CA1340263C - Expression of protein c analogues - Google Patents

Abstract

Genomic and cDNA sequences coding for a protein having substantially the same biological activity as human protein C and recombinant expression vectors comprising these sequences are disclosed. Methods are disclosed for producing a protein which has substantially the same biological activity as human protein C. The protein, which may be in the form of activated protein C, is produced by mammalian host cells transfected with a plasmid capable of integration in mammalian host cell DNA. The plasmid includes a promoter followed downstream by a DNA sequence which encodes a protein having substantially the same structure and/or activity as human protein C, the nucleotide sequence being followed downstream by a polyadenylation signal.

Description

~3S0263 EXPRE~SSION OF PROTEIN C

The present invention relates generally to plasma proteins and DNA sequences encoding them, and more speci-fically, to the expression o~ proteins having substantially the same structure and/or activity as human protein C or human activated protein C.

Protein C is a zymogen, or precursor, of a serine protease which plays an important role in the regulation of blood coagulation and generation of fibrinolytic activity in vivo. It is synthesized in the liver as a single-chain polypeptide which undergoes considerable processing to give rise to a two-chain molecule comprising heavy (Mr = 40,000) and light (Mr = 21,000) chains held together by a disu1-fide bond. The circulating two-chain intermediate is converted to the biologically active form of the molecule, known as "activated protein C" (APC), by the thrombin-mediated cleavage of a 12-residue peptide from the amino-terminus o~ the heavy chain. The cleavaqe reaction is augmented in vivo by thrombomodulin, an endothelial cell cofactor (Esmon and Owen, Proc. Natl. Acad. Sci. USA 78:
22~9-2252, 1981).
E'rotein C is a vitamin K-dependent glycoprotein which contains approximately nine residues of gamma-carboxyglutamic acid (Gla) and one equivalent of beta-hydroxyaspartic acid which are formed by post-translational modifications of glutamic acid and aspartic acid residues, respectively. The post-trans]ational formation of specific gamma-carboxyglutamic acid residues in protein C requires vitamin K. These unusual amino acid residues bind to calcium ions and are believed to be responsible for the interaction of the protein with phospholipid, which is required for the biological activity of protein C.
In contrast to the coagulation-promoting action of other vitamin K-dependent plasma proteins, such as factor VII, factor IX, and factor X, activated protein C
acts as a regulator of the coagulation process through the inactivation of factor Va and factor VIIIa by limited proteolysis. The inactivation of factors Va and VIlIa by protein C is dependent upon the presence of acidic phospho-lipids and calcium ions. Protein S has been reported toregulate this activity by accelerating the APC-catalyzed proteolysis of factor Va (Walker, J. Biol. Chem. 255:
5521-5524, 1980).
Protein C has also been implicated in the action ]5 of tissue-type plasminogen activator (Kisiel and Eujikawa, Behring Inst. Mitt. 73: 29-42, 1983). Infusion of bovine APC into dogs results in increased plasminogen activator activity (Comp and Esmon, J. Clin. Invest. 68: 1221-1228, 1981). Recent studies (Sakata et al., Proc. Natl. Acad.
Sci. USA 82: 1121-1125, 1985) have shown that addition of APC to cultured endothelial cells leads to a rapid, dose-dependent increase in fibrinolytic activity in the condi-tioned media, reflecting increases in the activity of both urokinase-related and tissue-type p]asminogen activators by the cells. APC treatment also results in a dose-dependent decrease in antiactivator activity.
Protein C deficiency is associated with recurrent thrombotic disease (Broekmans et al., New Eng. J. Med. 309:
340-344, 1983; and Seligsohn et al., New Eng. J. Med. 310:
559-562, 1984) and may result from genetic disorder or from trauma, such as liver disease or surgery. ~'his condition is generally treated with oral anticoagulants. Beneficial effects have also been obtained through the infusion of protein C-containing normal plasma (see Gardiner and Griffin in Prog. in Hematology, ed. Brown, Grune &
Stratton, NY, 13: 265-278). In addition, some investi-gators have discovered that the anti-coagulant activity of 13~02~

protein C is useful in treating thrombotic disorders, such as venous thrombosis (Smith et al., ~Crl' Publica~ion No. WO
85/00521). In some parts of the world, it is estimated that approximately 1 in 16,000 individuals exhibit protein C deficiency. Further, a total deficiency in protein C is fatal in newborns.
Finally, exogenous protein C has been shown to prevent the coagulopathic and lethal effects of gram negative septicemia (Taylor et al., J. Clin. ]nvest. 79:
918-925, 1987). Data obtained from studies with baboons suggest that protein C plays a natural role in protecting against septicemia.
While natural protein C may be purified from clotting factor concentrates (Marlar et al., B ood 59:
1067-1072) or from plasma (Kisiel, ibid.), it i5 a comp]ex and expensive process, in part due to the limited avail-ability of the starting materiaL and the low concentration of protein C in plasma. Furthermore, the therapeutic use of products derived from human blood carries the risk of disease transmission by, for example, hepatitis virus, cytomega]ovirus, or the causative agent of acquired immune deficiency syndrome (AlDS). In view of protein C's clini-cal applicability in the treatment of thrombotic disorders, the production of useful quantities of protein C and activated protein C is clearly invaluable.

Briefly stated, the present invention discloses L)NA sequences which code for proteins having substantially the same structure and/or biological activity as human protein C or human activated protein C. In one aspect of the present invention, the DNA sequence further codes for the amino acid sequence (Rl)n-R2-R3-K4, wherein R], R2, K3 and R4 are Lys or Arg and n = 0, 1, 2, or 3, at the cleav-age site of the light and heavy chains. In a preferredembodiment, the amino acid sequence at the cleavage site is Arg-Arg-Lys-Arg. In another aspect of the present inven-~263 tion, the protein includes the substitution of residue 158(Asp) with a non-acidic amino acid residue such as Ala, Ser, Thr or Gly. In a related aspect, the protein includes the substitution of residue 1.5~ (His) with a basic amino acid residue such as Lys or Arg. In another aspect, the protein includes the substitution of the l.,ys-Arg at residues 156-157 of native protein C with Lys-Lys or Arg-Arg.
Yet another aspect of the present invention is directed toward a DNA sequence which codes for a protein having subtantially the same biological activity as human protein C or human activated protein C, the sequence further encoding the pre-pro peptide of a protein such as factor VII, factor IX, factor X, prothrombin or protein S.
In addition, the present inventi.on discloses expression vectors capab]e of integration in mammal.ian host cell DNA, including a promoter followed downstream by a DNA
sequence which encodes a protein having substantially the same structure and/or activity as human protein C or human activated protein C as set forth above, transcription of the nucleotide sequence being directed by the promoter.
The nucleotide sequence is followed downsteeam by a polyadenylation signal. In one embodiment, the expression vector includes a selectable marker located between the nucleotide sequence and the polyadeny.lation signal., transcription of the selectable marker being directed by the promoter. The expression vector may also include a set of RNA splice sites.
A related aspect of the present invention dis-closes mammalian cells transfected to express a proteinwhich, upon activation, has substantially the same biologi-cal activity as human activated protein C. The mammalian cells are transfected with an expression vector capable of integration in mamma]ian host cell DNA, the expression vector including a promoter followed downstream by a DNA
sequence as described above. Within one embodiment, a selectable marker is also introduced into ~he cel.ls and 1 3 ~ 3 stably transfected cells are selected. Mammalian cells transfected to express a protein which has substantially the same biological activity as human activated protein C
are also disclosed. Preferred host cells for use within the present invention are COS, B~K and 293 cells.
A further aspect of the invention discloses a method for producing a protein which, upon activation, has substantially the same biological activity as human acti-vated protein C. The method comprises (a) introducing into a mammalian host cell an expression vector comprising a DNA
sequence as described above, which encodes a protein having substantially the same structure and/or activity as human protein C; (b) growing said mammalian host c:ell in an appropriate medium; and (c) isolating the protein product encoded by said DNA sequence and produced by said mammalian host cell. The protein product produced according to this method is also disclosed. A method for producing a protein which has substantially the same structure and/or bio-logical activity as human activated protein C is also disclosed.
The proteins described within the present invention may be used as active therapeutic substances, including use in the regulation of blood coagulation.
Further, these proteins may be combined with a physiologi-cally acceptable carrier and/or diluent to provide suitable pharmaceutical compositions.
Other aspects of the invention will become evident upon reference to the detailed description and attached drawings.

Figure 1 is a partial restriction map of the protein C cDNA in pHCA6L. The coding region is indicated by an open box.
Figure 2 illustrates the nucleotide sequence of the complete protein C cDNA and the deduced amino acid 13~026.~

sequence of protein C. A~rows indicate cleavage sites forremoval of the connecting dipeptide and activation peptide.
Figure 3 illustrates a restriction enzyme map of the genomic DNA coding for human protein C. Numbers below the line indicate length in kilobases (kb).
Figure 4 illustrates the complete genomic sequence, including exons and introns, of ~he human protein C gene. Arrowheads indicate intron-exon splice junctions.
The polyadenylation or processing sequences of A-T-T-A-A-A
and A-A-T-A-A-A at the 3' end are boxed. ~ , potential carbohydrate attachment sites ~ , apparent cleavage sites for processing of the connecting dipeptide ~ , site of cleavage in the heavy chain when protein C is converted to activa~ed protein C; ~, sites of polyadenylation.
Figure 5 illustrates a schematic two-dimensional model for the structure of human protein C.
Figure 6, which is shown o~ the same page as Figure 1, illustrates the subcloning of the 5' and 3' ~ortions of a protein C partial cDNA clone.

Figure 7 illustrates ~he removal of intron A ~rom the genomic clone, resulting in the fusion of exons I and II.
Figure 8 illustrates the fusion of exons I and II
to the 5'-most portion of the c~NA insert of Figure 1.
Figure 9 illustrates the construction of a plas-mid comprising the complete coding sequence for protein C.
Figure 10 illustrates the expression vector pD7C.
Symbols used are ori, ~he adenovirus 5 0-1 map unit sequence; E, the SV40 enhancer Ad2MLP, the adenovirus 2 major late promoter Ll-3, the ad~novirus 2 tripartite leader: 5'ss, 5' splice site 3~ss, 3~ splice site: pA, the SV~0 early polyadenylation signal; and ~, the deleted region of the pBR322 "poison" sequences.
Figure 11 illustrates the expression vector pD5(~C-DHFRr). DHFRr denotes the methotrexate resistant mutant dihydrofolate reductase gene sequence; pA denotes the SV40 late polyadenylation signal. Other symbols used are as described for Figure 10.

13~263 Figure 12 il1ustrates the expression vector pDX/PC. Symbols used are as described for ~igure 11.
Figure 13 illustrates the results o~ an assay for activated protein C on media sample5 from transfec~ed 293 cells.
Figure 14 illustrates the expression vectors pDX~PC962 and PC229/962.
Figure 15 illustrates the anticoagulan~ activity of protein C prepared according to certain embodiments of the present invention.

Best Mode for Carryin~ Out the Invention Prior to sctting forth the invention, it may be helpful to an understanding thereof to set forth de~ini-tions of certain terms to be used hereinafter.
Bioloqical ActivitY: A function or set of func-tions performed by a molecule in a biological context (i.e., in an organism or an ln v~tro facsimile). Biolog-ical activities of proteins may be divided into ca~alytic and ef~ector activities. Catalytic activities of the vitamin K-dependent plasma proteins generally involve the specific proteolytic cleavage of other plasma proteins, resulting in activation or deactivation of the substrate.
Effector activities include specific binding of the biologi-cally active molecule to calcium, phospholipids or other small molecules, to macromolecules, such as proteins, or to cells. Effector activity frequently augments, or is essential to, catalytic activity under physiological conditions.
For protein C, biological ac~ivity is ~haracter-ized by its anticoagulant and fibrinolytic properties.
Protein C, when activated, inactivates factor Va and factor VIIIa in the presence of phospholipid and ealcium. Protein S appears to be involved in the regulation of this function (Walker, ibid.). Activated protein C also enhances fibrinolysis, an effect believed to be mediated by the lowering of levels of plasminogen activator inhibitors (van r '''~ X

8 13~0263 Hinsbergh et al., Blood 65: 4~4-451, 1985). As more fully described below, that portion of protein C encoded by exons VII and VIII of the protein C gene is primarily responsible for its catalytic activities.
Pre-Pro Peptide: An amino acid sequence which occurs at the amino terminus of some proteins and is generally cleaved from the protein during translocation.
Pre-pro peptides comprise sequences directing the protein into the secretion pathway of the cell (signal sequences) and are characterized by the presence of a core of hydro-phobic amino acids. They may also comprise processing signals. As used herein, the term "pre-pro peptide" may also mean a portion of a naturally occuring pre-pro peptide.
Expression Unit: A DNA construct comprising a primary nucleotide sequence encoding a protein of interest, together with other nucleotide sequences which direct and control the transcription of the primary nucleotide sequence. An expression unit consists of at least the primary nucleotide sequence and a promoter sequence located upstream from and operably linked to the primary nucleotide sequence and a polyadenylation signal located downstream.
Additional genetic elements may also be included to enhance efficiency o~ expression. These elements include enhancer sequences, leaders, and mRNA splice sites.
Expression Vector: A DNA mo]ecule which contains, inter alia, a ~NA sequence encoding a protein of interest together with a promoter and other sequences which facili-tate expression of the protein. Expression vectors further contain genetic information which provides for their repli-cation in a host cell, either by autonomous replication or by integration into the host genome. Examples of expres-sion vectors commonly used for recombinant DNA are plasmids and certain viruses, although they may contain elements of both. They also may include a selectable marker.

As noted above, protein C is produced in the liver and requires vitamin K for its biosynthesis. Vitamin K is necessary for the formation of specific gamma-carboxy-glutamic acid residues in the amino-terminal region of the light chain. These amino acid residues are formed by a post-translational modification, and are required for calcium-mediated binding to phospholipid. In addition, protein C contains one beta-hydroxyaspartic acid residue which is also formed in a post-translational modification.
However, the role of this amino acid residue is not known.
Given the fact that the activity of protein C is dependent upon post-translational modifications involving the gamma carboxylation of specific g]utamic acid residues and cleavage to the two-chain form, and may also be dependent upon the hydroxylation of a specific aspartic acid residue, it is unlikely that an active product couLd be produced through the cloning and expression of protein C
.
n a mlcroorganlsm.
Accordingly, the present invention provides a method of producing a protein which is gamma-carboxylated and, upon activation, has the biological ac~ivity of human activated protein C through the use of mammalian cells transfected to stably express the protein.
The present invention further provides a method for producing a protein which is gamma-carboxylated and has the biological activity of human activated protein C
without the necessity for activation.
The light and heavy chains of bovine protein C
have been sequenced (FernLund and Sten~lo, J. Biol. Chem.
257: 12170-12179, 1982; and Stenflo and Fernlund, J. ~iol.
Chem. 257: 12180--12190, 1982). Isolation and characteriza-tion of human protein C have been described by Kisiel, J. Clin. Invest. 6~: 761-769, 1979. The anticoagulant _ activities of both the human and bovine enzymes were found to be highly species specific. Species specificity is believed to be mediated by protein S (Walker, Thromb. Res.
22: 321-327, 1981). However, the human and bovine proteins 131~0263 show considerable overa]l structural homology to each other and to other vitamin K-dependent plasma proteins, including prothrombin, factor VII, factor IX, and factor X. Similari-ties include the presence of the Gla residues in the light chain and the active site serine in the heavy chain, as well as other amino acid sequence homology in the amino-terminal region of the light chain.
Within the present invention, a ~gtll cDNA
library was prepared from human liver mRNA. 'l'his library was then screened with 125I-labeled antibody to human protein C. Antibody-reactive clones were further analyzed for the synthesis o~ a ~usion protein of B-gaLactosidase and protein C in the lgtll vector.
One of the clones gave a strong signal with the antibody probe and was found to contain an inserl: of approx-imately 1400 bp. DNA sequence analysis of the nNA insert revealed a predicted amino acid sequence which shows a high degree of homology to major portions of bovine protein C, as determined by Fernlund and Stenflo (J. Biol. Chem. 257:
12170-12179; J. Biol. Ch_m. 257: 12180-12190).
The DNA insert contained the majority of the coding region for protein C beginning with amino acid 6~ of the light chain, including the entire heavy chain coding region, and proceeding to the termination codon. Further, following the stop codon of the heavy chain, there werc 294 base pairs of 3' noncoding sequence and a poly (A) tail of 9 base pairs. The processing or po]yadenylation signal A-A-T-A-A-A was present 13 base pairs upstream from ~he poly (A) tail in this cDNA insert. This sequencc was one of two potential po]yadenylation sites.
The cDNA sequence also encodes the dipeptide Lys-Arg at position 156-157 (numbering of amino acids is shown in F'igure 2), which separates the light chain from the heavy chain and is removed during processing by prote-olytic cleavage resulting in secretion of the two-chain molecule. Upon activation of the two-chain molecule by thrombin, the heavy chain of human protein C is cleaved 11 13~0263 between arginine-169 and leucine-170, releasing the activa-tion peptide (Figure 2).
By a similar method, a second cDNA which lacked the coding sequence for the pre-pro peptide and the ~irst 23 amino acids of protein C was isolated. Using this cDNA
as a hybridization probe, the remainder of the coding sequence was obtained from a human genomic DNA library in A Charon 4A (Foster et al., Proc. Natl. Acad. Sci. USA 82:
4673-4677, 1985). Three different 1 Charon 4A phage were isolated that contained overlapping inserts for the protein C gene.
The positions of exons on the three phage clones were determined by Southern blot hybridization of digests of these clones with probes made from the 1400 bp cDNA
described above. The genomic DNA inserts in these clones were mapped by single and double restriction enzyme diges-tion followed by agarose gel electrophoresis, Southern blotting, and hybridization to radiolabeled 5' and 3' probes derived from the c~NA for human protein C, as shown in Figure 3.
DNA sequencing studies were performed using the dideoxy chain-termination method. As shown in Figure 4, the nucleotide sequence for the gene for human protein C
spans approximately 11 kb of DNA. These studies further revealed a potential pre-pro peptide of 42 amino acids.
The pre-pro sequence is cleaved by a signal peptidase following the Gly residue at position -25. Processing to the mature protein involves additional proteolytic cleavage following residue -1 to remove the amino-terminal propep-tide, and at residues 155 and 157 to remove the Lys-Arg dipeptide which connects the light and heavy chains. This final processing yields a light chain of 155 amino acids and a heavy chain of 262 amino acids.
The protein C gene is composed of eight exons ranging in size from 25 to 885 nucleotides, and seven introns ranging in size from 92 to 2668 nucleotides. ~xon I and a portion of exon II code for the 42 amino acid 13~0263 pre-pro peptide. The remaining portion of exon II, exon III, exon IV, exon V, and a portion of exon VI code for the light chain of protein C. The remaining portion of exon VI, exon VII, and exon VIII code for the heavv chain of protein C. The amino acid and DNA sequences for a cDNA
coding for human protein C are shown in Figure 2.
The introns in the gene for protein C are located primarily between various ~unctional domains. Exon II spans the highly conserved region of ~he pre-pro peptide and the gamma-carboxyglutamic acid (Gla) domain. Exon III includes a stretch of eight amino acids which connect the Gla and growth factor domains. Exons IV and V each represent a potential growth factor domain, while exon Vl covers a connecting region which includes the activation peptide.
Exons VII and VIII cover the catalytic domain typical of all serine proteases.
The amino acid sequence and proposed structure for human pre-pro protein C are shown in Figure 5. Protein C is shown without the Lys-Arg dipep-ide, which connects the light and heavy chains. The location of the seven introns (A through G) is indicated by solid bars. Amino acids flanking known proteolytic cleavage sites are circLed.
~ designates potential carbohydrate binding sites. The first amino acid in the light chain, activation peptide, and heavy chain start with number l. lhis numbering differs from that shown in Figures 2 and 4.
Carbohydrate attachment sites are located at residue 97 in the light chain and residues 79, 144, and 160 in the heavy chain, according to the numbering scheme of Figure 5. The carbohydrate moiety is covalently linked to Asn. In the majority of instances, the carbohydrate attachment environment can be represented by Asn-X-Ser or Asn-X-Thr, where X = any amino acid.
As noted above, protein C plays a regulatory role in the coagulation process. The catal~tic domain, encoded by exons VII and VIII, possesses serine protease activity which specifically cleaves certain plasma proteins (i.e., ... 1 3 ~a 2 ~ ;~

factors Va and VIlla), resultin9 in their deactivation. As a result of this selective proteolysis, protein C displays anticoagulant and fibrinolytic activities.
Due to the presence of intervening sequences in the genomic clone, merely joiining the genomic and cDNA
sequences to provide a comp]ete coding sequence is not sufficient for constructing an acceptable expression unit.
It is therefore necessary to delete these intervening sequences for reasons more fully described below if a genomic clone is used to construct the expression unit.
'I'he 5' coding region may also be obtained by alternative methods and consequently e1iminate the need to delete intervening sequences. The 5' coding region may be obtained by using probes derived from the existing c~NA or genomic clones to probe additional libraries. ~y this method, a full-length cDNA was iso~ated. Furthermore, the amino-terminal portions of the vitamin K-dependent pLasma proteins are responsible for their respective calcium bind-ing activities. It has been fo~_nd that, as a result o~
this functional homology, the calcium binding domains of these molecules may be interchanged and still retain the activity speci~ic to the catalytic domain of the resultarlt molecule. For cxampLe, as described inEuropean-~aten~ O~ice application 200,421, published December lO, 1986, the amino-terminal portion (calcium binding domain) o~
factor IX may be joined to factor Vl~ at amino acid 38 to produce a protein having the activity of factor VI r .
Factor VII, ~actor IX, factor X, prothrombin, and protein S
share this amino-terminal sequence homology with protein C.
Consequently, a cloned sequence comprising the 5'-coding region of the gene for any o~ these ~roteins might be substituted for the corresponding sequence of the protein C
gene. Additionally, suitable coding sequences may be synthesized based on the known amino acid sequences of severaL of the vitamin K-dependent plasma proteins or on the sequence of protein C disclosed herein. Techniques for .._.. . ..

1~ 13~02~3 producing synthetic nucleotide sequences are well known in the art. For example, a set of overlapping oligonucle-otides may be synthesized and annealed in pairs to yield double-stranded fragments with overlapping cohesive termini.
These fragments are then ligated as any restriction frag-ments would be. The resultant synthetic fragment is then ligated to the cDNA at a convenient restriction site. The junction sequence may be modified as necessary by oligo-nucleotide-directed mutagenesis.
When clones representing the entire coding sequence have been obtained, the appropriate regions may be joined, as necessary, to generate the desired coding sequence. Fragments obtained from one or more libraries are cut with appropriate restriction endonucleases and joined together enzymatical]y in the proper orientation.
Depending on the fragments and the particular restriction endonucleases chosen, it may be necessary to remove unwanted DNA sequences through a "loop out" process of dele-tion mutagenesis or through a combination o~ restriction endonuclease cleavage and mutagenesis. ~'he sequence so obtained should preferably be in the form of a continuous open reading frame, that is, that it lack the intervening sequences (introns) generally ~ound in higher eukaryotic genes. ~'he presence of introns in cloned genes may lead to aberrant splicing of messenger ~NA and~or reduced effi-ciency o~ gene expression or instability upon amplification when the gene sequence is introduced into a mammalian host cell. It is prererred that this coding sequence rurther encode a pre-pro peptide in order to facilitate proper processing and secretion of the protein C produced accord-ing to the present invention. The pre-pro peptide may be that of protein C or another secreted protein, such as factor IX, factor VII, or prothrombin.
Under some circumstances, it may be desirable to produce activated protein C directly, thereby removing the need to activate the protein product e;ther in vitro or in vi o. The cleavage sites involved in the maturation and ~-~ 15 13~263 activation of protein C are known (Foster and Davie, ibid.).
A sequence encoding AI~C may be constructed by deleting the region encoding the activation peptide through oligonucle-otide-directed deletion mutagenesis. The resultant protein will then become activated by removal of the Lys-Arg dipep-tide during normal proteolytic processing in the secretion pathway of the host cell. It has been found that proteins lacking the activation peptide are nevertheless properly processed by the host cells, resulting in secretion of activated protein C.
In order to enhance the proteolytic processing involved in the maturation of the recombinant protein C to the two-chain form, it may be desirable to modify the amino acid sequence around the processing site. Such modifica-tion has been found to facilitate the proper processing ofrecombinant protein C in transfected cells.
Efficient cleavage may increase the specific activity of the protein, given that single-chain protein C
is not known to be activated in the bloodstream. As previ-ously noted, this maturation process involves the remova~of the dipeptide Lys-Arg (amino acids 156-~57) (Foster and Davie, Proc. Natl. Acad. Sci. USA ~1: 4766-4770, 1984).
Modifications to the amino acid sequence in the vicinity of this processing site include the substitution and/or inser-tion of amino acids. One such group of modifications isthe alteration of the amino acid sequence to include the sequence (Rl)n-R2-R3-R4, wherein Rl, R2, R3 and 1~4 are Lys or Arg and n=0, 1, 2 or 3, in place of the native Lys-Arg dipeptide. A particularly preferred modification of this group is the sequence Arg-Arg-r,ys-Arg. rrhis sequence has been found to enhance processing of recombinant protein C
by about five-fold in transEected ~IIK cells. A second group of modifications includes the substitution of amino acid residue 154 (His) of native protein C with a basic amino acid residue (i.e., r.ys or Arg) to give a processing site sequence of the general formula Rl-X-R2-R3, wherein R], R2, and R3 are Lys or Arg, and X is an amino acid other ~ 16 13~02~

than r,ys or Arg, prefera~ly I.eu. A third group of modifica-tions includes substitution of the Asp res;due at positi.on 158 with a non-acidic amino acid residue. Use Or a small neutral amino acid, such as Ala, Ser, ~'hr or Gly is preferred. A fourth group of modifications includes the substitution of Lys-Lys or Arg-Arg for the Lys-Arg of native protein C. Combinations of these groups of modifica-tions may a]so be made. ~or example, amino acid 15~ may be substituted in a protein C mo]ecule contairlirlg a process-ing site having the sequence (Rl)n-R2-R3-R4. These modifi-cations can be used in producing wild-type protein C or activated protein C.
The coding sequence for protein C or activated protein C is then inserted into a suitab:le expression vector which is, in turn, used to transfect a mammalian cell line. Expression vectors for use in carrying out the present inventi.on wi]l comprise a promotor capable of directing the transcription of a roreign gene introduced into a mammalian cell. Viral promoters are preferred due t.o their effi.ciency in di.rect.ing transcription. A particu-larly preferred promotor is the major late promoter ~rom adenovirus 2. Such expression vectors wil-L also preferably contain a set of RNA splice sites located downstream from the promoter and upstream from the insertion site for the protein C sequence or within the protein C sequence itself.
Preferred RNA splice sites may be obtained from adenovirus and/or immunoglobulin genes. Also contained i.n the expres-sion vectors is a polyadenylatiorl signal, loc:ated down-stream of the insertion si.te. Viral polyadeny1ation signals are particularly preferred, such as the early or late polyadenylation signa]s from SV4~ or the polyadenyla-tion signal from the adenovirus 5 ~'Tb region. In a particu-larly pre.ferred embodiment, the expression vector also comprises a noncoding viral leader sequence, such as the adenovirus 2 tripartite leader, located between the promoter and the RNA splice si.tes. Preferred vectors may 17 1~402~3 also include enhancer sequences, such as the SV40 enhancer and the sequences encoding the adenovirus VA RNAs.
Cloned gene sequences may then be introduced into cultured mammalian cells by, for example, calcium phosphate-mediated transfection (Wigler et al., Cell 14:
725, 1978; Corsaro and Pearson, Somatic Cell Genetics 7:
603, 1981; Graham and Van der ~b, Virology 52: 456, 1973).
A precipitate is formed of the DNA and calcium phosphate, and this precipitate is applied to the cells. Some of the cells take up the DNA and maintain it inside the cell for several days. A small ~raction o~ these cells (typically 10-4) integrate the DNA into the genome. In order to identify these integrants, a gene that confers a selectable phenotype (a selectable marker) is generally introduced into the cells along with the gene of interest. Preferred selectable markers include genes that confer resistance to drugs, such as neomycin, hygromycin, and methotrexate.
Selectable markers may be introduced into the cell on a separate plasmid at the same time as the gene of interest, or they may be introduced on the same plasmid. 1~ on the same plasmid, the selectabLe marker and the gene of interest may be under the control of dif~erent promoters or the same promoter. In one embodiment, the selectable marker is placed on the same plasmid with the sequence encoding protein C such that both sequences are controlled by the same promoter, an arrangement known as a dicistronic message. Constructs of this type are known in the art (for example, European Patent Office publication 117,058). It may also be advantageous to add additional DNA, known as "carrier DNA," to the mixture which is introduced into the cells. After the cells have taken up the ~NA, they are allowed to gr-ow for a period of time, typically 1-2 days, to begin expressing the gene of interest. Drug selection is then applied to select for the growth of cells which are expressing the selectable marker in a stable fashion.
Clones of such cells may be screened for expression of protein C.

18 134-02~3 Preferred mammalian cell lines for use in the present invention include the COS, ~I~K and 293 ce]l lines.
In addition, a number of other cell lines may be used within the present invention, including Rat E~ep ~ (ATCC CRL
1600), Rat Hep II (ATCC CRL 1548), TCMK (ATCC CCL 139), Human lung (~TCC CCL 75.1), Human hepatoma (ATCC I~TB-52), Hep G2 (ATCC E1TB 8065), Mouse liver (ATCC CC 29.1) and DUKX
cells (Urlaub and Chasin, Proc. Natl. Acad. Sci. USA 77:
4216-4220, 1980).
I0 The 293 cell line (ATCC CRL 1573; Graham et al., J. Gen. Virol. 36: 59-72, 1977) is particularly preferred, due to its ability to efficiently process protein C to the two-chain form. This cell line is transformed with human adenovirus 5 DNA and has the Ad5 ElA gene integrated into its genome. Preferred expression vectors for use with 293 cells will include an adenovirus promoter. Neomycin resistance is a preferred selectable marker for use in 293 cells. A preferred BHK cell line is the tk- BHK cell line B~IK570 (Waech~er and Baserga, Yroc. Natl. Acad. Sci. USA
79: 1106-1110, 1982).
The copy number of the integrated gene sequence may be increased through amplification by using certain selectable markers (e.g., dihydrofolate reductase, which confers resistance to methotrexate). The selectable marker is introduced into the cells along with the gene of interest, and drug selection is applied. The drug concen-tration is then increased in a stepw;se manner, with selection of resistant cells at each step. ny selecting for increased copy number of cloned sequences, expression levels of the encoded protein may be substarltially increased.
Protein C produced according to the present inven-tion is preferably purified, as by affinity chromatography on an anti-protein C antibody column. Additional purifica-tion of the column eluate may be achieved by conventional chemical purification means, such as high-performance liquid chromatography (I~PLC).

~ 1340263 ~rotein C produced according to the present inven-tion may be activated by removal of the activation peptide from the amino terminus of the heavy chain. Activation may be achieved by incubating protein C in the presence of ~-thrombin (Marlar et al., Blood 59: 1067-1072, 1982), trypsin (Marlar et al., ibid.), Russell's viper venom factor X activator (Kisiel, ibid.) or the commercially available venom-derived activator Protac C (American Diagnostica).
To summarize the examples which fo]low, ExampLe 1 describes the cloning of ~NA sequences encoding human protein C. Example 2 describes the construction of a ~ull-length coding sequence for protein ~ from the sequences isolated in Example 1. Example 3 describes the construc-tion of expression vectors for the protein C ~NA. Example 4 describes the production of protein C using transfected mammalian cells. Example 5 describes a ~ull-1ength c~NA
encoding protein C and its expression in transfected mammalian cells. Example 6 describes the production of activated protein C in BIIK and 293 cells. ExampLe 7 describes the production of protein C from a precursor having a modified cleavage site. Example 8 describes the use of the Factor VII pre-pro peptide or the prothrombin pre-pro peptide to direct the secretion Or prol:ein C from transfected cells.

EXAMPLE:S
Restriction endonucleases and other L)NA modifica-tion enzymes (e.g., ~4 polynucleotide kinase, calf alka]ine phosphatase, Klenow DN~ polymerase, T4 polynucleotidc ligase) were obtained from ~e~hesda Research Iabora~ories (BRL) and New England E~iolabs and were used as directed by the manufacturer, unless otherwise noted.
Oligonucleotides may be synthesized on an Applied Biosystems Model 380 A l)NA synthesizer and purified by polyacrylamide gel electrophoresis on denaturing gels.
E. coli cells may be transformed as described by Maniatis et al. (Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, 1982). M13 and pUC cloning vectors and host strains were obtainëd from BRL.

Example 1 Cloning of DNA Sequences Fncoding l~uman Protein C
~ cDNA coding for a portion of human protein C
was prepared as described by E~oster and ~avie (ibid.).
~riefly, a lgtll cDNA library was pr~2pared from human liver mRNA by conventional methods. Clones were screened using 125I-labeled afIinity-purified antibody to human protein C, and phage were prepared from posit;ve clones by the plate lysate method (Maniatis et al., ibid.), followed by banding on a cesium chloride gradient. The cDNA inserts were removed using Eco RI and subcloned into plasmid pUC9 (Vieira and Messing, Gene 19: 259-268, 1982). Restriction fragments were subcloned in the phage vectors M13mplO and M13mpll (Messing, Meth. in Enzymo109y 10~: 20-77, L983) and sequenced by the dideoxy method (Sanger et al., L'roc. Natl.
Acad. Sci. USA 74: 5463-5467, 1977). A clone ~as selected which contained DNA corresponding to the known partial sequence of human protein C (Kisiel, ibid.) and encoded protein C beginning at amino acid 64 of the light chain and extending through the heavy chain and into the 3' non-coding region. This clone was designated ~IC1375. Asecond cDNA c]one coding for protein C from amino acid 24 was identified. The insert from this clone was subcloned into pUC9 and the plasmid designated p~C~6L (Eigure 1).
This clone encodes a major portion of protein C, including the heavy chain coding region, termination co~-30n, and 3' non-coding region.
']'he cnrlA insert from ~HCI375 was nick translated using Q_32p dN'I'P's and used to probe a tluman genomic library in phage ~ Charon 4A (Maniatis et al., Cell 15:
35 687-702, 1978) using the plaque hybridi%ation procedure o~
Benton and Davis (Science L96: 181-182, 1977) as modified by Woo (Meth. in Enzymology 68: 381-395, 1979). L'ositive - 134026.~

clones were isolated and plaque-purified (Foster et al., Proc. Natl. Acad. Sci. USA 82: 4673-4677, 1985) using a cDNA for human protein C (9) as the hybridization probe.
The cDNA started at amino acid 64 of human protein C and extended to the second polyadenylylation signal (9). It was radiolabeled by nick-translation to a specific activity of 8 x 108 cpm/~g with all four radioactive ([~-32P]dNTP) deoxynucleotides. The probe was denatured and hybridized to the filters at a concentration of 1 x 106 cpm/ml in a hybridization solution containing 6 x NaCl/Pj(1 x NaCl/Pi=
0.15 M NaCl/0.015 M sodium citrate, pH7.0),5 x Denhardt's solution (1 x = 0.02% polyvinylpyrrolidone/0.02%
Ficoll/0.02% bovine serum albumin), 0.1% sodium dodecyl sulfate, 100~g of yeast tRNA per ml, and 50% formamide at 42~C for 60 hr. The filters were washed in 1 x NaCl/P
containing 0.1% sodium dodecyl sulfate at 68~C for 1 hr and exposed to x-ray film for 16 hr. Positive clones were than isolated and plaque-purified. Phage DNA preparcd from posi-tive clones (Silhavy et al., in Exper1ments with Gene rusion~ Cold Spring Harbor Laboratory, l984) was digested with Eco RI or Bgl II and the genomic inserts purified and subcloned in pUC9. Insert restriction fragments were subcloned into Ml3 vectors and sequenced to confirm their identity and establish the DNA sequence o~ the entire gene.
The cDNA insert of p~C~6L was nick translated and used to probe the phage l Charon 4A library. One genomic clone was identified which hybridized to probes made from the 5' and 3' ends of the cDNA. This phage clone was digested with ~co RI and a 4.4 kb fragment, corresponding to the 5~ end of the protein C gene, was subcloned into pUC9. The resultant recombinant pLasmid was designated pHCR4.4. Complete DNA sequence analysis revealed that the insert in pHCR4. 4 comprised two exons of 70 and 167 base pairs separated by an intron of 1263 bp. The first exon encodes amino acids -42 to -l9; the second encodes amino acids -l9 to 37. Sequence analysis confirmed the DNA
sequence of the entire protein C gene.

21a 1 3 ~ 2 6 3 As noted above, it is then necessary to remove the intron in order to use a genomic clone to construct an acceptable coding sequence for use wi.thin the present invention.

~xample 2 Construction of a Fu.ll-Length Coding Seqllence ~or l~rc)teln C
A full-length coding sequence for ~rotein C, includ.ing the pre-pro peptide, is constructed by joining the appropriate fragments of the cDNA and genomic clones.
This is accomplished by removing the intron from the genomic clone (p~CR4.4) and joining the fused exons to ~he cDNA (from pl~CA61,) at convenient restriction si~es. lhe desired genomic:cDNA junction is then generated by looping .-- . . , 22 13~2~3 out unwanted sequences by oligonucLeotide-directed deletion mutagenesis.
Plasmid pHCA6L contains the protein C partial cDNA cloned in the Eco RI site of pUC9 (Figure 1). The cDNA insert is subcloned in two Lragments to prepare it for joining to the 5'-most coding region from the genomic c]one.
Plasmid pHcA6L is digested with F~co RI and Sal 1, and the reaction mixture is extracted with phenol and CHC13, then ethanol-precipitated. The resulting DNA fragments are resuspended in 1igation buffer, and T4 DNA ligase is added.
The ligation mixture is incubated at 15~C ~or 14 hours. An aliquot of the ligation mix is used to transform ~. coli JM83, and the cells are plated on LB agar containing X-gal.
White colonies are selected, and plasmid ~NA is prepared.
The DNA is analyzed by restriction enzyme digestion to identify clones containing the 3 ' portion of the cL)NA (ca.
1450 bp insert) and the 5' portion of the cDNA (ca. 65 bp insert). These clones are designated p9C3' and p9C5 ', respectively (Figure 6).
The 5' coding region missing from the cDNA is contained in exons I and 11 of the genomic clone pllCR~
This plasmid contains an insert of approximately 4400 base pairs and terminates on its 3 ' end at an ~co RI site located in intron B.
To remove the coding sequences from pHCR4.4, the plasmid is digested with PstI and Eco Rr and the resulting fragments are separated by e]ectrophoresis in an agarose ge]. l'he ca. 2540 bp ~ragmerlt containing exons r and 11 is isolated Lrom the gel and extracted with CTAB (Langridge, et al., Ana]yt. Biochem. 103: 264, 1980). This fragment, designated 5'P-R, is subcloned into pUC9 to produce plasmid p5 ' P-R ( ~ igu re 7).
l'he intron in p5'P-R (designated intron A), is removed in a two-step process (Figure 7). The plasmid is digested with Apa 1, which cleaves at a unique site in the intron and leaves 3' overhanging ends. The lineari7.ed plas-mid is then treated with na] 31 exonuc]ease or T4 polymer-~ 23 I3402~3 ase to remove approximately ~00 bp from each end and the resultant fragment ends are blunted with Sl nuclease. l'he linearized plasmid is recirculari.zed with ligase and used to transform E. coli JM83. Plasmid DNA is extracted and analyzed for the presence of the Sma I and Sst I restric-tion sites in intron A, and a plasmid having a Sma I-SstI
fragment reduced to 300-400 bp i.s chosen and designated p5'P~aR.
The remainder of intron A is removed by oligo-nucleotide-directed deletion mutagenesis, essentially as described by Zoller and Smith (Manual for Advanced Techniques in Mo~ecular Cloning Course, Cold Spring Harbor Laboratory, 1983) for the two-primer method. p5'P~aR is digested with ~st I and Eco RI, and the protein C fragment is subcloned into Pst I + Eco RI-digested M13mp9. Plus strand phage DNA is prepared as templ.ate and annealed to oligonucleotide mut-L (Table 1). 'I~his mutagenic oligo-nucleotide comprises sequences complementary to the exon I
and II sequences to be joined. The Ml3 universal sequenc-ing ~rimer is annealed 3'to mut-l. on the same template.
The primers are extended using DNA polymerase I (Klenow fragment) and nucleoside tr.iphosphates in the presence of T4 ~NA ligase. The resulting duplex ~NA circles are transformed into E. coli JMlO3 and the resulting pl.aques screened under stringent hybridi~ation conditions using the 32P-labe]ed mutagenic oligonucleotide as probe. VNA from positive p]aques is iso~ated and sequenced using oligo-nucleotide primer-l. (Table 1), which primes in exon Il, allow;ng the determination of the DNA sequerlce across the del.etion junction. A molecule having the correct inframe fusion of exons I and lI is selected. 'rhe Pst.[-EcoRL frag-ment is isolated from the M.l3 repl.icative forln by restric-tion endonuclease digestion and agarose gel electrophoresis and is subcloned into pUC9 to produce plasmid p5'I-II
(I~igure 7).
Referring to Figure 8, to join the 5' coding region to the cDNA, the ca. 1277 bp Pst l-Eco Rl fragment 13~0263 of p5'I-II is isolated from a Pst I + Eco RI digest of the plasmid and purified by agarose gel electrophoresis. The 65 bp 5'-most cDNA fragment is isolated from a Sal I + Eco RI digest of p9C5' and purified by electrophoresis on an acrylamide gel. The two fragments are ligated at the;r Eco Rl termini, and the resulting ca. 1330 bp Pst l-Sal fragment is subcloned into Pst I + Sal l-digested M13mp9 (Figure 8). ~lus strand phage DNA is prepared as template for oligonucleotide-directed deletion mutagenesis. Oligo-nucleotide mut-2 (Table 1) is annealed to the template, and oligonucleotide mut-3 (Table 1) is annea~ed upstream as second primer. The primers are extended as described above.
Oligonucleotide mut-2 directs the fusion of exon Il sequences encoding amino acids 23-26 to the cDNA at codon 27. The second primer (mut-3) ;ntroduces an ~co Rl site 35 bp upstream from the start of translation. 'rhe resulting phage are screened for the absence of Nco I and Xho I sites and for the presence of the introduced Eco Rl site. ~hage DNA showing the desired restriction pattern is sequenced using primer-2 ('l'able 1) to verify the presence of the correct junction between exon JI and the c~N~. ~'hage DNA
with the correct sequence is selected, and the ~st I-Sal I
fragment comprising the 5' coding region is isolated from the replicative form of the M13 recombinant phage. The fragment is purified by agarose gel electrophoresis and inserted into Pst I + Sal I-digested pUC9 to produce plasmid pC5'end.
Rererring to Eigure 9, plasmid pC5'end is digested with EcoRI and Sal 1, and the 5' protein C frag-ment is purified by agarose gel electrophoresis and extrac-tion with CTAB. The remainder of the cDN~ is iso:lated as a Sal I - F.co RI ~ragment from p9C3'. The two rragments are joined in a three-part ligation to Eco RI- digested pUC9.
The ligation mixture is used to transform E. coll JM83, the 35 cells are plated on LB + X-ga], and pLasmid DNA is iso~ated from white colonies. The resultant plasmid is designated 13~0263 pMMC. It contains ~he complete coding sequence for human protein C on a ca. 1500 bp Eco RI fragment.

T~BI,E I
Oligonucleotide Sequence mut-l 3'CGA CGA G~A crG AGT CAC AA5 mut-2 3'CTG AAG CTC c~r~ CGG T'l'C C'l'T TAA5 mut-3 5 G~ GGA ATT CTG AGC3 primer-l 5 TT'r GCG GA1' CCC CAG3 primer-2 5 CGA CGT GCT ~'GG ACC3 Example 3 Construction of Expression Vectors for ~'rotein C
The protein C-encoding insert is removed from pMMC as an Eco Rl fragment and inserted into a suitable mammalian cell expression vector. An exemplar~ vector is pD7, comprising the SV40 enhancer and the adenovirus 2 major late promoter and tripartite leader.
Ylasmid pD7 is generated ~rom plasmid pl)lll;RIII
(Berkner and Sharp, Nuc. Acids. Res. l3: 841-857, 1985).
l'he Pst I site immediately upstream from the DH~-R sequence in pDl~FRIIr was converted to a Bcl I site by digesting 10 ug Or plasmid with 5 UllitS of Pst I for l0' at 37"C in 100 uL buffer A (l0 mM 1'ris pl-l 8, 10 mM MgC]2, 6 mM NaCl, 7mM B-MSTt). The ~NA was phenol extracted, EtOI~ precipi-tated, and resuspended in 40 ul buffer ~ (50 mM 'rris p[l 8, 7 mM MgCl2, 7mM R-MSH) containing l0 mM dCTP and 16 units T4 DNA polymerase and incubated at 12~C for 60 minutes.
Following EtOII precipitation, the l)NA was ligated to 2.5 ug kinased Bcl I linkers in 14 ul bufrer C (l0 mM 'I'ris pH 8, l0 mM MgCl2, l mM D'l"l', 1.~ mM ATP) containing 400 uni~s ']'~
polynucleotide ligase for 12 hours at l2CC. Following phenol ex~raction and EtOI~ precipitation, the DNA was resuspended in 120 ul buffer D (75 mM KCl, 6 mM Tris pH 7.5, ]0 mM MgC12, l mM D~T), digesl-ed with 80 units Bcl I for 60 minutes at 50~C, then electrophoresed throuyh agarose. Form III plasmid DNA (10 ug) was isolated from the gel, and ligated in 10 ul buffer C containing 50 units T4 polynucleotide ligase for 2 hours at 12~C, and used to transform E. coli HB101. Positive colonies were identified by rapid DNA
preparation analysis, and plasmid DNA (designated pDHFR') prepared from positive colonies was transformed into dAM E.
coli.
Plasmid pD2' was then generated by cleaving pDHFR' (15 ug) and pSV40 (comprising Bam HI digested SV40 DNA cloned into the Bam HI site of pML-1) (25 ug) in 100 ul buffer D with 25 units Bcl I for 60 minutes at 50 degrees C, followed by the addition of 50 units Bam HI and additional incubation at 37~C for 60 minutes. DNA fragments were resolved by agarose gel electrophoresis, and the 4.9 kb pDHFR' fragment and 0.2 kb SV40 fragment were isolated. These fragments (200 ng pDHFR' DNA and 100 ng SV40 DNA) were incubated in 10 ul buffer C containing 100 units T4 polynucelotide ligase for 4 hours at 12~C, and the resulting construct (pD2') was used to transform E. coli RRI.
Plasmid pD2' was modified by deleting the "poison"
sequences in the pBR322 region (Lusky and Botchan, Nature 293:
79-81, 1981). Plasmids pD2' (6.6 ug) and pML-1 (Lusky and Botchan, ibid.) (4 ug) were incubated in 50 ul buffer A with 10 units each Eco RI and Nru I for 2 hours at 37~C, followed by agarose gel electrophoresis. The 1.7 kb pD2' fragment and 1.8 kb pML-1 fragment were isolated and ligated together (50 ng each) in 20 ul buffer C containing 100 units T4 polynucleotide ligase for 2 hours at 12~C, followed by transformation into E. coli HB101.
Colonies containing the desired construct (designated pD2) were identified by rapid preparation analysis. Ten ug of pD2 were then digested with 20 units each Eco RI and Bgl II, in 50 ul buffer A for 2 hours at 37~C. The DNA was electrophoresed through agarose, and the desired 2.8 kb fragment (fragment C) comprising the pBR322, 3' splice site and poly A sequences was isolated.
To generate the remaining fragments used in constructing pD3, pDHFRIII was modified to convert the Sac II
(Sst II) site into either a Hind III or Kpn I site. Ten ug pDHFRIII were digested with 20 units Sst II for 2 hours 27 13~026~

at 37~C, followed by phenol extraction and ethanol precipi-tation. Resuspended DNA was incubated in 100 ul buffer B
containing 10 mM dCTP and 16 units T4 ~NA polymerase for 60 minutes at 12~C, phenol extracted, dialyzed, and ethanol precipitated. DNA (5 ~g) was ligated with 50 ng kinased Hind III or Kpn I linkers in 20 ul bu~fer C containing 400 units T4 DNA ligase for 10 hours at 12~C, phenol extracted, and ethanol precipitated. After resuspension in 50 ul buffer A, the resultant plasmids were digested with 50 units llind III or Kpn I, as appropriate, and electro-phorescd through agarose. Gel-isolated DNA (250 ng) was ~igated in 30 ul buffer C containing 400 units ~4 DNA
ligase for 4 hours at 12~C and used to transEorm E. coli RRI. The resultant plasmids were designated pDllFRIlI (Hind III) and pDHFRIII (Kpn 1). A 700 bp Kpn I-Bgl II fragment (fragment A) was then purified from pDllFR-lII (Kpn I) by digestion with Bgl lI and Kpn I followed by agarose gel electrophoresis.
The SV40 erlhancer sequence was inserted into pDHERIII (l~ind III) as follows: 50 ug SV40 ~NA was incubated in 120 ul buffer A with 50 units llind III for 2 hours at 37~C, and the llind Ill C SV40 fragment (5089-968 bp) was gel purified. Plasmid pDIlFRIII (Hind I[l) (10 ug) was treated with 250 ng calf intestina] phosphatase for 1 hour at 37~C, pheno] extracted and ethanol precipitated.
The ]inearized plasmid (50 ng) was ligated with 250 ng llind III C SV40 in 16 ul buffer C for 3 hours at ]2~C, using 200 units ~4 polynucleotide ligase, and transformed into E. coli IIB101. A 700 base pair Eco RI-Kpn I ~ragment (frag-ment B) was then isoLated from this plasmid.
For the final construction of pD3, Eragments Aand B (50 ng each) were ligated with 10 ng fragment C with 200 units ~'4 polynucleotide ligase for 4 hours at 12~C, followed by transfection of E. coli RRI. Positive colonies were detected by rapid preparation analysis, and a large-scale preparation of pD3 was made.

~ lasmid pD3 i.s modified to accept the insertion of the protein C sequence by converting the Bcl I insertion site to an Eco RI site. lt i.s first necessary to remove the Eco RI site present in pD3 at the leftmost terlninus of the adenovirus 5 0-1 map unit sequences by converting it to a Bam HI s;.te via conventiona]. linkering ~rocedures. Brief-ly, the plasmid is digested wi.th Eco RI and the linearized DNA treated with T4 DNA polymerase and all four deoxynucle-otide triphosphates to generate blunt termini. The plasmid is then ligated to octonucleotides comprising the Bam HI
restriction site, the DNA digested wit.h Bam HI to remove excess linkers, and the fragment comprising the mamma]ian cell expression sequences is cloned into the ~am Hl site of pML-l. The resultant plasmid is transformed into E. coli EIB101, and plasmid DN~ is prepared and screened ~or the correct conversion. In a simil.ar manner, the Bcl I site ;s converted to an Eco Rl site using appropriate octonucle-otide linkers. The resultant vector is known as p~7. lhe 1.5 kb protein C Eco RI fragment from p~C is then inserted ;nto the Eco RI site of p~7 to produce the expression vector p~7C (Figure 1.0).
A vector enabLing expression of the protein C
sequence from a polycistronic message is constructed by using pD5, a plasmid similar to pn3 which contains a r)HFR
coding sequence lacking most of the 5' non--coding region.
The DHFR sequence is further modified to reduce its binding affi.nity to methotrexate.
The vector p~5 is constructed by a method ana]ogous to that described for pl)3, and difrers from pL)3 only in that a Bam l-II si~e is t:he site of insertion of heteroLogous DNAs, and that the Bc] I-Bam 11l SV~0 fragment containing the SV40 polyadenylat;on si.gnal is in the ]ate orientation.
The Dl-IFR sequence is modified by first digesting pD~lFRIII with Pst I and Sst I and isolating the 400 bp DIIFR
fragment. This is subcl<!ned in an M13 phage vector and mutageni7.ed as described by Simonsen and l.evinson (Proc.

29 13~02~3 Natl. Acad. Sci. USA 80: 2495-2499, 1983). Mutagenesis results in a single base pair change in the DIIFI~ sequence.
The altered fragment is then reinserted into pDHFRIII to produce plasmid pDHE'Rrl r I .
The 5' non-coding region of the DHFR sequence is then removed. Plasmid pDI~ERrITI is cleaved with Fnu 4HI, which cuts the plasmid at approximately 20 sites, then treated wi~-h T4 DNA polymerase and all four deoxynucleotide triphosphates to generate blunt termini. ~am HI linkers are ligated to the ends, and the mixture digested with Bam Hl and Nco 1. A 0.6 kb Bam HI-Nco I ~ragment comprising the DI~E'Rr cDNA is isolated. ~lasmid pDHE'RIIl is digested with Nco I and Bam HI and the 0.2 kb fragment comprising the SV40 polyadenylation signal is isolated. l'he poly-adenylation signal, in the early orientation, is then ligated to the DHE'Rr fragment. After digestion with Bam HI, the resultant Bam III fragment is then inserted into the Bam ~ll site of pl)5 and the ligation mixture used to trans-form E. coli IIB101. ~lasmid DNA is prepared and screened by restriction endonuclease digestion. A pLasmid having the DE~l;'Rr insert in the correct orientation ~or transcrip-tion from the Ad2 major late promoter is designated pD5(DIIERr).
To express protein C using plasmid pD5(DHFRr), pMMC is digested with ~co RI and the 1.5 kb protein C
fragment is isolated. The Eco Rt ttermini are converted to Bcl I termini by ]inkering. Plasmid pD5(DHFl~r) is partial-ly digested with Bam Ht to cleave it at the 5' end of the t)Hl~Rr sequence and is ligated to the protein C fragment.
Plasmid DNA is screened for the proper orientation and insertion of the protein C fragment. l'he resultant vector, designated pD5(~C-DH~'Rr), is illustrated in E'igure ]l.

Example 4 Expression of Protein C in Transfected Mamma]ian Cells Baby hamster kidney (BHK) cells (American Type Culture Col]ection accession number CCL10) are trans{ected 13402~3 with pD7C by calcium phosphate co-precipitation (Wigler et al., CeLl 14: 725, 1978; Corsaro and Pearson, Somatic Cell Genetics 7: 603, 1981; and Graham and Van der Eb, Virology 52: 456, 1973). The cells are grown at 37~C, 5~
C~2 in Dulbecco's medium (plus 10% heat-inactivated fetal calf serum and supplemented with L-glutamine and penicillin-streptomycin) in 60 mm tissue culture E'etri dishes to a confluency of 20%. A totaL of 10 ug Or ~NA is used to transfect one 60 mm dish: 3.75 ug of pD7C, 1.25 ug of pKO-neo (Southern and Berg, J. Mol. Appl Genet 1:
327-341, 1982) and 5 ug of salmon sperm DNA. rl'he ~N~s are precipitated in 0.3 M NaOAc, 75~ ethanol, rinsed with 70%
ethanol and redissolved in 20 ul 10 mM Tris-HCl p~-18, 1 mM
EDTA. The DNA is combined with 440 ul I~2O and 500 ul of 280 mM NaCl, 1.5 mM NaI~PO4, 12 mM dextrose, 50 mM ~IEP~S pll 7.12. Sixty ul of 250 mM CaC12 are added dropwise to the above mixture and the solution is allowed to stand at room temperature for 30 minutes. The solution is then added to the cells and the cells returned to 37~C for 4 hours. The medium is removed and 5 ml of 20~- DMSO in Dulbecco's with serum are added for 2 minutes at room temperature. The dish is then washed rapidly with 2 changes of medium and incubated in ~resh medium overnight. Twenty--four hours after the addition of the DNA, the medium is removed and selective medium (10 mg/ml of G418, 498 u/mg, Gibco, in Dulbecco's with serum) added. After approximately 10-13 days, individua] clones, representing cells that have incorporated the pKO-neo gene and are thus resistant to G418, are transferred to 96-welL plates and gr-own up for 30 protein assays in Dulbecco's plus 10~o fetal calf serum.
To assay for protein C, the medium is separated from the cells and cellular debris by centrifugation, and assayed for protein C polypeptide and biological activity.
The cells are removed ~rom the plates with trypsin, washed 35 with fresh medium, centrifuged and frozen at -20~C. For assay, the cell pellets are thawed in L)~S, pelleted, and 31 13~0263 resuspended in PBS containing 0.25% Triton X-100*. Samples are diluted and assayed for polypeptide and activity.
The enzyme-linked imml~nosorbent assay (ELISA) for protein C is done as follows: Affinity-purified polyclonal antibody to human protein C (100 ul of 1 ug/ml in 0.1 M Na2C03, pH 9.6), is added to each well of 96-well microtiter plates and the plates are incubated overnight at 4~C. The wells are then washed three times with PBS (5 mM phosphate buffer, pH 7.5, 0.15 M NaCl) containing 0.05% Tween-20*
and then incubated with 100 ul of 1% bovine serum albumin, 0.05%
Tween-20* in PBS at 4~C overnight. The plates are then rinsed several times with PBS, air dried, and stored at 4~C. To assay samples, 100 ul of each sample is incubated for 1 hour at 37~C in the coated wells and the wells are rinsed with 0.05% Tween-20* in PBS. The plates are then incubated for 1 hour at 37~C with a biotin-conjugated sheep polyclonal antibody to protein C (30 ng/ml) in PBS containing 1% bovine serum albumin and 0.05% Tween-20*. The wells are rinsed with PBS and incubated again for 1 hour at 37~C
with avidin conjugated to alkaline phosphatase in PBS containing 1%
bovine serum albumin and 0.05% Tween-20*. The wells are then rinsed with PBS, an alkaline phosphatase activity is measured by the addition of 100 ul of phosphatase substrate (Sigma 104; 600 ug/ml) in 10% diethanolamine, pH 9.8, containing 0.3 mM MgCl2. The absorbance at 405 nm is read on a microtiter plate reader.
Protein C biological activity is assayed by its ability to prolong the kaolin-cephalin clotting time of plasma following its activation as described in Example 5C.

Example 5 Expression of a Full Length cDNA Encoding Protein C
A. Isolation of cDNA. A genomic fragment containing an exon corresponding to amino acids -42 to -19 of the pre-pro peptide (Exon 1 in Figure 4) of protein C was isolated, nick translated, and used as a probe to * trade-mark . ~' ''! ~, 3 1 ~ 2 6 3 screen a cDNA library constru ted by the technique of Gubler and ~o~fman (Cene 2~: 263-269, L983) using mRNA ~rom HepG2 cells- This celL line was derived from human hepato-cytes and was previousl~ shown to synthesize protein C
(Fair and Bahnak, ~lood 64: l94-204, 1984). Ten positive clones comprising cl)NA inserted into the Eco RI site of phage ~gtll were isolated and screened with an oligonucle-otide probe corresponding to the 5' non-coding region o~
the protein C gene. One clone was also positive with this probe and its entire nucleotide se~uence was determined.
The cDN~ contained 70 bp oE 5' untrans~ated sequence, the entire coding sequence for human prepro-protein C, and the entire 3' non-coding region corresponding to ~he second polyadenylation site (Figure 2).
B. Expression Vector Construction. The expres-sion o~ protein C cr~NA was achieved in the vector pDX.
~'his vector was derived from pD3 (described in Example 3 above) and pD3', a vector identical to p~3 except that the SV40 polyadenylation signal (i.e., the SV40 Bam~ll 12533 bp~
to Bclr [2770 bpl fragment) is in the late orientation.
Thus, pD3' contains a Bam ~I site as the site of gene insertion.
To generate pDX, the Eco Rl site in pD3' was converted to a ~clI site by Eco RI cleavage, incubation with Sl nuclease, and subsequent ligation with ~cl linkers. DNA was prepared ~rom a positiveLy identified colony, and the 1.9 kb Xho I-Pst I fragment containing tho altered restriction site was prepared via agarose gel electrophoresis. In a second modification, Bcl ~-cLeaved pr~3 was ligated with kinased Eco Rr-Bcl ~ adaptors (constructed from oligonucleotidcs ZC525, 5'GCAA'l'TC'r3': and ZC526, 5'GATCAGAATTCC3') in order to generate an ~co RI
site as the position for inserting a gene into the expres-sion vector. Positive colonies were identified by restric-~ion endonuclease analysis, and DNA from this was used to isolate a 2.3 kb Xho ~-Pst I fragment containing the 13402i~3 modified restriction site. The two above-described DNA fragments were incubated together with T4DNA ligase, transformed into E. coli HB101 and positive colonies were identified by restriction analysis.
A preparation of such DNA, termed pDX, was then made. This plasmid contains a unique Eco R1 site for insertion of foreign genes.
The protein C cDNA was then inserted into pDX as an Eco Rl fragment. Recombinant plasmids were screened by restriction analysis to identify those having the protein C insert in the correct orientation with respect to the promoter elements and plasmid DNA (designated pDX/PC) was prepared from a correct clone (Figure 12). Because the cDNA insert in pDX/PC contains an ATG
codon in the 5' non-coding region (see Figure 2), deletion mutagenesis was performed on the cDNA prior to transfection and expression experiments. Deletion of the three base pairs was performed according to stAn~rd procedures of oligonucleotide-directed mutagenesis. The pDX-based vector cont~ining the modified cDNA was designated p594.
C. cDNA Expression. Plasmid p594 was transfected into COS-l (ATCC CRL1650), BHK and 293 cells by calcium phosphate precipitation. Four hours later, fresh culture media (supplemented with 5 ug/ml vitamin K) were added. At appropriate times (usually 48 or 72 hours), the culture media were harvested and the cells were collected and lysed.
The protein C secreted into the culture medium was assayed by ELISA using the same affinity-purified polyclonal antibody which was used in the initial identification of the cDNA
clones. Results of the assays of COS-l and 293 cells Table 2) showed that protein C was secreted from the transfected cells. It was found that 293 cells gave consistently higher levels of protein C than did COS cells.
To assess the extent of gamma-carboxylation of the recombinant protein, samples of the culture media were subjected to barium citrate precipitation, a process which selectively precipitates only gamma-carboxylated proteins from plasman (Bajaj et al., J. Biol. Chem. 256: 253-259, L~ X.

3 ~0 2 63 1981). Over 70~ of the protein C antigenic material could be precipitated with barium citrate.
The recombinant protein C was assayed for anti-coagulant activity by measuring its ability to prolong coagulation. Dialyzed media samples were treated with Protac C (American Diagnostica) to activate the protein C.
The sampl,es were then added to an ln vitro clotting assay (Sugo et al., J. Biol. Chem. 260: 10453, 19~5). Briefly, ul each of normal pooled human plasma, rabbit brain cephalin (10 mg/ml in TBS [50m~ ris pH 7.5, l,50 mM NaCl]) and kaolin suspension (5 mg/ml in TBS) were m;,xed in a si,liconized glass tube. After preincubat;on at 37~C for 2 minutes, 100 u1 of activated protein C diluted in 'rBS was added and the 37~C incubation was continued for an additional 2 minutes. Clotting was then initiated by the addition of 50 ul of 25 mM CaC12, and the clott~ng time was recorded. ~'he activity of the recombinant material was shown to be essentially the same as that of naturally occurring protein C.
Protein C produced by ~ransfected BHK and 293 cells was further analyzed by Western blotting. Media samples were electropIloresed on denaturing geJs and blots were prepared and probed with radiolabeled antibody to protein C. Results indi,cated that about 20Vo oE the protein C from BHK ce]ls was in the two-chain form, while about 90~O
of that from 293 cells was processed to the two-chain form.

I 3 ~0 2 63 TAt3 LE 2 TRANS IENT EXPRESS ION At'1D SECRETION OF PROTEIN C
IN COS-l and 293 CELLS

ng/ml pro~ein C
cells plasmid in media C~S-l none O
COS-I pS94 lO
293 none O
293 p59~ 50 Example 6 Expression of Activated Protein C
The cDNA sequence for protein C was altered by site-specific mutagenesis to delete the portion encoding the activation peptide. 'rhe altered sequence was then transfected into BHK and 293 cel.ls and stably transfected cells were selected. Active protein C was detected in 20 culture media samples from both cell lines.
~ o delete the activation peptide coding sequence, plasmid p594 was digested with Sst I and the ~880 bp frag-men- was purified and inserted into the Sst I siLe of M13mplO. The 12 activation peptide codons were deleted by 25 oligonucleotide-directed deletion mutagenesis (Zo]ler and Smith, DNA 3: 479-488, l984) using the mutagenic oligonucle-otide 5 CTGAAACGACTCATrGAT3 . Replicative form DNA was prepared from mutant phage cLones and digested with Sst I.
1'he protein C fragment ~840 bp) was isolated and insertcd into Sst I digested p59~. The r~sultant plasmi.ds were screened for proper orientation of the Sst I fragment by rcstriction mapping using Bgl II. A correct plasmid wa~
selected and designated pPC829. Plasmid prc829 was sequenced to verify the presence of the desired coding sequence-Plasmid pPc82s was co-trans~ec~ed into B~IK cells (wi.~h plasmid pSVDHFR (Lee et al., Na~ure 294: 228-232, '"' X

13~0263 1981)) and 293 cells (with pKO-neo (Southern and Berg, J. Mol. Appl. Genet. 1: 327-341, 1982)) by cal.cium phosphate coprecipitation (Graham and van der ~b, Virology 52: 456-467, 1973). After 48 hours, culture media were harvested and assayed for protein C by ELISA. Results are shown in Table 3. At the same time, cultures were split 1:5 into media containing 500 ug/ml of G418 (293 cells) or 250 nM methotrexate (BHK cells). After 10 days in the presence of se]ective media, stably transfected colonies were screened for protein C production by immuno~il.ter assay (McCracken and Brown, ~ioTechni.ques, 82-87, March/April 1984). Plates were rinsed with P~S or No Serum medium (Dulbecco's plus penicillin-streptomycin, 5 ug/ml vitamin K). Teflon~ mesh was then placed over the cells.
Nitrocellulose filters were wetted with P~S or No Serum medium, as appropri.ate, and placed over the mesh. After four hours' incubation at 37~C, filters were removed and placed in buEfer A (50 mM Tris pH 7.4, 5 mM l'l~llA, 0.05%
NP-40, 150 mM NaCl, 0.25% ge].atin) for 30 minut:.es at room temperature. ~'he filters were incubated for 1 hour at room temperature, with shaking, in biotin-labeled sheep poly-clonal antibody to protein C, ] ug/ml in buffer A. ~'ilters were then washed in buffer A and incubated 1 hour at room temperature, with shaking, in avidin-conjugated horseradish peroxidase (Boehringer-Mannheim), ]:1000 in buffer A.
Filters were washed in buffer B, then in ll2O, and incubated in color reagent (60 mg EIRP color development reagent l~io-~adl, 20 ml methanol, 100 ul ll2O2 in 100 ml 50 mM l'ris pll 7.4, 150 mM NaCl). The reaction was stopped by transfer-ring t.he fil.ters to ~12~
Positive colonies were picked and grown in se~ec-tive media (containi.ng 500 ug/m] G418 or 250 nM metho-trexate, as appropriate) for 10 days. Culture media were assayed for APC activity by chromogen;c assay. Media samples were added to microtiter wells containing 100 ul of 0.2 mM Spectrozyme PCa (American Diagnostica #336) in 50 mM
Tris pll 7.5, 150 mM NaCl. Plates were incubated at 37~C

37 1341)2~i3 and the A405 measured at various time intervals. Represen-tative results from one transfected 293 cell line (desig-nated 829-20) are shown in Figure 13. Media from positive colonies of line 829-20 consistently showed higher activity with the chromogenic substrate for APC than did control media which had been incubated with non-transfected 293 cells for the same length of time (10 days).

TABL,E 3 10 TRANSIENT EXPRESSION OF ACTIVATED ~ROTEIN C (El.ISA) Protein C
Cell r. i ne ng~'ml in Med~a RIIK 2.7 ~5 Example _7 Modification of the ~rotein C r'rocessing_S_te A. Site-Speci~ic Mutagenesis. To enhance the processing of single-chain protein C to the two-chain ~orm, two additional arginine residues were introduced into the protein, resulting in a cleavage site consisti.ng of four basic amino acids. 'I'he resul.tant mutant precursor of protein C was designated PC962. Tt contains the sequence Ser-llis-Leu-Arg-Arg-Lys-Arg-Asp at the cleavage site.
Processing at the Arg-Asp bond resu]ts in a two-chain protein C molecule.
The mutant molecule was generated by altering the c]oned cDNA by site-specific mutagenesis (essentially as described by Zoller and Sm.ith, DNA 3: 479-488, 198~, for the two-primer method) using the mutagenic oJigonuc]eotide %C962 (5 ~GTCACCTGAGAAGAAAACG~GACA3 ). Plasmid p594 was digested with Sst I and the approximately 87 bp r ragment was cloned into M13mpll and sinyle-stranded tempJate DNA
was isolated. ~ollowing mu-tagenesis, a correct clone was identified by sequencing. Replicative form DNA was isolated, digested with Sst I, and the protein C fragment 13~02~3 was inserted into Sst I-cut p594. Clones having the Sst I
fragment inserted in the desired orientation were identified by restriction enzyme mapping. The resulting expression vector was desi.gnated pnX/PC962 (Figure 14).

B. Expression and Characterization o~ Pro~ein C.
~l.asmid pl~X/PC962 was co-transfected i.nto tk- ~IIK cells with pSV2-DI~FR (Subramani et al., Mol. Ce~ Biol. 1:
854-864, 1981) by the calcium phosphate procedure (essen-tially as described by Graham and van der Eb, ibid.). Thetransfected cells were grown in I)ulbecco's modified Eagle's medium (MEM) containing 10% ~etal calr serum, lX PSN
antibiotic mix (Cibco 600-56~0), 2.0 mM L-glutamine and vitamin K (5 ug/ml). The cells were selected in 250 nM
methotrexate (MTX) for 14 days, and the resulting colonies were screened by the immunofilter assay (Exampl.e 6). Six o~ the most intensely reacting colonies were picked by cylinder cloning and grown individual]y i.n 10-cm plates.
When the cultures were n~arly confluent, ~rotein C produc-tion levels were measured by ELIS~. Results are given inTable 4.

'l'A~LE 4 Cl_ne Ce]l Number (x 10 7) FIISA n~/ml pg/cell/day 962-1 1.1 2500 2.20 -2 0.~ 1250 1.56 -3 l.2 l.350 1.]2 30 -4 1.2 550 0.~6 -5 1.2 1550 .l.30 -6 1.2 950 O.~o The clonc ~I~K/962-l was grown in larger scale culture, and several hundred micrograms o~ protein C were purified by af~inity chromatography on a colun~n prepared by coupling 7 mg of polyclonal sheep antibody against human 13~02i5~

protein C to 2 grams of CNBr-activated Sepharose* 4B/Pharmacia, Inc., Piscataway, NJ). Cell culture medium was applied to the column, the column was washed with 100 ml TBS, and the protein C was eluted with TBS containing 3 M KSCN or with pH 11.5 buffer. Western blot analysis demonstrated that the mutant protein C was approximately 95% in the two-chain form, compared to about 20% two-chain protein C obtained from BHK cells transfected with the native sequence.
Milligram quantities of protein C were purified from either stable BHK cell clones expressing the PC962 mutant protein or stable 293 cell clones expressing the wild-type protein C (p594 transfected cells) using a monoclonal antibody column specific for the calcium-induced conformation of protein C. Cell culture media were applied to the column in the presence of 5mM CaCl2, the column was washed with TBS containing 5mM CaCl2, and the protein C was eluted with TBS containing lOmM EDTA. The use of this purification method permitted purification of completely active protein C without exposure to denaturing conditions. The purified protein C was analysed by SDS/PAGE followed by silver staining and was shown to be 95% pure.
The BHK-produced PC962 protein was assayed for its ability to be activated to a form which shows both amidolytic and anticoagulant activities. Affinity-purified protein samples were exhaustively dialyzed against TBS, then activated by incubation at 37~C for 1 hour with 0.1 volume of 1 unit/ml Protax C (American Diagnostica). Amidolytic activity was measured by adding aliquots of the activation mixture to 100 ul of 1 mM protein C substrate (Spectrozyme* PCa, American Diagnostica) in a microtiter well and measuring the change in A405 over time using a microtiter plate reader. Anticoagulant activity of the activated protein C was assayed as described by Sugo et al. (Ibid.). The affinity-purified PC962 protein was ~e~nqtrated to be fully active in both amidolytic and anticoagulant assays. Elution from the antibody column with pH
11.5 *trade-mark r ' ~

~o 1340~3 buffer was shown to yi~ld a protein wi~h higher activity than that obtained using 3 M KSCN elution.
Clonal cell Lines from the pnX~'~C962 transfec~ion into B~{K cells were also isolated by a process of limi~ing dilution. One plate of MTX-selected colonies (approxi-mately 300 colonies) was trypsinized, counted, and re-plated into microtiter wells at an average of 0.5 cell/well. These were grown up in selective media contain-ing 250 nM MTX. About 50% of the wells contained colonies.
Wells containing identifiable colonies (1-2 mm diameter) were assayed by E:LIS~ for protein C level in the media.
For this assay, fresh medium was added to all the wells, allowed to incubate for 75 minutes, then removed and assayed. Five colonies which gave 7~-minute accumulations of greater than 50 ng/ml (corresponding to over l000 ng,ml/day) were split into l0-cm plates for larger scale culture. Protein C production levels for thes~ c1ones ranged from l.l to 2.8 pg/cell/day.
A second plasmid, designated PC229/962, was constructed by inserting ~he PC962 cDNA into plasmid Zem229.
Zem229 is a pUCI~-based expression vector conlaining a unique Bam HI site ~or insertion ~f foreign DNA between the mouse metallothionein-I promoter and SV40 Lranscription terminator. Zem229 also contains an expression unit of the SV40 early promoter, mouse dihydrofolate reductase gene, and SV40 terminator. An Eco ~I fragment containing the PC962 cDNA from pDX/PC962 was ligated, with Eco R~-Bam Hl syn~hetic oligonucleotide adaptors, to Zem229, which had been cut with Bam HI and treated wi~.h phosphatase. The resulting vector is PC962/229, illustrated in Figure J4.
P~asmid PC962/229 was transfected inLo ~ cel~s by the calcium phosphate method. CeJLs were cultured in Dulbecco's MEM containing 10% fetal calf serum and 5 ug/ml vitamin K. The 48-hour transient expression level from this transfection was approximately 25 ng/mL. After 2 days, the ~ransfected cells were split into selective media containing 250 nM M~rX and cultured for an additional l4 ~r 41 13~02fi3 days. Three plates from this transfection, containing approximately 200 colonies each, were screened by the immunofilter assay, and the 24 most intensely reacting colonies were picked by cylinder cloning. ~hese were grown individually in 10-cm plates, and their protein C produc-tion levels were measured. Colonies producing between 1.1 and 2.3 pg/cell/day w~re used for the production of stable protein C-producing cell lines.
Expression vector pDX/PC962 and plasmid pKO-neo were co-transfected by the calcium phosphate method into 293 cells. Transfected cells were split into media contain-ing 500 ug/ml G~18 after 48 hours. After 10 days in selec-tive media, immunofilter assays were done and two clones were picked by cylinder cloning. Protein C production was found to range from 1 to 2 pg/cell/day. The cultures were scaled up, and protein C was purified by immuno-affinity chromatography. Greater than 95~ of the protein C was found to be in the two-chain form.
The structure of the 962 mutant protein prepared from B~K and 293 celJs was compared to that of wild-type protein C from 293 cells and from plasma. Analysis by SDS/PAGE followed by silver staining showed that all the recombinant proteins contained heavy and light chains which co-migrated with those of the plasma protein. The wild-type protein C synthesized in 293 cells contained a significant amount (approximately 20~) o~ singJe-c}lain, unprocessed protein of Mr=66,000, whereas the mutant protein produced in either cell type was essentially completely processed to two chains. N-terminal sequence analysis showed that both the light and h~avy chains of the recombinant wild-type and BHK/PC962 mutant proteins were properly processed. The extent of gamma carboxylation of the recombinant proteins was measured by two distlnct ELISA
systems. The first system recognizes both gamma-carboxyl-ated and non-carboxylated forms of the protein, while the second utilizes specific antibodies which only recognize protein C which has undergone a gla-induced conformational ~' 13~2B3 change in the presence of calcium. Analysis indicated that approximately 60% of the recombinant protein C produced in BHK cells and 90~-95% of that produced in 293 cells was sufficiently gamma carboxylated to be recognized by the specific antibodies.
The three recombinant proteins were also analyzed for amidolytic and anticoagulant activity and the results were compared to the activity of plasma protein C. PC962 from BHK cells and wild-type protein C from 293 cells both showed full amidolytic activity. In the anticoagulant assay, protein C from ~HK cells had essentially the same specific activity as plasma protein C, whereas both wild-type and PC962 mutant proteins from 293 cells consistently exhibited approximately 40% greater specific activity.
C. Modification of Activat~d Protein C Process-ing Site. A DNA sequence encoding an activated protein C
precursor with the processing site sequence Arg-Acg-Lys-Arg was constructed by mutagenesis of the wild-type protein C
sequence. The resultant sequence was analogous to that of pPC962, but lacked the portion ecoding the activation peptide.
The protein C sequence pr~sent in plasmid p59~
was altered in a single mutagenesis to delete the codons for the activation peptide and insert the Arg-Arg codons at the processing site. Mutagenesis was performed on the 870 bp Sst I fragment from p594 essentially as described in Example 7A using an o~igonucleotide having the sequence 5' CGC AGT CAC C1'G AGA AGA AAA CGA c rc ATT GAT GGG 3'.
The mutagenized sequence was used to construct expression vector pDX/PC1058 and the vector was co-trans-fected into BIIK celLs as described in Example 7~3. rrhe protein was purified on a polyclonal antibody column eluted with pH 11.5 buffer.
The activity of the 1058 protein was compared to that of activated plasma protein C and activated PC962.
Plasma protein C and PC962 (5 ~g/ml) were activated by .........

1340 2~3 treatment with l/10 volume Protac C (American [)iagnostica) for ~wo hours. Anticoagulant activity was assayed by combining 50 ~1 human plasma with 50 ~l activated protein C
and incubating ~he mixtures at 38~C ~or 150 seconds. To this mixture was added 50 ~1 activat~d cephaloplastin (~merican Scientific Products, McGaw E'ark, Il,) and the mixture was incubated at 37~C for 300 seconds. One hundred ~1 of 20 mM CaC12 was added and the clotting time was recorded. Data are presented in Figure 15.
Example 8 Use of the Eactor Vll and Pro-hr~mbin Pre-Pro ~eptides to Secrete L~rotein C
The factor v11 pre-pro pep~ide was substituted for the protein C pre-pro peptide in an erfort to obtain higher yields of proper1y processed pro~ein C. These hybrid constructs are then inserted into suitable expres-sion vectors and transfected into mammalian celL lines.
A cDNA encoding fac~or VJ1 has been described (Hagen et al., Proc. NatL. Acad. Sci. U5A 83: 2412-2416, 1986). Clone ~HVII565 comprises the coding seyu~nce ror a 38 amino acid pre-pro peptjde. This coding se~uence was isolated as an Eco Rl-l~ha I fragment of 140 bp.
The protein C sequence was isolated from p594 by partial cleavage with Sst I and complete digestion with Eco RI. A 1540 bp iragment extending from the Sst I site at codon +7 to the Eco Rl site 3' to the c~NA was isolated.
The factor V~I and protein C sequences wore then joined by means of an oligonucleotide ~inker which com-pletes the coding sequence for amino acids -3 to -1 of the ractor VlI pre-pro peptide and amin~ acids 1-8 o~ protein C.
The linker was constructed from two oligonuc~eotides having the sequences 5 CCGGCGC~CCAACTCCTl'CCTGGAGGAGCT3 and S CCTCCAGGAAGGAGTTGGCGCGCCGGCG3 . The two oligonucleotides were annealed and joined, in a four-part ligation, to the ~actor Vl1 pre-pro sequence, protein C cDNA and pUC9 which had been cleaved with Eco RI and treated with bacterial '' X ~--*

13~026~

alkaline phosphatase. The ligated DNA was used to trans-form E. coli (JM 101). Plasmid DNA was prepared and screened for the presence of a 171~ bp Eco Rl fragment. A
correct clone was designated p7/C-10.
The factor VII/protein C fusion was expressed in 293 cells. The Eco Rl insert from plasmid p7/C-10 was ligated to Eco RI-digested pDX. The resulting expression vector was used to co-transfect 293 cells as previously described. Forty-eight hour expression levels were assayed by ELISA and compared to those of 293 cells transfected with the wild-type protein C expression construct and untransfected cells. Results are presented in Table 5.

Protein n~/ml Factor VII/protein C123 Wild-type protein C 187 Control <1 The prothrombin leader sequence was constructed from the oligonucleotides listed in Irable 6 and fused to the mature protein C coding sequence. The oligonucleotides were kinased by combining 50 ng of each oligonucleotide with 1 unit of T4 kinase in 20 ul of kinase buffer containing 1 mM ATP. The reaction was allowed to proceed at 37~C for 30 minutes, then the mixture was heated to 65~C
for 10 minutes to inactivate the kinase.~0 l'ABLE 6 ZC 1323 5' CCT CCA GGA AGG AGT TGG c rc GCC GGA 3' ZC 1324 5' CGC GTC CGG CGA GCC AAC TCC TTC CTG GAG GAG
CT 3' 13~-d2~3 ZC 1378 5' AAT TCC ACC ATG GCT CAT GTG AGA GGA CTG CAA
CTG CCT GGC TGC CTG GCT CTG GCT GCT CTG TGC AGC
C'rG GTG CAC AGC CAG CAT GTG TTC CTG GCT CCT CAG
CAG GCC AGG AGC CTG CTG CAA 3' ZC 1379 5' CGC GTT GCA GCA GGC TCC TGG CCT GCT GAG GAG
CCA GGA ACA CAT GCT GGC TGT GCA CCA GGC TGC ACA
GAG CAG CCA GAG CCA GGC AGC CAG GCA GT~r GCA GTC
CTC TCA CAT GAG CCA TGG TGG 3' The prothrombin leader was then assembled. Fifty ng of Eco RI, Sst I-cut M13mpl9 was combined with 2.5 ng each of the kinased oligonucleotides in 20 ul of lx ligase buffer containing 1 mM ATP and 4 units of T4 ligase. The mixture was incubated at 15~C for 48 hours and transformed into competent E. coli JM101 cells. A clear plaque was selected and phage DNA was prepared. DNA sequencing confirmed that the correct sequence had been constructed.
The prothrombin leader was then joined to the protein C sequence. RF DNA was prepared from the phage clone containing the synthesized leader and a 150 bp Eco RI-Sst I fragment was isolated. Plasmid p594 was digested to completion wth Eco RI and partially digested with Sst I
and the 1540 bp protein C fragment was recovered. These fragments were ligated with Eco RI-cut pDX and the ligation mixture was used to transform competent E. coli HB101 cells.
Plasmid DNA was isolated from transformant colonies and analyzed by restriction digestion to confirm that the fragments had been assembled in the correct orientation.
From the foregoing it will be appreciated that, although specific embodiments of the invention have been described herein for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the invention. Accordingly, the invention is not to be limited except as by the appended claims.

Claims (16)

1. A DNA sequence which codes for a protein C-like polypeptide or an activated protein C-like polypeptide, wherein said sequence codes for the amino acid sequence of protein C or activated protein as shown in Figure 2, said aminoacid sequence comprising an amino acid substitution selected from the group consisting of:
substitution of residues 156-157 with the amino acid sequence (R1)n-R2-R3-R4, wherein R1, R2, R3 and R4 are Lys or Arg and n=0, 1, 2 or 3;
substitution of residue 158 with a non-acidic amino acid residue selected from the group consisting of Ala, Ser, Thy and Gly;
substitution of residue 154 with a basic amino acid residue selected from the group consisting of Lys or Arg; and substitution of residues 156-157 with Lys-Lys or Arg-Arg.

2. The DNA sequence of claim 1 wherein the cleavage site at residues 156-157 is replaced with Arg-Arg-Lys-Arg.

3. A DNA sequence which codes for human protein C or human activated protein C as shown in Figure 2, said sequence further coding for the pre-pro peptide of a protein selected from the group consisting of factor VII, factor IX, factor X, prothrombin and protein S.

4. An expression vector capable of integration in mammalian host cell DNA, said expression vector including a promoter followed downstream by a DNA
sequence according to any of claims 1-3, said DNA sequence being followed downstream by a polyadenylation signal, wherein transcription of the DNA
sequence is directed by the promoter.

5. Mammalian cells transfected with an expression vector according to claim 4.

6. The cells of claim 5 wherein said cells are selected from the group consisting of COS, BHK, Rat Hep I, Rat Hep II, TCMK, Human lung, Human hepatoma, Hep G2, Mouse liver, DUKX and 293 cells.

7. A method for producing a protein C-like polypeptide or an activated protein C-like polypeptide, comprising:
introducing into a mammalian host cell an expression vector according to claim 4;
growing said mammalian host cell in an appropriate medium; and isolating the protein product encoded by said expression vector and produced by said mammalian host cell.

8. The method of claim 7 wherein said cell is selected from the group consisting of COS, BHK, Rat Hep I, Rat Hep II, TCMK, Human lung, Human hepatoma, Hep G2, Mouse liver, DUKX and 293 cells.

9. A method for producing a protein C-like polypeptide or an activated protein C-like polypeptide, comprising:
introducing into a BHK host cell an expression vector according to claim 4;
growing said BHK host cell in an appropriate medium; and isolating the protein product encoded by said expression vector and produced by said BHK host cell.

10. The method of claim 7, further comprising introducing into the host cell, with said expression vector, a selectable marker.

11. The method of claim 7, further comprising the step of activating the proteinproduct to produce a protein having substantially the same biological activity as human activated protein C.

12. The method of claim 8, further comprising the step of activating the proteinproduct to produce a protein having substantially the same biological activity as human activated protein C.

13. The method of claim 9, further comprising the step of activating the proteinproduct to produce a protein having substantially the same biological activity as human activated protein C.

14. The method of any of claims 11, 12 or 13, wherein the activating step comprises cleavage of the protein product with a protease selected from the group consisting of alpha-thrombin, trypsin, and venom activators.

15. The method of claim 8, further comprising introducing into the host cell, with said expression vector, a selectable marker.

16. The method of claim 9, further comprising introducing into the host cell, with said expression vector, a selectable marker.