BG60770B2 - Reduced human tissue plasminogenous activator - Google Patents

Abstract

Truncated human tissue plasminogen activator (t-pa), consisting primarily of amino acids 69-527, produced in significant quantities by recombinant DNA (rDNA) technology. The invention relates to a method for the production of t-pa without the contaminants with which it is normally associated in its natural cellular environment, for the preparation of expression vectors and various host cells. 1 claim

Description

The present invention is for human plasminogen activator, corresponding to that found in human serum and/or tissues, and for compounds and newly introduced forms as substitutes, as well as for the means and methods for its homogeneous production in therapeutically significant amounts.

The invention arises in part from the discovery of the DNA sequence and subsequently the amino acid sequence of human plasminogen activator. This discovery has enabled the reproduction of human plasminogen activator by the use of recombinant DNA technology, offering conditions for obtaining sufficient quantities of material to initiate and conduct animal and clinical trials, as a prerequisite for obtaining marketing authorization, unhindered by the necessary limitations inherent in the isolation methods used so far, including harvesting and extraction from existing cell cultures.

The publications and other materials used to illuminate the origin of the invention, as well as in individual cases to provide additional details about its operation, are attached in a bibliographical reference and, for convenience, are numbered and grouped in the appended bibliography.

A. Human tissue plasminogen activator

The fibrinolytic system of the body is in dynamic equilibrium with the coagulation system, and both maintain the normal state of the bloodstream. The coagulation system deposits fibrin as a matrix to restore the hemostatic status. The fibrinolytic system removes the fibrin mesh after the hemostatic state has been achieved. The fibrinolytic process is carried out by the proteolytic enzyme plasmin, which is derived from the plasma protein precursor plasminogen. Plasminogen is converted to plasmin by activation by an activator.

Currently, two activators are available in the pharmacy chain, streptokinase and urokinase. Both are indicated for the treatment of acute vascular diseases such as myocardial infarction, pulmonary embolism, deep vein thrombosis, peripheral arterial occlusion and other venous thromboses. These diseases pose a great danger to human health.

The etiological basis of these diseases indicates in some cases partial, and most often complete blockage of the blood vessel by a blood clot - thrombus or thromboembolus. Traditional anticoagulant therapy, such as that with heparin and coumarin, does not lead to a direct increase in the dissolution of thrombi or thromboembolus. The thrombolytic15 agents mentioned above, streptokinase and urokinase, have a wide and effective application. But each of them has strict contraindications. They do not have a high affinity for fibrin; therefore, both anticoagulants activate circulating and fibrin-bound plasminogen relatively indiscriminately. Plasmin (or fibrinolysin), formed in the circulating blood, is neutralized too quickly and loses its thrombolytic properties. Residual plasmin reduces some coagulation factors of protein origin, such as fibrinogen, Factor V and Factor VIII, which constitute the hemorrhagic potential. In addition, streptokinase is a potent antigen, and patients with high antibody titers respond poorly to treatment and may not remain on long-term therapy. Urokinase therapy is expensive, due to its isolation from human urine35 or tissue culture, and this is why it is not widely accepted in clinical practice. Urokinase has been the subject of many studies - see references 1-6.

The so-called plasminogen activators have been isolated from various human tissues, e.g. from uterine tissue, blood, serum - see references 7-11, and from cell culture ref. 94. The resulting compounds have been described (references 12, 13, 14-18). Plasminogen activators derived from these sources have been classified into two major groups: urokinase-type plasminogen activators (uPA) and tissue-type plasminogen activators (t-PA), based on differences in their immunological properties. (The abbreviations u-PA and t-PA were proposed at the XXVIII Congress of the International Committee on Thrombosis and

I hemostasis, Bergamo, Italy, July 27, 1982).

Recently, a human melanoma cell line was identified that secretes t-PA. Studies of this melanoma plasminogen activator have shown that it does not differ immunologically or in amino acid composition from plasminogen activator isolated from normal tissue cells (refs 19, 88).

The product was isolated in a relatively pure form, studied, and found to possess the properties of a highly active fibrinolytic agent (20).

Some studies (refs 95-98) using t-PA purified from a melanoma cell line have demonstrated its high affinity for fibrin compared with urokinase-type plasminogen activators. However, more intensive studies of human t-PA as a potential thrombolytic agent have been hampered by its extremely low concentration in blood, tissue extracts, vascular perfusate, and cell cultures.

It was considered that the use of recombinant DNA and related technologies was the most effective way to provide the necessary large quantities of high-quality human tissue plasminogen activator, substantially purified from other human proteins. Such materials would likely exhibit bioactivity that would allow for clinical application in the treatment of various cardiovascular conditions or diseases.

B. Recombination technology

DNA

Recombinant DNA technology has reached high levels. Molecular biologists are now able to recombine different DNA sequences with ease, creating new DNA molecules capable of producing replicated amounts of exogenous protein products in transformed microorganisms or cell cultures. Common methods and techniques involve ligation of “sticky” end fragments of DNA, producing efficient expression vectors suitable for partial transformation of microorganisms, thereby directing their synthesizing ability to a desired exogenous product. However, due to the individual characteristics of the starting products, the path remains tortuous and science has not yet made significant progress. In fact, anyone who promises successful results without a solid experimental basis risks failure.

The recombination of DNA from essential elements, i.e. the replication locus, one or more phenotypic selection characteristics, the expression promoter, the heterologous gene inserter and the residual vector, is usually carried out outside the host cell. The resulting recombinant replicative expression vector or plasmid is introduced into the cells by transformation, and large quantities of the recombinant vector are obtained from the cultured transformant. Where the gene is correctly inserted with respect to the segments that regulate the transcription and translation of the DNA-encoded information, the resulting gene expressor is capable of actually producing the polypeptide sequence for which the inserted gene codes, and the process is referred to as gene expression. The resulting product can be obtained by lysing, if necessary, the host cell, in the microbial system and subsequent appropriate purification from other proteins.

In practice, by using recombinant DNA technology, completely heterologous polypeptides can be expressed - so-called direct expression - or alternatively, a heterologous polypeptide can be expressed, synthesized as a segment of the amino acid sequence of a homologous polypeptide. In the latter cases, the expected bioactive product sometimes turns out to be bioinactivated during synthesis, the homologous/heterologous polypeptide being degraded in the extracellular space. See references (21) and (22).

In the same way, the rules of technique for the study of genetics and cell physiology have been established for both cell and tissue cultures. The means and methods for maintaining permanent cell lines, prepared by successful serial transfers from isolated normal cells, are known. For application in research, these cell lines are maintained on a solid support in a liquid medium or by

I growth in a suspension containing nutrients. Increasing the yield for large-scale production seems to pose only mechanical problems to be solved. See references (23) and (24). 5

Similarly, the biochemical properties of proteins are used in biotechnology. The cells producing the desired protein also produce hundreds of other proteins, endogenous products of cellular metabolism. These contaminating proteins, as well as other compounds, if not removed from the desired protein, can manifest as toxins if administered to animals or humans during the course of a therapeutic course with the desired protein. The technique of protein biochemistry offers the possibility of separation procedures suitable for the respective systems to obtain homogeneous products suitable for use. Protein biochemistry confirms the identity of the desired product and ensures the authenticity of the production, without alterations or mutations. This branch of science also includes the creation of drugs, preliminary studies and other necessary procedures to be performed before successful clinical trials.

The present invention is based on the discovery that recombinant DNA technology can be successfully used to produce human tissue plasminogen activator (t-PA), preferably by a direct method, and in quantities sufficient to initiate and conduct animal trials, as well as clinical trials, as required before sale. The resulting human t-PA is suitable for use, in all its forms, for the prevention or treatment of human beings in various cardiovascular conditions or diseases. Therefore, the present invention, in a very important aspect, is proposed as a method for the treatment of vascular diseases in humans by the administration of t-PA and its corresponding pharmaceutical forms.

The present invention comprises essentially pure human tissue plasminogen activator. The product is produced by genetic engineering from microorganisms or cell culture systems and provides opportunities for the production of human tissue plasminogen activator in a very efficient manner. In addition, depending on the host cell, the human tissue plasminogen activator may contain associated glycosylates to a lesser or greater extent than the natural product. In any case, T-PA does not contain the contaminants with which it is usually associated in the cellular environment.

The present invention also relates to replicative DNA expression carriers carrying gene structures encoding human tissue plasminogen activator in an expressed form, to colonies of microorganisms or cell cultures transformed therewith, or to microbial or cell cultures from such transformed colonies or cultures capable of producing human tissue plasminogen activator. In these aspects, this invention is directed to various processes for obtaining gene sequences, DNA expression carriers, colonies of microbial strains and cell cultures, and specific formations. The invention also includes the preparation of fermentation cultures from said microorganisms or cell cultures.

Figure 1 shows a 10% sodium dodecyl sulfate polyacrylamide gel electrophoresed from ³¹ 8-methionine labeled proteins precipitated with anti-t-PA immunoglobulin O secreted by melanoma cells with or without protease inhibitor.

Figure 2 shows electrophoresis of immunoprecipitated translation products from matrix ribonucleic acid (mRNA) fractions obtained from melanoma cells.

Figure 3 shows a hybridization pattern of 96 bacterial colonies transformed with DNA using a pool of ³² P-labeled 14-tag as a probe, prepared primarily from the 5 amino acid sequence of human t-PA.

Figure 4 is a restriction endonuclease map of the full-length human tPA cDNA.

Figures 5a, 5b and 5c show the nucleotide sequence and deduced amino acid sequence of the full-length human t-PA cDNA.

I

Figure 6 is a schematic construction of the expression plasmid pbecK1PA.

Figure 7 shows the results of the fibrin layer assay for fibrinolytic activity of E.coH cells transformed with pbecK1PA.

Figure 8 is a HPLC (high-performance liquid chromatography) record of human t-PA peptides from the digestive enzyme trypsin.

Figure 9 shows the construction of a plasmid encoding for direct expression of mature human t-PA in (E.coH).

Figure 10 shows the results of a fibrin layer sample tested for the fibrinolytic activity of human t-PA produced by E.coH.

Figure 11 shows the construction of a DHFR (mutant/or unaffected wild type)/tPA encoding plasmid suitable for transformation in mammalian tissue cell culture.

Figure 12 is a schematic diagram of human tissue plasminogen activator prepared by the method illustrated in E. 1.

A. Definitions

As used herein, “human tissue plasminogen activator” or “human t-PA” or “t-PA” means a non-native human (tissue-type) plasminogen activator produced by microbial or cell culture systems in bioactive forms comprising protease moieties and corresponding to those tissue plasminogen activators otherwise native to human tissue. The human tissue plasminogen activator protein obtained in this invention has been defined by a determined DNA gene and the associated amino acid sequence. It is understood that natural allelic variations exist and occur from individual to individual. These variations may be demonstrated by amino acid(s) difference(s) in the overall sequence or by deletions, substitutions, insertions, inversions or additions of amino acid(s) in the aforementioned sequence. In addition, the location and extent of glycosylation will depend on the nature of the environment surrounding the host cell.

The potential of recombinant DNA technology to produce derivatives of various human tissue plasminogen activators can be varied by single or multiple amino acid substitutions, deletions, additions or substitutions, for example, by site-directed DNA mutagenesis. This can be accomplished by preparing derivatives retaining an essential "knp£1e" region or a serine region. All such allelic variations or changes resulting in derivatives of human tissue plasminogen activator are included within the scope of this invention, as are other similar non-human (tissue-type) plasminogen activators with similar physical and biological properties. Human tissue plasminogen activator is prepared (1) by using methionine as its first amino acid, or (2) where the methionine is intracellularly or extracellularly degraded to use its normally first amino acid, or (3) together with another of its characteristic polypeptides or conjugated proteins, instead of the conventional characteristic polypeptide, the latter and the conjugate being specifically degraded in the intra- or extracellular space (see reference 21), or (4) by direct expression of the mature form without the need to degrade any extraneous and unnecessary polypeptides. The latter method is particularly important when the host cell cannot or does not efficiently remove the characteristic peptide, when the expression vector is directed to express the tissue plasminogen activator together with its characteristic peptide. In any case, the human t-PA thus produced, in its various forms, is extracted and purified to a level suitable for use in the treatment of many vascular conditions and diseases.

Furthermore, t-PA has forms that include a single-chain protein (1-chain) and a two-chain protein (2-chain). The latter is proteolytically derived from a single-chain compound. It is known from theory that the two-chain protein is associated with the resulting fibrin and that this proteolytic conversion from a single-chain to a two-chain component occurs at the locus of the plasminogen to plasmin conversion. The present invention proceeds from the proposal for the implementation of

I we from the 1-chain protein, as described, or for implementation from the 2-chain protein, as shown to be possible. The 2-chain protein can be obtained in vitro by conversion, after obtaining the 1-chain protein product. The so-called “Kιϊηβίε” region is positioned upstream of the serine region and is believed to play an important role in the binding of tissue plasminogen activator to the fibrin matrix; therefore, the observed specific activity of the presented human plasminogen activator is actually present with respect to thrombi. Tissue plasminogen activator contains enzymatically active regions corresponding to the natural product and the term human tissue plasminogen activator defines products comprising such regions alone or together with additional amino acid sequences throughout the molecule.

In summary, the human t-PA provided by this invention has a functional definition; that is, the ability to catalyze the conversion of plasminogen to plasmin, to bind fibrin, and is qualified as tPA based on the above immunological properties.

"Essentially pure form" is used to describe that state of t-PA obtained by the methods of the invention and purified from protein or substances normally associated with t-PA when produced by non-recombinant cells, i.e. in its "native" environment.

“DHFR-protein” means a protein having the properties to actively associate with Di-Hydro-Folate Reductase, and which is produced by cells capable of living in a medium deficient in Hypoxanthine, Glycine and Thymidine (-HGT medium). In general, cells deficient in DHFR protein are unable to grow in such a medium, while cells containing DHFR proteins grow.

“MTX-sensitive cells” refers to cells that are unable to grow in medium containing the DHPR inhibitor methotrexate (MTX). Thus, “MTX-sensitive cells” are cells that, unless genetically modified or otherwise supplemented, will not grow in medium suitable for growth when the MTX concentration is 0.2 μg/mL or greater. Some cells, such as bacteria, do not demonstrate MTX sensitivity because of their impermeability to MTX across the cell membrane, even though they contain the DHPR 5 protein. In general, cells containing such a DHPR protein will be sensitive to methotrexate if they are permeable to MTX.

"Wild-type" DHFR refers to dihydrofolate reductase that is normally found in individual organisms. Wild-type DHFR is normally sensitive in vitro to low concentrations of methotrexate (MTX).

"Low MTX-binding affinity DHFR protein" has a functional definition. It is a DHFR protein which, when generated in cells, enables the growth of MTX-sensitive cells in a medium containing at least 0.2 μβ/ml or more of MTX. It is known that such a definition depends on the ability of the organism to produce a "low MTX-binding affinity DHFR protein". However, as used in the context of this invention, a balance between these two mechanisms should not be of concern. The invention operates in accordance with the defined levels of MTX, and is not effective when overexpression is affected, in addition to the inherent nature of the DHFR produced. A conventional DHFR protein which is suitable for this definition is described in 1.5.Appl.5?er. No. 459,151, filed on 19 January 1983.

“Expression vector” refers to vectors that are capable of expressing DNA sequences, where the latter are operably linked to other sequences capable of selecting for their expression. It is assumed, although not precisely established, that these expression vectors can replicate in the host organism as episomes or as an integral part of the chromosomal DNA. A clear inability to replicate would render them effectively inactive. In short, “expression vector” has a functional definition and any DNA sequence that is capable of efficient expression of a precisely defined DNA coding configuration is included in this term, as applied to a specific sequence. In general, expression vectors in the application of the techniques of

I recombining DNA are often in the form of “plasmids”, which refers to double-stranded DNA structures which, in their vector forms, are not associated with the chromosome. In the present specification, “plasmid” and “vector” are used interchangeably, with plasmid being the more commonly used form of vector. However, it is intended to encompass such other forms of expression vectors which provide equivalent functions and which are now available for use.

“Recombinant host cells” refers to cells that have been transformed with construct vectors using recombinant DNA technology. As defined, t-PA is produced in amounts achieved by this transformation that are significantly less than the amounts that would be obtained from untransformed hosts. t-PA obtained from such cells may be designated as “recombinant t-PA”.

B. Host cell cultures and vectors

The vectors and methods disclosed by the invention are suitable for use in host cells from a wide range of prokaryotic and eukaryotic organisms.

In general, prokaryotes are preferred for cloning DNA sequences in the construction of vectors used in the invention. For example, E.coli K12 strain 294 (ATCC No. 31446) is particularly useful. Other microbial strains that may be used include E.coli strains such as E.coli B, E.coli X1776 (ATCC No. 31537). These examples are, of course, intended to be illustrative rather than limiting.

Prokaryotes can also be used for expression. The above-mentioned strains, as well as E. coli 3110 (P-, λ prototropene, ATCC No. 27325), bacilli such as Bacillus spp., and other enterobacteria such as Staphylococcus aureus or Streptococcus spp., and various types of phaeotropes can also be used.

Plasmids generally contain replicon and control structures, are derived from species compatible with the host cell, and are used in conjunction with those hosts. The vector usually carries the replication region, as well as marker structures that are capable of providing phenotypic selection in transformed cells. For example, E.coli is typically transformed using pBK322, and the plasmid is derived from a strain of E.coli (Borrow, et al., Sep 2: 95 (1977). pBK322 contains genes for resistance to ampicillin and tetracycline and thus provides all the means for identifying transformed cells. The plasmid pBK322, or other microbial plasmid, should also contain, or be modified to contain, promoters that can be used by the microbial organism to express its own proteins. These promoters are most commonly used in recombinant DNA sequences, including the β-lactamase (penicillinase) and lactose promoter systems (Cabney et al., No. 1978, 275: 617 (1978); Vaccaria et al., Sci., 198: 1056 (1977); (Groenewald, et al., Journal of Genetics 281:544 (1979)) and the tryptophan (trp) promoter system (Groenewald, et al., New England Journal of Medicine, 8:4057 (1980); EPO Appl., Publication No. 0036776). While these are the most commonly used, other microbial promoters have been discovered and used, and details of their nucleotide structures have been published, enabling those skilled in the art to operably link them to plasmid vectors (Schieppen, et al., Cell 20:269 (1980)).

In addition, in addition to prokaryotic and eukaryotic microbes, yeast cultures can also be used. Saccharomyces cerevisiae or common baker's yeast is the most commonly used species among eukaryotic microorganisms, although a large number of other strains are commonly available. For expression in Saccharomyces cerevisiae, the plasmid YEp7, e.g., (SpinsPsor, et al., Journal of Cell Biology, 282:39 (1979); Kerman et al., Cell Biology, 7:141 (1979); Txcsperg et al., Cell Biology, 10:157 (1980) is commonly used. This plasmid contains a tryp1 gene which provides a selectable marker for a mutant strain of yeast lacking the ability to grow in tryptophan, e.g., ATCC No. 44076 or PEP4-1 Sopec, Cell Biology, 85:12 (1977)). The presence of a tryp1 lesion as a feature of the yeast host genome provides an effective means for detecting transformations upon growth in

I absence of tryptophan.

Suitable constructs in yeast vectors include the promoters for 3-phosphoglycerate kinase (Henry, et al., 1. Biol., 255: 12073 (1980) or other glycolytic enzymes (Henry, et al., 1. Biol., 7:148 (1968); Hollmann, et al., Biol., 17: 4900 (1978) such as enolase, glyceraldehyde-3-phosphate dehydrogenase, phosphofructokinase, hexokinase, pyruvate decorcarboxylase, glucose-6-phosphate isomerase, 3-phosphoglycerate mutase, pyruvate kinase, triosephosphate isomerase, phosphoglucose isomerase, glucokinase. In suitable expression plasmids created, the specified constructs are associated with those genes that are linked in expression vector 3 from the expression sequence to achieve polyadenylation of mRNA. Other promoters that have additional advantages such as transcription controlling growth conditions are the promoter regions for alcohol dehydrogenase 2, isocytochrome C, acid phosphatase, degradative enzymes involved in nitrogen metabolism, and the aforementioned glyceraldehyde-3-phosphate dehydrogenase, and enzymes responsible for the absorption of maltose and galactose (Halc). Any plasmid vector containing a yeast-appropriate promoter, a replication locus, and a termination sequence is suitable for use.

In addition to microorganisms, cell cultures derived from multicellular organisms can also be used as hosts. In principle, any such cell culture is suitable, depending on whether it is a vertebrate or invertebrate culture. However, interest would be greatest for vertebrate cells and the propagation of such cells in culture (tissue culture) has become a routine procedure in recent years (Tiksdiye Silyere, Asabemps Pekx, Krize Pajerson, et al. (1973)). Examples of such useful host cell lines are the UEKO and Hela cell lines, Chinese hamster ovary (CHO) cell lines and the Y138, BNK, CO8-7 and MDO cell lines. Expression vectors for such cells typically include (if necessary) a replication locus, a promoter located upstream of the gene to be expressed, appropriate ribosome binding sites, RNA binding sites, a polyadenylation site, and transcriptional terminator sequences.

For use in mammalian cells, the control functions of expression vectors are often provided by viral material. For example, commonly used promoters are derived from polovirus, adenovirus 2, and most commonly spp. Uys 40 (Uys40). The early and late promoters from the 8Y40 virus are particularly useful because they are readily extracted from the virus as a fragment that also contains the 8Y40 viral replication locus (Perz, et al., Kary, 273: 113 (1978). Smaller or larger fragments of 8Y40 can also be used, the resulting fragments comprising approximately 250 bp of sequences extending from the Hind III region opposite the Bgl I located in the viral replication locus. In addition, it is also possible, and often desirable, to use a promoter or control sequence, usually associated with the desired gene sequence, the control sequences thus obtained being compatible with the host cell systems.

A locus for replication can also be provided by construction of a vector incorporating an exogenous vector, such as one derived from a sV40 or other viral (Polyvota, Adenovirus, etc.) source, or it can be obtained from a host cell with a chromosomal replication mechanism. If the vector is integrated into the chromosome of the host cell, the latter is often suitable.

In selecting a preferred host cell for transfection of the vectors of the invention which comprise a DNA sequence encoding both t-PA and DHFR proteins, it is appropriate to select the host according to the type of DHFR protein used. If a "wild-type" DHFR protein is used, it is preferred to select a host cell which is deficient in DHFR, allowing the use of the DHFR coding sequence as a marker for successful transfection in the selection medium which lacks hypoxanthine, glycine and thymidine. A suitable host cell in these cases is a DHFR-deficient Chinese hamster ovary (CHO) cell line, made and distributed as

I described by L. and C. (Proc. Man. Aad. Acn, 77:4216 1980), inserted in the bibliographic reference.

On the other hand, if a DHFR protein with low MTX binding affinity is used as a control sequence, it is not necessary to use DHFR-deficient cells. Because the DHFR mutant is resistant to methotrexate, MTX-containing media can be used as a selection medium, designed so that the host cells are themselves sensitive to methotrexate. Many eukaryotic cells that are able to absorb MTX are sensitive to methotrexate. One such cell line is that of CHO, CHO-K1 ATCC No CCb 61.

The experiments described below describe the use of E.coH, with the lac and trp promoter systems, and the use of CHO cells as host cells and expression vectors that include the 8Y40 replication locus as a promoter. However, it would be desirable to use analogous techniques to construct expression vectors for expression of the desired protein structures in alternative prokaryotic or eukaryotic host cell cultures.

Satisfactory amounts of t-PA are produced by cell culture, but later improvements using secondary coding sequences have offered increased levels of production. The secondary coding sequence includes dihydrofolate reductase (DHFR), which is treated with an external control parameter, such as methotrexate, and this allows for control of expression by controlling the concentration of methotrexate (MTX).

C. Methods used

If the host cells used are devoid of poorly permeable cell barrier membranes, transfection is carried out by the calcium phosphate precipitation method as described by Krahn and Waan de Eb, 52:456 (1973). However, other methods of DNA introduction, such as nuclear injection or protoplast fusion, can also be used.

If prokaryotic cells or cells containing a substantial cell wall are used, the preferred method of transfection is calcium treatment, using calcium chloride, as described by Coblen, E. et al. Proc. Ni. Asad. 5α. (II5A), 69:2110 (1972).

Standard ligation techniques are used to construct suitable vectors containing the desired code and control sequences. Isolated plasmids or DNA fragments are digested, annealed, and religated in the desired forms to the forms imposed by the plasmids. Digestion is performed with restriction enzyme(s) in a suitable buffer. In general, about 1 μl of plasmid or DNA fragments is applied with 1 unit of enzyme in about 20 μl of buffer solution. (Manufacturers offer suitable buffers and substrates for specific restriction enzymes). Incubation time is generally 1 hour at 37°C. After incubation, protein is removed by extraction with phenol and chloroform, and nucleic acid is recovered from the aqueous fraction by ethanol precipitation.

If “blun1” ends are required, the preparation is treated for 15 minutes at 15°C with 10 units of polymerase I (Kleenov), phenol-chloroform for extraction and ethanol for precipitation.

Size separation of digested fragments was performed on a 6% polyacrylamide gel, as described by Soedde, J. A., New Astd. Kez, 8:4057 (1980), and noted in the bibliography.

For ligation, approximately equimolar amounts of the desired components, compatible to ensure an exact match, are treated with about 10 units of T4 DNA ligase per 0.5 μl of DNA. (When cleaved vectors are used as components, it is possible to use cleaved vectors to prevent re-ligation by pre-treating them with bacterial alkaline phosphatase.)

For analysis, to confirm the correctness of the plasmid sequence, the ligation mixtures were used to transform E. coli K12 strain 294 (ATCC 31446) and transformants selected for atrHIII or laccase resistance, where appropriate. Plasmids were obtained from transformants analyzed by restriction and/or combination of the Mes51P8, e( a1, Nikleide) method.

I

Aaάx. Ke$. 9:309 (1981) or by the method of Mahat, e! ak MeίΗοόκ ΐη Επγότου^υ, 65:499 (1980).

Amplification of DHPR protein coding sequences was accomplished by growing host cell cultures in the presence of approximately 20-500,000 nM of methotrexate, a suitable inhibitor of DHPR activity. The effective concentration range is, of course, highly dependent on the nature of the DHPR gene, the protein, and the characteristics of the host cell. It is clear that no universally defined upper or lower limits can be established. Suitable concentrations of other folic acid analogs or other compounds that inhibit DHPR may also be employed. MTX, however, is convenient and effective to use.

D. General Description of the Preferred Embodiments of the Invention

Human tissue plasminogen activator was prepared according to the following protocol:

1. Human melanoma cells actively producing tissue plasminogen activator were cultured by confluence.

2. Cell pellets from such cell cultures were extracted in the presence of ribonuclease inhibitors to isolate total cytoplasmic RNA.

3. An oligo-άT column isolated all messenger RNA (mRNA) in the polyadenylated form. This mRNA was size fractionated by acid urea agarose gel electrophoresis.

4. The gel fraction containing tissue plasminogen activator-specific RNA was identified as follows: RNA from each gel fraction was translated in rabbit reticulocyte lysate in vitro with the addition of canine pancreatic microsomes. The resulting translation products were immunoprecipitated with a specific IβO antibody.

5. The appropriate RNA (21 to 248) was converted with the corresponding single-stranded complementary DNA (cDNA) from which a double-stranded cDNA was obtained. After poly -6C addition, the latter was inserted into a vector, such as a plasmid carrying one or more phenotypic markers.

6. The vectors thus obtained were used to transform bacterial cells, providing a clonal cDNA library. A pool of radiolabeled synthetic deoxy oligonucleotides complementary to codons for known amino acid sequences in t-PA, such as the 814-teg5 pool.

A A C

5' -dTS( )CA( )TA ( )TCSCA-3 “

C C T (complementary to sequences encoding a known n-111-amino acid sequence: tryptophan-glutamic acid-tyrosine-cysteine-aspartic acid (NA-Y-O)) was obtained and used to probe the colony library.

7. From positive cDNA clones, plasmid DNA was isolated and structured.

8. Structured DNA encoding m-PA was then added in vitro for insertion into a suitable expression vector, which was used to transform a suitable host cell to grow in cell culture and produce the desired human tissue plasminogen activator.

9. Thus produced, human tissue plasminogen activator has 251 amino acids in its enzyme sequence - protease region and a "kpp£1e" containing structure, which is thought to be involved in fibrin binding.

The above-described procedure is effective for the preparation of pure t-PA. The methods of the invention using an additional coding sequence sensitive to methotrexate allow the production in host cell cultures of t-PA with antigenic activity in amounts greater than 0.1 pg per cell per day. With appropriate application of amplification conditions, yields greater than 20 pg per cell per day can be achieved.

Placed in alternative conditions, the gene expression levels achieve production of more than 9 x ^10'6 PE (Pyu^b unyz) per cell per day, and with appropriate amplification, more than 18 x ^10'4 PE per cell per day of m-PA can be produced.

The advantages obtained in such an aspect are in the invention of methotrexate, as

I preparation, which is normally fatal to cellular uptake, and makes possible cell growth in a medium of controlled levels of MTX, by amplification of the gene encoding the DHFR coding structure (Zserman, Koblep T. et al., 1978; Scherer, L.B. et al., 1972; Scherer, D.E. et al., 1976)).

The importance of this aspect of the invention is that amplification of the DHFR gene can cause amplification of associated sequences that encode other proteins. This occurs in the case where the associated protein is hepatitis B surface antigen (HBxA®) (Clinic.L. e! a1, Proc.

AcA. 5c, 79: 1815 (1982); protein from E. coli HSPRT (Kidney, Co., et al., J. Mol. Appl., 1:165 (1981)); and endogenous sequences from the DHFR/ZU40 plasmid combination (Kidney, K.P. et al., J. Mol. Biol. 159:601 (1982)).

Other mechanisms involved in methotrexate resistance include a reduction in the binding affinity of the DHPR protein, such that it is less susceptible to methotrexate (Phinid. W.P. et al., 1976).

It was found that both the genes for the "wild" type DHFR and for the DHFR protein, which is resistant to MTX, due to its own reduced binding capacity, are amplified by the presence of MTX. In principle, this aspect of the invention relates to the use of the amplification effect of the DHFR structure on the associated protein coding sequences to provide a control mechanism that allows for an increase in the expression levels of t-PA sequences, in the presence of MTX or due to prior treatment of transformed cells with MTX.

F. Examples

The following examples are intended to illustrate, but not to limit, the invention. In these experiments, cultures of E. coli and CHO cell lines suitable for the type of DHFR protein coding sequence were used as host cell cultures. However, other eukaryotic and prokaryotic cells are also suitable for the methods of the invention.

E.1 Expression of the human t50 gene

PA in E.coI

E.1.A Legend to the figures

Fig. 1 is an autoradiogram of δίδ PAOE - 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis of immunoprecipitated [ ³⁵ δ] -methionine-labeled protein(s) secreted by human melanoma cells for 3 hours at room temperature, in the presence (lane b) or in the absence (lane a) of the protease inhibitor aprotinin. After immunoprecipitation with tissue plasminogen activator specific IβO, three white bands were observed (lane a), with molecular weights of approximately

65000, 63000 and 35000. In the presence of protease inhibitor, however, species with a molecular weight of 35,000 were not observed. When preimmune serum was used (lane c) no products were immunoprecipitated. To the left of lane a are shown the transitions and molecular weights of protein standards marked with ^I C.

Figure 2 depicts the gel immunophoresis of immunoprecipitated translation products from RNA fractions isolated from an acid urea agarose gel. A large band was observed in fractions 7 and 8 after translation in the presence of canine pancreatic microsomes followed by immunoprecipitation with tissue plasminogen activator specific for IβO. This band has a molecular weight of 63,000 daltons. The size of the mRNA migrating in fractions 7 and 8 is approximately 21 to 248. The positions of the ribosomal RNA markers that were determined after RNA urea gel electrophoresis and visualized by ethanedioic bromide are indicated above the corresponding gel band.

Fig. 3 presents the hybridization pattern of 96 bacterial colonies with

A A C

P-cGGS( )CA( )TA( )TSCCA(\UEU-C-O) probe CST individual transformants were grown in microtiter medium, microorganism imprinted on nitrocellulose membrane. Colonies were then lysed, bacterial DNA fixed and filters were hybridized with ³² P-14-teg (\ν-E-Y-Ό-ϋ) probes. Filters were pro11

I washed to remove unhybridized sample and exposed to X-ray film. This autoradiogram is representative of the configurations obtained with 48 individual filters (4600 individual colonies). An example of a positive tissue plasminogen activator cDNA clone on filter number 25 is indicated as E10 (arrow).

Figure 4 is a restriction endonuclease map of the full-length t-PA cDNA. The number and fragments resulting from restriction endonuclease cleavage were determined by electrophoresis through a 6% acrylamide gel. The positions of the regions were confirmed by nucleic acid sequencing (shown in Figure 5). The coding region of the largest open frame is the sequence of the putative signal peptide, while the dotted region represents the sequence of the putative mature tissue plasminogen activator (527 amino acids). The 5' end of the mRNA is on the left and the 3' end is on the right.

Figures 5a, 5b, 5c illustrate the nucleotide sequence and deduced amino acid sequence of the full-length human tissue plasminogen activator cDNA. The transcript is depicted as a continuous sequence of 35 amino acids (-35 to -1), representing its mature configuration. This 35-amino acid sequence is believed to consist of a hydrophilic “pro” sequence, a transcript of the (+1) series of the mature protein, of about 12 to 15 amino acids, in a sequence preceded by a “conventional” hydrophobic signal (extending from 5' to -35). This type of pre-prosequence of secreted proteins has been described previously, for example with preproalbumin. According to this theory, all secreted tissue plasminogen activator molecules will start with a serine (+1) as the amino-terminal. Another theory suggests that the hydrophilic sequence may be involved in the function of tissue plasminogen activator in a manner analogous to that observed for plasminogen, where a 10,000 dalton peptide can be cleaved from the amino terminal portion of native plasminogen (C1i-plasminogen), resulting in smaller molecules with a novel amino terminus, modeled after Lux-plasminogen. Lux-plasminogen is more readily activated to plasmin and also has a greater affinity for fibrin than C1i5 plasminogen. Plasmin has been shown to catalyze the conversion of C1i-plasminogen to Lux-plasminogen. This type of control mechanism results in a “positive feedback” mechanism. The first 10 plasmin molecules form degraded fibrin and also result in the generation of plasminogen and molecules that are more easily activated and bind more tightly to their substrate than native plasminogen. The result is a more rapid degradation of fibrin. The hydrophilic peptide of tissue plasminogen activator may be involved in a similar mechanism, its degradation resulting from modified binding of the enzyme to fibrin. In any case, the 35 amino acid sequence is considered to be a presequence of the mature protein.

Figure 6 is a schematic construction of the tissue plasminogen activator expression plasmid p1ekaK1PA. The starting plasmid pbekaPA25E10 was first digested with P3H to isolate a 376 bp fragment, which was then digested as shown in the figure.

Figure 7 shows the results of the fibrin layer assay for fibrinolytic activity of the expression product derived from U1a pbecK1PA in transformed cells.

Figure 8 is a HPLC (high-performance liquid chromatography) trace of human tissue plasminogen activator peptides from the digestive enzyme trypsin (Absorption at 210 nm). The arrow identifies the peak corresponding to the peptide used to construct the nucleotide probe applied from the colony library. The peptide represented by this peak was resolved with respect to its entire sequence: L-T-λY-E Y-C O-Y-P-5-C-5-T-C-C-L.

The sequences of the other major peaks were also sequenced and found to confirm the correct amino acid sequence of human tissue plasminogen activator. The single-letter peptide code for the amino acids is as follows:

I

Ahr ϋ Asparaginova k-na No I Isoleucine Tg T Threonine Yes b Leucin 5eg 5 Series Tug Υ Tyrosine C1i It is Glutamine K-na Pe It is Phenylalanine Rgo P Proline Hyuk b Histidine C1y With Glycine Buh K Lysine A1a A Alanine Ατβ K Arginine Dry With Cysteine Tgr \ν Tryptophan ν ₃ ι V Valine C1p Ω Glutamine Me! m Methionine Ahp N Asparagine

Figure 9 shows the construction of a plasmid encoding for direct expression of mature human m-PA in E. coli. 50 μg of plasmid pPA17 was digested with SαiZα1, HαiΙΙ and Hα1 and electrophoresed on a 6% polyacrylamide gel. Approximately 0.5 μg of the 55 bp SαiZα1-Hα1 fragment was amplified. Similarly, approximately 3 μg of the 263 bp Hα1-Hα1 fragment was purified from 80 μg of clone pPA25E10 as the first isolated 300 bp PχίΙ - Hα1 fragment. All digestions were performed at 37°C for one hour and the reaction products were resolved and electroeluted from a 6% polyacrylamide gel. The two indicated deoxyoligonucleotides 5' OAATTCATSTSTTATCAAST (I) and 5' CATSASTTSATAACASATO (II) were synthesized by solid phase phosphotriester (51). 100 pmol of oligonucleotide II were phosphorylated in 30 μl of a reaction mixture containing 60 mM Tph (pH 8), 10 mM MβO ₂ , 15 mM β -mercaptoethanol and 50 μl of [γ ³² P]ATP (AtermxNat 5000 SP mtoP ¹ ), 12 units of T4 polynucleotide kinase were added and the reaction was continued at 37°C for 15 minutes, 1 μl of 10 mM ATP and 12 units of T4 kinase were added and the reaction was continued for an additional 30 minutes. After extraction with phenol/CHCl _{3 ,} the phosphorylating oligomer II and the 5' hydroxyl oligomer I were combined with 0.5 μβ of the eluted 55 bp 8αiZα1-Hbα1 fragment and ethanol precipitated. These fragments were ligated at room temperature for 4 hours in 60 μΐ of 20 mM Tris-HCl (pH 7.5), 10 mM _{MgCl 2} 10 mM dithiothreitol, 0.5 mM ATP and 1000 units of T4 DNA ligase. The mixture was digested for 1 hour with 48 units of Nol, 20 units of EcoK1 and 40 units of BβIII (to eliminate polymerization by ligation of the cleaved 8αiZα1 termini) and electrophoresed on a 6% gel. A 338 bp product (approximately 0.1 μβ) was recovered by electroelution. The remaining t-PA coding structures (amino acids 111-528) were isolated on a 1645 bp fragment by digesting plasmid pPA25E10 with Ig1 and BβIII. Plasmid pbe1PA1gr103 is a derivative of plasmid pbe1PA25 (52) in which the EcoK1 region distal to the Bβ1P A gene has been removed (53). Three μβ of pbe1PA1gr103 were digested with 20 units of EcoK1 and 20 units of BβIII for 90 min at 37°C, electrophoresed on a 6% polyacrylamide gel, and the large (4200 bp) vector fragment was recovered by electroelution. For the final construct, 80 ng of the EcoK1-BβIII fragment was ligated with 100 ng of the 1645 bp NaΓΓ-BβIII fragment and 20 ng of the 338 bp EcoK1-Nz1 fragment for 10 hours at room temperature. Plasmin produces a clear lysed zone in the fibrin layer and the area of this zone can be correlated with the amount of tissue plasminogen activator in the sample.

E.1.B. Source of tissue plasminogen activator mRNA

Human melanoma cells (Bocnex Melanoma) cell line, ATCC number SKB9607) were used. Melanoma cells were grown in confluent monolayer culture in 100 ml of Echarex Melanoma Exhepta Mesal with added sodium bicarbonate (0.12 percent final concentration), 2 mM glutamine, and 10 percent heat-inactivated fetal calf serum. To confirm that melanoma cells actively produce human tissue plasminogen activator, human melanoma cells were grown to confluence in a 24-well microtiter plate. Depending on the presence or absence of 0.33 μM protease inhibitor aprotinin, the cells were washed once with phosphate buffered saline and 0.3 ml of sulfur-purified methionine. 75 μΙ of [ ³⁵ 8] methionine were added and the cells were labeled for 3 hours at 37°C. At the end of the three-hour labeling period, the medium was removed from the cells and treated with either tissue plasminogen activator-specific IβΟς or pre-immune serum for immunoprecipitation (54). The immunoprecipitation products were visualized by electrophoresis on a 10 percent 5C5-acrylamide gel. The layer was fixed, dried, and subjected to fluorography.

E.1.B Isolation of messenger RNA and size fractionation.

Total RNA from melanoma cell cultures was extracted according to the method of (55). The cells were pelleted by centrifugation then resuspended in 10 mM NaCl, 10 mM Tris-HCl pH 7.5,

1.5 mM MgCl ₂ . The cells were lysed by the addition of NP-40 (1 percent final concentration) and the nuclei were pelleted by centrifugation. The supernatant contained total RNA, which was further purified by repeated extraction with phenol and chloroform. The aqueous phase was made 0.2 M in MgCl and then total RNA was precipitated by the addition of 2 volumes of ethanol. Oligo-άT cellulose chromatography was used to purify mRNA from the total RNA obtained (54). Typical yields from 10 grams of cultured melanoma cells were 5 to 10 milligrams of total RNA and 50-200 micrograms of Polγ(A)β1 mRNA.

Fractionation of the PolyA ⁺ mRNA (200 μβ) (56) was performed by urea-agarose gel electrophoresis. The agarose gel layer (57, 58) was composed of 1.75 percent agarose, 0.025 M sodium citrate, pH 3.8, and 6 M urea. Electrophoresis was performed for 7 hours at 25 milliamperes and 4°C. The gel was then fractionated with a razor blade. The individual layers were melted at 70° and extracted twice with phenol and once with chloroform. The fractions were then ethanol precipitated and subsequently analyzed by in vitro translation in rabbit reticulocyte lysate systems, BeFexba Keserach Eab. (59, 60) mixed with canine pancreatic microsomes as follows: Translations were performed using 25 μL of [ ³⁵ 5] methionine, 500 nanograms of each RNA gel layer in a final volume of 30 μL containing 25 mM HEPE5, 43.8 mM potassium chloride, 10 mM creatinine phosphate. 19 amino acids 50 mM each, 1.1 mM magnesium chloride, 16.6 mM ethylenediaminetetraacetic acid (EDTA), 0.16 mM dithiothreitol, 8.3 mM hemin, 16.6 μβ/ηιΐ creatine kinase, 0.33 mM calcium chloride, 0.66 mM ESTA, 23.3 mM sodium chloride.

Incubation was carried out at 30°C for 90 minutes. Canine pancreatic microsomal membranes were prepared from crude microsomes using EDTA to remove ribosomes (61) and nuclease treated as previously described (62) and placed in translation mixture at a final concentration of 7 A ₂₆₀ units/ml. Translation products from the immunoprecipitated translation products were analyzed by electrophoresis on a 10 percent sodium dodecyl sulfate polyacrylamide gel as previously described (63). The decolorized gel layers were fixed, dried, and subjected to fluorography (64).

The resulting translation products from each gel fraction were immunoprecipitated with rabbit anti-human tissue plasminogen activator specific IβO. A major immunoprecipitated polypeptide band was observed in the translation of RNA fractions 7 and 8 (migrated from 21 to 245), having a molecular weight of approximately 63,000 daltons. This band was not observed when preimmune IβO was used for immunoprecipitation, suggesting that these polypeptides were tissue plasminogen activator specific.

E.1.D. Preparation of a library of colonies containing tissue plasminogen activator sequences.

Five μβ of gel fractionated mRNA (gel layer 7 mRNA) was used to prepare double-stranded cDNA by standard procedures (52, 65, 66). The cDNA was fractionated on a 6 percent polyacrylamide gel. cDNA larger than 350 base pairs in length (125 ηβ) was electroeluted. 30 ng of the cDNA was extended with deoxy (C) residues using terminal deoxynucleotidyl transferase (67) and renatured with 300 ng of plasmid pBK322 (68), which is similarly linked to deoxy (C) residues of the PxI site (67). The renatured mixture was then transformed into E. coli K12 strain 294 (ATCC No. 31446). Approximately 4600 transformants were obtained.

E.1.E. Obtaining a probe from DNA

Purified human tissue plasminogen activator was obtained using the procedures described in the accompanying literature reference (19, 20).

The molecule was scanned to locate the regions most suitable for making synthetic probes, as follows:

To make the protein suitable for digestion by trypsin, it was reduced and carboxymethylated. 2 ml of tissue plasminogen activator was first dialyzed through 0.01 percent Tris 80 overnight at room temperature. The lyophilized protein was then dissolved in 12 ml of 0.56 M TP8-HCl buffer (pH 8.6), 8 molar in urea and 5 mM EDTA. The disulfide bonds were reduced by the addition of 0.1 ml of D-mercaptoethanol. This reaction was carried out under nitrogen for 2 hours at 45°C. The reduced disulfides were alkylated to the carboxymethyl derivative by the addition of 1.0 ml of 1.4 M iodoacetic acid in 1N NaOH. After 20 minutes at room temperature the reaction was stopped by dialysis through 0.01 percent Tadaep 80, for 18 hours at room temperature and lyophilization.

The resulting lyophilized carboxymethylated protein was redissolved in 3 ml of 0.1 M sodium phosphate buffer (pH 7.5). Trypsin (TRPC) was added (ratio 1 to 50) and dissolved at 37°C. Aliquots (0.1 ml) were taken at 3 hours, 6 hours and 12 hours. A second addition of trypsin was made at 12 hours. After 24 hours the reaction was stopped by freezing the sample until it could be injected into the HPLC. The progress of the digestion was determined by 808 gel of the aliquots. All gels were empty, except for a faint band at the 3 hour fraction. This indicated that the 24 hour digestion was complete and no large peptides remained.

A sample (0.5 ml) was injected into a high-resolution ABTEX C-8 5 μ ultrasphere in two runs. An acetonitrile gradient was prepared in a stepwise manner (1 to 5 percent in 5 minutes, 5 to 35 percent in 100 minutes, 3550 percent in 30 minutes). In one of the two preparation runs, the eluent was monitored at two wavelengths (210 nm and 280 nm). The absorbance at both wavelengths was used to indicate tryptophan-containing peptides.

Peptide peaks, most likely due to their tryptophan content or their use for other reasons, were sequenced first. This allowed the sequence determination of most of the tryptophan peptides. After sequencing about 25 of the most likely peptide peaks, all sequence data that were subject to regulation were pooled to obtain a preliminary model of the primary structure of tissue plasminogen activator. From this data and this model, a number of possible probes were localized.

E.1.E. Identification of bacterial clones containing tissue plasminogen activator cDNA sequences.

Colonies were individually inoculated into the wells of microtiter plates containing LB (93) + 5 μg/ml tetracycline and stored at -20°C after addition of DMSO to 7%. Two replicates of the colony library were grown on nitrocellulose filters and DNA from each colony fixed to the filter according to the method of Orr and Hoepner (69).

The probe

A A C

P-labeled-TC( ) CA( )TA( )TCSCA was prepared (from synthetic oligomer) OU-E-U-S-P) 14-tag pool as described above. Filters containing 4600 transformants were prehybridized for 2 hours at room temperature in 50 mM sodium phosphate pH 6.8, 5 x 88C (80), 150 µg disrupted salmon sperm DNA, 5 x solution

I of Oenlabrk (85) 10 percent formamide and then hybridized with 50 x 10 ⁶ per minute of the labeled sample in the same solution. After incubation overnight at room temperature in 6 x 88C, 0.1 percent 8δ for 30 minutes, 5 times in 2 x 88C and then exposed to KODAK HK-5 X-ray film with Οιπονηΐ τίβηνηηβ Πηιχ intensifying screens for 16 hours.

Plasmid DNA was isolated by a rapid method (71) from all colonies showing a positive hybridization reaction. cDNA inserts from these clones were then sequenced after the subcloned fragments were subcloned into the M13 vector tp7 (73) and according to the chemical procedure of Mahat et al. (74). Figure 3 shows filter number 25 showing the hybridization configuration of a positive tissue plasminogen activator clone. The cDNA insert in clone 25E10 was demonstrated to encode tissue plasminogen activator DNA by comparison of its amino acid sequence with the peptide sequence (see above) obtained from purified tissue plasminogen activator and by expression product produced by E. coli, as described in more detail below. The cDNA insert from clone 25E10 (plasmid pPA25E10) was 2304 base pairs in length with the longest open reading frame encoding a protein of 508 amino acids (M>¥ 56,756) and containing 772 bp of 30 3' untranslated region. This cDNA clone lacks an N-terminal coding sequence.

Bacterial clone E. coli (pPA25A10), as well as plasmid sample pPA25E10, were deposited in Atentop Ture Cillière 35 Collegio and received numbers 67587 and 40401, respectively.

E.1.G. Direct expression of a human tissue plasminogen activator clone in E. coli.

As in Fig. 6, 50 μg of pPA25E10 were digested with P$1 I and a 376 bp fragment isolated by electrophoresis on a 6 percent polyacrylamide gel. Approximately 3 μg of this fragment was isolated from the gel by electroelution, digested with 30 units of δάε I for 1 hour at 37°C, extracted with phenol and chloroform, and precipitated with ethanol. The resulting δάε I sticky ends were extended to “δ1un(” ends by the addition of 5 units of DNA polymerase I (fragment of

(Kienoquin) and 0.1 mM each of δATP, δCTP, δOTP, δTTP to obtain a reaction mixture and subsequent incubation at 4°C for 8 hours. After extraction with phenol and chloroform, the DNA was digested with 15 units of Iar I for 2 hours and the reaction mixture was electrophoresed on a 6 percent polyacrylamide gel. Approximately 0.5 μg of the desired 125 bp “B1inG” end of Iar I fragment was isolated. This fragment encodes amino acids numbered 69 to 110 in the mature full-length tissue plasminogen activator protein.

To isolate the 1645 bp Ig I - Bgl II fragment, 30 μg of pPA25E10 was digested with 30 units of Ig I and 35 units of Bgl II for 2 hours at 37°C and the reaction mixture electrophoresed on a 6 percent polyacrylamide gel. Approximately 6 μg of the desired 1645 bp Ig I - Bgl II fragment was recovered.

Plasmid rdecaK18KS, which is a derivative of plasmid p8KSex16 (79), in which the Eco K1 site is proximal to the 1g promoter and distal to the 8KS gene, was removed by repair with DNA polymerase I (28), and the self-complementary oligodeoxynucleotide AATTATCAATTCAT (synthesized by the phosphotriester method) was inserted into the remaining Eco K1 site immediately adjacent to the Xba I site. 20 μ% of p0decaK18KS was digested to completion with Eco K1, extracted with phenol and chloroform, and precipitated with ethanol. The plasmid was then digested with 100 units of nuclease 81 at 16°C for 30 minutes in 25 mM sodium acetate (pH 4.6) 1 mM Znα ₂ and 0.3 M LaCl to create a “b1in1” end with an ATC sequence. 40 Following extraction with phenol and chloroform and ethanol precipitation, the DNA was digested with Bam HI, electrophoresed on a 6 percent polyacrylamide gel, and the large (4300 bp) vector fragment was recovered by electroelution.

The expression plasmid was assembled by ligation together with 0.2 μg of vector, 0.06 μg of 125 bp “B111” end-B1 fragment and 0.6 μg of 1645 bp B1 - B1 II fragment with 10 units of T4 DNA ligase for 7 hours at room temperature and used to transform E. coli strain 294.

I (ATCC No. 31446) for ampicillin resistance. Plasmid DNA was obtained from 26 of the colonies and digested with Xba I and Eco K1. 12 of these plasmids contained the desired 415 bp Xba I-Eco K1 and 472 bp Eco K1 fragments. DNA sequence analysis verified that several of these plasmids had the ATC (Antithymocyte Globulin) stimulation code correctly placed at the start of amino acid number 69 (serine). One of these plasmids, p(lekaK1PA), was tested and produced the desired tissue plasminogen activator (Figure 7). The bacterial clone E.coP (lekaK1PA) and also a sample of the plasmid plekaK1PA have been deposited in the Atencana Toure CiPure Collection and acquired numbers 67585 and 40400, respectively.

F. 1.3. Full-length tissue plasminogen activator cDNA

a) Preparation of a colony library containing N-terminal tissue plasminogen activator sequences: 0.4μς of the synthetic oligonucleotide

5' TTCTSCACCACCACCCCC 3' was used to prime a 7.5μς gel fraction of No 8 mRNA (above) to prepare double-stranded cDNA using standard procedures (65, 66). The cDNA was size fractionated on a 6% polyacrylamide gel. A fraction larger than 300 base pairs (36 bp) was electroeluted. The 5 bp cDNA was elongated with deoxy (C) radicals using terminal deoxycytidyl transferase (67) and renatured with 50 bp from plasmid pBK322 (68) , which had been similarly ligated with deoxy (O) radicals at the PxI site (67). The renatured mixture was then transformed into E. coli K12 strain 294. Approximately 1500 transformants were obtained.

b) 3H-brem hybridization of human genomic DNA: Since the cDNA priming reaction was performed using a synthetic fragment that hybridized 13 base pairs from the N-terminal of clone pPA25E10, an unsuitable restriction fragment became available in this 29 base pair region (which includes a 16-tag sequence) for screening cDNA clones. Therefore, it was necessary to isolate a human tissue plasminogen activator genomic clone to identify all primary amplified cDNA clones containing tissue plasminogen activator coding sequences with an N-terminal.

The first step in this process involved establishing that a single homologous tissue plasminogen activator gene was present in human genomic DNA. To determine this, 801Hep hybridization was performed. In this procedure (77), 5 μg of high molecular weight human lymphocyte DNA (obtained as in 80) was digested to completion with various restriction endonucleases, electrophoresed on a 1.0 percent agarose gel (81), and dried on a nitrocellulose filter (77). A ³² P-labeled DNA probe was prepared (76) from the 5' end of the cDNA insert of pPA25E10 (230 bp Hpa II - Ksa 1 fragment) and hybridized (82) to the nitrocellulose filter. 35 x 10 ⁶ per minute of the sample were hybridized for 40 hours and then washed as described (82). Two endonuclease digestion structures provided only one hybridizing DNA fragment: B ₈ 1 H (5.7 Kbp) and Pνυ II (4.2 Kbp). Two hybridizing DNA fragments were observed with Hmc II (5.1 Kbp and 4.3 Kbp). Taken together, these data suggest the presence of a single tissue plasminogen activator gene in the human genome, and that this gene contains at least one intervening sequence.

c) Screening of a human lambda phage library for tissue plasminogen activator genes.

The strategy used to identify lambda phage recombinants carrying tissue plasminogen activator genes consisted of detecting a homologous nucleotide with a radioactive probe prepared from the tissue plasminogen activator cDNA from pPA25E10. One million recombinant lambda phage were plated on Petri dishes with 50 µg/15 cm2 of DPF at a density of 10,000 µg/15 cm2, and a nitrocellulose filter was prepared for each plate by the method of Bremsdorf and OaU15 (78). According to a standard procedure (83). ^{A 32} P-labeled DNA probe was prepared from a 230 base pair Hpa II - Ksa I fragment localized 34 base pairs from the 5' end of plasmid pPA25E10. Each nitrocellulose filter was prehybridized at 42°C for 2 hours in 50 mM sodium phosphate (pH 6.5), 5 x 85C (77), 0.05 μg/ml sonicated salmon sperm DNA, 5 x Pentall solution (84),

I percent formamide and then hybridized with 50 x 10 ⁶ for one minute of the labeled probe in the same solution containing 10 percent sodium dextran sulfate (85). After overnight incubation at 42°C, the filters were washed 4 times at 50°C in 0.2 x 88C, 0.1 percent 5ϋδ for 30 minutes, once at 2 x 58C at room temperature, and then exposed to CoOac HK-5, X-ray film with a Cyrop! Cgopeh intensifying screen overnight. 10 A total of 19 clones were obtained that were hybridized with the probe. Phage DNA was prepared as previously described (86) from 6 recombinants. Lambda clone C was selected for the production of fragment Pνυ 11 for 15 colony screening. 30 µl of DNA was digested with Pvll for 1 hour at 37°C, electrophoresed on a 1.0 percent agarose gel. A 4.2 kbp fragment with previously demonstrated tissue plasminogen activator sequence content was electroeluted and purified. The ³² P-labeled probe was prepared by standard colony hybridization procedures (83) as described below.

c) Screening of a colony library 25 for 5' tissue plasminogen activator sequences.

Colonies were transferred from the petri dishes and plated onto nitrocellulose filters, and DNA from each colony fixed to the filter, 30 according to the procedure of O'Neill-Hoppen (69). ³² P-labeled probe was prepared by priming calf thymus (83) 4.2 kb Pvll II fragment from an isolated tissue plasminogen activator lambda genomic clone. Filters containing 1500 transformants were hybridized with 112 x 10* cpm of ³² P - genomic Pvll II fragment. Hybridization was for 16 hours, using the conditions described by Pvll et al. (82). The filters were extensively washed and then exposed to Ko<1ak HK-5 X-ray film with διυροηΐ ΕίβΗίνηηβ-P1x amplifying screens for 16-48 hours. 18 colonies clearly hybridized to the genomic probe. Plasmid DNA was isolated from each of these colonies and 45 were ligated to nitrocellulose filters and hybridized to ^{the 32} P-labeled synthetic oligonucleotide (16-reg) used for the original priming reaction. Of the eighteen clones, seven hybridized to the kinase 16-reg. From sequence analysis after subcloning the fragments into the sh13 vector tp ⁷ (73), one clone (pPA17) was shown to contain the correct 5' N terminal region of tissue plasminogen activator, signal leader sequence and 84 bp 5' untranslated region. From the two clones pPA25E10 and pPA17 the complete nucleotide sequence (Fig. 5) and restriction endonuclease configuration (Fig. 4) of the full-length tissue plasminogen activator clone were determined.

The bacterial clone E.coH (pPA17) and also a probe from plasmid pPA17 have been deposited in the Archaeology Library under accession numbers 67586 and 40402, respectively.

The native tissue plasminogen activator molecule has the potential to be stabilized by 17 disulfide bridges based on homology with other protease series. There are four potential N-glycosylation sites, three localized “kpp£1e” regions, the α5P ₁₁₇ , α8P ₁₈₄ , α8P ₂₁₈ regions, and one potential light chain region, α5P _448. Variations in the structure of the oligosaccharide ligands may be responsible for the different molecular forms (65,000 and 63,000 mol.m.).

E.1.I. Direct expression in E.coH of a full-length tissue plasminogen activator cDNA clone.

The reconstruction of the complete coding sequence was possible thanks to the use of a common Hba1 restriction endonuclease site, separated from the two partial clones pPA17 and pPA25E10. A 55 bp 8ai3A1-Hba1 restriction fragment, corresponding to amino acids 5-23, was isolated from plasmid pPA17. The 8ai3A1 restriction site was located at codon four of the putative mature coding sequence and was used to remove the signal peptide coding region. A 263 bp Hba1NηγI fragment (coding for amino acids 24-110) was also isolated from plasmid pP25E10. Two synthetic deoxyoligonucleotides were designed with the restored codes for amino acids 1-4, incorporated into the ATC (antithymocyte glob.) translational promotion codon and created at the EcoK1 binding site. These three fragments were then ligated together to form a 338 bp fragment encoding amino acids 1-110. This fragment and the 1645 bp NβγΙ-Β^ΙΙΙ fragment from pPA25E10 were then ligated

I between the EcoK1 and BβIII regions of plasmid p(-PA1gr103 (53) to yield the expression plasmid p(-PA1gr12). The cloned t-PA gene was transcribed under the control of a 300 bp fragment of the E.coI 1gr operon, which contains the 1gr promoter, operators 8H1ne - Oal£arpo sequence of the 1gr leader peptide, but lacks the leader peptide ATC promoter codon (52).

A sample of plasmid p1-PA1gr12 was deposited in the Atepsap Toure Silighe Society and acquired number 40404.

E.1.J. Sequencing analysis

The sequencing analysis was based on the degradation of Ebana (83B). The sample was placed in a beaker of a Vescman 890B or 890C sequencer. Polyurene TM (poly N,N,N',N ¹ -tetramethyl-1H-trimethylenehexamethylene diammonium diacetate) was used as the carrier in the beaker (63C). The sequencer was modified with a cold trap and several program changes to reduce background peaks. The reagents were Vescman sequence 0.1 M 9-arbo buffer, phenylisothiocyanate, and heptafluorobutanoic acid.

Selected Edman cycles were manually converted to 2-anilino-5-thiazolinone derivatives. The 1-chlorobutane was dried under nitrogen. Then 1.0 N hydrochloric acid in water was added to the 2-anilino-5-thiazolinone and heated to 70°C for 10 minutes to convert to 3-phenyl-2-thiohydanthione (PTX derivative). The PTX amino acid residue was then dissolved in 50% acetonitrile and water and injected into a high-pressure, reversed-phase liquid chromatograph. Each PTX amino acid was then identified by comparison with the retention times of a standard mixture of PTX amino acids, which were introduced into the conversion vial and treated in the same manner as the sequencer cycle.

E.1.K. Assays for detection of tissue plasminogen activator expression

1. direct analysis of plasmin formation

a) Theory

A sensitive assay for tissue plasminogen activator can be obtained by monitoring the tissue plasminogen activator catalyzing the conversion of plasminogen to plasmin. Plasmin is an enzyme for which chromatographic assays exist. These assays are based on the proteolytic cleavage of a tripeptide from a chromophore group. The extent of cleavage is directly dependent on both the specificity and the concentration of the protease being tested. The basis of the assay is the determination of the amount of plasmin formed by incubation of a tissue plasminogen activator containing solution with a solution of plasminogen. The greatest amount of activator forms the greatest amount of plasmin. Plasmin is measured by monitoring its degradation by the chromatogenic substrate 82251 (obtained from Karl Orr, Inc., Chicago, IL).

b) Procedure

A multiple of the sample volume was mixed with 0.10 ml of 0.7 ng/ml plasminogen (in 0.05 M Tris HCl, pH 7.4, containing 0.012 M NaCl) and the volume was made up to 0.15 ml. The mixture was incubated at 37°C for 10 minutes, 0.35 ml of 82251 (1.0 mM solution in the above buffer) was added, and the reaction was continued for 30 minutes at 37°C. Glacial acetic acid (25 μL) was added to terminate the reaction. The samples were centrifuged and the absorbance was measured at a wavelength of 405 nm. Quantification of activity was achieved by comparison with a standard urokinase solution. The activity was recorded in P100£B units, with 90,000 P100£B units being equivalent to the activity exhibited by 1 mg of purified tissue plasminogen activator.

2. Indirect analysis of plasmin formation

a) Theory

A sensitive assay for tissue plasminogen activator activity has been proposed (87). The assay is based on the determination of plasmin formation by measuring the extent of plasmin degradation of fibrin in an agar medium containing fibrin and plasminogen. Plasmin produces a prominent zone of lysis in the fibrin plaque. The area of this zone of lysis can be correlated with the amount of tissue plasminogen activator in the sample.

b) Procedure

Following the procedure of Sgapensh Iregpo and Kesb (87), the petri dishes were incubated19

I wound 3.5 hours at 37°C and the lysis zone was measured. Quantification was achieved by comparison with a standard urokinase solution.

E.1.L. Determination of tissue plasminogen activator activity

1. Bacterial growth and sample preparation.

A colony of E. coli containing the plasmid (p<1e11aK1PA”) was inoculated into a tube containing 5 ml of EB medium with 20 μg/ml ampicillin. The cells were left overnight at 37°C. An aliquot of this culture was diluted 1:100 into 300 ml of M9 medium containing 20 μg/ml ampicillin. The cells were left in the flask at 37°C for 4 hours, and an absorbance at 550 nm of 0.419 was obtained. The tryptophan analogue indolacrylic acid was added to a concentration of 30 μg/ml. The cells were incubated for 90 minutes with a running absorbance at 550 nm of 0.628. The cells were separated by centrifugation and resuspended in 0.8 1 ml of 0.01 M Tnx, pH 8.0, containing 0.01 M EDTA. The resulting suspension was stirred at room temperature for 18 hours. The sample was centrifuged and the supernatant was assayed for tissue plasminogen activator activity.

2. Establishing activity

Table 1 shows the results of the activation of plasminogen by E.coH extract. The activity generated depends on the presence of plasminogen. This activity was not affected by preimmune rabbit serum, but was inhibited by antiserum raised against tissue plasminogen activator derivatives from melanoma cells. This demonstrates that E.coH extracts produce plasminogen, affecting the activity that was inhibited by antibodies against tissue plasminogen activator.

Fig. 7 shows the result of the fibrin Petri dish assay for fibrinolytic activity. A standard amount of urokinase was added to the center row at concentrations from left to right, of 0.24, 0.14, 0.10, 0.05, 0.02 PLU units. The bottom row is of samples of natural tissue plasminogen activator, with the same amount of enzyme in each well. The wells 15 contain, from left to right, tissue plasminogen activator, antiplasminogen activator plus tissue plasminogen activator antibodies. The wells of the top row contain 8 μl of recombinant tissue plasminogen activator extracts from E.co. H. In the first well is extract only, in the second there is preimmune serum added, and in the third there are tissue plasminogen activator antibodies added. It is apparent that the pre-immune serum does not affect either natural or recombinant tissue plasminogen activator, and that tissue plasminogen activator antibodies inhibit the activity of natural, such as E. coli extracts. Based on urokinase standards, the extracts contained slightly less than 2.5 PKU units per milliliter. This compares favorably with the value obtained in Table 1 of 1.3 PKU units per milliliter.

Table 1 sets out the results of the analyses carried out in the manner described above in E.1.K.1.6.:

Table 1

E.coN from cultures containing rbeKaShRA

Plasminogen activation by extracts of

Sample A ⁴⁰⁵ Activity Calculated Philippine <%> units in t! Extract (without plasminogen) 0.043 (0) Extract 0.451 (100) 1.3 Extract Plus preimmune serum 0.477 106 - Extract Plus anti t-PA antibodies 0.079 9

I

Figure 10 presents the results of a fibrin layer assay performed with extracts from 10 L of E. coli fermentation cultures containing a tissue plasminogen activator expression plasmid. The fibrinolytic activity of the tissue plasminogen activator containing extract is shown in Fig. 10 (panel A). This fibrinolytic activity was inhibited by anti-mPA IβO (panel C), but not by preimmune IβO (panel B) or anti-urokinase IβC (panel D), and no evidence of activity was seen from the extract prepared from cells containing the leukocyte interferon plasmid pbe1PA103 as a control (panel H).

E.2 Preparation of t-PA using DHFR protein with low MTX binding activity.

E.2.A Vector construction

The sequence encoding human tissue plasminogen activator (t-PA) was inserted into an expression plasmid containing a mutant DHFR with low MTX binding activity, described in application, υ.δ. 55er. No. 459151 filed on January 19, 1983, corresponding to E. coli Plasminogen Activator.

No. 117060 (noted in the bibliographic reference), carried out with the following procedure (see Fig. 11):

Three fragments of the partially overlapping plasmids, pPA25E10, pPA17, and PPA1rP12 (above) were prepared as follows: Plasmid pPA17 was digested with Diβe I, using Kevron DNA polymerase 1, and cut with Psa I; an approximately 200 bp fragment containing the 5'-terminal t-PA sequence was isolated. The second t-PA fragment was obtained by digesting pPA25E10 with Psa I and Bsa II and isolating an approximately 310 bp fragment. The third t-PA fragment was obtained by digesting pPA25E10 with 40 Bsa I and Bsa II and isolating an approximately 1645 bp fragment that contained in addition to the t-PA coding region, several 3' untranslated sequences.

Plasmid pE342, which expresses HBV surface antigen (also referred to as pHB348-E), was described by Leuchson et al., patent application Ser. No. 326,980, filed December 3, 1981, now abandoned, in continuation of application Ser. No. 603,539, filed April 24, 1988, corresponding to Ergoran Palleni Appl. No. 73,656, which is mentioned in the bibliographic reference. (Briefly, the 5'1 Przez, N.Y. 1980), (which contains the gene encoding HBSAβ) into the plasmid pMB (Bizuk et al., 293:79 (1981) at the EcoK1 and BamHk regions (pMB is a derivative of pBK322 which has a deletion eliminating sequences which are inhibitors of plasmid replication in monkey cells). The resulting plasmid (pK1-Bβ1) was then linearized with EcoK1, and the 348 bp fragment, representing the locus region of 8Y40, was introduced into the EcoK.1 region of 8Y40. The locus fragment can insert in another orientation. This fragment encodes both the early and late 8Y40 promoters with an addition to the 30 locus of replication. HBV genes can be expressed under the control of one of the promoters, depending on this orientation (pHB5348-E representing HB3 expressed under the control of the early promoter). pE342 was modified by partial digestion with EcoK1, filling in the cleavage site using Kleptovyn DNA polymerase I, and ligating the plasmid backbone together, thus removing the EcoK1 site encoding the 8Y40 locus in pE42. The resulting plasmid, designing pE3420eIaK1, was digested with EcoK1, filled in using Kleptovyn DNA polymerase I, and cut with Baz HI. After acrylamide gel electrophoresis, the approximately 3500 bp fragment was electroeluted, extracted with phenol-chloroform, and precipitated with ethanol as above.

The thus obtained p342E 3500 bp vector, as well as the above-described t-PA fragments comprising approximately 2160 bp, were ligated together using standard techniques. A plasmid containing the three t-PA coding fragments in their own orientation was isolated, characterized and modeled - pE342-t-PA. This plasmid was digested with 8ac II and treated with alkaline phosphatase (BKb). To obtain the sequence of DHFR, an approximately 1700 bp fragment was generated by 8ac II cleavage from pENEK. (pENEK is a plasmid expressing a mutant DHFR described in U.S. Patent No. 459151 (supra). This fragment was ligated into the pE342-t-PA plasmid to create pETRAEK400, and a plasmid that is analogous to pENEK except for the HBxAβ coding region was transferred via cDNA sequences from t-PA.

E.2.B Expression and amplification of the mPA sequence pETRAEK400 (pETRAEK) was transfected into two (Rr-CHO-OiX B11 cells and ONEK+CHO-K1 (ATCC CCB61) cells by the method of Cram and Waan <ler Eb. Transformed ΔHΘ- cells were selected by growth in medium deficient in glycine, hypoxanthine and thymidine. Transformed ΔHΘR+ cells were selected by growth in 100 ηM MTH. Colonies that grew in the appropriately selected medium were isolated using cloning rings and propagated in the same medium for several generations.

For amplification, cells from colonies were split in medium containing 5 x ¹⁰⁴ , ¹⁰⁵ ,

2.5 x 10 ⁵ , 5 x 10 ⁵ and 10 ⁶ ηΜ MTX and passaged several times. The cells were plated at very low (10 ² - 10 ³ cells/medium) density in 10 cm dishes and the resulting colonies were isolated.

E.2.B Analytical methods

The expression of t-PA in the transfected amplified colonies can be conventionally analyzed by methods similar to those set forth in E.1.K.1.6.

Co-amplification of DHFR and t-PA sequences was analyzed by isolating DNA from confluent monolayer colonies as follows: confluent monolayer colonies in 150 ml Petri dishes were washed with 50 ml sterile PBS (phosphate-buffered saline) and lysed by the addition of 5 ml of 0.1 percent δίδ (sulfur-buffered saline), 0.4 M CaCl ₂ , 0.1 M EDTA, pH 8. After

After 5-10 minutes, the mixture was separated, extracted with phenol and chloroform, and precipitated with ethanol. The DNA was resuspended in 1 ml! (for 150 ml petri dishes) 10 mM Tris-HCl pH 8, 1 mM EDTA (TE), KCl was added to 0.1 μg/ml, and the solution was incubated for 30 minutes at 37°C. δίδ was then added to 0.1% and pronase (at 8µg) was also added to 0.5 μg/ml. After 3-16 hours of incubation at 37°C, the solution was again extracted with phenol and chloroform, sequentially, and precipitated with ethanol. The DNA pellet was resuspended in 0.5 ml of water and digested with restriction enzymes. Approximately 5-10 μ% of the digested DNA was electrophoresed in an agarose gel [1% agarose in Tnx-acetate buffer (40 mM Tnx, 1 mM EDTA, adjusted to pH 8.2 with acetic acid)]; Croix, et al., J. Biol. Cell. 257: 7887 (1982)). After the bromophenol blue dye had migrated down the gel, the gel was stripped and stained with ethanoic bromide. After visualization of the DNA with ultraviolet light, the DNA was transferred from the gel to nitrocellulose filters as in the procedure of Sollbern (J. Biol. 98:503. (1975)). The filters were then hybridized with a labeled translation probe made from a 1700 bp 5ac11 fragment of pENEK (obtained and hybridized as described above), or from an approximately 1970 bp Bgl II fragment of pETREK.

E.3 Production of t-PA in combination with wild-type DHFR protein

E.3A. Vector Construction In a manner analogous to the method used in the construction of pETREK, a plasmid pETREK containing a DNA sequence encoding wild-type DHFR was constructed. The construction is similar to that described in Example E.2.A., except that plasmid pENEK as the source of the gene sequence for the DHFR protein was replaced by plasmid pE342.HBV.E400.D22, described in U.S. Patent No. 459152, filed January 19, 1983, now Patent No. 4713339, issued December 15, 1987, corresponding to U.S. Patent No. 117058, mentioned in the bibliographic reference. The plasmid ρΕ342.ΗΒν.Ε400.΋22 is the same as ρΕΝΕΚ, except for one single different base pair, a mismatch between the wild type and the DHFR mutant. The resulting plasmid pΕΤΡΕΚ is analogous to pΕΤΡΕΚ, with 22

It was concluded that the DNA sequence encoding the wild-type DHFR was substituted for that of the mutant. A sample of the plasmid pETRPK was deposited in the Atepsap Toure Synige SoNesOop and received accession number 40403.

E.3B Expression of the m-PA sequence pETPPK was used to transfect DHFR deficient CHO cells (mylaib and Cbaxm) using calcium phosphate precipitation according to the method of Cbaxm and Wana de Eb. Twenty-one colonies that grew on selective medium (HCT) were assayed by detecting plasmin formations resulting from the degradation of fibrin in agar plates containing fibrin and plasminogen, as described by Cbaxm-Plperno, et al. Tech. Med., 148: 223 (1978).

Four of the most strongly positive clones were then quantitatively analyzed for plasmin formation, according to the method outlined in E.1.K.1.6.

From such quantitative determination it was found that the four tested clones showed the same or comparable secretory capacity of t-PA into the medium, expressed as units/cell/day. Subclones were obtained by transfer inoculation of two of the clones into separate petri dishes containing -HCT medium. Two of the obtained subclones, 18B and 1, were used for further analysis.

E.Z.V. Amplification and secretion levels of t-PA.

The above mentioned clones were plated at 2 x ¹⁰⁰ cells per 100 ml petri dishes in 50 nM MTX to aid amplification. The surviving cells, when assayed as above, yielded in all cases about 10 times the unamplified amounts of tissue plasminogen activator activity. Two of these clones were selected for further study and were designated 1-15 and 18B-9.

Next, subclone 1-15 was amplified by plating 2 x 10 ⁵ cells in 100 ml petri dishes containing 500 nM MTX. Analysis of the thus amplified cells showed a further increase in the yield (of about 3 fold) of t-PA when quantified according to the method in C.1.C, levels were in the range of 7 x 10 -* units/cell/day. A portion of these amplified cells was then transferred and maintained in the presence of 10,000 nM MTX. Subclones 1-15 and 18B-9 were then tested after being maintained for approximately 1-2 months under the conditions shown in Table 3.

Table 3

Cell line Growth conditions ηβ t-PA/cell/day* 1-15 _5(M 500 nM MTX 28.5 x 10 ³ 1-15 ₅₀₀ 500 nM MTX 26.0 x 10 ³ 15 _5й0 (-HST medium, without MTX) 8.3 x 10 ³ 1-15 ₅₀₀ (-HST medium, without MTX) 18.0 x 10 ³ 1 ”1^10000 YulM MTX 29.3 x 10 ³ 1 -15! ooooh Yu^M MTX 49.0 x 10 ³ 18B-9 50 nM MTX 14.3 x 10 ³ 18B-9 50 nM MTX 14.4 x 10 ³ 18B-9 (-HST medium, without MTX) 14.3 x 10 ³ 18B-9 (-HOT medium, without MTX) 14.4 x 10 ³ 1 (-HST medium, without MTX) 1.0 x 10 ³ 1 (-HST medium, without MTX) 0.7 x 10 ³

I* t-PA in the cell medium was assayed quantitatively by radioimmunoassay as follows: Purified t-PA and purified iodinated t-PA marker derived from melanoma cells were serially diluted to include concentrations from 12.5 to 400 ng/ml in a buffer solution containing phosphate buffered saline, pH 7.3, 0.5 bovine serum albumin, 0.01 percent Tc/eem 80, and 0.02 percent NηN ₃ . Appropriate dilutions of the samples were added to the radioactive labeled proteins to be assayed. The antigens were incubated overnight at room temperature in the presence of a 1:10,000 dilution of the 1&C fraction of rabbit anti-t-PA antiserum. The antigen/antibody complex was precipitated by absorption with goat anti-rabbit IβO immunobeads (VuKab) for two hours at room temperature. The beads were washed with saline diluent, centrifuged for 10 minutes at 2000 x β at 4°C. The supernatants were discarded and the radioactivity of the precipitates was measured. Concentrations were determined by comparison with a standard.

The cell lines are as follows: cell line “1” is an unamplified clone from the original group of four clones. “115 _50θ ” is an amplified subclone from cell line “1” that was initially amplified in 50 nM MTX to give 1-15 and then transferred for further amplification in 500 nM MTX. 1-15 _1θθθθ is a subclone of 1 - 15 _50θ that was subsequently amplified in the presence of 10,000 nM MTX. Cell line 18B-9 is a subclone of the original group of four clones that was amplified with 50 nM MTX.

The CHO cell line 1-15 _50θ sample was deposited in the Atepsap Tuge Sikige SoIeskop and received accession number SKB 9606.

All of the amplified cells showed increased levels of t-PA secretion, well above those shown by the unamplified cells. If an unamplified culture secreted t-PA more than 0.5 ρβ/cell/day, amplification gave 50 ρβ/cell/day.

G. Pharmacological structures

The compounds of the present invention can be formulated according to known methods for preparing pharmaceutically usable structures, wherein the human tissue plasminogen activator product is combined in a mixture with a pharmaceutically acceptable carrier.

Suitable carriers and the technology for their preparation, including other human proteins, including human serum albumin, are described, for example, in Kernpp£1on'5 PHARMACASEiTlCAb δΟΙΕΝΟΕδ by E.W.Mahrin, mentioned in the incorporated bibliographic reference. Such structures will contain an effective amount of protein, together with a suitable amount of the carrier, to produce pharmaceutically acceptable compositions suitable for effective administration.

For example, human tissue plasminogen activator could be administered parenterally to individuals suffering from cardiovascular disease. The dosing and dosing regimen could be paralleled to those currently in use in clinical trials of other cardiovascular thrombolytic agents, i.e., about 440

IU/kg body weight, with an intravenous loading dose followed by a continuous intravenous infusion of approximately 440 IU/kg/hour for 12 hours, in patients suffering from pulmonary embolism.

As an example of appropriate dosage determination of essential homogeneous human tissue plasminogen activator for parenteral administration, a vial containing 25,000 IU of tissue plasminogen activator activity, 25 mg of tabacco, and 45 mg of NoCl can be reconstituted with 5 mL of sterile Dextrose Injection and mixed with an appropriate volume of 0.9 percent sodium chloride or 5 percent benzalkonium chloride for intravenous administration.

3. Detailed description of recombinant human t-PA

The structure of the partial embodiment of human t-PA obtained in the experiments described so far has been investigated in some detail, both by elucidation of the gene coding sequence and by known biochemical techniques for protein analysis. The current understanding of protein structure is illustrated in Fig. 12.

Also, SoPen et al. (88) demonstrated that the double-chain human t-PA is formed by the proteolytic cleavage of a single-chain molecule into two disulfide-linked polypeptides. The current state of knowledge allows us to conclude that the heavy chain (30882 mol. wt.) is derived from the NH ₂ terminal portion and that the light chain (28126 mol. wt.) includes a COOH-terminal region. The N-terminal structure of the double-chain molecule suggests that the double-chain form is generated by the cleavage of a single arginyl-isoleucine bond (Fig. 12; drawn arrow).

The primary structure of the heavy chain portion of human t-PA (Fig. 12) reveals a high degree of sequence homology with the so-called “kppk1e” regions of plasminogen (89) and prothrombin (40 and 41). The “Kppk1e” region refers to a characteristic triadic disulfide structure originally found in the “pro”-fragment of prothrombin, first described in detail by Ma£pi$$on et al. (91, 92). From the primary sequence of t-PA, two so-called “kppk1e” regions, of 82 amino acids each, are evident, showing a high degree of homology with the 5 “kppk1e” regions of plasminogen. The remaining N-terminal 91 amino acids share a low degree of homology with the “conventional” “kppk1e” region. However, one may speculate that this region may also adopt a structure containing multiple disulfide bonds, as the 11 additional cysteine residues found here.

The catalytic region of the light chain of human t-PA, the so-called serine protease region, as in other serine enzymes, is most likely formed by histadine ₃₂₂ , asparagine _37] , and serine ₄₇₈ residues. Moreover, the amino acid sequences surrounding these residues are highly homologous to the corresponding parts of other serine proteins, such as trypsin, prothrombin, and plasminogen.

Although reference has been made to selected embodiments, it should be understood that the present invention is not limited thereto, but rather constitutes a reasonable list of the appended claims.

Claims (1)

Claims A recombinant human tissue plasminogen activator consisting essentially of the amino acid sequence 69-527 depicted in Figures 5a, 5b and 5c.