DD279687A5 - PROCESS FOR THE PRODUCTION OF ERYTHROPOIETIN - Google Patents

Abstract

The invention relates to processes for the production of erythropoietin which are used as medicaments. The invention provides a method for producing erythropoietin. The object of the invention is to provide a method for the production of erythropoietin available. According to the invention, the procedure involves culturing host cells in a suitable medium which contain a DNA sequence as shown essentially in FIG. 3 B, which is effectively bound to an expression control sequence and thus produced by the cells and the medium Erythropoietin is separated.

Description

For this 45 pages drawings

Field of application of the invention

The invention relates to a process for the production of erythropoietin. This is used as a medicine, for example for the treatment of anemia.

In particular, the present invention relates to cloned genes of human erythropoietin that provide high levels of expression, expression of said genes, and in vitro production of human erythropoietin.

Characteristic of the known state of the art

Erythropoietin (abbreviated hereafter as EPO) is a circulating glycoprotein that stimulates erythrocyte formation in higher organisms (see Carnot et al., Comgst., Rend. 143, 384 [15061]. Therefore, EPO is sometimes referred to as the electropoietic stimulating factor.

The lifetime of human erythrocytes is about 120 days. This means that approximately V120 of the total erythrocytes are destroyed daily in the reticulocutothelial system, while at the same time a relatively constant number of erythrocytes are produced daily to keep the concentration of erythrocytes constant (Guyton, Textbcok of Medical Physiclocy, p.56- 60, WBSaundersCo., Philadelphia 1976).

Erythrocytes are formed in the maturation and differentiation of the erythroblasts in the bone marrow. EPO is a factor that affects less differentiated cells and induces their differentiation into erythrocytes (Guyton, supra). EPO is a promising therapeutic agent for the clinical treatment of anemia or, in particular, renal anemia. The use of EPO is unfortunately not yet familiar in practical therapy due to its low availability.

In order to be able to use EPO as a therapeutic agent, consideration should be given to the possible antigenic effect, and therefore, EPO should preferably be prepared from raw material of human origin. For example, human blood or urine can be used by patients suffering from aplastic anemia or similar diseases and excreting high levels of EPO. However, these raw materials are only available in limited quantities (cf.

For example, White et al., Rec. Progr. Horm. Res. 16,219 [19601; Espada et al., Biochem. Med. 3,475 [19701; Fisher, Pharmacol.

Rev. 24,459 [1972] and Gordon, Vitam. Horm. (N.Y.) 31,105 [19731].

The manufacture of EPO products has generally been accomplished by concentrating and purifying urine from high EPO level patients suffering, for example, aplastic anemia or similar diseases (see U.S. Patents 4,397,840, 4,303,650 and 3,865,801). , The limited availability of such urine presents an obstacle to the practical use of EPO. Therefore, it is desirable to produce EPO urinary specimens from healthy people. The problem with using urine from healthy people is the low content of EPO compared to that of anemic patients. In addition, the urine of healthy humans contains certain inhibitory factors that act against electropoiesis in sufficiently high concentrations so that a satisfactory therapeutic effect can only be obtained after thorough purification of the EPO.

EPO can also be obtained from the blood plasma of sheep. The separation of EPO from the blood plasma gives sufficiently strong and stable water-soluble preparations (see Goldwasser, Control Cellular Dif., Develop., Part A, pp. 487-494, Alan R. Liss, Inc., N.Y. [1981]). However, ovine EPO could have an antigenic effect in humans.

Although EPO is a desirable therapeutic agent, current isolation and purification techniques from natural sources are not adequate for mass production of this compound. Sugimoto et al. US Pat. No. 4,377,513 describes a method for the mass production of EPO, which comprises the in vivo multiplication of human lymphoblastic cells, such as Namalwa, BALL-1, NALL-1, TALL-1, and IBL.

The production of EPO via other genetic engineering methods is described in the literature. However, neither a complete disclosure nor the chemical nature of the product has been published so far. The present application provides complete disclosure for the mass production of proteins and demonstrates the biological properties of the proteins and the human EPO. Such techniques also make it possible to produce proteins that differ chemically from authentic human EPO but have similar (and in some cases better) properties. Such proteins having biological properties of human EPO will be referred to hereinafter as EPO, irrespective of whether or not they are chemically identical to EPO.

Object of the invention

The aim of the invention is to provide a new process for the production of erythropoietin.

Explanation of the essence of the invention

The invention has for its object to provide a method for the production of erythropoietin available. The appropriate eypression vectors for the production of EPO, expression cells, purification schemes and related processes are described in the present invention.

A method of producing a human cDNA clone expressing biologically active erythropoietin is characterized in that

a) purified erythropoietin protein is digested by trypsin.

In accordance with the present invention, it is anticipated that host cells are grown in an appropriate medium containing a DNA sequence substantially as shown in Figure 3 B, which is operably linked to an expression control sequence and thus derived from the cells and erythropoietin produced in the medium is separated. Preferably, eukaryotic host cells are grown in an appropriate medium containing a DNA sequence as shown substantially in Fig. 4C, which is operably linked to an expression control sequence, and which thus is derived from the Cells and the medium produced erythropoietin is separated. According to the invention, the host cells are mammalian cells, for example 3T3 cells or Chinese hamster ovary cells (CHO). Furthermore, according to the invention, the DNA sequence in question is contained in a vector which also contains the bovine papilloma virus DNA.

As will be further described below, EPO was obtained in partially purified form, completely purified and digested with trypsin to produce specific fragments. These fragments were purified and sequenced. From these sequences, EPO oligonucleotides were assigned and synthesized. These oligonucleotides were used to screen a human gene map from which an EPO gene was isolated.

The accuracy of the EPO gene was determined by its DNA sequence, which was consistent with many sequenced tryptic protein fragments. A portion of the genomic clone was then used to show by hybridization that EPO could be detected in human fetal 20 week old mRNA. A human fetal liver cDNA library was set up and tested in the screening procedure. Three EPO cDNA clones were obtained (after the screening of 750,000 recombinants). Two of these clones were full-length, as deduced from the complete coding sequence and the essentially untranslated sequence of the 3 'and 5' ends. This cDNA was amplified both in SV-40 virally transformed Mf cells (COS-1 cell line; Gluzam, Cell 23, 175-182 [1981]) and in Chinese hamster ovary (CHO cell line, Urlaub G. and Chasin LA, Proc. Natl. Acad. Sci. USA 77, 4216-4230 (1980)). The EPO produced by COS cells is in vitro and in vivo biologically active EPO. The EPO produced by CHO cells is biologically active in vitro and has not been tested in vivo.

The EPO cDNA clone has an interesting open reading frame of 14-15 amino acids (aa) with initiator and terminator of 20-30 nucleotides (nt) above the coding region. A representative sample of E. coli transfected with the cloned EPO gene was obtained from the American Type Culture Collection, Rockville, Maryland, and is available under the designation number ATCC 40153.

Brief description of the drawings

Figure 1: represents base sequence of an 87 base pair exon of the human EPO gene.

Figure 2: shows the detection of EPOmRNA in the human fetal liver mRNA.

Figure 3a: shows the amino acid sequence of an EPO protein derived from the nucleotide sequence of lambda HEPOFL 13

was derived

 Figure 3 b: shows the nucleotide sequence of the EPO cDNA in lambda HEPOFL 13 and the derivatives derived therefrom

amino acid sequence

 Figure 4a: shows the relative positions of DNA inserts of four independent human

EPO genomic clones.

Figure 4 b: shows a map of the aparent intron and exon structures of the human EPO gene.

Figure 4c: shows a DNA sequence of the EPO gene shown in Figure 4b.

Figure bc, 5b

and 5c: show the construction of the vector 91023 (B).

Figure 6: shows the SDS-polyacrylamide gel analysis of the EPO produced in COS-1 cells compared to

nativenEPO.

Figure 7: shows the nucleotide and amino acid sequence of the EPO clone, lambda HEP0FL6.

Figure 8: shows the nucleotide and amino acid sequence of the EPO clone lambda HEPOFLB. -

Figure 9: shows the nucleotide and amino acid sequence of the EPO clone lambda HEPOFL 13.

Figure 10: shows a schematic representation of the plasmid pRKL-4.

Figure 11: shows a schematic representation of the plasmid pdBPV-MMTneo (342-12).

More detailed description

The present invention relates to the cloning of EPO genes and the production of EPO by the in vitro expression of these genes.

The patent literature and the scientific literature are overflowing with processes that are applicable to the production of recombinant products. These techniques generally involve the isolation or synthesis of the desired gene sequence and the expression of this sequence in prokaryotic or eutrophic cells by the techniques customary to those skilled in the art. Once oin certain gene il ^! s. If goi is unified and inserted (ie cloned) into a transfer vector, its availability in larger amounts is M. (Jer vector is transferred with its cloned gene into a suitable microorganism or cell line, eg Lakteri ° n! (efe, mammalian cells, such as COS-1 (monkey kidney), CHO (ovary of the Chinese hamster), cell lines of insects, and oak trees, in which the vector is replicated as soon as the microorganism or cell line becomes and that the vector can be isolated by conventional methods.

In this way, a continuously renewable source of the gene is delivered allowing for further manipulation, modification, and transfer to other vectors or other locations within the same vector. Expression can often be achieved by transferring the cloned gene (in the correct direction and reading frame) to the appropriate site of a transfer vector such that translation from a prokaryotic or eukaryotic gene results in the synthesis of a protein precursor. This includes the amino acid sequence encoded by the cloned gene linked to methionine or to an amino-terminal sequence of the prokaryotic or eukaryotic gene. In other cases, the transcriptional and translational initiation signals may be provided by a corresponding genomic fragment of the cloned gene. A variety of specific protein cleavage techniques can be used to cleave the protein precursor at a desired point so as to cleave the desired amino acid sequence, which can then be purified by conventional methods. In some cases, the protein containing the desired amino acid sequence is produced without the need for specific cleavage techniques, and thus can also be delivered by cells into the extracellular food medium.

Isolation of a gene clone of the human EPO

Human EPO has been isolated from urine of patients with aplastic anemia in a highly purified form, as described below. This purified EPO was completely digested by trypsin. In this case, fragments were obtained, which were separated by reversed-phase HPLC. From the gradient fractions, the individual fragments were recovered and subjected to microanalysis. The sequences of the Egyptian fragments are underlined in Figure 3 and are discussed in more detail below. Two of the amino acid sequences, Val-Asn-Phe-Tyr-Ala-Trp-Lys and Val-Tyr-Asn-phe-Leu-Arg were selected for the construction of oligonucleotide probes (yielding a 17-nucleotide and 32-fold degenerate pool of oligonucleotides, respectively) an 18-nucleotide and 128-fold degenerate pool of oligonucleotides, from the earlier tryptic fragments, and two 14-nucleotide, 48-fold degenerate pools of the latter trypic fragments). The 32-fold degenerate 17-mer pool was used to screen a human genomic DNA library in a Ch4A vector (22), using a modification of the Woo-and-O'Malley insitu to prepare the filters for screening Reinforcement Procedure (47).

The Arabic numbers (1) - (61) used in the text refer to the publications listed in numerical order at the end of this application.

The phage which hybridize to the 17-mer nucleotides were pooled, pooled into small groups and screened with the 14-mer and 18-mer pools. The phage which hybridize to the 17mer, 18mer and 14mer pools were purified by plaque technique and fragments were subcloned into M13 vectors for sequencing by the dideoxy chain termination method of Sanger and Coulson (23) (1977). The sequence of the region that hybridizes to the 32-fold degenerate 17-mer nucleotide in one of the clones is shown in Figure 1. This DNA sequence contains within an open reading frame the nucleotides that could accurately encode the tryptic fragment used to derive the 17-mer pool of oligonucleotides. Furthermore, analysis of the DNA sequence shows that the 17mer hybridization region was contained within an 87 bp exon bounded by potential splice acceptors and donor sites. The positive confirmation that these two clones (hereinafter referred to as lambda HEP01 and lambda HEPO 2 bezeicl.net) are EPO genome clones was obtained by sequencing additional exons containing additional code information obtained via tryptic fragmentation.

Isolation of EPO cDNA clones

Northern analysis (56) of human fetal (20 week old) liver mRNA was performed using a 95 nucleotide-long single-stranded sample prepared from an M 13 clone that was part of the 87-bp -Fxons contains. As shown in Figure 2, a strong signal could be detected in fetal liver mRNA. Accurate identification of this band as EPO mRNA was achieved by using the same sample for screening the bacteriophage lambda cDNA library of fetal liver mRNA (25). Several hybridization clones were obtained at a frequency of approximately 1 per 250000 recombinants. The complete nucleotide and deduced amino acid sequences for the ^ e clones (lambda HEPOFL 13 and lambda HEPOFL8) are shown in Figures 7 and 8. The EPO coding information is contained within 594 nucleotides in the 5 'half of the cDNA, which includes a very hydrophobic 27 amino acid initial and the final protein 166 amino acids in length. The assignment of the N-terminus of the finished protein is based on the N-terminal sequence of the protein excreted in the urine of persons with aplastic anemia (see Figure 1, Goldwasser [26], Sue and Sytkowski | 27] and

Yanagawa [21]). Whether this N-terminus (Ala-Pro-Pro-Arg) represents the current N-terminus of the circulating EPO, or

whether liver or urinary splittings occur is currently unknown. The amino acid sequences underlined in Figure 3 indicate those tryptic fragments or parts of the N-terminus for which the protein sequence is eri; 'ten was. The deduced amino acid sequence exactly matches the tryptic fragments that were sequenced, confirming that the isolated genes encode human EPO.

Structure and sequence of human EPO genes

Figure 4a shows the relative positions of DNA inserts of 4 independent human EPO genome clones. Hybridization analyzes of these cloned LNAs with oligonucleotide probes and with various samples prepared from two different classes of EPO cDNA clones confirm the position of the EPO gene throughout the region of the 3.3kb region, as shown in Figure 4a Full sequence analyzes of this region (see Example 4) and comparison with the cDNA clones provide the map of the intron and exon structures of the EPO genes (see Figure 4b) .The EPO gene is present in five exons Part of exon I, all exons II, III and IV, and part of exon V contain the protein coding information The remainder of exon I and V encodes the 5 'and 3' untranslated sequences.

Transient expression of EPO in COS cells

To demonstrate that biologically active EPO can be expressed in an in vitro cell culture system, COS cell expression studies were performed (58). The vector p91023 (B) used for the transient studies is described in Example 5. This vector contains the adenovirus major late promoter, an SV40 polyadenylation sequence, an SV40 origin of replication, an SV40 enhancer and the adenovirus VA gene. The cDNA insert lambda HEPOFL13 (see Figure 8) was inserted into a p91023 (B) vector, below the adenovirus major late promoter. This new vector is identified as pPTFL 13.

Twenty-four hours after transfection of this construct into the M6 strain of COS-1 cells (Horowitz et al., J. Mol. Appl. Genet. 2, 147-149 [1983]), the cells were washed and the medium against serum-free medium replaced. The cells are harvested 48 hours later. The degree of delivery of EPO to the culture supernatant was then examined by quantitative radioimunoassay for EPO (55). As shown in Table 1 (Example 6), immunologically reactive EPO was expressed. The biological activity of EPO produced by COS-1 cells was also examined. In a separate experiment, the vector containing λ: PO cDNA of lambda HEP0FL13 was transfected into CCS-1 cells and the media harvested as described above. EPO was then quantitated in the medium quantitatively by one of two in in vitro biological assays, ³ H-trymidine and CFJ-E (12,29), and by one of the two in vivo assays (in hypoxic mice and fasting rats [30,31]) (see Table 2, Example 7).

These results show that biologically active EPO is produced in COS-1 cells. In Western blots, it was demonstrated by means of a polyclonal anti-EPO antibody that EPO produced by COS cells has a mobility on SDS-polyacrylamide gels identical to that of the native EPO prepared from human urine was (Example 8). Therefore, it can be concluded that the extent of glycosylation of COS-1 produced EPO is similar to that of the native EPO.

Different vectors containing other promoters may also be used in COS cells or in other mammalian cells or eukaryotic cells. Examples of such other promoters for the purposes of the present invention include SV40 early and late promoters, the mouse metallothionein gene promoter, the promoter found in retroviruses of birds or mammals, the baculovirus polyhedron gene promoter and other. Examples of other cell types for the purposes of the present invention include E. coli, yeast, mammalian cells, such as. B. CHO (ovary of Chinahamsters), C127 (monkey epithelium), 3T3 (mouse fibroblasts), CV-1 (African green monkey kidney) and insect cells, such as. B. those of Spodoptera frugiperda and Drosophila metanogaster. These alternating promoters and / or cell types can effect regulation of the time or degree of expression of F.PO by producing a cell-specific type of EPO or the growth of large quantities of EPO-producing cells from less expensive, but more easily controllable Conditions allow.

An expression system that has the merit of mammalian expression but takes less time to produce a high-level expression cell line is composed of an insect cell line and a DNA virus that proliferates in that cell line. The virus is a nuclear polyhedrosis virus. It has a double-stranded circular DNA genome of 128kb. The nucleocapsid is rod-shaped and exists in two forms, the non-occluded form in the membrane-bound virus and in an occluded form enveloped in the infected nucleus by a protein crystal. These viruses can usually be propagated in in vitro insect cell cultures and are alterable by all routine animal-virological methods. The cell culture medium is usually a nutripriming solution of 10% fetal calf serum. Virus growth is initiated in vitro when an unoccluded virus (NOV) enters a cell and travels to the nucleus where it reproduces. Replication takes place at the core. During the initial phase (8-18 hours after infection) nucleocapsids are accumulated in the nucleus and then germinate as NOV through the plasma membrane, thus spreading the infection through the cell culture. Some nucleocapsids then remain in the cell nucleus (more than 18 hours after infection) and are occluded in a protein matrix (known as the polyhedral inclusion body, PIB), which is non-infectious in cell culture a protein known as polyhedrin (molecular weight 33kd).

Each PIB has an approximate diameter of 1 mm and can contain several hundred PIB per nucleus. Certainly, a large amount of polyhedrin will be produced later in the information cycle (approximately 25% of the total cellular protein level).

Since the PIBs play no role in the in vitro replication cycles, the polyhedrin gene can be deleted from the virus chromosome without a lasting effect on the in vitro viability of the virus chromosome. By using the virus as the expression vector, we have replaced the coding region for the polyhedrin gene with the foreign DNA to be expressed by placing it under the control of the polyhedrin promoter. This results in a non-PIB-forming virus phenotype.

This system has been used by many researchers, notably Pennock et al. and Smith et al., Pennock et al. (Gregory D. Pennock, Charles Shoemaker and Lois K. Miller, Mol. Cell. Biol. 3, 399-406 [1984]) reported the high level of expression of a bacterial protein, β-galactosidase, when placed under the control of the polyhedrin promoter , Another nuclear expression vector derived from the polyhedrosis virus has been described by Smith et al. Cellular Biol., May 16, 2156-2165 [19831]. The authors demonstrated the effectiveness of their vector through the expression of human β-interferon. The synthesized product was found as a glycosylated form and secreted by insect cells as expected. Example 14 describes modifications of the plasmid containing the

Autographa californica nuclear polyhedrosis virus (AcNPV) polyhedron gene, which allows for easy insertion of the EPO gene into the plasride so that it is under the transcriptional control of the polyhedron promoter. The resulting DNA is co-transfected into insect cells with intact wild type AcNPV chromosomal DNA. A genetic recombination experiment results in the replacement of the AcNPVC polyhedrin gene region by the DNA of the plasmid. The resulting recombinant virus can be identified from the viral progeny by the presence of the DNA sequences of the EPO gene. This recombinant virus was expected to produce EPO after reinfection of insect cells.

Examples of EPO expression in CHO, C127 and 3Γ3 and insect cells are given in Examples 10 and 11 (CHO), 13 (C127 and 3T3) and 14 (insect cells).

The biologically active EPO produced by prokaryotic or eukaryotic expression of the cloned EPO genes of the present invention may be used by physicians and / or veterinarians for the in vivo treatment of mammals. The amount of active ingredients will, of course, depend on the severity of the condition to be treated, on the route of administration chosen, on the specific activity of the active EPO, and will ultimately be decided by the attending physician or veterinarian. The amount of active EPO determined by the attending physician will hereinafter be referred to as the "EPO-treatment-effective amount." For example, the treatment of induced hypoproliferative anemia, which is associated with chronic renal failure in sheep, has been reported over the period 15-40 days a daily effective amount of EPO of 10 U / kg was found (see Eschbachetal, J. Clin Invest 74, 344 [1984]). The active EPO can be administered in any way dependent on the condition to be treated. Preferably, the EPO is injected into the bloodstream of the mammal to be treated It will be readily appreciated by one skilled in the art that the preferred route of administration may vary depending on the nature of the condition to be treated, although it is possible to: st, the active EPO in pure form or as it is preferably administered as a pharmaceutical composition.

The pharmaceutical compositions of the subject compounds, for both veterinary and human use, comprise the active EPO protein as described above together with one or more pharmaceutically acceptable carriers and optionally other therapeutic ingredients. The excipients must be compatible with other ingredients of the composition and not harmful to the patient. The composition should not contain any oxidizing agents and other substances known to be incompatible with peptides. The compositions may conveniently be packaged in unit dosage forms and may be prepared by many methods known in the pharmaceutical art. All methods include the step of mixing active ingredient with the carrier material composed of one or more excipients. In general, the compositions are prepared by uniformly and vigorously mixing active ingredient with liquid carriers or finely divided solid carriers or both, and then, if necessary, converting the product to the desired form of the composition.

The compositions for pirtereral administrations suitably comprise sterile aqueous solutions of the active ingredient and solutions which are preferably isotonic with the blood of the recipient. Such compositions may conveniently be prepared by dissolving the solid active ingredient in water to produce an aqueous solution, and under sterile conditions this solution may be administered in unitary or multi-dose dosage forms, e.g. B. in sealed ampoules or containers.

The EPO cDNA used herein includes the mature EPO cDNA gene preceded by an ATG 3odon and EPO cDNA encoding allelomorphic variations of the EPO protein. An allelomorphic variation is shown in Figures 3a and 3b. The EPO protein contains the 1 methionine derivative of the EPO protein (Met-EPO) and allelomorphic variations of the EPO protein. The mature EPO protein represented by the sequence in Figure 3 a begins with the sequence Ala-Pro-Pro-Arg ..., the beginning of which is characterized by the addition of "1" in Figure 3. The Met EPO would start with the sequence Met-Ala-Pro-Pro-Arg ...

embodiment

The following examples are provided for a better understanding of the present invention, the true scope of which is set forth in the appended claims. Likewise, variations of the stated methods are possible, which are included in the present invention. All temperatures are given in ⁰ C and shd uncorrected. The icon for Micron or Micro, z. Microliter, micromol, etc., is "μ", ζ. B. μΙ, etc.

Example 1: Isolation of a genome clone of EPO

EPO from the urine of patients with aplastic anemia was purified essentially according to the method developed by Miyake et al., J. Biol. Chem. 252, 5558 (1977), except that the phenol treatment was not performed and by heat treatment at 80 ° C (five minutes) to inactivate neuraminidase. The final step of purification was by fractionation on a C-4 Vydac HPLC acid (The Seoaratious Group), adding a 0-95% acetonitrile gradient with 0.1% trifluoroacetic acid (TFA) for 100 minutes was created. The EPO fraction was determined by gel electrophoresis and N-terminal sequence analysis (21,26,27) of the main fractions. The EPO was eluted at about 53% acetonnitrl and constituted approximately 40% of the protein oar, which was separated by reversed-phase HPLC. The EPO containing Fraktioner were concentrated by rotary evaporation to ΙΟΟμΙ, adjusted to pH 7.0 with Amoniumbiearbonat with 2% TPCK-treated trypsin (Worthington) for 18 hours completely digested at 37 ⁰ C. The tryptic digest products were then subjected to reversed-phase HPLC as described above. The optical density at 280 nm and 214 nm was measured. Well-separated peaks were ooprotected to near dryness and subjected directly to N-terminal amino acid sequence analysis (59) (Applied Biosystems Model 480A Gas Phase Sequenator). The resulting sequences are underlined in Figure 3. As previously described, two of these tryptic questions were selected for the synthesis of oligonucleotide probes. Auo of the sequence Val-Asn-Phe-Tyr-Ala-Trp-Lys (amino acids 46-52 in Figure 3) became 17mer with 32-fold degeneracy

AAA

TTCCANGCGTAGAAGTT and an 18mer with 128fold degeneration AAA

CCANGCGTAGAAGTTNAC

manufactured. From the sequence Vel-Tyr-Ser-Asn-Phe-Leu-Arg (amino acids 144-150 in Figure 3) two pools of 14mers were degenerate, each 32-fold

TTGN TT

TACACCTAACTTCCT and TTGN TT

TACACCTAACTTCTT

which differ in the first position of the leucine codon. The oligonucleotides were 5 'end with ³² P by polynucleotide kinase (New England Biolabs) and ³² P-ATP (New England Nuclear). The specific activity of the oligonucleotides varied between 1000 and SOOOCi / mmol. In the screening procedure of the human genomic DNA library in the bacteriophage lambda (Lawn et al., [22]), a modification of the in situ propagation method was carried out as originally described by Woo et al. (47) (1978). Approximately 3.5 x 10 ^s phage were plated at a density of 6000 phage on petri dishes with a diameter of 150mm (NZTYM medium) and incubated at 37 ⁰ C until plaques of about 0.5 mm were visible. After cooling to 4 ° C for one hour, duplicate replica plaques were transferred to nylon membranes (New England Nuclear) and were used. Power incubated at 37 ° C on fresh NZCYM plates. The filters were then denatured and neutralized by washing with 0.5 N NaOH / 1 M NaC 1 and 0.5 M Tris (pH 8) 1 M NaC 1 for 10 minutes. (1 h 5 χ SSC, 0.5% SDS) and the cell residue on the FiH .- / surface removed by gently scratching with a damp cloth after it, following heating in a vacuum at 8O ⁰ C for two hours, the filters were washed. This scratching reduces the adhesion of the sample to the filters. The filters were then washed with water and heated for 4 to 8 hours at 48'C in 3m tetramethylammonium chloride, 10 mM NaPO ₄ (pH 6.8), 5 Denier's, 0.5% SDS and 10 mM EDTA prehybridized. The ³ V labeled 17mer was then added at a concentration of 0.1 pmol / ml and hybridization was carried out for 72 hours at 48 ° C. After hybridization, the filters were washed extensively with 2 χ SSC (0.3m NaCl / 0.03m Na citrate, pH 7) at room temperature, then for one hour in 3m TMACI / 10mM NaPO ₄ (pH 6.8) at room temperature and then for a period of 5-15 minutes of both hybridization temperatures. After two days of autoradiography with a reinforcing screen, approximately 120 strong doublet signals were recorded. The positives were picked, pooled in eight, replated and screened using one half of the 14-mer pool on each of the two filters and the 17-mer on the third filter. The conditions for application to the plates and the hybridization are described above, however, the hybridization of the 14mer was carried out at 37 ⁰ C. After autoradiography, the sample was transferred from the 17mer filter to 50% formamide, and after 20 minutes at room temperature, the filter was rehybridized at 52 ° C with the 18-mer sample. Two independent phages hybridized in all three samples. The DNA of one of these phages (hereinafter referred to as lambda HEP01) was completely digested with 3au3A and used in M13 for the subsequencing DNA sequence analysis, inc'em dia dideoxy chain termination method of Sanger and Coulson (23) (1977). The nucleotide sequence and deduced amino acid sequence of the open reading frame coding for the EPO tryptic fragment (underlined region) is described in this application, the intron sequences are indicated in lower case, the Εχυη sequences (87 nucleotides) are in Capital letters specified. Coincident sequences between splice acceptor (a) and dnor (d) sites are underlined (see Figure 4c).

Example 2: "Northern" analysis of human fetal liver mRNA

5μg of human fetal liver mRNA (prepared from a 20 week old fetal liver) and an adult liver mRNA were electrophoresed in a 0.8% agarose-formaldehyde gel and analyzed by the method of Derman et al. , Cell, 23, 731 (1981) to nitrocellulose. A single-stranded sample was then prepared from the M13 template containing the insert shown in Figure 1. The Prirnei was a 20mer derived from the same tryptic fragment as the original 17-mer sample. The sample was prepared as previously described by Anderson et al., PNAS, (50) (1984), except that after digestion with Sma I (to produce the desired sample of 95 nucleotides in length, the Contains 74 nucleotides of the coding sequence), the small fragments of M 13 template were purified by chromatography on a Sepharose CI4B column with 0.1 η NaOH / 0.2m NaC 1. The filter was hybridized to approximate x 10 ⁶ cpm of this sample for 12 hours at 68 ⁰ C, washed twice mitSSCbei68 ° C and exposed for six days an intensifying screen. A single marker mRNA of 1200 nucleotides (indicated by an arrow) was run as a control in the adjacent row (see Figure 2).

Example 3: Fetal Liver cDNA

A sample identical to Example 2 was prepared and used for the screening of a fetal liver cDNA library prepared by the standard plaque screening method (Benton Davis, Science (541 (1978) in the vector Lambda-Ch 21A Toole et al , Nature, [25] [1984]). Three independent positive clones (hereinafter referred to as lambda HEPOFL611350 bp1, lambda HEP0FL8 [700bpl, and lambda HEPOFL13 [1400 bp]) were isolated and then screened for 1 χ 10 ^e spots The entire insert of lambda HEPOFL 13 and lambda HEPOFL6 was sequenced and then subcloned into M13 (see Figures 9 and 7, respectively) of lambda HEPOFL8 only parts were sequenced and the remainder identical to the other two clones (See Figure 8.) The 5'- and 3'-untranslated sequences are indicated by lowercase letters and the coding region is represented by uppercase letters.

With reference to Figures 3 a and 3 b, the deduced amino acid sequence depicted below the nucleotide sequence is numbered beginning with number 1 for the first amino acid of the finished protein. The putative precursor peptide is also given. Cysteine residues in the finished protein are additionally indicated by the designation SH and possible N-glycosylation sites by a star. The underlined amino acids indicate such residues, which by

N-terminal protein sequencing or by sequencing the tryptic fragments of EFO as described in Example 1. Partial solid lines indicate residues in the amino acid sequence of certain tryptic fragments that could not be uniquely determined. The cDNA clones lambda HEP0FL6, lambda HEP0FL8 and lambda HEPOFL 13 have been deposited and are available from the American Type Culture Collection, Rockville, Maryland under the designation numbers ATCC 40156, ATCC 40152 and ATCC 40153.

Example 4: Genome structures of the EPO genes

The relative sizes and positions of the four independent genome clones (lambda HEP01, 2,3, and C) from the Haelll / Alu I library are shown by the overlapping lines in Figure 4a. The more solid line indicates the position of the EPO gene. A scaling (in Kb) and the positions of the known cleavage sites for the restriction endonucleases are given. The EPO gene-containing region was completely sequenced from both strands using exonuclease III generated series of quenching within this region. A schematic representation of five exons encoding EPO mRNA is shown in Figure 4 b. The exact 5 'limit of exon I is currently unknown. The protein coding part of these exons is shaded. The complete nucleotide sequence of this region is shown in Figure 4c. The known boundaries of each exon are outlined by vertical bars. The genome clones lambda HEP01, lambda HEPO2, lambda HEPO3 and lambda HEPO6 have been deposited and are available from the American Type Culture Collection of Rockville, Maryland under the designation numbers ATCC 40154, ATCC 40155, ATCC 40150 and ATCC 40151.

Example 5 Construction of the Vector ρ91023 (B)

The transformation vector used was pAdD26 SVpA (3) as described by Kaufman et al., Mol. Cell. Biol. 2,1304 (1982). The structure of this vector is shown in Figure 5a. This plasmid contains a murine dihydrofolate reductase (DHFR) cDNA gene which is under the transcriptional control of the adenovirus 2 (Ad 2) major late promoter. A 5 'splice site is indicated in the adenovirus DNA and a 3' splice site derived from an immunoglobulin gene is located between the Ad2 maj-r late promoter and the DHFR coding sequence. The SV40 early polyadenylation site is located in the strand below the DHFR coding sequence. The prokaryotic-derived section of pAdD26SVpA (3) is derived from pSVÜd (Mellon et al., Cell 27, 279 (1981)) and does not contain the pBR322 sequences known to inhibit replication in mammalian cells (Lusky et al., Nature 293 , 79 [19811].

pAdD26SVpA (3) was transformed into the plasmid pCVSVL2 (see Figure 5a). pAdD26SVpA (3) was transformed into the plasmid pAdD26SVpA (3) (d) by deleting one of the two potl sites into pAdD26SVpA (3). This was achieved by partial digestion with PstI using an enzyme suppression to give a subpopulation of linearized plasmids in which only a single PstI site was cleaved followed by treatment with Kien jw, ligation recirculation and screening for PstI site deletions located at the 3 'end of the SV 40 polyadenylation sequence.

The adenovirus tripartite leader and the virus-associated genes (VA genes) were inserted into pAdD26SVpA (3) (d) (see Figure 5a). First, pAdD26SVpA (3) (d) was cleaved with PvuII to obtain a linear molecule opened within the 3 'region of the three elements comprising the tripartite leader. Then, pJAW43 (Zain et al., Cell, 16,851 [1979]) was digested with Xhol, treated with Klenow, digested with Pvu II, and the 140 bp fragments containing the second part of the third leader were purified by electrophoresis on acrylamide gel (6 % in Tris-borate buffer; Maniatis et al., supra). The 140bp fragment was then joined to the Pvull-digested pAdD26SVpA (3) (d). The joined product was used to transform E. coli to tetracycline resistance and colonies were screened for the foundation-Hoguess procedure by a ^'2 p-labeled probe was used which hybridized to the 140bp fragment. A DNA from positively hybridized colonies was prepared to test whether the reconstructed Pvu II-Stolle was a 5 'or 3' end of the inserted 140 bp DNA specific for the second or third adenovirus Iate leaders , The correct orientation of the Pvu II site is the 5 'end of the 140bp slot. This plasmid is referred to as pTPL in Figure 5a.

The SV40 Ava II D fragment containing the SV40 enhancer sequence was obtained by digestion of SV40 DNA with Ava II, the ends were blunt ended with the Klenow fragment of Po II, Xhol-Linkei " linked to the fragments, digested with Xhol to open the XhoI sites, and the four largest (D) fragments isolated by gel electrophoresis This fragment was then ligated to Xhol-cut pTPL to yield the plasmid pCVSVL2-TPL Orientation of the SV40-D fragment in pCVSVL2-TPL was such that the SV40 late promoter was in the same orientation as the adenovirus major-iate promoter.

To convert the adenovirus-associated (VA) genes into the pCVSVL2-TPL, a plasmid pBR322 was first constructed which contained the adenovirus type 2 HIND Ill-B fragment. The adenovirus type 2 DNA was digested with Hind III and the B fragment was isolated by gel electrophoresis. This fragment was inserted into pBR 322 previously digested with Hind III. After transformation of E. coli into ampicillin resists7, the recombinants were screened for insertion of the Hind W-B fragment and the inserted orientation determined by digestion with restriction enzymes. pBR322 AdHindIII-B contains the adenovirus type 2 HindIII B fragment in the orientation shown in Figure 5b. As shown in Figure 5 b, the VA genes are usually obtained from the plasmid pBR322-Ad Hind HI-B by digestion with Hpa I by adding EcoR 1 linker and digested with EcoR 1, followed by the 1, 4 kb fragment is recovered. The fragment with the protruding EcoR 1 ends is then linked to the EcoR 1 site of PTL, which was previously digested with EcoR 1. After transformation of E. coli HB101 and selection for tetracycline resisteriz, a screening procedure was performed on the colonies by filter hybridization to DNA specific for the VA genes. The DNA was prepared from positively hybridized clones and characterized by digestion with restriction endonucleases. The resulting plasmid is designated p91023.

As shown in Figure 5c, the two EcoR 1 sites in p91023 are removed by completely cutting p91023 with EcoR 1 to produce two DNA fragments, one at about 7 kb, the other at about 1.3 kb. The latter fragment contained the VA genes. The ends of both fragments were filled in using the Klenow fragment of Poll and the two fragments were then joined together. A plasmid p91023 (A) containing the VA genes, which is similar to p91P? 3, but does not contain the two EcoR 1 sites, was prepared by both the base stone hogness screening procedure with the VA gene fragment also identified by conventional restriction site analysis.

The single PstI site in p91023 (A) was removed and replaced with EcoR 1 site. p91023 (A) was completely cut with PstI and treated with the Klenow fragment of Poll to give fresh ends. EcoR 1 linkers were linked to the smooth PstI sites of p91023 (A). The linear p91023 (A) with the EcoR 1 linkers at the smooth PstI sites was separated from the unlinked linkers and digested to completion with EcoR 1 and recombined. A plasmid p91023 (B) was obtained (see Figure 5c) and characterized with a structure similar to p9102JiA), but with an EcoR 1 site in place of the earlier PstI site. Plasniid p91023 (B) has been deposited and is available from the American Type Culture Collection of Rockville, Maryland under the number ATCC 39754.

Examples:

The cDNA clones (lambda HEPOFL6 and lambda HEPOFL13, see Example 3) were inserted into plasmid p91023 (B) to give pPTFL6 and pPTFL13. 8 μg of each purified DNA was then used to transfer 5 × 10 ⁶ COS cells using the DEAE-dextran method (see below). After 12 hours, the cells were washed and treated with chloroquine (0.1 mM) for two hours, washed again and exposed to 10 ml of medium containing 10% fetal calf serum for 24 hours. The medium was then exchanged for 4 ml of serum-free medium and allowed to go 48 hours later. The production of immunologically active EPO was quantified using a Rauioimmunoassays according to the method Sherwoou -> d Goldwdsser (55). The antibody was obtained from Dr. Judith Sherwood provided. The iodinated tracer was prepared from the purified EPO (see Example 1). The sensitivity of the assay is about 1 ng / ml. The results are shown in Table 1.

Table 1 Conc. On in the medium vector released EPO (ng / ml) 330 pPTFL13 31 pPTFL6

pPTFL 13 has been deposited and is available from the American Type Culture Collection of Rockville, Maryland under the number ATCC 3999.

Example 7:

The EPO cDNA (lambda HEPOFL13) was inserted into the p91023 (B) vector, transferred to COS-1 cells and harvested as above (Example 6), except that the chloroquine treatment was omitted.

In vitro biologically active EPO was measured using either a mouse colony-forming assay with mouse fetal liver cells as a source of CFU-E or the ³ H -timidine uptake assay using spleen cells with phenylhydrazine-injected mice. The sensitivity of this detection is about 25mU / ml. Biologically active EPO was measured in vivo either by the hypoxic mouse or starving rat method. The sensitivity of this detection is about 100mU / ml. In both tests, no activities were detected in blind experiments. The results of the EPO expressed by the clone HEPOFL13 are shown in Table 2, in which the activities are expressed in U / ml when used as a standard in commercial EPO (Toyobo Incorp.).

Table 2 Aue .τιit Typl EPO cDNA transferred cells excreted EPO Detection activity

RIA 100ng / ml

CFU-E 2 <0, EU / ml

³ H-Thy 3.1 <1,8U / ml

Hypoxic mouse 1 U / ml

starving rat 2 U / ml

Example 8: SDS-polyacrylamide gel analysis of EPO from COS cells

180 ng of EPO released into the medium of COS cells transferred with EPO (lambda HEPOFL 13) cDNA in vector 91023 (B) (see above) were seeded on a 10% SDS-Laemlli polyacrylamide Gel electrophoretically separated and electrotransferred to nitrocellulose paper (Towbin et al., Proc. Natl. Acad. Sei., USA 76,4350 [1979]). The filter was checked with anti-EPO antibodies as described in Table 1, washed, and republished with ¹²⁵ L-staph A protein. The filter was autoradiographed for two days. Native pure EPO as described in example 1 was electrophoretically separated either before (line B) or after iodination (line C) (see Figure 6). The markers used include ³⁵ S-methionine labeled serum albumin (68000D) and ovalbumin (45000D).

Example 9: Construction of RK1-4

The BamHI-Pvull fragment from the plasmid PSV2DHFR (Subramani et al., Mol. Cell. Biol., 1854-864 (1981)) which encodes the SV40 early-region adjacent to the murine dihydrofolate fleductase (DHFR) gene. Promoter, an SV40 enhancer containing small t-antigen intron and the SV40 polyadenylation sequence was isolated (fragment A) The remaining fragments were obtained from vector p91023 (A) (see above) as follows:

p91023 (A) was digested with Pst I at the single Pst I side next to the adenovirus promoter to linearize the plasmid and either linked to synthetic PstI or to EcoRI converters and recircularized (making the PstI-EcoRI sites PstI was generated at the original PstI site; 91023 [B ']) or treated with the large fragment of DNA polymerase I to disrupt the PstI sites and linked to the synthetic EcoR 1 linkers and recircularized (thereby an EcoR 1 site is generated at the original PstI site, 91023 [B]).

Each of the two resulting plasmids 91023 (B) and 91023 (P ') were digested with Xba and EcoR I to form two fragments (F and G). By joining fragment F from p91023 (B) and fragment G from p91023 (B ') and fragment G out

p91023 (B) and fragment F from p91023 (B ') were generated two neuo plasmids containing either an EcoR I-Pst I site or a PstI-EcoRI site at the original PstI site. The plasmid containing the PstI-EcoRI site, where the PstI site is closest to the adenovirus js-major-late promoter, was designated p91023 (C).

The vector p91023 (C) was completely digested with Xhol and the resulting linearized DNA with protruding ends was transformed with a large fragment of E. coli DNA polymer pse 1 into a blunt-ended DMA. To this DNA was bound a 340 bp HindIII-EcoRI fragment containing the SV40 enhancer prepared as follows:

The SV40 Hind III-Pvu II fragment containing SV40 origin or replication and enhancer was inserted into plasmid c-lac (Little et al., Mol. Giol. Med. 1,473-488 [1983] ). The c-lac vector was prepared by digesting c-lac DNA with RamH I, then filling the protruding ends with a large fragment of DNA polymerase 1 and digesting the DNA with Hind III. The resulting plasmid (c-SVHP 1 ac) regenerated] the BamH I site by joining the blunt ends of Pvu II. The EcoR I-Hind III fragment was prepared from c-SVHPIac and ligated to EcoR I-Hind III. Fragment of PSVOd bound (Mellou et al., Supra) containing the plasmid origin of replication, and the resulting plasmid PSVHPOd was selected. The 340bp EcoR I-Hind HI fragment of PSVHPOd containing the SV40 origin / E, hancer was then prepared, blunt-ended both ends with a large fragment of DNA polymerase, and Xhol-digested p91023 (c) vector as described above.

The resulting plasmid (p91023 [C] / Xho / blunt end plus EcoRI / HindIII / blunt end, SV40 origin plus enhancer) in which the orientation of the HindIII H-EcoR I fragment was such that the BamHI site within this fragment was closest to the VA gene was designated pES105. The plasmid pES 105 was digested with BamH I and Pvu II and also with Pvu II alone, and the SamH I-Pvu II fragment containing the adenovirus major late promoter (fragment B) and the Pvu II fragment. Fragment containing the plasmid within the resistance genes (tetracycline resistance) and other sequences (Fragment

C) were isolated. Fragments A, B and C were linked and the plasmid shown in Figure 10 b was isolated and referred to as RK1 -4. The plasmid RK1 -4 was deposited with the American Type Culture Collection, Rockville, Maryland and is available there under the assignment number ATCC 39940.

Example 10 Expression of EPO in CHO Cells (Method I)

The DNA (20 μg) from the plasmid pPTFL13 (see Example 6) was digested with the restriction endonuclease C1 al (to linearise the plasmid) and was digested with Cla-1 digested DNA from the plasmid pAdD26SVp (A) 1 ( 2pg) containing an intact dihydrofoiate reductase (DHFR) gene regulated by the adenovirus major late promoter (Kaufmann and Sharp, Mol and Cell Biol., 2, 1304-1319 (1982)) .This linked DNA was used to transfect DHFR-negative CHO cells (DUKX-B II, Chasin LA and Urlaub G., Proc Natl Acad., 77, 4216-4220 [1980]), and after two days of culturing, the cells were transfected containing at least one DHFR gene selected in nucleoside-free alpha medium and 10% dialyzed fetal calf serum were added After two weeks growth in selective media, the colonies were removed from the original plates, assembled in groups of 10 to 100 colonies, replated and grown in nucleoside-free alpha medium The supernumerary media The pools grown prior to methotrexate selection were tested for EPO using an RIA.

Positive EPO production pools were cultured in the presence of methotrexate (0.02 UM), then subcloned and retested. EPO-Cla4alpha4.02-7, a single subclone from the EPO-Cla4alpha4.02 pool, delivers 460ng / ml of EPO into medium ah containing 0.02 μΜ MTX (see Table 3). EPO-Cla4alpha4.02-7 is the cell line of choice for EPO production and has been deposited with the American Type Culture Collection (assignment number ATCC CRL8695). At present, this clone undergoes stepwise selection with increasing concentration of MTX and is expected to yield cells producing even higher concentrations of EPO. For pools negative in the RIA, methotrexate-resistant colonies obtained from the countercultures grown in the opposite form of methotrexate (0.02μΜ) were rechecked for EPO by RIA. Those cultures that were not positive were subcloned and cultured in further increasing concentrations of methotrexate.

Stepwise methotrexate (MTX) selection was achieved by repetitive cell culture cycles in the presence of increasing concentrations of methotrexate and selection of surviving cells. After each run, EPO in the culture supernatant was measured by RIA and by in vitro biological activity. The concentrations of methotrexate used in each incremental enhancement were 0.02M, 0.1M and 0.5M. As shown in Table 3, after one hour of selection in 0.02M MTX, significant concentrations became apparent from EPO into the culture medium.

Table 3

Concentration of EPO released into the medium

Prcbe test alpha medium harvest 0.02 pMmethotrexa * ^: n the

alpha-medium harvest

4a4Pool RIA 17 mg / ml 50ng / ml

4a4Single

(.02-7) - 460ng / ml

Example 11 Expression of EPO in CHO Cells (Method 2)

The DNA of clone lambda HEPOFL13 was digested with EcoRI and the small R1 fragment containing the EPO gene was subcloned into the EcoRI site of plasmid RK1-4 (see Example 10). This DNA (RKFL13) was then used to directly transfect the DHFR-negative CHO cells (without digestion). The selection and amplification was carried out as described in Example 10.

The RDFL13 DNA was also inserted into CHO cells by protoblast fusion and microinjection. The plasmid RKFL13 has been deposited and is available from the American Type Culture Collection, Rockville, Maryland, under the designation number ATCC 39989.

Colony Pool A RIA 3 H-Thy single Kolonieklon RIA ; .02c-Z) ³ H-Thy Mikroinji- RIA zierterPool (DEPO-I) ³ H-ThV

Table 4 Concentration of EPO released into the medium

Sample assay alpha-medium-one 0, (^ MMethotrexatin

the alpha-medium harvest

3ng / ml 42ng / ml (pool)

150ng / ml (clone) 1.5U / ml

90ng / ml 5.9 U / ml 60ng / ml 160ng / ml

1.8 U / ml

The preferred single colony clone has been deposited and is available from the American Type Culture Collection, Rockville, Maryland, under the designation number ATCC CRL8695.

Example 12: Expression of the EPO genome clone in COS-1 cells

The vector used for expression of the EPO genome clone is pSVod (Mellon et al., Supra). The DNA of pSVOd was completely digested with HindIII and the ends were blunt-ended with the large fragment of DNA polymerase-I. The EPO genome clone lambda HEP03 was completely digested with EcoRi and HindIII and the 4.0kb fragment containing the EPO gene was isolated and converted to blunt ends as described above. The nucleotide sequence of this fragment from the HindIII site to the region above the polyidenylation signal is shown in Figure 4. The EPO gene fragment was inserted into the pSVOd plasmid fragment. Correctly constructed recombinants in both directions could be isolated and verified. Plasmid CZ2-1 has the EPO gene in orientation "a" (i.e., the 5 'end of EPO is closest to the SV40 origin) and the plasmid CZ1-3 has the opposite orientation (orientation

"B").

Plasmids CZ1-3 and CZ2-1 were transfected into COS-1 cells as described in Example 7, the medium harvested and tested for immunologically active EPO. Approximately 31 ng / ml of EPO was detected in the culture supernatant of CZ2-1.

16-31 ng / ml of CZ1-3.

The genome clones HEP01, HEP02 and HEP06 can be inserted into COS cells for expression in a similar manner.

Example 13: Expression in C127 and CHO cells and construction of pBPVEPO A plasmid containing the EPO cDNA sequence under the transcriptional control of the metallothionein promoter of mice and which is linked to the complete bovine papilloma virus DNA , was made as follows:

pEP049f

Plasmid SP6 / 5 was obtained from Promega Biotec. Purchased. This plasmid was digested to completion with EcoRI and the 1340bp EcoRI fragment of lambda HEP0FL13 was inserted by DNA ligase. The resulting plasmid, in which the 5 'end of the EPO gene was closest to the SP6 promoter (determined by 3g11 and HindIII digestion), was designated pEPO49f. In this orientation, the BamHI site in the PSP6-5 poly linker is directly adjacent to the 5 'end of the EPO gene.

pMMTneoABPV

The plasmid pdBPV-MMTneo (342-12) (Law et al., Mol. And Cell Biol. 3, 2110-2115 [1983]), shown in Figure 11, was completely digested with BamHI to generate two fragments A large fragment of about 8kb length containing the BPV genome and a smaller fragment of about 6.5kb in length containing the pML2 origin of replication and the ampicillin resistance gene, the methallothionein promoter, contains the neomycin resistance gene and the SV40 polyadenylation signal. The digested DNA was reclassified by DNA ligase and the plasmids containing only the 6.8kb fragment were identified by digestion with EcoRI and BamHI restriction endonucleases. One of these plasmids was called pMMTneoABPV.

pEP015a

pMMTneo BPV was completely digested with BglII. pEPO49f was completely digested with BAMHI and BglII and the approximately 700bp fragments containing the amino acid EPO coding region were prepared by gel isolation. The Bgl II digested pMMTneo BPV and the 700bp BamHI / BG1II EPO fragments were linked and the resulting plasmids containing the EPO cDNA were identified by colony hybridization with the oligonucleotide d (GGTCATCTGTCCCCTGTCC) probe specific for the EPO gene , Of the plasmids that were positive in the hybridization analysis, one (pEP015a) having the EPO cDNA in the orientation such that the 5 'end of the EPO cDNA was closest to the metallothionein promoter was digested with EcoRI and Kpnl identified.

pBPVEPO

The plasmid pEP015a was completely digested by BamHI to linearize the plasmid. The plasmid pdBPVMMTneo (342-12) was also completely digested with BamHI to generate two fragments of 6.5 and 8kb. The 8kb fragment containing the entire bovine papilloma virus genome was isolated (gel). pEP015a / BamHI and the 8kb BamHI fragment were linked together, and a plasmid (pBPV-EPO) containing the BPV fragment was identified by colony hybridization using an oligonucleotide probe d (P-CCACACCCGGTACACA-OH), which is specific for the BPV genome. Digestion of pBPV-EPO DNA with HindIII indicated that the direction of transcription of the BPV genome was the same as the direction of transcription of the meiallothionein promoter (as in pdPBV-MMTneo [342-12], cf. Figure 11). The plasmid pdBPV-MMTneo (342-12) is available from the American Type Culture Collection, Rockville, Maryland, under the designation number ATCC 37224.

expression

The following methods were used for the expression of EPO.

Method I

DNA-pBPV-EPO was prepared and approximately 25pg were used to transfer approximately 10 ⁶ C127 (Lowy et al., J. of Virol. 26, 291-298 [1978]) - CHO-Zelleri using the standard technique of calcium phosphate precipitation (Grahm et al., Virology 52, 45 & 467 [19731].

Five hours after transfection, the transfection medium was removed, the cells quenched with glycerol, washed, and fresh alpha medium supplemented with 10% fetal bovine serum. Forty-eight hours later, the cells were trypsinized and partitioned 1:10 in DME medium m: ¹ 500 μg / ml G418 (Southern et al., Mol. Appl. Ganet., 1,327-341,1982) and the cells split for two incubated for up to three weeks. G418-resistant colonies were individually isolated in microtiter plates and grown in the presence of G418. The cells were then washed, fresh medium supplemented with 10% fetal Rin Jerserum, and the media harvested 24 hours later. The conditioned media were tested by radioimmunoassay and in vitro biological experiments and were positive for EPO.

Method Il

C127 or CHO cells were cotransfected with 25 μg of pBPV-EPO and 2 μg of pSV2nco (Southern et al, supra, see Method I). This represents an approximately 10-fold molar excess of pBPV-EPO-Nac. In transfection, the procedure is the same as in method I.

Method II!

C 127 cells were transfected with 30 pg of pBPV-EPO transferred (see. Method I ". After transfection unddur Auftei'ung \ ^Λ · Λ0), the medium was alie three days with fresh replacement. After about two weeks points were transformed from pBPV Cells were visualized Individual spots were separated into 1-well microtiter plate wells, grown to a sub-confluent monolayer, and assayed for EPO activity or antigenicity in the conditioned medium.

Example 14 Expression in Insect Cells Construction of pIVEV EP0FL13

The plasmid vector pIVEV has been deposited and is from the American Type Culture Collection, Rockville, Maryland, lower. the assignment number ATCC 39991 available. The vector was modified as follows:

pIVEVNI

pIVV was used with EcoRI to lihearize the plasmid, using a large fragment of DNA polymerase I, the ends were blunt-ended and a single NotI linker GGCGGCCGCC

CCGCCGGCGG was inserted by joining the blunt ends. The resulting plasmid is called pIVEVNI.

pIVEVSI

pIVV was digested with SmaI to linearize the plasmid and a single sfil linker GGGCCCCAGGGCCC

CCCGGGTCCCCGGG was inserted by joining the blunt ends. The resulting plasmid was named pIVEVSI.

pIVEVSIBgKp

The plasmid pIVEVSI was digested with KPnI to linearize the plasmid, and about 0-1000 bp was removed from each end by digestion with the double-stranded exonucloase Ba 131. All of the resulting ends, which were not completely smooth, were blunt-ended using the large fragment of DAN polymerase I and the poly-linker

Xhol Xbal

I j "

Iii '

BgLII EJco ^ I CIaI Kpnl

¹ AGATC afcG / LGÄA üWtAG'aTCG AT * GG TACC ¹ TCTAGAGCTCTTAAGATGTAGCTACCATGG

is inserted by joining the smooth ends. The poly linker was inserted in both directions. A plasmid in which the poly-linker is oriented so that the BglII site within the poly-linker is closest to the polyhedron gene promoter is termed pIVEVSIBgKp. A plasmid in which KpnI sites within the poly linker are closest to the polyhedron gene promoter is termed pIVEVSIKpBg. The number of base pairs deleted between the original KPnI site in pIVEVSI and the polyhedron promoter was not determined.

plEIVSIBgKp has been deposited and is available under the designation number ATCC 33988 from the American Type Culture Collection, Rockville, Maryland.

plVEVSIBgKpN

pIVVNI was completely digested with KPnI and PstI to obtain two fragments. The larger fragment containing the plasmid origin of replication and the 3'-eno of the polyhedron gene was prepared by gel isolation (fragment A). pIVI and KPn were completely digested with pVSLIIBgKp to obtain two fragments and a small fragment containing the polyhedron gene pro-thioro. The poly-linker was prepared by gelation (fragment B). Fragments A and B were then ligated together by DNA ligase to form the new plasmid pIVVSIBgKpNI, which contains a partially deleted polyhedron gene into which a poly-linker has been inserted, and which also contains the NotI site (replace the disrupted EcoRI site) and an SHI site flanking the polyhedron gene region.

plVEPO

plVEVSI BgKpNI was completely digested with EcoRI to linearize the plasmid and the 1340 bp EcoRI fragment of lambda HEPOFL13 was inserted. The plasmids containing the EPO gene in the orientation so that the 5 'end of the EPO gene was closest to the polyhedron promoter and the 3' end of the polyhedron gene were identified by digestion with Bgl II. One of these plasmids in the orientation described above was called plVEPO.

Expression of EPO in insect cells

Large quantities of the plVEPO plasmid were produced by transformation of the E. coli strain JM101-tgi. The plasmid DNA was isolated by the cleared-lysate technique (Maniatis and Fritsch, Cold Spring Harbor Manual) and further purified by CsC1 centrifugation The L-1 wild-type autographa californica polyhydrosis virus (AcANPV ) was prepared by phenol extraction of Viir, particles and subsequent CsCl purification of the viral DNA.

These two DNAs were then cotransfected into spod.-frugiperda cells IPLB-SF-21 (Vaughn et al .. In Vitro, Vol. B, pp. 213-217 [1977]) using the method of calcium phosphate transfection (Potter and Miller, 1977). For each plate of cells to be cotransfected, 1 μg of the wild type AcNPV DNA and 10 μg of plVEPO were used. The plates were incubated at 27 ^° C. for five days. The supernatant was then harvested and EPO expression in the supernatant was confirmed by radioimmunoassay and by in vitro biological methods.

Example 15: Purification of EPO

The COS cell conditioned media (121) with EPO concentrations up to 200 pg / l were concentrated to 600 ml using ultrafiltration membranes with a cut-off of 10000 molecular weight, e.g. B. a Millipore Pellikan membrane. The tests were performed by means of an RIA as described in Example 6.

The supernatant from the ultrafiltration was diafiltered against 4 ml of a 10 mM sodium phosphate buffer (pH 7.0).

The concentrated and diafiltered conditioned media contained 2.5 mg EPO (380 mg total protein). The EPO solution was further concentrated to 186 ml and the precipitated proteins were spun down at 120000xg for 30 minutes. The supernatant containing 2.0 mg of EPO was adjusted to pH 5.5 with 50% acetic acid, stirred at 4 ° C for 30 minutes, and the precipitate was spun down at 13,000xg for 30 minutes.

Carbonylmethyl-Sepharose chromatography

The supernatant from centrifugation (20 ml) containing 200 μg of EPO (24 mg total protein) was applied to a column packed with CM-Sepharose (20 ml) and equilibrated with 10 mM sodium acetate (pH 5.5). Subsequently, the column was washed with 40 ml of the same buffer. The CMO-Sepharose bound EPO was eluted after 100 ml of a gradient of NaU (0-1) in 10 mM sodium phosphate (pH 5.5). The fractions containing EPO (total 50 μg of 2 mg total protein) were pooled and concentrated to 2 ml using an Amicon YMIO ultrafiltration membrane.

Reversed-phase HPLC

The concentrated fractions after column chromatography containing EPO were further purified by reversed-phase HPLC over a Vydac C-4 column. The EPO was designed with a flow rate of 1 m! min was applied to the column equilibrated with 10% solution B (solution A was 0.1% CF ₃ COOH in water, solution B was 0.1% CF ₃ COOH in CF ₃ CN). The column was washed with 10% of solution B for 10 minutes and the EPO was eluted with a linear gradient of B (10-70%, 60 min). The EPO-containing fractions were pooled (approximately 40 μg of EPO from 120 μg total protein) and lyophilized. The lyophilized EPO was reconstituted in 0.1 M Tris-HCl buffer (pH 7.5) with 0.15 M NaCl and rechromatographed on a reversed-phase HPLC. The fractions containing EPO were pooled and analyzed by SDS-polyacrylamide (10%) gel electrophoresis (Lameiü, UK Nature). The pooled EPO fractions contained 15.5pg of EPO (25mg total protein).

Claims (5)

1. A process for the production of erythropoietin, characterized in that A) host cells are grown in an appropriate medium containing a DNA sequence substantially as shown in Fig. 3B, which is operably linked to an expression control sequence and which is thus separated from the cells and the medium produced erythropoietin or B) eukaryotic host cells are cultured in an appropriate medium containing a DNA sequence as shown substantially in Fig. 4C, which is operably linked to an expression control sequence, and which thus separates erythropoietin produced by the cells and the medium becomes. 2. An ancestor according to claim 1, characterized in that the host cells are mammalian cells. 3. The method according to claim 2, characterized in that the mammalian cells are 3T3 cells. 4. The method according to claim 2, characterized in that mammalian cells are ovarian cells of the Chinese hamster (CHO). 5. The method according to claim 2, characterized in that the DNA sequence in question is included in a vector which also contains the bovine papilloma virus DNA.