DK175431B1 - DNA encoding human tissue factor inhibitor - used in study of coagulation cascade for agents which inhibit factor Xa and Factor VIIA-TF - Google Patents

Abstract

DNA sequence encoding human tissue factor inhibitor (TFI) is claimed. Also claimed is human TFI having a specified aminoacid sequence.

Description

DK 175431 B1 j 'i

Antibodies which bind tissue factor inhibitor (TFI) fragments, compositions comprising these antibodies, and method for using an immunoaffinity matrix comprising such antibody to purify a polypeptide from a biological fluid 5

The present invention relates to a coagulation inhibitor termed tissue factor inhibitor (TFI) and, alternatively, as a lipoprotein-associated coagulation inhibitor (LACI). More particularly, the invention relates to a cDNA clone representing substantially the entire TFI compound.

10

BACKGROUND OF THE INVENTION

The coagulation process that occurs in mammalian blood involves two distinct systems - the so-called internal and external systems. The latter system is activated when the blood is exposed to tissue thromboplastin (factor III), hereinafter referred to as tissue factor (TF). Tissue factor is a lipoprotein found in the plasma membrane of a whole range of cell types, with the brain and lungs particularly rich in it. Upon contact with TF, plasma factor VII or its activated form, factor VIa, forms a calcium-dependent complex with TF, which then proteolytically activates factor X to factor Xg, and factor IX to factor IXa.

Early studies on the regulation of TF-initiated coagulation showed that incubation of TF (in crude tissue thromboplastin preparations) with serum inhibited its activity in vitro and prevented its lethal effect when infused into mice. Further studies of Hjort, Scand. J. din.

25 Lab. Invest. 9, Suppl. 27, 76-97 (1957) confirmed and further developed previous work in this field and led to the conclusion that an inhibitory part of serum recognizes the factor VII-TF complex. Consistent with this hypothesis are the facts that inhibition of TF occurring in plasma requires the presence of Ca 2+ (which is also necessary for the binding of factor VII / VIIa to TF) and that inhibition can be prevented and / or reversed. about by chelating divalent cations with EDTA. Recent studies have shown that not only factor VIIa but also catalytically active factor Xa and an additional factor are required for the achievement of TF inhibition in plasmids.

I 2 I

In ma or serum. See Broze and Miletich, Blood 69,150-155 (1987) and Sanders I

et al., Ibid. 66, 204-212 (1985). This additional factor, herein defined as I

In Tissue Factor Inhibitor (TF!) And Alternatively as Lipoprotein-Associated Coagulation I

In onshore inhibitor (LACI) is present in barium-absorbed plasma and appears to be I

I 5 associated with lipoproteins, since TFI functional activity is secreted by lipopro-I

In the tein fraction that floats on top when the serum is centrifuged at a density I

I of 1.21 g / cm 3. According to Broze and Miletic, supra, and Proc. Natl. Acad. Sci. USA I

In 84,1886-1890 (1987), HepG2 cells (a human hepatoma cell line) secrete I

In an inhibitory portion having the same properties as the TFI contained in I

In plasma. IN

EP EP Publication No. 300,988 discloses a purified tissue factor inhibitor (TFI) secreted by HepG2 cells. It was found to be available in two forms, in a TFI1 migrating at about 37000-40000 daltons and a TFI2 migrating at 25000-26000 daltons as determined by sodium dodecyl sulfate polyacrylamide I gel electrophoresis (SDS-PAGE ). A partial N-terminal amino acid sequence for TFI was mapped as:

I 1 15 I

XX-Glu-Glu-Asp-Glu-Glu-His-Thr-lle-lle-Thr-Asp-Thr-Glu-16 27 I 20 Leu-Pro-Pro-Leu-Lys-Leu-Met-His-Ser Phe- (Phe) -Ala I where XX had not been determined. The contents of this application are considered to be incorporated in this specification by reference.

BRIEF DESCRIPTION OF THE INVENTION

In accordance with the present invention, the complete code sequence of a cDNA clone substantially representing full-size tissue factor inhibitor (TFI) has been developed.

Initially, human placental and fetal liver Agt11 cDNA libraries were screened with a polyclonal rabbit antiserum against a purified TFI. Immunologically positive clones were further screened for 125 I factor Xa binding activity. Seven clones that were immunologically and functionally active were obtained.

The longest clone, placenta-derived ΛΡ9, was 1.4 kilobases (kb) long, while ___; _; · · - ϊίτ-.τ ··.;., ^. ^ - · τβπτ. - · "ν · Τ '. ♦: * - · ··? =? ·? ► ♦ _Jaig.jal.'. I'T'ry Ύ ^. ^ Rx.T'J.vim DK 175431 B1 3 - The other six had a length of 1.0 kb. Partial DNA sequencing showed that the 1.0 kb clones had sequences identical to part of the longer 1.4 kb clone. Nucleotide sequence analysis showed that AP9 consisted of 1432 base pairs (pb) cDNA. insertion comprising a 5 'non-coding region of 133 bp, an open 5 reading frame of 912 bp, a stop codon and a 3' non-coding region of 384 bp.

The cDNA sequence encodes a 31,950 dalton protein of 276 amino acids, which comprises 18 cysteine residues and 7 methionine residues. The translated amino acid sequence shows that a signal peptide of about 28 amino acids precedes the mature TFI protein. It should be noted that the "mature" TFI is defined to include both TFI and methionyl TFI due to the translational ATG codon of the AP9 clone described herein.

There are three potential N-linked glycosylation sites in the TFI protein of the sequence Asn-X-Ser / Thr, wherein X may be any of the common 20 amino acids. These sites are at amino acid positions Asn 145, Asn 195 and Asn 256 when the first methionine residue after the 5 non-coding region is designated amino acid position +1.

The translated amino acid sequence of TFI shows several distinguishable domains, including a highly negatively charged N-terminal, a highly positively charged carboxy-terminal, and an intermediate portion consisting of three homologous domains with sequences typical of Kunitz-type enzyme inhibitors. Based on a homology study, TFI appears to be a member of the basal protea signal inhibitor gene superfamily.

The original source of the protein material to produce cDNA clone AP9 25 was human placental tissue. This tissue can be extensively obtained after birth by common surgical methods. AgtII (Iac5 nin5 c1857 S100) used herein is a well known and widely available lambda phage expression vector.

Its construction and restriction endonuclease map are described by Young and Davis, Proc. Natl. Acad. Sci. USA 80, 1194-1198 (1983).

Northern blot analysis revealed that the following liver-derived cell line Chang liver cell, HepG2 hepatoma and SK-HEP-1 hepatoma, all contained two major mRNAs (1.4 and 4.4 kb) that hybridized with the TFI cDNA.

I DK 175431 B1

The cloning of the cDNA for TFI and the development of its entire protein sequence and I structural domains referred to herein allows detailed structural functional analyzes and provides a basis for studying its biosynthetic regulation. Thus, the invention is important for the medical science of the study of coagulation with respect to agents capable of inhibiting factor Xg and factor VIIa / TF enzyme complex.

Specifically, the present invention relates to antibodies which bind fragments H of tissue factor inhibitor (TFI), as defined in claims 1-4.

H In addition, the invention relates to particular embodiments of these antidotes, as defined in claims 6-9.

Further, the invention relates to compositions comprising these antibodies, as defined in claims 4 and 5, and to a method for using an immunoaffinity matrix comprising such an antibody to purify a polypeptide from a biological fluid as defined in claim 10.

I DETAILED DESCRIPTION OF THE INVENTION

The invention will now be described in more detail with reference to the drawing, in which Figure 1 shows the screening of AgtII clones with 125 I factor Xa. Cloned phage lysates (0.1 ml) were stained on a nitrocellulose paper by suction using a dot blot apparatus. The nitrocellulose paper was then probed with 125 I factor Xa and autoradiographed as described below. Clones that appear as dark spots are positive clones that bind in 125 I factor Xa. Control EgtII (lower right corner) and other clones do not bind to 125 I factor Xg.

In FIG. Figure 2 shows a partial restriction map and sequencing strategy for the AP9-1 insertions. The scale below indicates the nucleotide position. The thick I bar represents the coding region. The thin bars represent I 5 'and 3 non-coding regions. The restriction endonuclease sites were confirmed by degradation. The arrows show the overlapping M13 clones I used to sequence the cDNA.

. ...... π - * - * - rvr- - - ~ ^. - y - - - "" ^ '' -. 'V · »·» ·: - MUt ^ DK 175431 B1 5

FIG. 3 shows the nucleotide sequence and the translated amino acid sequence of I

the human TFI cDNA. Nucleotides are numbered on the left and amino acids on the right. The underlined sequences have been independently confirmed by amino acid sequence analysis of the purified TFI protein 5 and two Vs protease + trypsin degraded peptides. Amino acid + 1 was assigned to the first methionine residue after a stop codon in the 5 'non-coding region. Potential N-linked glycosylation sites are marked with asterisks.

FIG. 4 is a graph showing the charge distribution of the amino acid sequence in TFI. Charges are calculated from the first residue to the end residue and 10 shown at the end residue. Thus, the value in the end position is the summation of all charges from the first residue to the end residue, and the difference in the charges between the end and the j residue (j> i) is the net charge of the fragment from the at the end to the j-th rest.

j

FIG. Figure 5 is a graph showing the hydrophobicity profile of TFI. The hydrophobicity profile was analyzed by a computer program in which the hydrophobicity index of the amino acid residues is defined as the depth to which an amino acid residue is buried within a protein (from X-ray crystallographic data) [Kidera et al., J. Protein Chem. . 4, 23-55 (1985)]. The hydrophobicity profile along the sequence was smoothed using the program 20 ICSSCU in the IMSL Library [IMSL Library Reference Manuai, 9th Edition, Institute for Mathematical and Statistical Subroutine Library, Houston, Texas (1982)].

FIG. 6 shows a line arrangement of the basal protease inhibitor domains of TFI with other basal protease inhibitors. All sequences except TFI were obtained from the National Biomedical Research Foundation Protein Sequence Da-25 tabase (Georgetown University, Washington, D.C., U.S.A., released June 13, 1987). 1. Bovine basal protease inhibitor precursor; 2. Bovine colostrum trypsin inhibitor; 3. Bovine serum basal protease inhibitor; 4. Edible snail isoin inhibitor K; 5. Red Sea Turtle Basal Protease Inhibitor (amino acids only 1-79 indicated); 6. Western Sandworm Poison-Basal Protease Inhibitor I; 7th

Ring-neck poison basal protease inhibitor II; 8. Cape cobra venom-basal protease inhibitor II; 9. Russell's venomous poison-basal protease inhibitor II; 10. Sandworm poison-basal protease inhibitor III; 11. Eastern green mamba venom-basal protease inhibitor In homologue; 12. Black mamba venom-basal protease inhibitor B; 13. Black I DK 175431 B1 mamba venom-basal protease inhibitor E; 14. Black mamba venom-basal protease I inhibitor I; 15. Black mamba venom-basal protease inhibitor K; 16. β-1- I Bungarotoxin B chain (minor); 17. β-1-Bungarotoxin B chain (larger); 18. β-2-Bungarotoxin B chain; 19. Horse inter-α-trypsin inhibitor [amino acids 1- 577 (1); 58-123 (2)]; 20. Pig inter-α-trypsin inhibitor [amino acids 1-57 (1); I 58-123 (2)]; 21. Bovine inter-α-trypsin inhibitor [amino acids 1-57 (1); 58- I 123 (2)]; 22. Human α-1 microglobulin / inter-α-trypsin inhibitor precursor [amino acids 227-283 (1); 284-352 (2)]; 23. TFI [amino acids 47-117 (1); 118- I 188 (2); 210-280 (3)]. Spaces were included in 16,17,18 to achieve best compliance. Standard one-letter codes for amino acids have been used.

In FIG. Figure 7 shows Northern blot analysis of RNAs from 3 liver-derived cell lines. 10 µg of poly (A) + RNA was used per Lane. Lane 1: Chang liver cell; lane 2: In SK-HEP-1 hepatoma cell; lane 3: HepG2 hepatoma cell.

In this specification description, standard biochemical nomenclature is used wherein the nucleotide bases are designated as adenine (A); thymine (T); guanine (G); and In cytosine (C). Corresponding nucleotides are, for example, deoxyguanosine 5'-triphosphate (dGTP). As is common for the convenience of the structural

In tureile mapping of a DNA nucleotide sequence, only one strand is shown wherein A

I 20 on one strand represents T on its complementary strand, and G represents I C. Amino acids are shown either by three letter or one letter abbreviations in the following manner: I Abbreviated designation Amino acid IA Ala Alanine I 25 C Cys Cysteine ID Asp Aspartic acid E Glu Glutamic Acid F Phe Phenylalanine | G Gly Glycine 30 H His Histidin I I Ile Isoleucine I K Lys Lysine I L Leu Leucin 7 DK 175431 B1 M Met Methionine N Asn Asparagine P Pro Prolin Q Gin Glutamine 5 R Arg Arginln S Ser Serin T Thr Threonin V Val Valin W Trp Tryptophan iosο Y_lyr__

Commonly available restriction endonucleases described herein have the following sequences and (indicated by arrows) cleavage patterns:

EcoR1 GAATTC

15 GI I AA G

r i

Ssp1 AAT AAT

TTATAA

20 T

in

Cla1 ATCGAT

TAGCTA

T

25 i

Alu1 AGCT

TCGA

t l

30 Stu1 AGGCCT

'TCCGGA T

In order to illustrate specific preferred embodiments of the invention, the following examples of preparative laboratory work are set forth below.

I DK 175431 B1 IN EXAMPLE 1

materials

Human placental and fetal liver cDNA libraries were obtained from Clonetech.

Protoblot immuno-screening equipment was purchased from Promega Biotech. Restriction enzymes were from New England Biolabs. Calf intestinal alkaline phosphatase, T4 DNA ligase, DNA polymerase I (Klenow), exonuclease III and S1 nuclease were from Boehringer Mannheim. dNTP was from P.L. Biochemicals. 5'- [α-35S] -thio-dATP (600 Ci / mml) was from Amersham. Sequencing equipment (see

quenase) was from United States Biochemicals. Chang liver cells (ATCC CCL

I 10 13) and HepG2 hepatoma cells (ATCC HB 8065) were obtained from the American Type Culture Collection. SK-HEP-1 hepatoma cells originally originated from a liver adenocarcinoma of G. Trempe of the Sloan-Kettering Institute for Cancer Research in 1971 and are now widespread and readily available.

I 125l Factor Xg was prepared by radiolabelling using lodogen. The specific activity was 2000 ppm / ng. More than 97% of the radioactivity could be precipitated with 10% trichloroacetic acid (TCA). The iodinated protein retained> 80% of its catalytic activity against Spectrozyme Xg (American Diagnostica Product).

In an anti-TFI Ig "Sepharose® 4B" column was prepared as follows: An I peptide (designated TFI peptide) containing a sequence corresponding to amino acid I sequence 3-25 of the mature TFI was synthesized under the use of

Biosystems solid phase peptide synthesis system. The TFI peptide (5 mg) was conjugated to 10 mg of Keyhole elbow peel hemocyanin by glutaraldehyde. Two New Zealand breed white rabbits were immunized by intradermal injection I 25 of a homogenate containing 1 ml of Freund's complete adjuvant and 1 ml of conjugate (200 µg TFI peptide). One month later, the rabbits were re-injected with a homogenate containing 1 ml of Freund's incomplete adjuvant and 1 ml of conjugate (100 µg conjugate). Antiserum was collected weekly for three months and further injections were performed once a month. To isolate specific antibodies to the TFI peptide, antiserum was chromatographed on a

In TFI peptide "Sepharose 4B" column. The column was washed with 10 volumes of PBS

I (0.4 M NaCl-0.1 M benzamidine-1% Triton® X-100 "), and the same solution I without Triton X-100". The antibody was eluted with 0.1 M glycine / HCl, pH 2, 2, and ____. ..- .. _:, .ν -is _ · - ^ ^ ΠΤιΐ -> DK 175431 Β1 9 then immediately neutralized by the addition of 1/10 volume of Tris-OH and dialyzed against brine. antibody was coupled to cyanobromide-activated "Sepharose 4B" following the manufacturer's (Pharmacia) instructions and used to isolate TFI from the cell culture substrate.

5 Chang liver cells were cultured by the method previously described by Broze and Miletich, Proc. Natl. Acad. Sci. USA 84, 1886-1890 (1987). The conditioned substrate was chromatographed on the anti-TFI-Ig "Sepharose 4B" column.

The column was washed with 10 volumes of PBS-1% Triton X-100 "and PBS. The bound TFI was eluted with 0.1 M glycine HCl, pH 2.2. Immunoaffinity-isolated io TFI was further purified by preparative sodium dodecyl sulfate-polyacrylamide gel. Electrophoresis (Savant's apparatus). Amino acid analysis of the final product showed the same amino-terminal sequence as the TFI isolated from HepG2 cells as described in the co-pending Danish patent application PA1988 04135. The isolated Chang liver cell TFI was then used to immunize The antiserum obtained had a titer of about 100 pg / ml in the dual immunodifusion assay, which was used in the immuno-screening of AgtII cDNA libraries.

Methods 2 o Isolation of cDNA clones

Methods to scan placental and fetal liver cDNA libraries with antibody, plaque purification and production of λ-phag lysate and DNA were as described by Wun and Kretzmer, FEBS Lett. 1.11-16 (1987). The antiserum was pre-adsorbed with BNN97 Agt11 lysate and diluted 1/500 to screen the library.

Screening for factor Xa binding activity

Recombinant proteins induced with isopropyl-β-thio-galactoside from immunopositive λ-phag isolates or from control Agt11 were screened for factor Xa binding activity. The α-phage lysates (0.1 ml) were filtered through a nitro cellulose paper using a dot blot apparatus (Bio Rad). The nitrocellulose paper was then immersed and stirred in phosphate buffered saline containing I DK 175431 B1

I ίο I

5 mg / ml bovine serum albumin and 2.5 mg / ml bovine gamma globulin at room temperature for 1 hour. The solution was replaced with 125 I factor Xa (1.0 x 106 cmp / ml) dissolved in the same solution added 0.1 mg / ml heparin and stirring was continued for an additional hour. Then the nitrocellulose paper was washed with phosphate buffered saline containing 0.05% "Tween® 20". The wash buffer was replaced every five minutes, four times. The nitrocellulose paper was then air dried and prepared for autoradiography using "Kodak I XR5" film. The film was developed after a week of exposure.

Preparation of Poly (A) + RNA and Northern Blotting ίο Total RNAs were prepared from cultured Chang liver cell, HepG2 hepatoma cell and SK-HEP-1 hepatoma cell using the sodium perchlorate extraction method described by Lizardi and Engelberg, Anal. Biochem. 98, 116-122 (1979). Poly (A) + RNA preparations were isolated by batch adsorption on oligo (dT) cellulose (P-L Biochemical, type 77F) using the manufacturer's recommended procedure. For Northern blot analysis, 10 µg of each poly (A) + RNA was treated with glyoxal [Thomas, Methods Enzymol.

In 100,255-266 (1983)] and subjected to agarose gel electrophoresis in a buffer containing 10 mM sodium phosphate, pH 7.0. Bethesda Research Laboratory's RNA ladder was used as a molecular weight marker. The RNAs were dipped into a nitrocellulose paper which was then heated for 2 hours at 80 ° C.

In the Insertion DNA of the ÅP9 clone, radiolabelled with 32

H translation and used as a probe [Maniatis et al., Molecular Cloning: A

In the Laboratory Model, Cold Spring Laboratory, Cold Spring Harbor, N.Y. (1982)].

In the duct, 5 x 10 6 cpm of the probe was hybridized in 5 ml of a solution containing 50% formamide, 5X SSC, 50 mM sodium phosphate, pH 7.0, 250 μg / ml denatured salmon sperm DNA and 1X Denhardt's solution at 42

At 16 ° C for 16 hours. The filter was washed in 0.1% sodium dodecyl sulfate (SDS), 2X SSC

I at room temperature 3 times, each time for 5 minutes, and in 0.1% SDS, 0.2 X

In SSC at 50 ° C twice, each time for 5 minutes. The nitrocellulose paper was air dried and then autoradiographed for 3 days at -70 ° C using Kodak XAR-5 film and intensifier screen.

11 DK 175431 B1

Other recombinant DNA methods

Preparation of cloned AgtII DNA, subcloning into pUC19 plasmid and M13mp18 vector, generation of deletion by exonuclease I II degradation and DNA sequencing by the dideoxy method [Sanger et al., Proc. Natl. Acad. Sci. USA 5 83, 6776-6780 (1977)] was carried out as described by Wun and Kretzmer, supra.

The FASTP program written by Lipman and Pearson, Science 227, 1435-1441 (1985) was used to identify homologous families of proteins from the National Biomedical Research Foundation Sequence Data Bank (released June 13, 1987) and to align the sequences with the homologous family.

RESULTS

Screening of cDNA libraries

A variety of cell lines were screened for the presence of TFI in the conditioned substrate and it was found that several liver-derived cell lines, Chang-15 liver cell, HepG2 hepatoma and SK-HEP-1 hepatoma, secrete TFI in culture. Initially, an antiserum against TFI was used to screen a human fetal-liver AgtII cDNA library (106 plaque-forming units) and 15 immunologically positive clones were obtained. Then, the same method was used to screen a placental AgtII cDNA library. Out of 106 plaque-forming units, 10 immunologically positive clones were obtained. These clones were plaque purified and the lysates of the purified clones tested for TFI functional activity. The isopropylthiogalactoside-induced phage lysates were absorbed onto the nitrocellulose paper and screened for the 125 I factor Xa binding activity. FIG. Figure 1 shows that some of these immunologically positive clones showed the ability to bind 125 I factor Xg on the nitrocellulose paper. In total, three out of 15 immunologically positive fetal liver clones and 4 out of 10 immunologically positive placental clones showed 125 I factor Xa binding activity. These immunologically and functionally positive clones were digested with EcoR1 and the size of the insertions was determined by gel electrophoresis. One clone from the placental library (AP9) had an insertion of approximately 1.4 kb, while all the other clones contained insertions of approximately 1.0 kb. Partial DNA sequencing has shown that the 1.0 kb clones contain sequences that are identical.

I DK 175431 B1 I

I 12 I

You happen to have part of the longer 1.4 kb placental clone (λΡ9). ΛΡ9 therefore became you

In selected for complete sequencing. IN

In Nucleotide sequence and predicted protein sequence of TFI cDNA isolate I

In the AP9 clone, restriction mapping, M13 subcloning and seI were subjected

In the quantization as indicated in FIG. 2. The entire sequence was determined on I

In both strands using the exonuclease III deletion method [Henikofff, Ge- I

In Ne 28, 351-359 (1984)] and was found to consist of 1432 bases in length. Se- I

In the figure shown in FIG. 3. It contains a 133 '5' non-coding region

In bases, an open reading frame of 912 nucleotides, and a 3 'non-coding region I

In 10 of 387 nucleotides. The first ATG occurs at nucleotide 134 of sequence I

In late TAGATGA, which was closely followed by another ATG at nucleotide 146 in I

In the sequence ACAATGA. These may be the starting sequences, even if they are you

In different from the proposed consensus sequence for starting eukaryotic ribo- I

In which, ACCATGG [Kozak, Cell 44, 283-292 (1986)]. 28 amino acids precede I

I 15 for a sequence corresponding to the N-terminus of the mature protein. The length and I

In the composition of the hydrophobic portion of these 28 amino acids are typical of I

In signal sequences [Von Heijne, Eur. J. Biochem. 133, 17-21 (1983); J. Mol. IN

In Biol. 184, 99-105 (1985)]. A signal peptidase may cleave at Ala28-I

In Asp2g to form a mature protein. The predicted sequence of mature I

I 20 TFI consists of 276 amino acids containing 18 cysteine residues and 7 methio-I

In nineteen. The calculated mass of 31,950 daltons based on the deductible I

In the protein sequence of mature TFI is somewhat lower than the 37-40 kDa I rated

I by sodium dodecyl sulfate-polyacrylamide gel electrophoresis of isolated protein, I

I and the difference probably reflect the glycosylation contribution to the mobility of I

In the natural protein. The deduced protein sequence corresponding to the mature I

In protein, three potential N-linked glycosylation sites contain se- I

In the sequence Asn-X-Thr / Ser (amino acid positions 145, 195 and 256). Amino-I

In sequence analysis of purified whole TFI and two isolated proteolytic fragments I

You exactly match the protein I deduced from the cDNA sequence

In 30 sequence (Fig. 3, underlined), indicating that the isolated cDNA clone I

In TFI codes. The 3 non-coding region is A + T rich (70% A + T). Neither I

In the consensus polyadenylation signal, AATAAA [Proudfoot and Brownlee, Na- I

In Tours 252, 359-362 (1981)], or the poly A tail was found in this clone,

13 DK 175431 B1 is also due to the artificial loss of part of the 3-terminal part during the construction of the library.

Charge distribution, hydrophobicity / hydrophilicity and internal homology

The translated amino acid sequence of TFI contains 27 lysine residues, 17 arginine residues, 11 aspartic acid residues and 25 glutamic acid residues. The charge distribution along the protein is at most uneven as indicated in FIG. 4. The signal peptide region contains two positively charged lysine residues with 26 neutral residues. The amino-terminal region of the mature protein contains a high negatively charged stretch. Six of the first 7 residues are either aspartic acid or glutamic acid, which is closely followed by two more negatively charged amino acids below before a positively charged lysine residue appears. The center portion of the molecule is generally negatively charged. At the carboxy terminal there is a high positively charged portion. Amino acid bands 265-293 of TFI contain 14 positively charged amino acid residues, including a sequence of 6 consecutive arginine-lysine residues.

The predicted hydrophilicity / hydrophobicity profile of the TFI protein is given in FIG. 5. The signal peptide contains a high hydrophobic region as expected. Rather, the rest of the molecule appears to be hydrophilic.

The translated amino acid sequence of TFI contains several special domains. In addition to the highly negatively charged N-terminal domain and the high 20 negatively charged C-terminal domain, the center portion consists of three homologous domains that have the typical Kunitz-type inhibitors (see below).

Homology to other proteins

By searching the National Biomedical Research Foundation Sequence Database 25, it was found that the N-terminal domain and C-terminal domain of TFI do not show significant homology to other known proteins. However, the three internal homologous domains are separately homologous to sequences of other basal protease inhibitors, including basal bovine pancreatic protease inhibitor (aprotinin), basal venom protease inhibitors, and inter-α-trypsin inhibitors (Fig. 6). Remarkably, the disulfide bond structure is

I DK 175431 B1 I

I 14 I

In strongly preserved in all these three inhibitors. Based on these homologies, it is you

Clearly, TFI belongs to the basal protease inhibitor gene superfamily. IN

In Northern blotting I

In Poly (A) + RNAs were purified from TFI-producing liver-derived cell lines,

In 5 Chang liver cells, HepG2 hepatoma cells and SK-HEP-1 hepatoma cells. Po- I

The ly (A) + RNAs were resolved by denaturing agarose gel electrophoresis, I

I over-dipped on a nitrocellulose paper and probed with 32P-labeled I

In TFI cDNA (λΡ9). As shown in FIG. 7, two major hybridizations were observed

In ring bands corresponding to mRNAs of 1.4 kb and 4.4 kb, in all three tested I

In 10 cell lines. Several other cell lines were examined which did not produce I

In detectable amounts of TFI and in which no hybridization I was found

In the ring with the probe (data not shown). IN

In various other examples, after reading this description, I will be included

In illuminating to those skilled in the art without going outside the scope of the invention. Such you

In 15 further examples, it is also to be considered included in the invention as defined

In court by the following claims. IN

Claims (10)

1. Isolated and purified antibody having a binding factor for tissue factor inhibitor (TFI) with an amino acid sequence as shown in FIG. 3 and binds to a tissue factor inhibitor (TFI) region selected from: 5 (a) the N-terminal amino acids 29-55 or 31-53 of purified mature TFI as shown in FIG. 3; (b) the C-terminal amino acids 265-304 in purified mature TFI as shown in FIG. 3; (c) a V8 protease plus trypsin-degraded peptide of TFI with the sequence of amino acids 82-88 as shown in Figs. 3; (D) a V8 protease plus trypsin-degraded peptide of TFI with the sequence of amino acids 153-166 as shown in Figs. 3; (e) amino acids 54-104 or 47-117 of FIG. 3, which forms Kunitz domain 1 of the TFI; (f) amino acids 125-175 or 118-188 of FIG. 3, which forms the Kunitz domain 15th in the TFI; and (g) amino acids 217-267 or 210-280 of FIG. 3, which forms the Kunitz domain 3 of the TFI. An isolated and purified antibody according to claim 1, which has a tissue factor inhibitor binding specificity (TFI) having an amino acid sequence as shown in FIG. 3 and which does not bind a polypeptide selected from bovine basal protease inhibitor precursor, bovine colostrum trypsin inhibitor, bovine serum basal protease inhibitor, edible snail isoin inhibitor, red sea turtle basal protease inhibitor, western basal protease inhibitor II, Cape cobra venom-basal protease inhibitor II, Russell's venom venom-basal protease inhibitor II, sandy venom venom-basal protease inhibitor III, eastern green mamba venom-basal protease inhibitor I homolog, black mamba venom-basal protease inhibitors B, E, I and K, β-1-bungarotoxin B chain (minor), β-1-bungarotoxin B chain (larger), β-2-bungarotoxin B chain, equine inter-α I trypsin inhibitor (amino acids 1- 57 and 58-123), swine inter-α-trypsin inhibitor II (amino acids 1-57 and 58-123), bovine inter-α-trypsin inhibitor (amino acids I 1-57 and 58-123), and human α-1- microglobulin / inter-α-trypsin inhibitor precursor (amino acids 2 27-283 and 284-352). I I 5 i An antibody having a binding specificity for a polypeptide comprising I in one or more of the following regions of TFI: I (a) the N-terminal amino acids 29-55 or 31-53 of purified mature TFI as shown in I of FIG. 3; In (b) the C-terminal amino acids 265-304 of purified mature TFI as shown in FIG. 3; In (c) a V8 protease plus trypsin digested peptide of TFI with the sequence of I I amino acids 82-88 as shown in Figs. 3; In I (d) a V8 protease plus trypsin degraded peptide of TFI with the sequence of I I amino acids 153-166 as shown in Figs. 3; In amino acids 54-104 or 47-117 of Figs. 3, which forms the Kunitz domain 1 I 'TFI; In (f) amino acids 125-175 or 118-188 of FIG. 3, which forms the Kunitz domain I 2 of the TFI; and I (g) amino acids 217-267 or 210-280 of FIG. 3, which forms the Kunitz domain I 20 3 in the TFI. A composition comprising an antibody according to claim 1,2 or 3 and an effective carrier, diluent or adjuvant. A composition comprising an antibody according to claim 1,2 or 3 and an I solution. DK 175431 B1 The antibody of claim 1,2 or 3 which is a polyclonal antibody. The antibody of claim 1, 2 or 3 which is purified on a TFI peptide * column. 5 The antibody of claim 1,2 or 3 conjugated to an immunoaffinity matrix. The antibody of claim 8, wherein said immunoaffinity matrix is "Sepharose® 4B". 10 A method of using an immunoaffinity matrix according to claim 8 or 9 to purify a polypeptide from a biological fluid. 4 t