WO1994020537A1 - Non-glycosylated tfpi analogues - Google Patents

Abstract

Novel, non-glycosylated TFPI analogues containing the first and second Kunitz domain and lacking the third Kunitz domain have a prolonged half life compared with corresponding glycosylated TFPI analogues.

Description

NON-GLYCOSYLATED TFPI ANALOGUES

Field of the invention

The present invention relates to non-glycosylated tissue factor pathway inhibitor (TFPI) analogues.

Background of the invention

Blood coagulation is a complex process involving many acti¬ vating and inactivating coagulation factors. Anticoagulant proteins are known to be important for regulation of the coa¬ gulation process and anticoagulants are thus important in the treatment of a variety of diseases, e.g. thrombosis, myocardial infarction, disseminated intravascular coagulation etc.

Thus heparin is used clinically to increase the activity of antithrombin III and heparin cofactor II. Antithrombin III is used for the inhibition of factor Xa and thrombin. Hirudin is used for the inhibition of thrombin and protein C may be used for the inhibition of factor V and factor VIII. Anticoagulant proteins may also be used in the treatment of cancer.

Coagulation can be initiated through the extrinsic pathway by the exposure of tissue factor (TF) to the circulating blood (Y. Nemerson, Blood 21 (1988) 1-8) . Tissue factor is a protein cofactor for factor Vll/VIIa and binding of tissue factor en¬ hances the enzymatic activity of factor Vila (FVIIa) towards its substrates factor IX and factor X.

Recently a new anticoagulant protein, the tissue factor pathway inhibitor (TFPI) has been isolated (G.J. Broze et al., Proc. Natl. Acad. Sci. 81 (1987) 1886-1890).

On a molar basis TFPI has been shown to be a potent inhibitor of TF/FVIIa-induced coagulation (R.A. Gra zinski et al., Blood 21 (1989) 983-989) . TFPI binds and inhibits factor Xa (FXa) and the complex between TFPI and FXa inhibits TF/FVIIa (Rapaport, Blood 21 (1989) 359-365) . TFPI is especially interesting as an anticoagulant/antimetastatic agent as many tumor cells express TF activity (T. Sakai et al., J. Biol. Chem. 264 (1989), 9980- 59988) and because TFPI shows anti-Xa activity like antistatin which has antimetastatic properties.

TFPI has been recovered by Broze et al. (supra) from HepG2 he- patoma cells (Broze EP A 300988) and the gene for the protein has been cloned (Broze EP A 318451) . A schematic diagram over 0 the secondary structure of TFPI is shown in copending patent application No. 07/828,920 (WO 91/01253). The amino acid sequence of TFPI with its natural 28 amino acid signal peptide (Sequence ID Number 1 and 2) is shown in Fig. 1 where the N- terminal amino acid Asp is given the number 1. The protein 5 consists of 276 amino acid residues and has, in addition to three inhibitor domains of the Kunitz type, three potential glycosylation sites at position Asnll7, Asnl67 and Asn228. The molecular weight indicates that some of these sites are gly¬ cosylated. Furthermore, it has been shown that the second 0 Kunitz domain binds FXa while the first Kunitz domain binds FVIIa/TF (T.J. Girard et al., Nature 338 (1989) 518-520). TFPI has also been isolated from HeLa cells (WO 90/08158) and it was shown that HeLa TFPI binds heparin.

In European patent application No. 0 487 591 certain TFPI 5 analogues are described as retaining this TFPI activity as well as anti Xa activity although parts of the molecule have been deleted. Furthermore, these analogues show a much lower af¬ finity for heparin than full-length TFPI, making them more useful as therapeutic agents than the native molecule. The TFPI analogues furthermore have a longer half life as compared with native TFPI which will further reduce the amount of active ingredients for the medical treatment. These TFPI analogues are thus characterized in having TFPI activity but with no or low heparin binding capacity unde physiological conditions (pH, ionic strength) .

The term "low heparin binding capacity" is meant to indicate binding capacity of about 50%, more preferably of about 25% an most preferably less than about 10% of that of native TFPI a physiological pH and ionic strength.

It was thus shown that the heparin binding capacity i substantially lost when the sequence from amino acid residu number 162 to amino acid residue number 276 is deleted from th TFPI molecule. It was therefore concluded that the hepari binding domain is located in this part of the TFPI molecule an it was assumed that the heparin binding domain comprises a least a region from Arg246 to Lys265 near the C-terminal end o the TFPI molecule which is rich in positively charged amin acid residues.

In European patent application No. 0 487 591 it is reporte that active TFPI analogues lacking C-terminal parts of th molecule surprisingly are expressed in good yields in yeast. These TFPI analogues contain at least the first and secon Kunitz domain and lack part of the C-terminal end of the nativ TFPI molecule, more specifically the third Kunitz domain fro amino acid Cysl89 to amino acid Cys239 and a substantial par of the amino acid sequence from Lys240 to Met276. By " substantial part" is meant from about 70% to 100%.

SUMMARY OF THE INVENTION

The present invention relates to a non-glycosylated TFPI analogue containing at least the first and second Kunitz domai and lacking the third Kunitz domain and a substantial part o the amino acid sequence from amino acid Lys240 to Met276 o native TFPI, said analogue being modified at either or both o the two N-glycosylation triads Asnll7-Glnll8-Thrll9 and Asnl67- Asnl68-Serl69 to avoid N-glycosylation.

The present invention is based on the surprising finding that non-glycosylated TFPI,,..₁₆₁ has the same activity as the glycosylated TFPI_,.₁₆₁ and on the finding that a glycosylation mutant of TFPI.,..,^ has pharmacokinetics which makes it very suitable for use as an anticoagulant for infusion. The half- life is significantly increased compared with the glycosylated variant produced in yeast. Therefore the amount of active ingredient in the pharmaceutical preparation can be reduced. On the other hand the half life is still sufficiently short to obtain a reasonably fast clearance of the protein in case of bleeding complications seen in some patients suffering from thrombosis.

It is well known from the literature that glycosylation modifications may change the pharmacokinetics of proteins (see e.g. P. Stanley, Glycobiology 2_. (1992) 99-107) . However, it is very difficult to predict the consequences of glycosylation changes. The biological activity of the protein may be changed and the biological half life may either be increased or decreased by such modifications.

The TFPI analogues may also contain a Ser residue as the N- terminal residue for efficient cleavage of a signal peptide by a signal peptidase. Thus, the N-terminal in the TFPI molecule may be replaced by a Ser or an additional Ser may be inserted adjacent to the original N-terminal residue. The TFPI analogues may furthermore lack part of the N-terminal sequence of native TFPI such as the sequence from amino acid residue 1 to 24.

Thus the present invention also relates to a non-glycosylated TFPI analogue containing at least the amino acid sequence from Phe25 to Glu 148 of the native TFPI molecule and lacking the third Kunitz domain from amino acid Cysl89 to amino acid Cys239 and a substantial part of the amino acid sequence from Lys240 to Met276 of the native TFPI molecule, said TFPI analogue being modified at either or both of the two N-glycosylation triads Asnll7-Glnll8-Thrll9 and Asnl67-Asnl68-Serl69 to avoid N- glycosylation.

More specifically the present invention relates to a non- glycosylated TFPI analogue containing at least the amino acid sequence from Aspl to Glul48 of the native TFPI molecule and lacking the third Kunitz domain from Cysl89 to Cys239 and a substantial part of the amino acid sequence from Lys240 to Met276 of the native TFPI molecule, said TFPI analogue being modified at either or both of the two N-glycosylation triads Asnll7-Glnll8-Thrll9 and Asnl67-Asnl68-Serl69 to avoid N- glycosylation.

In a particularly preferred embodiment the present invention relates to a non-glycosylated TFPI analogue lacking the amino acid sequence from Glnl62 to Met276 of the native TFPI molecule in yeast, said analogue being modified at the N-glycosylation triad Asnll7-Glnll8-Thrll9 to avoid N-glycosylation.

The modification of the N-glycosylation triad may be in the form of a deletion and/or substitution of one or more of the three amino acid residues of the triad. Asn in position 117 may thus be replaced by any other naturally occurring amino acid residue; Gin in position 118 may be replaced by Pro or Asp; Thr in position 119 may be replaced by any other naturally occurring amino acid residue except Ser or may be deleted and Asnll7, Glnllδ and Thr 119 may all be deleted. In the second N- glycosylation triad (Asnl67-Asnl68-Serl69) , Asnl67 may be replaced by any naturally occurring amino acid or may be deleted; Asnl68 may be replaced by Pro or Asp or may be deleted; Serl69 may be replaced by any naturally occurring amino acid residue except Thr or may be deleted and Asnl67, Asnl68, Serl69 may all be deleted. The modification according to the present invention is intended to cover any combination of such modifications. In a further aspect the present invention is related to a DNA sequence encoding the novel, non-glycosylated TFPI analogues.

The present invention furthermore relates to recombinant expression vectors comprising DNA sequences pemitting gene expression, including a promoter and a terminator, functionally fused to a DNA sequence encoding the TFPI analogue and capable of expressing the TFPI analogue according to the invention in a transformed or transfected eukaryotic host cell.

In a still further aspect, the present invention relates to eucaryotic cells containing a recombinant expression vector as defined above and to a method of making the novel, non- glycosylated TFPI analogues which process comprises culturing a eukaryotic cell line as defined above in a suitable nutrient medium under conditions permitting the expression of the TFPI analogues and recovering the resulting TFPI analogues from the culture.

Detailed description of the invention

The cDNA for the native TFPI has been cloned and sequenced (T.- C. Wun et al., J. Biol. Chem. 263 (1988) 6001-6004). DNA sequences encoding the TFPI analogues according to the present invention may be constructed by altering TFPI cDNA by site- directed mutagenesis using synthetic oligonucleotides in accordance with well-known procedures (cf. Sambrook et al., Molecular Cloning: A Laboratory Manual, 1989, Cold Spring Harbor, NY) .

The DNA sequence encoding the TFPI analogue of the invention may also be prepared synthetically by established standard methods. Thus, oligonucleotides may be synthesized by phosphoamidite chemistry in an automatic DNA synthesizer, purified, annealed, ligated and cloned in suitable vectors. The expression vector may be any vector which may conveniently be subjected to recombinant DNA procedures, and which is capable of expressing the TFPI analogues in the selected eukaryotic cell. Thus, the vector may be an autonomously replicating vector, i.e. a vector which exists as an extrachromosomal entity, the replication of which is independent of chromosomal replication, e.g. a plasmid. Alternatively, the vector may be one which, when introduced into a host cell, is integrated into the host cell genome and replicated together with the chromosome(s) into which it has been integrated.

In the expression vector, the DNA sequence encoding the TFPI analogue will be operably connected to a suitable promoter sequence. The promoter may be any DNA sequence which shows transcriptional activity in the host cell of choice and may be derived from genes encoding proteins either homologous or heterologous to the host cell. Examples of suitable promoters for directing the transcription of the DNA encoding the TFPI analogues of the invention in mammalian cells are the SV 40 promoter (Subramani et al., Mol.Cell Biol. 1 (1981) 854-864), the MT-1 (metallothionein gene) promoter (Palmiter et al.. Science 222 (1983) 809-814), and the adenovirus 2 major late promoter or the CMV (cytomegalovirus IE1) promoter (Henninghausen et al., EMBO J. 5 (1986) 1367-1371). Suitable yeast promoters include promoters from yeast glycolytic genes

(R.A. Hitzeman et al. , J.Biol.Chem. 255 (1980) 12073-12080; T.

Alber and G. Kawasaki, J.Mol.Appl.Gen. 1 (1982) 419-434) or alcohol dehydrogenase genes (T. Young et al., in Genetic

Engineering of Microorganisms for Chemicals (Hollaender et al., eds.), Plenum Press, New York, 1982, pp 335-361) or other highly expressed genes. Specific examples are the TPI1 promoter (T. Alber and G. Kawasaki, op.cit and US patent 4,599,311), or the ILV5 (J.G.L. Petersen and S. Holmberg, Nucl. Acids Res. JL4_. (1986) 9631-9651) promoter. The DNA sequence encoding the TFPI analogues may also be operably connected to a suitable terminator sequence which show transcription termination activity in a host cell. Examples of suitable terminators may be the human growth hormone terminator (Palmiter et al., op.cit.) . For yeast the terminator sequences may be derived from the 3 ' untranslated regions of yeast genes such as TPI1 (T. Alber and G. Kawasaki, op. cit.) and ILV5 (J.G.L. Petersen and S. Holmberg, op. cit.) . The vector may further comprise elements such as polyadenylation signals, transcriptional enhancer sequences and translational enhancer sequences.

It is preferred to express TFPI analogues in host cells that can secrete the analogues into the culture media. To direct the TFPI analogues into the secretory pathway of the host cell, a secretory signal sequence is operably linked to the TFPI analogue DNA sequence. The secretory signal should preferably be cleaved in vivo, e.g. by a signal peptidase or in yeast by the yeast KEX2 protease (D. Julius et al., Cell 2 (1984) 1075- 1089) during export of the fusion protein to allow for secretion of a TFPI analogue having the correct N-terminal amino acid. A suitable signal sequence for mammalian cells is the t-PA signal sequence (Friezner et al., J. Biol. Chem. 261 (1986) 6972-6985) . Suitable secretory signals for yeast include the α-factor prepropeptide (J. Kurjan and I. Herskowitz, Cell 10 (1982) 933-943; U.S. Patent No. 4,546,082 and EP 116,201), the PH05 signal peptide (WO 86/00637) , secretory signal sequences derived from the BAR1 gene (U.S. Patent No. 4,613,572 and WO 87/002670), the SUC2 signal peptide (M. Carlson et al., Mol. Cell. Biol. 1 (1983) 439-447) and the human serum albumin prepropeptide (A. Dugaiczyk et al., Proc. Natl. Acad. Sci. USA, 79 (1982) 71-75) . Alternatively, a secretory signal sequence may be synthesized according to the rules established, for example, by G. von Heijne (Nucl. Acids Res. .14. (1986) 4683- 4690) . Examples of synthetic secretory signal sequences are described in WO 89/02463 and WO 92/13065. Suitable yeast vectors include YRp7 (K. Struhl et al., Proc. Natl. Acad Sci. USA 76 (1987) 1035-1039), YEpl3 (J.R. Broach et al.. Gene 8 (1979) 121-133), POT vectors (U.S. Patent No. 4,931,373), pJDB249 and pJDB219 (J. Beggs, Nature 275 (1978) 5104-109) and derivatives thereof. Such vectors will generally include a selectable marker, which may be one of any number of genes that exhibit a dominant phenotype for which a phenotypic assay exists to enable transformants to be selected. Preferred selectable markers are those that complement host cell 0 auxotrophy, provide antibiotic resistance or enable a cell to utilize specific carbon sources, and include for yeast the genes LEU2 (Broach et al., op.cit.), URA3 (D. Botstein et al.. Gene 8 (1979) 17-24), HIS3 (K. Struhl et al., op.cit.) or POT1 (US Patent No. 4,931,373). For mammalian cells suitable 5 selectable markers are the gene coding for dihydrofolate reductase (DHFR) or one which confers resistance to a drug, e.g. neomycin, hygromycin or methotrexate.

The host cell may be any eukaryotic cell which is capable of producing the TFPI analogues and is preferably a mammalian cell 0 or a yeast cell. Examples of suitable mammalian cell lines are the COS (ATCC CRL 1650 and 1651), BHK (ATCC CRL 1632, ATCC CCL 10) or CHO (ATCC CCL 61) cell lines.

The yeast host cell may be any yeast species which is capable of producing the TFPI analogue. Examples of suitable yeast host 5 cells include strains of Saccharomyces spp., Schizosaccharo- mvces spp. Kluweromyces spp. , Pichia spp. and Hansenula spp. , in particular strains of Saccharomyces cerevisiae.

Methods of transfecting mammalian cells and expressing DNA sequences introduced in the cells are described in Kaufman and 0 Sharp, J. Mol. Biol. 159 (1982) 601-621; Southern and Berg, J. Mol. Appl. Genet. 1 (1982) 327-341; Loyter et al.. Proc. Natl. Acad. Sci. USA 29 (1982) 422-426; and Wigler et al., Cell 14 (1978) 725. Techniques for transforming yeast are well known in the literature, and have been described for instance by Beggs (op.cit.) . The genotype of the host cell will generally contain a genetic defect that is complemented by the selectable marker present on the expression vector. Choice of a particular host and selectable marker is well within the level of ordinary skill in the art. To optimize production of heterologous proteins, it is preferred that the host strain carry a mutation, such as the yeast pep4 mutation (E.W. Jones, Genetics 5_. (1977) 23-33) , which results in reduced proteolytic activity.

The recombinant expression vector of the invention may further comprise a DNA sequence enabling the vector to replicate in the host cell in question. An example of such a sequence is the yeast 2-micron sequence and the SV40 origin (for mammalian cells) .

The procedures used to ligate the DNA sequences coding for the TFPI analogues of the invention, the promoter and the terminator, respectively, and to insert them into suitable vectors containing the information necessary for replication, are well known to persons skilled in the art (cf. Sambrook et al. , supra) .

The transformed or transfected host cells are grown according to standard methods in a growth medium containing nutrients required for growth of the particular host cells. A variety of suitable media are known in the art and generally include a carbon source, a nitrogen source, essential amino acids, vitamins, minerals and growth factors. The growth medium will generally select for cells containing the DNA construct by, for example, drug selection or deficiency in an essential nutrient which is complemented by the selectable marker on the DNA construct or co-transfected with the DNA construct. Suitable growth conditions for yeast cells, for example, include culturing in a medium comprising a nitrogen sourc (e.g. yeast extract or nitrogen-containing salts) , inorgani salts, vitamins and essential amino acid supplements a necessary at a temperature between 4°C and 37°C, with 30° being particularly preferred. The pH of the medium i preferably maintained at a pH greater than 2 and less than 8, more preferably pH 5-6.

The medium used to culture mammalian cells may be an conventional medium suitable for growing mammalian cells, suc as a serum-containing or serum-free medium containin appropriate supplements. Suitable media are available fro commercial suppliers or may be prepared according to publishe recipes (e.g. in catalogues of the American Type Cultur Collection) .

The TFPI analogues will preferably be secreted to the growt medium and may be recovered from the medium by conventiona procedures including separating the host cells from the mediu by centrifugation or filtration, precipitating th proteinaceous components of the supernatant or filtrate b means of a salt, e.g ammonium sulphate, followed b purification by a variety of chromatographic procedures, e.g. ion exchange chromatography, affinity chromatography, or th like.

The present invention also relates to a pharmaceutica composition comprising a TFPI analogue of the inventio together with a pharmaceutically acceptable carrier or diluent. In the composition of the invention, the TFPI analogue may b formulated by any of the established methods of formulatin pharmaceutical compositions, e.g. as described in Remington' Pharmaceutical Sciences. 1985. The composition may typically b in a form suited for systemic injection of infusion and may, a such, be formulated with sterile water or an isotonic saline o glucose solution. The compositions may be sterilized b conventional lyophilized preparation being combined with the sterile aqueous solution prior to administration. The composition may contain pharmaceutically acceptable auxiliary substances as required to approximate physiological conditions, such as buffering agents, tonicity adjusting agents and the like, for instance sodium acetate, sodium lactate, sodium chloride, potassium chloride, calcium chloride, etc. The concentration of the TFPI analogue of the invention may vary widely, i.e. from less than about 0.5%, such as from 1%, to as much as 15-20% by weight. A unit dosage of the composition may typically contain from about 0.1 to about 100 mg of the TFPI analogue of the invention.

The pharmaceutical preparations may be in a buffered aqueous solution with appropriate stabilizers and preservatives. The solution may be heat treated and may be contained in ampoules or in carpoules for injection pens. Alternatively the stabilized solution may be freeze dried and contained in ampoules or in two chamber injection systems with freeze dried substance in one chamber and solvent in the other chamber.

The TFPI analogue of the invention is contemplated to be advantageous to use for the therapeutic applications suggested for full-length TFPI. These include, but are not limited to treatment of patients with coagulation disorders or cancer as described in European patent application No. 0 487 591. A specific coagulation disorder which may be treated with the TFPI analogue is disseminated intravascular coagulation (DIC) which is a common and serious complication occurring in patients with sepsis, trauma, burns, haemolytic anemia, metastatic cancer, etc. DIC is characterized by fibrin deposition in various organs, e.g. kidneys, and by a decrease in the levels of coagulation factors (fibrinogen, FVII etc.), coagulation inhibitors (e.g. antithrombin III) and platelets. It may result from exposure of tissue factor (TF) on the surface of various cells, e.g. monocytes and endothelial cells leading to formation of complexes between TF and activated factor VII (FVIIa) resulting in activation of the coagulation system by the extrinsic pathway. Due to the longer half-life in the circulation, it is estimated that a lower dosage of the TFPI analogue may be required to obtain the same anticoagulant effect as with full-length TFPI.

Brief description of the Drawings

The present invention is further described with reference to the drawings in which

Fig. 1 shows a synthetic gene and the corresponding amino acid sequence for human TFPI including the signal peptide.

Fig. 2 shows DNA sequences and corresponding amino acid sequences for the prepropeptide of human serum albumin pp_HSA (Sequence ID Number 3 and 4) and the synthetic secretion signal 212spx3 (Sequence ID

Number 5 and 6) fused to the N-terminal of the mature form of TFPI. Only the three N-terminal amino acids of TFPI are shown.

Fig. 3 shows the synthetic gene for TFPI₁.₁₆₁-ll7Gln fused to the synthetic secretion signal 212spx3 (Sequence ID

Number 7 and 8) ,

Fig. 4 shows restriction site maps of plasmid pY-ppTFPI161 and plasmid pP-212TFPI161-117Q (the map of the third expression plasmid described in this study, pP- 212TFPI161, is similar to that of pP-212TFPI161-

117Q) , and

Fig. 5 shows a Western analysis of secreted TFPI₁.₁₆₁ and its

Asnll7 to Gin substitution. Lane 1, molecular weight markers; lane 2, purified TFPI.,..,^ secreted from transformant YNG452[pY-ppTFPI161] (2 U) ; lane 3, affinity-purified TFPI₁._l61-ll7Gln from transfor ant E18[pP-212TFPI161-117Q] (2 U) . Immunodetection was performed with polyclonal anti-TFPI antibodies.

Fig. 6 illustrates the construction of the URA3-2_ yeas expression plasmid pYES-GykTFPI161-117Q encoding a fusion protein consisting of the synthetic secretion peptide yk and TFPI₁.₁₆₁-117Q. The sizes of the plasmids are given in base pairs. Only relevant restriction endonuclease sites are shown.

Fig. 7 illustrates the construction of five URA3-2u yeast plasmids derived from pYES-GykTFPI161-117Q for expression of secreted unglycosylated two-domain TFPI analogues with different polypeptide lengths. The four TFPI expression plasmids not depicted by drawings are very similar to pYES-GykTFPI161-117Q and pYES-GykT21-161-Q differing only in the coding region for the TFPI precursors. The sizes of the plasmids are given in base pairs. Only relevant restriction endonuclease sites are shown.

The invention is further described in the following examples which are not in any ways intended to limit the scope of the invention as claimed.

Experimental Part

Materials and Methods

Standard recombinant DNA techniques were carried out as described (T. Maniatis et al.. Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, 1982). Synthetic oligonucleotides were synthesized by the phosphora- idite method using an Applied Biosystems DNA Synthesizer Model 380B.

Restriction endonucleases and T4 DNA ligase were obtained from New England Biolabs. Modified T7 DNA polymerase (Sequenase) was obtained from United States Biochemicals. Restriction endonucleases and other enzymes were used in accordance with the manufacturers recommendations. pBS-i- (Stratagene) was used as cloning vector for construction of the synthetic TFPI gene by cloning of synthetic DNA fragments.

E. coli strains XL-1 Blue (Stratagene) and MC1061 (M.J.

Casadaban and S.H. Cohen, J. Mol. Biol. H (1980) 179-207) were used as bacterial recipients for plasmid transformations and as hosts for propagation and preparation of plasmid DNA.

Strains of Saccharomyces cerevisiae used as hosts for expression of TFPI analogues were the two diploids E18 (MATa/MAT Δtpi: :LEU2/Δtpi: :LEU2 Ieu2/leu2 +/his4 pep4-3/pep4- 3) (US Patent No. 4,931,373) and YNG452 (MATα/MATα ura3- 52/ura3-52 Ieu2-Δ2/Ieu2-Δ2 +/his4 pep4-Δl/pep4-Δl) . The latter was derived from strain JC482 (J.F. Cannon and K. Tatchell, Mol. Cell. Biol. 2 (1987) 2653-2663).

Yeast expression vectors used for expression of TFPI analogues in yeast were of the POT-type (US Patent No. 4,931,373) or the URA3-LEU2d-2μ plasmid pAB24 (P.J. Barr et al., in Proc. Alko Symp. on Industrial Yeast Genetics (Korkola and Nevalainen, eds.) Found. Biotech. Industr. Ferment. Res. 5_ (1987) 139-148).

DNA sequences were determined by the dideoxy chain termination method (Sanger et al., Proc. Natl. Acad. Sci. 7_4 (1977) 5463- 5467) using double stranded plasmid DNA as template and ³P- or ³⁵S labelled primers and Sequenase. SDS polyacrylamide gel elecrrophoresis under reducing conditions was performed according to U.K. Laemmli (Nature 227 (1979) 680-685) using 12.5% separating gels. Protein was stained with Coomassie Brillant Blue R-250 (Sigma) .

Western blot analysis was carried out after electrophoresis as described by J. Mikkelsen and J. Knudsen (Biochem. J. 248 (1987) 709-714) . Proteins were stained with Auro Dye and TFPI was immunodetected by rabbit antiserum raised against full length human TFPI (O. Nordfang et al., Thromb. Haemostas. j56_ (1991) 464-467) .

Affinity purification of the TFPI analogues was carried out from culture supernatants by affinity chromatography using polyclonal anti-TFPI immunoglobulin G coupled to Sepharose (O. Nordfang et al., Biochemistry 10_. (1991) 10371-10376).

N-terminal sequence analysis was carried out by automated Edman degradation using an Applied Biosystems 470A gas-phase sequencer. Analysis by on-line reverse-phase HPLC was performed for the detection and quantification of the liberated PTH amino acids from each sequence cycle.

TFPI activity was measured in a chromogenic microplate assay, modified according to the method of Sandset et al., (Thromb. Res. 4_7 (1987) , 389-400) . Heat treated plasma pool was used as a standard. This standard is defined as containing 1 U/ml of TFPI activity. Standards and samples were diluted in buffer A (0.05 M Tris-HCl, 0.1 M NaCl, 0.1 M Na-citrate, 0.02% NaN₃, pH 8.0) containing 2 g/ml polybrene and 0.2% bovine serum albumin. FVIIa/TF/FX/CaCl₂ combination reagent was prepared in buffer A and contained 1.6 ng/ml FVIIa (Novo Nordisk A/S), human tissue factor diluted 60 fold, 50 ng/ml FX (Sigma) and 18 mM CaCl₂. The assay was performed in microplate strips at 37°C. 50 μl of samples and standards were pipetted into the strips and 100 μl combination reagent was added to each well. After 10 minutes incubation, 25 μl of FX (3.2 μ.g/ml) was added to each well and after another 10 minutes 25 μl of chromogenic substrate for FXa (S2222) was added 10 minutes after the ad¬ dition of substrate. The reaction was stopped by addition of 50 μl 1.0 M citric acid pH 3.0. The microplate was read at 405 nm.

5 The inhibitory activity of TFPI analogues in the extrinsic pathway of coagulation was measured in PT clotting assay using human plasma and diluted human thromboplastin (O. Nordfang et al. , op.cit. ) .

Example 1

° Expression of TFPI_{1 161} and TFPI_{1 161}-ll7Gln in yeast

In EP Patent Application No. 0487591, a synthetic gene coding for human TFPI with its 28 amino acid signal peptide was described. The DNA sequence was derived from the published sequence of a cDNA coding for human TFPI (Wun et al., J. Biol. 5 Chem. 263 (1988) 6001-6004) . The synthetic gene was assembled by the step-wise cloning of synthetic restriction fragments into plasmid pBS(+). The resulting gene was contained on a 922 base pair (bp) Sail restriction fragment. The gene had 26 silent nucleotide substitutions in degenerate codons as 0 compared to the cDNA resulting in fourteen unique restriction endonuclease sites in order to facilitate the introduction of mutations in TFPI as well as the in-frame insertion of new secretion signals at the N-terminal of mature TFPI. The DNA sequence of the 922 bp Sail fragment and the corresponding 5 amino acid sequence of human TFPI (pre-form) is shown in Fig. 1.

Using standard DNA manipulation technology the coding sequences for TFPI analogues according to the present invention were constructed from the synthetic TFPI gene by replacing portions 0 of the TFPI gene with appropriate synthetic DNA fragment. The DNA fragments were annealed oligodeoxynucleotides synthesized by phosphoramidite chemistry. Resulting plasmids were propagated in E. coli and the nucleotide sequences verified by DNA sequencing. TFPI.,..₁₆₁ was expressed as a fusion protein with an N-terminal addition of 24 amino acids corresponding to the prepropeptide of human serum albumin (Fig. 2) , or with the synthetic secretion sequence 212spx3 (Fig. 2) in order to obtain secretion and in vivo processing of the secretion sequences in yeast resulting in TFPI₁.₁₆₁ with the correct N- terminal being delivered to the growth medium. The analogue TFPI₁.₁₆₁ has one consensus-site, Asnll7, for the addition of N- linked carbohydrate characteristic of many eukaryotic cells like mammals and fungi, including yeast (M.A. Kukuruzinska et al., Ann. Rev. Biochem. 55 (1987) 915-944). In order to avoid N-glycosylation of TFPI.,_..,^, we constructed a gene for a variant, TFPI₁.₁₆₁-ll7Gln in which the AAT-codon for Asnll7 (see Fig. 3) had been changed to CAA coding for Gin. In this case, only the synthetic secretion sequence 212spx3 was used for secretion of correctly processed TFPI-analogue.

In order to express the TFPI genes in S. cerevisiae. two yeast expression plasmids for TFPI,.^,, pY-ppTFPI161 and pP- 212TFPI161, were constructed, while one plasmid, pP-212TFPI161- 117Q, was constructed for TFPI₁.₁₆₁-ll7Gln. The restriction site maps for two of these plasmids are shown in Fig. 4. High-level expression was achieved by placing the genes behind the strong constitutive promoters of the TPIl or ILV5 genes of S. cerevisiae. Transcription termination sequences were derived from the same genes. Plasmid pY-ppTFPI161 (Fig. 4) was based on the yeast vector pAB24 (P.J. Barr et al., in Proc Alko Symposium on Industrial Yeast Genetics, Korhola and Nevalainen, eds., Found. Biotech. Industr. Ferment. Res. 5. (1989) 139-148), while plasmids pP-212TFPI161 and pP-212TFPI161-117Q (Fig. 4) were based on a vector of the POT-type (G. Kawasaki and L. Bell, US patent 4,931,373). All expression plasmids carried, in addition to selective markers for transformation of the plasmids into suitable host strains of S. cerevisiae. DNA sequences of the 2-micron plasmid of yeast for high plasmid- copy numbers in yeast.

Plasmid pY-ppTFPI161 was transformed into strain YNG452 derived from strain JC482 (J.F. Cannon and K. Tatchell, Mol. Cell. Biol. 2 (1987) 2653-2663)) under selection for uracil independence (H. Ito et al., J. Bacteriol. 153 (1983) 163-168). Plasmids pP-212TFPI161 and pP-212TFPI161-117Q were transformed into strain E18 selecting for ability to grown on media with glucose as the carbon source by complementation of the disrupted triose phosphate isomerase gene with the POT-marker (P.R. Russell, Gene 40 (1985) 125-130.

In subsequent expression studies, individual transformants were grown in shake flasks at 30°C containing liquid selective media, namely synthetic complete medium lacking uracil for the YNG452 transformants and rich glucose medium with the E18 transformants. Upon growth to stationary phase, the cells were removed by centrifugation and the supernatant media analyzed for secreted TFPI activity using the chromogenic assay. The result of this experiment is shown in Table 1.

Table 1

Shake-flask expression studies with yeast transformants secreting TFPI,.^ and TFPI₁.₁₆₁-ll7Gln

Transformant TFPI analogue Secreted TFPI activity (U/10⁸ cells)

The results shown in Table I shows that significant levels of TFPI activity secreted to the growth medium were observed when

TFPI 1-161 was expressed in strain YNG452 as a fusion to the prepropeptide of HSA. However, more than a 10-fold increase in secreted activity was seen when the same analogue was expressed in strain E18 as a fusion to the 212spx3 prepropeptide, and this increase in activity was unaffected by substitution of Asnll7 to Gin.

Example 2 R Reellaattiivvee <anticoagulant activities of TFPI_{1 161} and TFPI., ₁₆₁-ll7Gln in a PT-clotting assay

In order to compare the anticoagulant properties of TFPI_1.161 and the corresponding Asnll7 to Gin substitution, the two analogues were partially purified by affinity-chromatography (O. Nordfang et al., Biochemistry 3_0 (1991) 10371-10376; 0. Nordfang et al., Thromb. Haemostas. 66 (1991) 464-467) from the culture supernatants of transformants YNG452[pY-ppTFPI161] and E18[pP- 212TFPI161-117Q] , respectively, and their ability to inhibit extrinsic pathway coagulation determined in a PT clotting assay with human plasma (O. Nordfang et al., Biochemistry lϋ (1991) 10371-10376) . Anticoagulant units were normalized to chromogenic TFPI activity (P.M. Sandset et al., Thromb. Res. 4 (1987) 389-400) , which was assumed to be similar for the different TFPI forms. Both activity determinations were calculated as the mean of measurements carried out in duplicate. TFPI in human plasma which was used as standard was defined to have a relative anticoagulant activity of 1. By this method we determined a relative anticoagulant activity of about 0.03 for both TFPI,.,^ and TFPI₁.₁₆₁-ll7Gln (Table II). Thus, the amino acid substitution did not affect the anticoagulant activity. Their activity is about 50-fold lower than full length TFPI expressed in BHK cells (A.H. Pedersen et al., op.cit) .

Table II

Relative anticoagulant activity of TFPI analogues

TFPI polypeptide TFPI PT-anti- Relative anti¬ (units/ml) activity coagulant coagulant

(U/ml) activity activity

TFPI 1-161

TFPI₁.₁₆₁-ll7Gln

TFPI (from BHK)

Example 3

Purification and molecular characterization of secreted TFPI_{1 1 1} and TFPI_{1 1 1}-ll7Gln

In order to characterize the two TFPI analogues, transformants YNG452[pY-ppTFPI161] and E18[pP-212TFPI161-117Q] were grown in pilot- or laboratory-scale fermentors in fed-batch processes with glucose, and the analogues purified from the supernatant medium by a combination of ion exchange chromatography and gel- filtration. Subsequent analysis of similar activity amounts (P.M. Sandset et al., op.cit) by SDS-polyacrylamide gel electrophoresis and Western blotting using an antiserum raised against an N-terminal peptide of TFPI (A.H. Pedersen et al., op.cit.) showed 1) that TFPI,.,^ and TFPI₁.₁₆₁-ll7Gln had very similar specific activities on a molar basis, as judged from the staining intensities (Fig. 5) ; 2) an apparent molecular mass of about 25 kDa for TFPI.,.^.,, while that of TFPI₁.₁₆₁-ll7Gln was about 22 kDa. These results strongly suggested that TFPI₁.₁₆₁ contained N-linked oligosaccharides which were not present in TFPI₁.₁₆₁-ll7Gln, and that neither the amino acid substitution nor the modification of Asnll7 by N-linked glycosylation affected the FXa/TF/FVIIa-dependent activity of TFPI.,..,^.

In order to assure that the observed difference in apparent molecular mass for TFPI.,.₁₆₁ and TFPI₁._l61-ll7Gln was not a result of improper processing of the HSA and 212spx3 prepropeptides during secretion of the TFPI analogues in yeast, the N-terminal amino acid sequences were determined on purified preparations of the analogues or on proteolytic fragments thereof. In both cases the expected N-terminal sequence for mature TFPI was obtained.

Example 4

TFPI₁.₁₆₁ and TFPI₁.₁₆₁-ll7Gln were prepared as described in the preceding examples and dissolved in 10 mM glycylglycine, 100 mM NaCl, 30 g/1 mannitol, pH 7.0 to a concentration of 1 mg/ l. Six female rabbits (New Zealand) with a mean weight of 2.56 kg were anaesthetized with pentobarbital sodium. Test compounds were administered via a catheter placed in vena jugularis, and blood samples were obtained from a catheter placed in a. carotis on the opposite side. The first 5 ml of blood were discarded. Two groups of rabbits were treated with either TFPI,,. ₁₆₁ or TFPI₁._l61-ll7Gln and 1.8 ml blood samples were obtained at t = -10, 2, 5, 10, 20, 30, 45, 80, 120 and 180 minutes. TFPI activity was measured in the chromogenic activity assay. Alpha- and beta half-lives, clearance and mean residence time were calculated by non-liniar regression by using a two- compartmental model. The fittings were performed by use of the SIMPLEX procedure written in a program adopted from K. Yamaoka et al., (A pharmacokinetic analysis program ( ulti) for microcomputer. J. Pharm. Dyn. 4. (1981) 879) . The following pharmacokinetic parameters were obtained:

Table III

Pharmacokinetics of glycosylated and unglycosylated TFPI₁.₁₆₁ in rabbits.

As seen in Table III, the β half life of the unglycosylated variant was increased 3-4 fold compared with the glycosylated form and the clearance rate was reduced by half. Thus the amount of TFPI 1-161 needed to keep a steady state plasma level will be reduced two fold by using the unglycosylated variant (117Gln) . However, the clearance rate is sufficient to obtain clearance in case of bleeding complications.

Example 5

Expression of different non-glycosylated TFPI analogues lacking the third Kunitz domain in yeast

In order to produce TFPI analogues of different polypeptide length and characterized by lacking both N-linked glycosylation and the third Kunitz domain, the following TFPI analogues were expressed in yeast:

TFPI₁.₁₆₁-ll7Gln

^TFPI1-16CT^ll7Gln TFPI₁._u9-ll7Gln

TFPI₂₁.₁₆₁-ll7Gln

TFPI₂₁.₁₆₀-ll7Gln

TFPI₂₁.₁₄₉-ll7Gln.

The six TFPI analogues were expressed as fusion proteins with 10 an N-terminal addition of a 54 amino acid, Kex2-cleavable synthetic secretion peptide, denoted yk. The secretion sequence consisted of the putative 21 amino acid signal peptide of the aspartyl protease of S. cerevisiae encoded by the YAP3 gene (M.

Egel-Mitani, H.P. Flygenring and M.T. Hansen, Yeast 6. (1990) 15127-137) followed by the synthetic leader peptide from amino acids -33Gln to -lArg of the 212spx3 secretion sequence

(Fig.2) .

As a first step in the construction the yeast plasmids encoding the analoques, a yeast plasmid encoding TFPI₁.₁₆₁-ll7Gln fused to

20 the yk secretion sequence was constructed in the following manner (Fig.6). Plasmid pP-212TFPI161-117Q (Fig.4 and Fig.6) was digested with restriction endonucleases Sphl and Xbal in a double digestion, and the 1.1 kb Sphl-Xbal fragment consisting of the TPIl promoter and the coding region for the 212spx3

25 secretion peptide/TFPI₁.₁₆₁-H7Gln fusion protein (Fig.3) was isolated. The fragment was inserted into a URA3-2/- yeast vector pYES21. This vector is a derivative of yeast vector pYES2.0 (Stratagene) from which the GAL1 promoter had been removed. The resulting plasmid, pYES-212TFPI161-117Q, consisted of a TPIl

30 promoter/CYCl terminator expression cassette with the gene for the 2l2spx3/TFPI₁.₁₆₁-ll7Gln precursor, 2-micron sequences for high copy-number replication in yeast, the yeast URA3 gene for plasmid selection in ura3 mutants, the jø-lactamase gene for selection of ampicillin resistant clones in E. coli. the ColEl origin of replication in E. coli, and the fl origin for recovery of single-stranded DNA plasmid from superinfected J . coli strains (Fig.6). The plasmid was cleaved with EcoRI and PflMI in order to remove the coding region for the modified α-amylase signal peptide of the 212spx3 secretion sequence. Subsequent insertion of a synthetic double-stranded EcoRI-PflMI oligonucleotide with codons for the YAP3 signal peptide created the DNA sequence encoding the yk secretion sequence fused to the N-terminus of TFPI₁.₁₆₁-ll7Gln. The gene fusion was assured by DNA sequencing around the EcoRI and PflMI sites. Finally, the TPIl promoter fragment was replaced by a DNA fragment containing a 0.44 kb fragment of the promoter for the glyceraldehyde-3-phosphate dehydrogenase gene GPP (G3PDA) of S. cerevisiae (J.P. Holland and M.J. Holland, J. Biol. Chem. 254 (1979) 9839-9845; G.A. Bitter and K.M. Egan, Gene 12. (1984) 263-274) with an Sphl site inserted immediately upstream of position -452 and an EcoRI site immediately downstream of position -12, as this promoter was expected to be somewhat stronger than the TPIl promoter. The resulting 6.4 kb yeast plasmid pYES-GykTFPI161-117Q with the gene for the yk/TFPI^,^- 117 Gin fusion protein under the control of the GPP promoter is shown in Fig. 6. The plasmid harbored a convenient distribution of relatively few restriction endonuclease sites, thus facilitating the construction of plasmid derivatives encoding the other TFPI analogues.

To construct yeast plasmids encoding C-terminal truncations of the TFPI.,_₁₆₁-117 Gin precursor, pYES-GykTFPI161-117Q was digested with Xhol and Xbal (Fig. 7) . This digestion removed the coding sequence for amino acids 142 to 161 in TFPI₁.₁₆₁ (see Fig.3; an Xbal site is located 6 nucleotides downstream of the translational stop codon) . Insertion of a synthetic double- stranded Xhol-Xbal oligonucleotide restored the TFPI coding sequence with a stop codon after amino acid 160Gly resulting in a plasmid encoding ykTFPI₁.₁₆₀-H7Gln (plasmid pYES-GykT160-Q) . Similarly, insertion of a different synthetic double-stranded Xhol-Xbal oligonucleotide introduced a stop codon after amino acid 149Asp resulting in a plasmid encoding ykTFPI₁. ₉-H7Gln (plasmid pYES-GykT149-Q) .

To construct yeast plasmids with N-terminal truncations of TFPI₁.₁₆₁-ll7Gln, the coding sequence for the ykTFPI₁.₁₆₁-H7Gln fusion was isolated as a 0.68 kb EcoRI-Xbal fragment from pYES- GykTFPI161-117Q and inserted into the polylinker region of plasmid pUC19 (C. Yanisch-Perron, J. Viera and J. Messing, Gene 33 (1985) 103-119) (Fig.7). Cleavage of the resulting plasmid pUC19-ykT161-Q with Ncol and Nsil removed the DNA sequence encoding the C-terminal 3 amino acids of the synthetic leader peptide and amino acids 1 to 21 of TFPI. A double-stranded synthetic Ncol-Nsil oligonucleotide was inserted in order to restore the codons in the leader region and amino acid 21 of TFPI resulting in an in-frame gene fusion linking the C- terminal LysArg of the secretion leader to amino acid 21Leu of TFPI (plasmid pUC19-ykT21-161-Q) . In-frame fusion was verified by DNA sequencing around the Ncol and Nsil sites. The ykTFPI₂₁. ₁₆₁-117 Gin coding sequence was isolated from the plasmid as a 0.62 kb EcoRI-Xbal fragment and reinserted behind the GPP promoter by cloning into pYES-GykTFPI161-117Q digested with EcoRI and Xbal. The resulting yeast plasmid, pYES-GykT21-161-Q, is shown in Fig.7.

Yeast plasmids with C-terminal truncations of the precursor ykTFPI₂₁.₁₆₁-H7Gln were constructed in a manner similar to the plasmids for C-terminal truncations of the ykTFPI_l.₁₆₁-H7Q described above. Thus, pYES-GykT21-161-Q was digested with Xhol and Xbal and the two different synthetic double-stranded oligonucleotides inserted in order to introduce stop codons after amino acids 160Gly or 149Asp. This resulted in plasmids pYES-GykT21-160-Q and pYES-GykT21-149-Q, encoding the TFPI_21.160- 117Gln and TFPI₂₁._U9-ll7Gln fusions, respectively (Fig.7).

The pYES plasmids encoding the six TFPI analogues fused to the yk secretion sequence were transformed into the haploid S. cerevisiae strain YNG318 (genotype MATα ura3-52 Ieu2-Δ2 his4 pep4-Δl; an isogenic derivative of strain JC482 (J.F. Cannon and K. Tatchell, Mol. Cell. Biol. 2 (1987) 2653-2663)). The plasmids were introduced by the alkali cation transformation procedure (H. Ito, Y. Fukuda, K. Murata and A. Kimura, J. Bacteriol. 153 (1983) 163-168) selecting for growth of transformant colonies on agar-containing medium lacking uracil. As a control plasmid not encoding TFPI polypeptide was used pYES21. Reisolated transformants were grown to stationary phase in 50 ml of synthetic complete medium lacking uracil (SC-ura) for 3 days at 30 °C with shaking. After measuring of the cell densities (A₆₀₀) , the cultures were centrifuged, and FXa/TF/FVIIa-dependent chromogenic TFPI activity in the resulting supernatant media determined. The result is shown in Table IV.

Table IV.

Shake flask expression study with six non-glycosylated two-domain TFPI analogues expressed in yeast strain YNG318. Secreted TFPI activity (calculated as units TFPI per ml and normalized to cell density at the time of harvest) is the mean of 2-3 independent growth experiments. Cell densities of 7-13 (A₆₀₀) were obtained, nt, not tested.

Plasmid TFPI-117 Gin Secreted TFPI analogue activity (U/A₆₀₀-ml )

pYES-GykTFPI161-117Q TFPI₁.₁₆₁-ll7Gln 0.35 pYES-GykT160-Q TFPI₁.₁₆₀-ll7Gln 0.57 pYES-GykT149-Q TFPI_1.149-ll7Gln nt. pYES-GykT21-161-Q TFPI_{21 161}-ll7Gln 0.33 p YES -Gy kT21 - 160 -Q ^TFPI21-160-^{l l7Gln} 0.36 pYES-GykT21-149-Q TFPI₂₁. ₉-ll7Gln 0.21 pYES21 none <0.004

As seen in Table VI the tested TFPI-117 Gin analogues are effectively secreted by yeast transformants, and they are produced in active form. The 2-3 fold difference in activity levels observed for the analogues may be due to differences in e.g. gene expression levels, the amounts of TFPI polypeptides secreted, different specific activities, or it may reflect experimental variance. In order to compare directly the antithrombotic properties, the analogues can be purified from the culture supernatants and characterized further.

SEQUENCE LISTING

(1) GENERAL INFORMATICN:

(i) APPLICANT: Petersen, Jens G. Litske Nordfang, Ole Juul Søren Erik Bjørn

(ii) TITLE OF INVENTION: Method for Making non-glycosylated TFPI Analogues

(iii) NUMBER OF SEQUENCES: 8

(iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE: Nσvo Nordisk of North America, Inc.

(B) STREET: 405 Lexington Avenue, Suite 6200

(C) CITY: New York

(D) STATE: N. Y.

(E) COUNTRY: United States of America

(F) ZIP: 10174-6201

(V) CEMFUTER READABLE FORM:

(A) MEDIUM TYPE: Floppy disk

(B) COMPUTER: IBM PC compatible

(C) OPERATING SYSTEM: PC-DOS/MS-DOS

(D) SOFTWARE: Patentln Release #1.0, Version #1.25

(vi) CURRENT APPLICATION DATA:

(A) APPLICATION NUMBER:

(B) FILING DATE:

(C) CLASSIFICATION:

(vii) PRIOR APPLICATION DATA:

(A) APELICAΠON NUMBER: US 07/828,920

(B) FILING DATE: 27-JAN-1992

(viii) ATTORNEY/AGENT _NPORMA_TON:

(A) NAME: Agris, Cheryl H.

(B) REGISTRATION NUMBER: 34086

(C) REFERENCE/DOCKET NUMBER: 3967.000-US

(ix) TE_EOO-__m_AT_ON INFOPMATICN:

(A) TELEPHONE: 212 867 0123

(B) TELEFAX: 212 867 0298 (2) INFORMATION FOR SEQ ID NO:l:

(i) SEQUENCE CHARA<TERIS_TCS:

(A) LENGTH: 928 base pairs

(B) TYPE: nucleic acid

(C) STRAN EPNESS: single

(D) TOPOLOGY: linear

(ii) MDLECULE TYPE: DNA

(ix) FEATURE:

(A) NAME/KEY: CDS

(B) LOCATION: 8..919

(ix) FEATURE:

(A) NAME/KEY: sig_peptide

(B) LOCATION: 8..91

(ix) FEATURE:

(A) NAME/KEY: mat_peptide

(B) LOCATION: 92..919

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:l:

GTCGACC ATG ATT TAC ACA ATG AAG AAA GTA CAT GCA CT TGG GCT AGC 49 Met lie Tyr Thr Met Lys Lys Val His Ala Leu Trp Ala Ser -28 -25 -20 -15

GTA TGC CTG CTG CTT AAT CTT GCC CCT GCC CCT CTT AAT GCT GAT TCT 97 Val Cys Leu Leu Leu Asn Leu Ala Pro Ala Pro Leu Asn Ala Asp Ser -10 -5 1

GAG GAA GAT GAA GAA CAC ACA ATT ATC ACA GAT ACG GAG CTC CCA CCA 145 Glu Glu Asp Glu Glu His Thr lie lie Thr Asp Thr Glu Leu Pro Pro 5 10 15

CTG AAA CTT ATG CAT TCA TTT TGT GCA TTC AAG GCG GAT GAT GGG CCC 193 Leu Lys Leu Met His Ser Phe Cys Ala Phe Lys Ala Asp Asp Gly Pro 20 25 30

TGT AAA GCA ATC ATG AAA AGA TTT TTC TTC AAT ATT TTC ACT OGA CAG 241

Cys Lys Ala lie Met Lys Arg Phe Phe Phe Asn lie Phe Thr Arg Gin

35 40 45 50

TGC GAA GAA TTT ATA TAT GGG GGA TGT GAA GGA AAT CAG AAT OGA TTT 289 Cys Glu Glu Phe lie Tyr Gly Gly Cys Glu Gly Asn Gin Asn Arg Hie 55 60 65

GAA AGT CTG GAA GAG TGC AAA AAA ATG TGT ACA AGA GAT AAT GCA AAC 337 Glu Ser Leu Glu Glu Cys Lys Lys Met Cys Thr Arg Asp Asn Ala Asn 70 75 80 AGG ATT ATA AAG ACA ACA CTG CAG CAA GAA AAG OCA GAT TTC TGC TTT 385 Arg He He Lys Thr Thr Leu Gin Gin Glu Lys Pro Asp Phe Cys Phe 85 90 95

TTG GAA GAG GAT CCT GGA ATA TGT OGA GGT TAT ATT ACC AGG TAT TTT 433 Leu Glu Glu Asp Pro Gly He Cys Arg Gly Tyr He Thr Arg Tyr Phe 100 105 110

TAT AAC AAT CAG ACA AAA CAG TGT GAA AGG TTC AAG TAT GGT GGA TGC 481 Tyr Asn Asn Gin Thr Lys Gin Cys Glu Arg Fhe Lys Tyr Gly Gly Cys 115 120 125 130

CTG GGC AAT ATG AAC AAT TTT GAG ACA CTC GAG GAA TGC AAG AAC ATT 529 Leu Gly Asn Met Asn Asn Fhe Glu Thr Leu Glu Glu Cys Lys Asn He 135 140 145

TGT GAA GAT GGT COG AAT GGT TTC CAG GTG GAT AAT TAT GGT ACC CAG 577 Cys Glu Asp Gly Pro Asn Gly Fhe Gin Val Asp Asn Tyr Gly Thr Gin 150 155 160

CTC AAT GCT GTT AAC AAC TCC CTG ACT COG CAA TCA ACC AAG GTT CCC 625 Leu Asn Ala Val Asn Asn Ser Leu Thr Pro Gin Ser Thr Lys Val Pro 165 170 175

AGC CTT TTT GAA TTC CAC GGT CCC TCA TGG TGT CTC ACT CCA GCA GAT 673 Ser Leu Fhe Glu Fhe His Gly Pro Ser Trp Cys Leu Thr Pro Ala Asp 180 185 190

AGA GGA TTG TGT OGT GCC AAT GAG AAC AGA TTC TAC TAC AAT TCA GTC 721 Arg Gly Leu Cys Arg Ala Asn Glu Asn Arg Fhe Tyr Tyr Asn Ser Val 195 200 205 210

ATT GGG AAA TGC CGC CCA TTT AAG TAC TCC GGA TGT GGG GGA AAT GAA 769 He Gly Lys Cys Arg Pro Phe Lys Tyr Ser Gly Cys Gly Gly Asn Glu 215 220 225

AAC AAT TIT ACT AGT AAA CAA GAA TGT CTG AGG GCA TGC AAA AAA GGT 817 Asn Asn Fhe Thr Ser Lys Gin Glu Cys Leu Arg Ala Cys Lys Lys Gly 230 235 240

TTC ATC CAA AGA ATA TCA AAA GGA GGC CTA ATT AAA ACC AAA AGA AAA 865 Fhe He Gin Arg He Ser Lys Gly Gly Leu He Lys Thr Lys Arg Lys 245 250 255

AGA AAG AAG CAG AGA GTG AAA ATA GCA TAT GAA GAA ATT TTT GTT AAA 913 Arg Lys Lys Gin Arg Val Lys He Ala Tyr Glu Glu He Fhe Val Lys 260 265 270

AAT ATG TGAGTCGAC 928

Asn Met

275 (2) INFORMATION FOR SEQ IP NO:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 304 amino acids

(B) TYPE: amino acid (0) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(xi) SEQUENCE OESCRIFTTON: SEQ IP NO:2:

Met He Tyr Thr Met Lys Lys Val His Ala Leu Trp Ala Ser Val Cys -28 -25 -20 -15

Lsu Leu Leu Asn Leu Ala Pro Ala Pro Leu Asn Ala Asp Ser Glu Glu -10 -5 1

Asp Glu Glu His Thr He He Thr Asp Thr Glu Leu Pro Pro Leu Lys 5 10 15 20

Leu Met His Ser Fhe Cys Ala Fhe Lys Ala Asp Asp Gly Pro Cys Lys 25 30 35

Ala He Met Lys Arg Phe Fhe Fhe Asn He Fhe Thr Arg Gin Cys Glu 40 45 50

Glu Fhe He Tyr Gly Gly Cys Glu Gly Asn Gin Asn Arg Fhe Glu Ser 55 60 65

Leu Glu Glu Cys Lys Lys Met Cys Thr Arg Asp Asn Ala Asn Arg He 70 75 80

He Lys Thr Thr Leu Gin Gin Glu Lys Pro Asp Fhe Cys Phe Leu Glu 85 90 95 100

Glu Asp Pro Gly He Cys Arg Gly Tyr He Thr Arg Tyr Fhe Tyr Asn 105 110 115

Asn Gin Thr Lys Gin Cys Glu Arg Phe Lys Tyr Gly Gly Cys Leu Gly 120 125 130

Asn Met Asn Asn Fhe Glu Thr Leu Glu Glu Cys Lys Asn He Cys Glu 135 140 145

Asp Gly Pro Asn Gly Phe Gin Val Asp Asn Tyr Gly Thr Gin Leu Asn 150 155 160

Ala Val Asn Asn Ser Leu Thr Pro Gin Ser Thr Lys Val Pro Ser Leu 165 170 175 180

Phe Glu Fhe His Gly Pro Ser Trp Cys Leu Thr Pro Ala Asp Arg Gly 185 190 195

Leu Cys Arg Ala Asn Glu Asn Arg Fhe Tyr Tyr Asn Ser Val He Gly 200 205 210 Lys Cys Arg Pro Phe Lys Tyr Ser Gly Cys Gly Gly Asn Glu Asn Asn 215 220 225

Phe Thr Ser Lys Gin Glu Cys Leu Arg Ala Cys Lys Lys Gly Phe He 230 235 240

Gin Arg He Ser Lys Gly Gly Leu He Lys Thr Lys Arg Lys Arg Lys 245 250 255 260

Lys Gin Arg Val Lys He Ala Tyr Glu Glu He Phe Val Lys Asn Met 265 270 275

(2) INPORMATTON FOR SEQ IP NO:3:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 81 base pairs

(B) TYPE: nucleic acid

(C) STRANOEPNESS: single (P) TOPOLOGY: linear

(ii) MOLECULE TYPE: DNA

(ix) FEATURE:

(A) NAME/KEY: COS

(B) LOCATION: 1..81

(ix) FEATURE:

(A) NAME KEY: sig_peptide

(B) LOCATION: 1..72

(ix) FEATURE:

(A) NAME/KEY: mat_peptide

(B) LOCATION: 73..81

(xi) SEQUENCE DESCRIPTION: SEQ IP NO:3:

ATG AAG TGG GTT ACT TTC ATC TCT TTG TTG TTC TTG TTC TCT TCT GCT 48 Met Lys Trp Val Thr Fhe He Ser Leu Leu Phe Leu Phe Ser Ser Ala -24 -20 -15 -10

TAC TCT AGA GGT GTT TTC AGG AGG GAT TCT GAG 81 Tyr Ser Arg Gly Val Phe Arg Arg Asp Ser Glu -5 1

(2) INFORMATION FOR SEQ IP NO:4:

(i) SEQUENCE CHA-ACTERISTTCS:

(A) LENGTH: 27 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear (ii) MOII-CULE TYPE: protein

(xi) SEQUENCE DESCRIFTTCN: SEQ ID NO:4:

Met Lys Trp Val Thr Phe He Ser Leu Leu Fhe Leu Fhe Ser Ser Ala -24 -20 -15 -10

Tyr Ser Arg Gly Val Fhe Arg Arg Asp Ser Glu -5 1

(2) INFORMATTON FOR SEQ ID NO:5:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 231 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: DNA

(ix) FF__[URE:

(A) NAME/KEY: CDS

(B) LOCATION: 76..231

(ix) FEATURE:

(A) NAME/KEY: sig_peptide

(B) LOCATION: 76..222

(ix) FEATURE:

(A) NAME/KEY: mat_peptide

(B) LOCATTON: 223..231

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:5:

GAATTCATTC AAGAATAGTT CAAACAAGAA GATTACAAAC TAT_AATTTC AIACACAATA 60

TA CGATTA AAAGA ATG AAG GCT GTT TTC TTG GTT TTG TCC TTG ATC GGA 111 Met Lys Ala Val Fhe Leu Val Leu Ser Leu He Gly -49 -45 -40

TTC TGC TGG GCC CAA CCA GTC ACT GGC GAT GAA TCA TCT GTT GAG ATT 159 Phe Cys Trp Ala Gin Pro Val Thr Gly Asp Glu Ser Ser Val Glu He -35 -30 -25

COG GAA GAG TCT CTG ATC ATC GCT GAA AAC ACC ACT TTG GCT AAC GTC 207 Pro Glu Glu Ser Leu He He Ala Glu Asn Thr Thr Leu Ala Asn Val -20 -15 -10

GCC ATG GCT AAG AGA GAT TCT GAG 231

Ala Met Ala Lys Arg Asp Ser Glu -5 1 (2) INFORMATION FOR SEQ ID NO:6:

(i) SEQUENCE CHA-AC_____7n.CS:

(A) LENGTH: 52 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(xi) SEQUENCE DESCRIPEICN: SEQ ID NO:6:

Met Lys Ala Val Ihe Leu Val Leu Ser Leu He Gly Fhe Cys Trp Ala -49 -45 -40 -35

Gin Pro Val Thr Gly Asp Glu Ser Ser Val Glu He Pro Glu Glu Ser -30 -25 -20

Leu He He Ala Glu Asn Thr Thr Leu Ala Asn Val Ala Met Ala Lys -15 -10 -5

Arg Asp Ser Glu

1

(2) LNPORMATTON FOR SEQ ID NO:7:

(i) SEQUENCE C3__^CIΕR_ST_CS:

(A) LENGTH: 726 base pairs

(B) TYPE: nucleic acid

(C) STRANDEENESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: DNA

(ix) FEATURE:

(A) NAME/KEY: CDS

(B) IOCATTON: 76..705

(ix) FEATURE:

(A) NAME/KEY: sig_peptide

(B) L0CATT< N: 76. .222

(ix) FEATURE:

(A) NAME/KEY: mat_peptide

(B) KICATTON: 223. .705

(xi) SEQUENCE DESCRIETION: SEQ ID NO:7:

GAATTCATTC AAGAATAGTT <_AAAC_AAGAA GATTACAAAC TATCAATTTC AIACACAATA 60

TAAAOGATTA AAAGA ATG AAG GCT GTT TTC TTG GTT TTG TCC TTG ATC GGA 111 Met Lys Ala Val Fhe Leu Val Lsu Ser Leu He Gly -49 -45 -40 TTC TGC TGG GCC CAA CCA GTC ACT GGC GAT GAA TCA TCT GTT GAG ATT 159 Fhe Cys Trp Ala Gin Pro Val Thr Gly Asp Glu Ser Ser Val Glu He -35 -30 -25

COG GAA GAG TCT CIG ATC ATC GCT GAA AAC ACC ACT TTG GCT AAC GTC 207 Pro Glu Glu Ser Leu lie He Ala Glu Asn Thr Thr Leu Ala Asn Val -20 -15 -10

GCC ATG GCT AAG AGA GAT TCT GAG GAA GAT GAA GAA CAC ACA ATT ATC 255 Ala Met Ala Lys Arg Asp Ser Glu Glu Asp Glu Glu His Thr He lie -5 1 5 10

ACA GAT ACG GAG CTC CCA CCA CTG AAA CTT ATG CAT TCA TTT TGT GCA 303 Thr Asp Thr Glu Leu Pro Pro Leu Lys Leu Met His Ser Fhe Cys Ala 15 20 25

TTC AAG GOG GAT GAT GGG CCC TGT AAA GCA ATC ATG AAA AGA TTT TTC 351 Phe Lys Ala Asp Asp Gly Pro Cys Lys Ala He Met Lys Arg Fhe Fhe 30 35 40

TTC AAT ATT TTC ACT OGA CAG TGC GAA GAA TTT ATA TAT GGG GGA TGT 399 Phe Asn He Fhe Thr Arg Gin Cys Glu Glu Fhe He Tyr Gly Gly Cys 45 50 55

GAA GGA AAT CAG AAT OGA TTT GAA ACT CIG GAA GAG TGC AAA AAA ATG 447 Glu Gly Asn Gin Asn Arg Fhe Glu Ser Leu Glu Glu Cys Lys Lys Met 60 65 70 75

TGT ACA AGA GAT AAT GCA AAC AGG ATT ATA AAG ACA ACA CTG CAG CAA 495 Cys Thr Arg Asp Asn Ala Asn Arg He He Lys Thr Thr Leu Gin Gin 80 85 90

GAA AAG CCA GAT TTC TGC TTT TTG GAA GAG GAT CCT GGA ATA TGT CGA 543 Glu Lys Pro Asp Phe Cys Phe Leu Glu Glu Asp Pro Gly He Cys Arg 95 100 105

GGT TAT ATT ACC AGG TAT TTT TAT AAC CAA CAG ACA AAA CAG TGT GAA 591 Gly Tyr He Thr Arg Tyr Fhe Tyr Asn Gin Gin Thr Lys Gin Cys Glu 110 115 120

AGG TTC AAG TAT GGT GGA TGC CTG GGC AAT ATG AAC AAT TTT GAG ACA 639 Arg Fhe Lys Tyr Gly Gly Cys Leu Gly Asn Met Asn Asn Fhe Glu Thr 125 130 135

CTC GAG GAA TGC AAG AAC ATT TGT GAA GAT GGT COG AAT GGT TTC CAG 687 Leu Glu Glu Cys Lys Asn He Cys Glu Asp Gly Fro Asn Gly Fhe Gin 140 145 150 155

GTG GAT AAT TAT GGT ACC TGAAGATCCT CTAGAGTCGA C 726

Val Asp Asn Tyr Gly Thr 160 (2) INPOFMATTON FOR SEQ ID NO:8:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 210 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

( i) SEQUENCE DESCRIPTION: SEQ ID NO:8:

Met Lys Ala Val Fhe Leu Val Leu Ser Leu He Gly Fhe Cys Trp Ala -49 -45 -40 -35

Gin Pro Val Thr Gly Asp Glu Ser Ser Val Glu He Pro Glu Glu Ser -30 -25 -20

Leu He He Ala Glu Asn Thr Thr Leu Ala Asn Val Ala Met Ala Lys -15 -10 -5

Arg Asp Ser Glu Glu Asp Glu Glu His Thr He He Thr Asp Thr Glu 1 5 10 15

Leu Pro Pro Leu Lys Lu Met His Ser Phe Cys Ala Phe Lys Ala Asp 20 25 30

Asp Gly Pro Cys Lys Ala He Met Lys Arg Phe Fhe Fhe Asn He Fhe 35 40 45

Thr Arg Gin Cys Glu Glu Fhe He Tyr Gly Gly Cys Glu Gly Asn Gin 50 55 60

Asn Arg Fhe Glu Ser Leu Glu Glu Cys Lys Lys Met Cys Thr Arg Asp 65 70 75

Asn Ala Asn Arg He He Lys Thr Thr Leu Gin Gin Glu Lys Pro Asp 80 85 90 95

Fhe Cys Phe Leu Glu Glu Asp Pro Gly He Cys Arg Gly Tyr He Thr 100 105 110

Arg Tyr Fhe Tyr Asn Gin Gin Thr Lys Gin Cys Glu Arg Fhe Lys Tyr 115 120 125

Gly Gly Cys Leu Gly Asn Met Asn Asn Fhe Glu Thr Leu Glu Glu Cys 130 135 140

Lys Asn He Cys Glu Asp Gly Pro Asn Gly Phe Gin Val Asp Asn Tyr 145 150 155

Gly Thr 160

Claims

1. A non-glycosylated TFPI analogue containing at least the first and second Kunitz domain and lacking the third Kunitz domain from amino acid Cysl89 to amino acid 5 Cys239 and a substantial part of the amino acid sequence from Lys240 to Met276 of the native TFPI molecule, said TFPI analogue being modified at either or both of the two N- glycosylation triads Asnll7-Glnll8-Thrll9 and Asnl67-Asnl68- Serl69 to avoid glycosylation.

102. A non-glycosylated TFPI analogue according to claim 1 containing at least the amino acid sequence from Phe25 to Glu 148 of the native TFPI molecule and lacking the third Kunitz domain from amino acid Cysl89 to amino acid Cys239 and a substantial part of the amino acid sequence from Lys240 to

15 Met276 of the native TFPI molecule.

3. A non-glycosylated TFPI analogue according to claim 1 containing at least the amino acid sequence from Aspl to Glul48 of the native TFPI molecule and lacking the third Kunitz domain from amino acid Cysl89 to amino acid Cys239 and

20 a substantial part of the amino acid sequence from Lys240 to Met 276 of the native TFPI molecule.

4. A non-glycosylated TFPI analogue lacking the amino acid sequence from amino acid residue Glnl62 to Met276 of the native TFPI molecule, said TFPI analogue being modified at 5 the glycosylation triad Asnll7-Glnll8-Thrll9.

5. A TFPI analogue according to claim 4 wherein Asnll7 has been replaced by another amino acid residue.

6. A TFPI analogue according to claim 5 wherein Asnll7 has been replaced by Gin.

7. A TFPI analogue according to claim 4 wherein Thrll9 has been replaced by another amino acid residue except Ser.

8. A TFPI analogue according to claim 4 wherein Glnllδ has been replaced by Pro or Asp.

59. DNA sequence encoding a non-glycosylated TFPI analogue according to any of claims 1-8.

10. A recombinant expression vector comprising DNA sequences permitting gene expression, including a promoter and a terminator, functionally fused to a DNA sequence 0 according to claim 9.

11. Eucaryotic cell containing a recombinant expression vector according claim 10.

12. A pharmaceutical composition comprising a non- glycosylated TFPI analogue according to any of claims 1-8 5 together with a pharmaceutically acceptable diluent or carrier.

13. A composition according to claim 12 in unit dosage form comprising 0.1-100 mg of the TFPI analogue.

14. A composition according to claim 12 or 13 for the 0 prophylaxis or treatment of coagulation disorders or cancer.

15. A composition according to claim 14, wherein the coagulation disorder is disseminated intravascular coagulation.

16. Use of a non-glycosylated TFPI analogue according 5 to any of claims 1-8 for the preparation of a medicament for the prophylaxis or treatment of coagulation disorders or cancer.

17. Use according to claim 16, wherein the coagulation disorder is disseminated intravascular coagulation.