. (1980) Anal.

MacDonald, R. J., Swift, G. H., Przybyla, A. E., and Chirgwin, J.

Aviv, H., and Leder, P. (1972) Proc. Natl. Acad. Sci. U. S. A. 69,

D’Alessio, J. M., Noon, M. C., Ley, H. L., and Gerard, G. F.

Huynh, T. V., Young, R. A., and Davis, R. W. (1985) in DNA

U. S. A. 78,2249-2252

(1977) Biochemistry 16,5824-5831

J. Biol. Chem. 258, 1914-1920

T. S. (1984) Blood 6 3 , 486-489

15432-15438

7 1,728-733

Haycraft, D. L. (1980) Blood 55,271-275

Biochem. 105,304-310

M. (1987) Methods Enzyrnol. 152,219-227

1408-1412

(1987) Focus 9 , 1-4

Cloning: A Practical Approach (Glover, D. M., ed) pp. 49-78, IRL, Oxford

Feinberg, A. P., and Volgelstein, B. (1983) A d . Biochem. 132,

Wen, D., Dittman, W. A., Ye, R. D., Deaven, L. L., Majerus, P. 6-13

23.

24. 25. 26.

27.

28.

29.

30.

31.

32.

33. 34.

35.

36.

37.

38.

39.

W., and Sadler, J . E. (1987) Bioc,kmistry 26,4350-4357

Acad. Sci. U. S. A. 74,5463-5467 Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl.

Tabor, S., and Richardson, C. C. (1987) Proc. Natl. Acad. Sci.

Favaloro, J., Treisman, R., and Kamen, R. (1980) Methods En-

Kelly, K., Cochran, B. H., Stiles, C. D., and Leder, P. (1983) Cell

Parnes, J. R., and Seidman, J. G. (1982) Cell 29, 661-669 Lehrach, H., Diamond, D., Wozney, J. M., and Boedtker, H.

Chomczynski, P., and Sacchi, N. (1987) Anal. Biochem. 162,

Kessler, S. W. (1981) Methods Enzymol. 73, 442-459 Laemmli, U. K. (1970) Nature 227,680-685 Bonner, W. M., and Laskey, R. A. (1974) Eur. J. Biochern. 46,

Salem, H. H., Maruyama, I., Ishii, H., and Majerus, P. W. (1984) J. Biol. Chem. 259,12246-12251

Bolton, A. E., and Hunter, W. M. (1973) Biochem. J. 133, 529- 539

Cooper, J. A., Sefton, B. M., and Hunter, T. (1983) Methods Enzymol. 99,387-402

Jackman, R. W., Beeler, D. L., VanDeWater, L., and Rosenberg, R. D. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 8834-8838

Caput, D., Beutler, B., Hartog, K., Thayer, R., Brown-Shimer, S., and Cerami, A. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,

Harris, L. J., Remmers, E. F., Brodeur, P., Riblet, R., D’Eusta- chio, P., and Marcu, K. B. (1983) Nucleic Acids Res. 11,8303- 8315

U. S. A. 84,4767-4771

zymol. 65, 718-749

35,603-610

(1977) Biochemistry 16,4743-4751

156-159

83-88

1670-1674

Brown, M. S., and Goldstein, J. L. (1986) Science 232, 34-47 Goldstein, J. L., Anderson, R. G. W., and Brown, M. S. (1979)

Nature 279,679-685 Davis, C. G., Goldstein, J. L., Sudhof, T. C., Anderson, R. G. W.,

Russell, D. W., and Brown, M. S. (1987) Nature 326,760-765 Stenflo, J., Lundwall, A., and Dahlback, B. (1987) Proc. Natl.

Acad. Sci. U. S. A. 84.368-372 Stenflo, J., Ohlin, A., Owen, W. G., and Schneider, W. J. (1988)

J. Biol. Chem. 263, 21-24 Russell, D. W., Schneider, W. J., Yamamoto, T., Luskey, K. L.,

Brown, M. S., and Goldstein, J. L. (1984) Cell 37,577-585 Shaw, G., and Kamen, R. (1986) Cell 46, 659-667

15822 The Structure and Function of Mouse Thrombomoddin 40. Treisman, R. (1985) Cell 42,889-902 Acud. Sci. U. S. A. 81,2016-2020 41. Jackman, R. W., Beeler, D. L., Fritze, L., Soff, G., and Rosenberg, 49. Davis, R. J., Johnson, G. L., Kelleher, D. J., Anderson, J. K.,

R. D. (1987) Proc. Natl. Acud. Sei. U. S. A. 84,6425-6429 Mole, J. E., and Czech, M. P. (1986) J. Biol. Chem. 261,9034- 42. Shirai, T., Shiojiri, S., Ito, H., Yamamoto, S., Kusumoto, H., 9041

Deyashiki, Y., Maruyama, I., and Suzuki, K. (1988) J. Biochem. 50. Zerial, M., Suomalaninen, M., Zanetti-Schneider, M., Schneider,

43. Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra, R. J., Rahms- 51. Rothenberger, S., Iacopetta, B. J., and Kuhn, L. C. (1987) Cell 103,281-285 C., and Garoff, H. (1987) EMBO J. 6,2661-2667

dorf, H. J., Jonat, C., Herrlich, P., and Karin, M. (1987) Cell 49,423-431 49,729-739 52. Davis, R. J., and Meisner, H. (1987) J. Bid. Chern. 262, 16041-

44. Klausner, R. D., Harford, J., and van Renswoude, J. (1984) Proc. 16047

45. Aggarwal, B. B., and Eessalu, T. E. (1987) J. B i d . Chern. 262, 745

46. Lee, L. S., and Weinstein, I. B. (1978) Science 202, 313-315 (1986) Proc. Nutl. Acad. Sci. U. S. A. 83, 3460-3464 47. Salomon, D. S. (1981) J. Biol. Chem. 256,7958-7966 55. Moore, K. L., Andreoli, S. P., Esmon, N. L., Esmon, C. T., and 48. May, W.S., Jacobs, S., and Cuatrecasas, P. (1984) Proc. Natl. Bang, N. U. (1987) J. Clin. Znuest. 79, 124-130

Nut!. A d . Sci. U. S. A. 81, 3005-3009 53. Nawroth, P. P., and Stern, D. M. (1986) J. Exp. Med. 163, 740-

16450-16455 54. Nawroth, P. P., Handley, D. A., Esmon, C. T., and Stern, D. M.

the structure and function of mouse thrombomodulin · training grant t32 hlbi 07088 from the...

Documents