SK386392A3 - Interferon alpha - Google Patents

Abstract

The objects of the invention are O-glycosylated IFN-alpha, a process for producing the same and the use of the O-glycosylated proteins as medicinal drugs.

Description

The invention relates to interferon alpha, in particular O-glycosylated interferon alpha, in particular with IFN-alpha2 valsts.

r. of an immunological or biological nature, a process for the manufacture of the substance and pharmaceutical compositions containing it.

BACKGROUND OF THE INVENTION

Since the discovery of interferon approximately 30 years ago, their properties as mediators of intercular communication have been extensively investigated. Initially, the designation of the various types of these substances was derived from the cells in which they originated, e.g. leukocyte IFN, β-fibroblast IFN and the like. With increasing knowledge of their structure, a new terminology was introduced. At present, four types of interferon are distinguished, Interferon alpha, beta, gamma and omega, with alpha, beta and omega interferons belonging to group 1 of interferon, which are close in structure and properties.

IFN-gamma is produced in lymphocytes that have been stimulated by antigens or mitotic agents. The amino acid chain, which has no homology to the Group 1 chain, contains two potential glycosylation sites.

IFN alpha, beta and omega are synthesized by a variety of cells in response to viral infections or following induction of double-stranded RNA.

In the case of IFN-alpha, it is inherent to the whole family of proteins. A6 has so far discovered at least 14 functional genes that code for different types of IFN-alpha. These proteins are closely related and their amino acid chains are approximately 80 to 90% homologous. With the exception of IFN-α1al4, no N-glycosylation site (Asn-X-Ser / Thr) is included in any of the other amino acid chains of IFN-alpha. Thus, glycosylation is not feasible except for IFN-alpha1, but O-glycosylation is the subject of discussion (Labdon et al., Ar. Biochem. Biophys. 232, 422426, 1984).

A large number of naturally occurring proteins are modified after translation, one of the most common modifications being glycosylation. Glycoproteins are membrane bound or soluble or are found in both intracellular and extracellular matrix. The function of glycosylation is still under review. For the time being, it is only clear that glycans can protect proteins from proteolytic degradation and are also responsible in many cases for cell-to-cell metabolism. Furthermore, they have an apparent effect on protein binding and contribute to the stability and structure of the molecule. Also, the solubility of proteins depends on the carbohydrate chains.

It is necessary to distinguish between N- and O-glycosylated proteins. The N-glycans can bind exclusively to the ASN triplet -ASN-X-SER / THR-, wherein X can be any amino acid except PRO or GLU. However, this requirement for protein structure is only one of many requirements since not all potential glycosylation sites will be occupied by carbohydrate. For the O-glycans, no precisely defined structure is available. However, O-glycans appear to be predominantly synthesized in regions rich in PRO--, SER-, and THR-. The steric availability of these structures evidently creates sites that are prerequisites for O-guycosylation. The effect of gylcosylation on the pharmacokinetics and immunological properties of a protein cannot be excluded either. Short reports have already been reported (Gibbon et al., Lancet, 335, 434-437, 1990) that 4 out of 16 patients treated with recombinant human GM-CSF (a granulocyte-macrophage colony-stimulating factor) produced in yeast have formed against this protein antibody. These antibodies have been shown to react with epitopes that are free in the endogenous GM-CSF protected by O-glycosylation, but are recombinantly available in the recombinant protein.

To date, no carbohydrate moieties have been detected in IFN-alpha2, nor has O-glycosylated IFNalpha been isolated. Various compositions containing IFN-alpha and recombinant IFN-alpha2 are used as medicines for viral and cancer diseases. Since IFN-alpha2 is produced in E. coli and cannot therefore be glycosylated, it appears that the carbohydrate moiety is of no importance for in vivo biological activity. Recently, however, there have been reports of patients who, after long-term treatment with recombinant, in E. coli produced IFN-alpha2, have developed antibodies against this substance (e.g. Figlin et al., Semin. Haemoatol. 1988, 25, 9-15).

It is therefore an object of the present invention to provide IFN-alpha2 of a new type.

SUMMARY OF THE INVENTION

SUMMARY OF THE INVENTION The present invention provides a process for the production of interferon alpha2 by introducing a DNA strand encoding said compound into a separate expression vector, which is then transfected into a multicellular cell. upon cultivation of such modified cells, unexpectedly IFN-alpha2 proteins are obtained which clearly differ in molecular weight from the known recombinant IFN-alpha2.

Cultures of multicellular organisms, particularly vertebrate or insect cell cultures, are suitable for the production of the novel interferons of the invention. Examples are vertebrate lines, for example VERO cells, HeLa cells, CHO cells, WI38 cells, BHK cells, COS-7 cells, MDCK cells or mouse myeloma cells. Vectors for expression in these cells include, where appropriate, a replication site, a promoter and, optionally, an RNA splicing site, a polyadenlation site, and a transcription termination chain. The control functions of such expression vectors usually originate from viral material. Useful promoters can be found, for example, in pclyoma virus, Acovovirus 2, SV40 virus, and preferably from cytomegalovirus CKV. The necessary replication site can be created directly in the construction of the vector, for example, a site for replication from SV4C virus, pclyoma virus, acenovirus, VSV or PBV, or the chromosomal replication mechanism of the producer cell. In particular, the latter can be used to integrate the vector into the chromosome of the producer cell.

Preferably, two expression vectors, constructed in part from the plasmid, are used in the method of the invention. These expression vectors contain, according to the invention, a multiple cloning site for the storage of heterologous DNA chains and can be propagated preferably in Ξ. coli in multiple copies using ampicillin resistance. In order to allow expression of heterologous genes in ssvc cells, the plasmids for expression according to the invention preferably contain a promoter / enhancer from cytomegalovirus (M. 3oshart et al., Cell 41,

1985, 512-530). In order to allow autonomous replication of plasmids for expression of the invention to a greater number of copies and thereby a high level of transient expression in the respective cell lines, e.g. COS-7 or adenovirus transformed cell line 293 (ATCC CRL 1573), the origin of replication from SV40 is used. . To obtain permanent transformed cell lines and to subsequently amplify the methotrexate expression cassette, the modified hamster minigene (prcmotor with coding region and first intron) for dihydrofolate reductase (DHFR) is used as a selectable label. In order to facilitate the retrieval of single-stranded plasmid DNA for the superinification of transformed yeast phage bacteria, for example R408 or M13K07, to facilitate the analysis of the strand and the mutagenesis of the plasmid DNA, the plasmids according to the invention preferably comprise an insertion region.

Μ13. If, in accordance with another preferred embodiment, the T7 promoter is placed in front of the multiple cloning site, the generation of RNA transcripts is allowed in vitro.

In particular, plasmids for the expression of DAD-CMV13, pAD-CMV15 and in particular pAD-CMV19 are preferred according to the invention. The preparation of these plasmids is given in Example 1.

In order to achieve improved expression and facilitate directed cloning of the cDNA coding for IFN-alpha2, the cDNA coding for 0FN-alpha2 can be modified by PCR in the 5-non-coding region by substituting the 5'-non-coding strand for this region. human globin beta mRNA regions (Lawn et al., Cell 21, 1980, 647-651). Simultaneously, restriction enzyme sites are created at both ends of the cDNA to facilitate controlled cloning. This change results in an unexpectedly substantial increase in expression.

The modified IFN-alpha2 cDNA is then inserted into the corresponding restriction enzymes of the digested plasmids for expression, preferably into the plasmids pAD-CMV-19. The resulting plasmid for expression of IFN-alpha2 is then transfected with suitable mammalian cells and cultured in a suitable and nutrient medium. The culture supernatant is then purified in a very gentle manner in a known manner. Preferably, purification is performed by affinity chromatography using monoclonal antibodies against IFNalpha2. Preferred monoclonal antibodies are EBI-1 or EBI-10 or equivalents thereof. The preparation of these highly specific antibodies has already been described (Acolf G.R., J. Gen. Virol. 68, 1669-1676, 1987, Adolf et al., J. Cell Physiol. Suppl.

2, 61-68 (1982). The methods used have also been described (Secher and Búrke, Nature 285, 446-450, 1980, Adolf et al., J.

Biol. Chem. 265, 9290-9295, 1990, Adolf et al., Biochem. J., 276, 511-518, 1991. Particularly preferred is the method of purification according to EP 203 382, in which there is no need to disrupt cells. Reverse phase HPLC was used to characterize recombinant IFNalpha2 produced in mammalian cells. The N-terminus and C-terminus were also analyzed. To compare tulle always use recombinant, v v. ccli obtained by IFN-alpha2. Based on the electrophoretic monitoring of the SDS-gel, it can be demonstrated that recombinant IFN-alpha2 derived from mammalian cells has a higher molecular weight than the same substance obtained from Ξ. coli. After treatment of both recombinant interferons with sodium hydroxide, the molecular weight of the interferons produced in mammalian cells was reduced to an IFN-alpha2 of Ξ. coli. Thus, it is apparent that recombinant IFN-alpha2 expressed in mammalian cells must be glycosylated. Upon identification of the glycopeptide by chain analysis and peptide mapping, it was shown that a glycosyl group is attached to the threonine residue at position 105 (TKE-106). Comparing these results with those obtained using wild-type IFN-alpha2 from leukocyte stimulated with viruses, it was shown that both the glycosylation site and the oligosaccharide portion of the molecule are identical.

The invention also relates to a purification process which does not contain any steps in which the substituents on IFN-alpha2 could be altered or cleaved. For this purification, highly specific monoclonal antibodies are used, taking care to ensure that the pH of the alkaline region is not greater than 8.0 throughout the procedure.

Natural human IFN-alpha2 was isolated. from the leukocyte interferon using a highly specific monoclonal antibody. After two successive purification steps using affinity chromatography, the purity of the resulting protein was greater than 95%. In the chain analysis, the expected N-terminal chain was shown, with CYS as the first amino acid only indirectly.

So far, three variants of IrN-alphaZ are known which differ in amino acid residues at positions 23 and 34 and tc:

IFN-alpha2a, containing ² -LYS and ^o4 KIS (formerly referred to as

Le IFA (Goeddel et al., NAture 290, 20-26, 1981), infra

IFN-alpha2b containing ²³ ARG and ³⁴ HIS (Streuli et al., Science,

209, 1343-1347, 1980) and IFN-alpha2c, comprising ²³ ARG and ³⁴ ARG (formerly referred to as IFN-alpha2Arg, Dworkin-Rastl et al., J.

Interferon REs. 2, 575-585, 1982, 3odo and Maurer-Fogy, The Biology of The Interferon System 1985, Stewart II, W.E. and Schellehus, H. Hrsg., 59-64, 1986.). In the isolated interferon, only ARG could be detected at position 23, eliminating the presence of IFN-alpha2a. The amino acid at position 34 was unambiguous histidine, so that the isolated interferon was an IFN-alpha2b variant. However, it is also possible to obtain the other two variants, depending on the cellular material used as the starting material. In addition to IFN-alpha2b, IFN-alpha2c is known to be present in Namalwa cells. In the case of interferon from E. coli, which was used as. the comparative compound was an IFN-alpha2c variant.

Reverse phase HPLC analysis of naturally purified IFN-alpha2 showed that two peaks were observed which eluted earlier in the column than recombinant E. coli IFNalpha2c. Strong homogeneity of natural IFN-alpha2 was also observed using SDS-PAGE. All proteins contained in native IFN-alpha2 had significantly higher apparent molecular weight than recombinant E. coli IFN-alpha2c. All of the previously established types of IFN-alpha, with the exception of IFN-alpha1, contained no site for N-glycosylation, i. !

-ASN-X-THR / SER-. Thus, N-glycosylation can also be avoided in IFN-alpha2. Similar structural conditions in the molecule are not known for O-glycans. Therefore, it cannot be excluded that the isolated IFN-alpha2 is glycosylated.

Since the O-glycans can be cleaved from the protein under mild alkaline conditions, the respective fractions have been processed under mild alkaline conditions. In both cases, this reaction leads to a decrease in the molecular weight to the molecular weight of recombinant IFN-alpha2c from E. coli, which is an indication of O-glycosylation of the molecule.

Experiments with neuraminidase and O-glycanase for peak 2 of FIG. 14 also resulted in a decrease in apparent molecular weight to the weight of IFN-alpha2c from E. coli, which is also an indication of O-glycosylation. The results of this gradual degradation of glycan by the action of neuraminidase and O-glycanase show that peak 2 heterogeneity was conditioned by different N-acetylneuraminic acid content (NeuAc = sialic acid). Introduce the three belts (Fig. 20, spot 4)! a disialylated, monosialylated and non-sialylated form of native IFN-alpha2 (21,000, 20,000 and 19,000 molecular weight, respectively). The lightest form of IFN-alpha2 could only be degraded by reaction with O-glycanase. Since O-glycanase cleaves only the unsubstituted Gal (phi-3) disaccharide GalNAc, the reaction demonstrates that, besides both sialylated forms, an asialoform exists in IFN-alpha2.

The apparent molecular weight of the peak 1 material could be reduced enzymatically. Incubation with neuraminidase did not, as expected, lead to a decrease in apparent molecular weight. That is, the disaccharide nucleus must be substituted other than by NeuAc, so that the substituent cannot be cleaved by O-glycanase.

Comparison of the peptide map after cleavage of wild-type and recombinant IFN-alpha2 expressed in trypsin in E. coli was able to identify the glycopeptides from peaks 1 and 2. The chains of these glycopeptides showed that the residue ΤΗΗθθΤΗΗ is a glycosylation site.

In addition to enzymatic reactions, the structure of the natural IFN-alpha2 oligosaccharides provided information on glycopeptide monitoring using a mass spectrometer. Interpretation of the mass spectra together with the results obtained using SDSPAGE showed that natural IFN-alpha2 contains at least four different glycan structures: at the top 2 neutral Gal (fil-3) GalNAc disaccharide, whose structure is most likely due to the high specificity of O-glycanase , in addition, a mono- and disialylated variant; at the top 1, a neutral oligosaccharide consisting of two hexose and two N-acetylexosamine units. As a structure of these tetrasaccharides, Gal- (Gal-GlcNAc-) GalNAc is also contemplated in analogy with the frequently described O-glycans already described.

In accordance with the invention, it has been unexpected for the first time to prepare O-glycosylated IFN-alpha2 in a highly pure form. This interferon is O-glycosylated at the amino acid residue of threonine in flesh 106 (^ ⁰⁶ THR). The oligosaccharides obtainable at this position are neutral disaccharides

Gal (B1-3) GalNAc, its mono- and di-dialylated variant and the neutral tetrasaccharide Gal- (Gal-GlcNAc-) GalNAc.

This O-glycosylated IFN-alpha2 can be processed in a manner similar to recombinant IFN-alpha2, which has been formulated in E. coli into pharmaceutical compositions, which can then be used in all known indications for the use of the known form of IFN-alpha2.

The compounds of the present invention can be used to treat viral infections and cancer in the form of pharmaceutical compositions containing an effective amount of IFN, optionally in admixture with an inorganic or organic, solid or liquid, pharmaceutically acceptable carrier.

Pharmaceutical compositions for parenteral, for example, intramuscular, subcutaneous or intravenous administration in humans are preferred. Isotonic aqueous solutions or suspensions,

The protein-containing proteins of the invention optionally together with a carrier and optionally with excipients such as stabilizers, emulsifiers, co-solvents, salts for controlling pH and osmotic pressure, preservatives and / or crosslinking agents. These pharmaceutical compositions can be prepared in a manner known per se, for example by mixing the proteins of the invention, a pharmaceutical carrier and excipients, optionally lyophilizing and reconstituting them before use.

The dosage of the pharmaceutical compositions depends on the disease to be treated, the patient's weight, age and body condition, the final decision will always depend on the attending physician, who will also choose the route of administration:

According to the invention, for the first time, a pharmaceutical composition comprising O-glycosylated L-interferon-aifa2 suitable for the antiviral and antineoplastic properties of IFN-alpha2 for the treatment of viral and tumor diseases can be prepared.

The invention will be further elucidated with reference to the accompanying drawings.

Description of the drawing

Fig. 1: Construction of plasmid pCMV + SV40

Fig. 2: Construction of plasmid pAD-CMVIOA

Fig. 3: Construction of plasmid pAD-CMV15

Fig. 4: Construction of plasmids pAD-CMV13 and pAD-CMV19

Fig. 5: Construction of a plasmid for expression of pAD19B-IFN

Fig. 6: HindIII / XbaI-plasmid insert for expression of pAD19B-IFN

Fig. 7; DNA strand of plasmid pAD-CMV19

Fig. 8: Construction of plasmid pCMV-SV40

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

9. Construction of plasmids pSV2gptDKFRMut2

10; Construction of plasmids pAD-CMV1 and pAD-CMV2

11: DNA strand of plasmid pAD-CMV1

12: Affinity chromatography of interferon from human leukocyte using monoclonal antibodies

13: Human IFN-alpha ELISA:

(0) Reference material containing recombinant human IFN-alpha2c (O) Leukocyte interferon (starting material) (D) eluate fraction A (ô) eluate fraction B

14: Reverse phase HPLC (b) natural IFN-alpha2 (a) IFN-alpha2c from E. coli

15: Amino acid chain of IFN-alpha2C

15: SDS-PAGE of wild-type IFN-alpha2 before and after reaction with neuraminidase and O-glycanase.

(1) peak 1, untreated (2) peak 1, after reaction with saeuraminidase (3) peak 1, after reaction with neuraminidase and O-glucanase (4) peak 2, without processing (5) peak 2, after reaction with neuraminidase (6) peak 2, after reaction with neuraminidase and O-glucanase (7) IFN-alpha2C from E. coli

17: SDS-PAGE of wild-type IFN-alpha2 (peak 2 of Figure 14b) (1) before O-glycanase reaction (2) after O-glycanase reaction

18: SDS-PAGE of wild-type IFN-alpha2 (top 1 and 2) and E. col IFN-alpha2c after incubation with 0.1 M NaOH.

(1) E.coli-IFN-alpha2 (3) peak 1 (5) peak 2:

(2) comparator! peak samples 1 without processing (4) comparative peak samples 2 without processing

Fig. 19: Comparative peptide map E: coli-IFN-alpha2C and wild-type IFN-alpha2.

(1) the top 1 of FIG. 14b (2) of peak 2 of FIG. 14b h, these peaks are derived from non-glycosylated peptides whose retention time was the same.

Fig. 20: SDS-PAGZ of wild type IFN.alpha2 and IFN-alpha2C from E. coli (1) molecular weight labeling (2) E-coli-IFN-alpha2C (3) wild type IFN-alpha2, peak 1 of FIG. 14b (4) natural IFN-alpha2, peak 2 of FIG. 14b Color: blue coomassie

Fig. 21: Reverse phase HPLC for (a) CHO-IFN-alpha2c (b) E. coli ± FN-alpha2c

Fig. 22: Comparative peptide maps (a) CHO-IFN-alpha2c, peak 1 (b) CHO-IFN-alpha2c, peak 2 (c) E. coli-IFN-alpha2c

Fig. 23: SDS-gel electrophoresis (SDS-PAGE) for CHO-IFN-alpha2 Stains 1 and 8: molecular weight labeling (1000 each) Stains 2 to 4: non-reducing conditions

Stains 5 to 7: reducing conditions Stains 2a 5: peak 1 of CH0-IFN-alpha2c Stains 3 and δ: peak 2 of CH0-IFN-alpha2c Stains 4 and 7: E. coli-IFN-alpha2c Upper gel: all IFN- 4 µg stains Bottom gel: all IFN-stains 1 µg staining: coomassie blue

Fig. 24: SDS-gel electrophoresis (SDS-PAGE) for CHO-IFNalpha2c and for E. coli-IFN-alpha2C before and after incubation with 0.1 M NaOH.

Stains 1 and 8: molecular weight labeling (1000 unit scale)

Stains 2, 4, 6: unprocessed samples

Streaks 3, 5, 7: samples after incubation with 0.1 M NaOH

Spots 2.3: E. coli-IFN-alpha2c

Spots 4.5: peak 1 of CHO-IFN-alpha2c Spots 6.7: peak 2 of CH0-IFN-alpha2C

All IFN materials were loaded at 1.5 µg under non-reducing conditions.

Color: blue coomassie

The following examples are intended to illustrate the invention.

DETAILED DESCRIPTION OF THE INVENTION

Example 1

Plasmid construction for expression of pAD-CMV13, pAD-CMV15 and PAD-CMV19

From parts of the plasmid for expression (pCDM8, Seed and Aruffo, Proc. Natl. Acad. Sci. USA 84, 1987, 8573-8577, B. Seed, Nature 329, 1987, 840-842), Invitrogen Inc., San Diego, CA, pSV2gtpDHFR2O, EP 321842 and pBluescript KS- (Short et al., Nucleic Acids Res. 11, 1988, 5521-5540, Stratagene, La Jolla, CA) were constructed new plasmids containing a multiple cloning site to store a heterologous DNA chain and which can be resisted against ampicillin in E. coli to produce a large number of copies. The intergenesis region of M13 allows the construction of a single stranded DNA plasmid after superinfection of transformed bacteria with yeast phage (e.g., R408 or M13K07), thereby facilitating the analysis of the strand DNA and the mutagenesis of the plasmid DNA. The T7 promoter upstream of the multiple cloning site allows the generation of an RNA transcript in vitro; In mammalian cells, heterologous genes are expressed under the assistance of a cytomegalovirus (CMV) transcriptional promoter / enhancer according to M. Boshart et al., Cell 41, 1985, 521-530. The origin of replication from SV40 allows autonomous replication of the plasmid for expression to a high copy number and hence a high level of transient expression in the respective cell lines (e.g. SV40 transformed cells such as COS-7) in the adenovirus transformed cell line 293 (ATCC CRL1573). The transformed minigenic hamster (coding region promoter and first intron) for dihydrofolate reductase (DHFR) serves as a selectable label for permanent transformed cell lines and subsequent amplification of the methotrexate expression cassette.

Polymerase reaction (PCR) acquisition of vector and promoter moieties

The plasmid pBluescript KS- was linearized with HindIII and 100 ng of DNA was used in 100 µl PCR mix (Saiki et al., Science 239, 1988, 487-491). The reaction medium used is 50 mM KCl, 10 mM tris-HCl, pH 8.3, below

1.5 mM MgCl _2> 0.01% gelatin, 0.2 mM each of the four desoxynucleoside triphosphates (dATP, dGTP, dCTP and dTTP) and 2.5 units of Taa.-polymerase per 100 µl. 50 pmol of synthetic oligonucleotides EBI-1785 (5 * -GGAATTCAGCCTGAA-TGGCGAATGGG-3 ') and oligonucleotide EBI2134 (5'-CACTGAACTCGAGCAGC-TGCGTTGCTGGCGTTTTTCC-3') were used as primers.

After 5 minutes of denaturation at 94 ° C, more than 10 PCR cycles are performed (single cycle conditions: 40 seconds at 94 ° C, 45 seconds at 55 ° C, 5 minutes at 72 ° C using the instrument

Perkin Elmer Cetus Therma Cycler). Oligonucleotides delimit!

the M13 intergenic region; and the origin of replication (ori) with the beta-lactamase gene. At the same time, a single XhoI and PvuII enzyme treatment site is created at the end of the ori and an EcoRI site at the other end. The reaction mixture was then deproteinized with phenol and chloroform and the DNA precipitated with ethanol. The resulting DNA was digested with XhoI and EcoRI, and a 2.3 kb fragment was isolated after agarose gel electrophoresis. Under identical conditions. with 50 ng of SacII linearized pCDM8 amplified by PRC using EBI-2133 (5'-GGTCACTGTCGACAT-TGATTATTGACTAG-3 ') and EBI-1734 (5'-GGAATTCCCT-AGGAATACAGAGGG-3') oligonucleotide. Oligonucleotides bind to the origin of the promoter / enhancer chain and contain a SalI cleavage site (EBI-2133) or bind at the end of the SV40 polyadenylation site (EBI-1734). The PCR product was digested with SalI and EcoRI and the 1.8 kb DNA fragment was isolated from an agarose gel.

Both digested PCR products were ligated with T4 DNA ligase and used to transform E-coli HB101. The plasmid of the desired structure (FIG. 1) was designated pCMV + M13.

The SV40 origin of replication (SV40 ori) is isolated from plasmid pSV2gptDHFR2O (EP 321842). For this purpose, said plasmid is digested with HindIII and PvuII and the ends of the DNA are filled with a large fragment of E. coli DNA polymerase (Klenow enzyme) in the presence of all four desoxynucleoside triphosphates. The resulting 0.36 kb DNA fragment was isolated from an agarose gel and ligated to the EcoRI linearized pCMV + M13 vector. The plasmid obtained after transformation of E. coli HB101 and containing the SV40 ori in the same orientation as the beta-lactamase gene and the CMV promoter, was designated pCMV + SV40 (FIG. 1).

Plasmid pCMV + SV40 was digested with EcoRI and BamHI, and the DNA ends were filled with Klenow enzyme. Then she was:

i i

i í

i i

F

DNA purified by phenol / chloroform extraction and ethanol precipitation. The DNA fraction was circulated with T4-DNA-Igase and the plasmid obtained after E.coli transformation was designated pAD-CMV10 (Fig. 2). The remainder of the plasmid pCMV + SV40 DNA was dephosphorylated by incubation with alkaline phosphatase, and the 4.4 kb vector was isolated from an agarose gel.

Plasmid pSV2gptDHFR-Mut2 (Example 4, Fig. 9) containing a modified hamster minigene for dihydrofolate reductase (DHFR), from which the restriction enzyme sites EcoRI, PstI, BglII, BamHI and Kppen were deleted by targeted mutagenesis. and PstI and the DNA ends were filled by incubating for 20 minutes at 11 ° C with 5 units of T4-DNA polymerase (Reaction medium: 50 mM tris-HCl pH8.0, 5 mM MgCl 2 → 5 mM dithiothreitol, 0, 1 mM each of the four desoxynucleotide triphosphates, 50 µg / ml bovine serum albumin). The 2.4 kb DNA fragment containing the mutated DHFR gene was isolated from an agarose gel and bound to the above described pCMV + SV40. The plasmid obtained after transformation of E. coli and containing the DHFR gene in the same orientation as the CMV promoter was designated pAD-CMVIOA (FIG. 2).

Using a pAD-CMV1 expression plasmid (Example 4, FIG. 10) containing between the multiple cloning site and the intron chain polyadenylation signal, a number of variants were constructed, differing in the number and placement of introns relative to the multiple cloning site ( Fig. 4).

In pAD-CMV13 (Fig. 4), the SV40 t-antigen was removed between the multiple cloning site and the polyadenylation site. Plasmid pAD-CMV15 (FIG. 3) contains a synthetic intron between the CMV promoter and the multiple cloning site and the t-antigen intron SV40 between the multiple cloning site and the polyadenylation signal. Plasmid pAD-CMV19 (FIG. 4) contains only one intron between the CMV promoter and the multiple cloning site.

Starting from 100 ng of HindIII-linearized plasmid pAD-CMV1 and 50 pMol of oligonucleotide EBI2625 (5 '-CACTGATCTAGAGATATCTTGTTTATTGCAGCTTATAATGG-3 *) and EBI-1857, (5'-GGCAAGGGGCUI 100) Performing the PCR as described above and performing a 10 PCR cycle (40 seconds at 94 ° C, 45 seconds at 55 ° C, 90 seconds at 72 ° C) amplify the 1.26 kb DNA fragment. E3I-2525 binds shortly before the SV40 polyadenylation signal (position 1280 in plasmid pAD-CMV1) and additionally contains a restriction enzyme digestion site XbaI and EcoRV. EBI-1857 binds to complementary DNA strand on the first intron of the following DHFR minigene (position 2525 in plasmid pAD-CMV1). The PCR product was isolated by extraction of the phenol / chloroform mixture and then the DNA was precipitated with ethanol. The DNA was then digested twice with XbaI and BglII, and a 0.32 kb fragment was isolated from the agarose gel and ligated to the pAD-CMV1 vector (5.8 kb) digested with the same enzymes. The plasmid obtained by transformation of E. ccli H3101 (FIG. 4) was then designated pAD-CMV13.

The splice-modified donor chain following the CMV prcmotor (M. Boshart et al., Cell 41, 1985, 521-530) was then joined by SOE-PCR (Splicing by Overlapp Extension, SN Ho et al., Gene 77, 1989, 51). -59) with a spliced acceptor site in the first intron of the human beta-globin gene (LAwn et al., Cell 21, 1980, 647-651), followed by a site for multiple cloning from plasmid pAD-CMV1. In addition, 100 ng of plasmid pGJ7 was amplified (G. Jahn et al., J. Virology 49, 1984, 363-370) containing a promoter and transcriptional enhancer from human cytomegalovirus strain AD169 (Boshart et al., Cell 41, 1935, 521). -530) using 50 pmoles each of the above EBI-2133 oligonucleotide and EBI-2586 oligonucleotide (5 '-GCAGAGAGAGTCAGTGCCTATCAGAAACCCAAGAGTCTTCTCTATAGGCGGTACTT-ACCTGACTCTTG-3') in 100 µl of PCR reaction for 40 sec. ° C, 45 seconds at 45 ° C and 90 seconds at 72 ° C). The last 24 bases of EBI-2586 correspond to the CMV chain (in reverse orientation) and the previous beta-globin intron base, 18 being exactly in reverse order complementary to the EBI2585 oligonucleotide chain, and the overlapping strands generate SOE-PCR. The PCR products were resolved on an agarose gel and a 0.8 kb DNA fragment was isolated (FIG. 3). 100 ng of plasmid pAD-CMV1 was amplified in the same way using EBI-2585 oligonucleotides (5'-GCACTGACTCTCTCTGCCTATTGGTCTATTTTC'CCACCCTTAGGCT-GCTGGTGCTTAACTGGCTTATCG-3 *) and EBI-2112 (5 -GTTGGGTGGGTGGGGGGGGGGGGGGGGGGGGGGGGGGGTGGGTGGGTGGTGGTGGGTGG) size 0,2 kb. Oligonucleotide EBI2585 contains the last 45 bases of beta-globin (its intron) and the five following bases, as well as 17 bases at the 3 'terminus, which are precisely capable of hybridizing to positions 511-627 of plasmid pAD-CMV1. The EBI-2112 oligonucleotide is based on a complementary DNA strand at positions 743 to 760 in the pAD-CMV1 plasmid chain. One-tenth of the isolated 0.8 kb DNA fragment and 1/30 of the 0.2 kb DNA fragment were mixed in an additional 100 µl PCR medium (SOE-PCR) and amplified using 50 µM each of EBI 2133 and EBI oligonucleotides each. 2112 at 30 PCR cycles (40 seconds at 94 ° C, 45 seconds at 45 ° C and 2 minutes at 72 ° C). The reaction was terminated by extraction of the phenol / chloroform mixture and the DNA was precipitated with ethanol. The 5-PCR product was phosphorylated by treatment with T4-polynucleotide kinase (reaction buffer: 70 mM tris-HCl pH7.6, 10 mM MgCl2 5 mM dithiothreitol, 1 mM ATP) and finally digested with XbaI. The DNA was subjected to agarose gel separation and a 0.98 kb fragment was isolated. Plasmid pAD-CMV10 was digested twice with Pvull and XbaI, and a portion of the CMV promoter-free vector was isolated from the agarose gel. This plasmid vector was ligated to a 0.97 kb DNA fragment with a CMV promoter and an intron and multiple cloning enhancer and the whole was used to transform E. coli HB101. Plasma DNA and this new DNA containing oligonucleotides are obtained from the transformants obtained

ΕΒΙ-2112, ΕΒΙ-2586 and ΕΒΟ 1773 (5'-GGTCGACATTGATTATTGACTAG-3 ') were analyzed by a chain dideoxy cleavage (F. sanger et al.

Proc. Natl. Acad. Sci. USA 74, 1977, 5463-5467) using modified T4 DNA polymerase (S. Tabor and C.C. Richardson, Proc. Natl.

Acad. Sci. USA 84, 4767-4771; Sequenase, United Sates Biochemical Corp.). The expected chain plasmid (Fig. 3) was designated pAD-CMV15.

Plasmid pAD-CMVIOA was digested twice with SpeI and BglII and the portion of the CMV promoter-free vector was isolated from the agarose gel. Plasmid pAD-AMV15 was digested twice with SpeI and HindIII, and a 0.8 kb fragment containing the CMV promoter and the synthetic intron was isolated. Plasmid pAD-CMV13 was digested twice with HindIII and BglII, and a 0.36 kb fragment containing a multiple cloning site, a polyadenylation signal from the SV40 initial strand portion, and a portion of the DHFR hamster promoter region was isolated. These three DNA fragments were ligated with T4 DNA ligase and used to transform E. coli HB101. Plasmid DNA was obtained from the resulting transformants and analyzed by digestion with various restriction enzymes. The plasmid with the desired structure was then designated pAD-CMV19 (FIGS. 4 and 5).

Example 2

Preparation of modified cDNA for huIFN-alpha2c,

1F7 cDNA encoding human IFN-alpha2c (E.Dworkin- ·

Rasti et al., J. Interferon Res. 2, 1982, 575-585, E-Dworkin-

Rasti et al., Gene 21, 1983, 237-248) has been modified by PCR in the 5 ' -incode region so that the 5-non-coding region of the human beta-globin mRNA has been cut in place of the 5-non-coding region (Lawn et al. 1980, 647-651). As a result of this change in the 5'-non-coding region, there is a substantial increase in expression, presumably due to more efficient translation initiation.

At the same time, we are introducing sites of action at both ends of the cDNA

I

Ϊ

I (

which results in subsequent controlled cloning of cDNAs in plasmids for expression.

100 ng of plasmid 1F7 linearized with EcoRI was amplified using 50 pmol of oligonucleotide EBI2747 (5'-CTTCAGAAGCTTACATTTGCTTCTGACACAACTGTGTTCACTAGCAACCT-CAAACAGACAGGATGGCCTTGACCTTTCTCTTTTC-3TTCTCTTTTC-3TTTCTCTTTTC-3TTTCTC 20 cycles (40 seconds at 94 ° C, 45 seconds at 55 ° C and 90 seconds at 72 ° C). EBI2747 contains, after the HindIII cleavage site, a 5 * non-coding region of human beta-globin mRNA followed by the first 22 lanes, which encode the human IFNalpha2c signaling peptide (huIFN-alpha2c), the start codon is underlined. EBI-2744 binds to a complementary strand at the end of the fetus which encodes huIFN-alpha2c (the stopcodone is underlined) and contains an XbaI cleavage site. The reaction is stopped by extraction of the phenol / chloroform mixture and the DNA is precipitated with ethynol. The PCR product was digested with HindIII and XbaI and a 0m54 kb fragment was isolated from the agarose gel (FIGS. 5 and 7. Plasmid pAD-CMV19 was also digested twice with HindIII and XbaI and then ligated to the cDNA fragment. obtained after transformation of E. coli HB101 are grown for plasmid DNA preparation One of the plasmids obtained was completely analyzed for the detection of the HindIII-XbaI deposited region except for a single swap base (CTG for TTG) in the 8th codon of the signal peptide (but this did not The plasmid for expression of secreted and O-glycosylated human IFN-alpha2c was designated as pAD19B-IFN (Fig. 6).

Description of the part of the pAD-CMV19 plasmid chain (Fig. 5)

Bases up to 21 binding site for oligonucleotide ΕΞΙ 2133

Base up to 590 Transcription enhancer and cytomegalovirus promoter 722 to 740 cytomegalovirus intron chain (spliced donor)

strings) 741-805 Human beta-globin intron chain (spliced acceptor) 836-853 T7 promoter. 862-922 a site for multiple cloning 923-1055 polyadenylation sites SV40 1056-1953 the promoter and the 5 'non-coding region of the hamster DHFR gene 1954-2039 DHFR exon 1 2040-2333 DHFR intron 1 2151-2168 EBI-1857 binding site 2344-2821- -DHFR exons 2 to 6 of the coding region 2822-3474 DHFR s'-non-coding region 3475-3812 SV40 origin of replication (SV40 ori) 3813-6055 plasmid pBluescript 3813-4291 Intergenic region M13 (M13 ori) 4423-5283 beta-lactamase, coding region 6038-6062 the EBI-2134 binding site Example 3

Transient expression of huIFNalpha2c in higher eukaryotic cells

cell g Approximately 10 cells (293, human embryonic cells

the kidney, transformed with part of the adenovirus AD5 genome; F. L. Graham et al., J. Gen. Virol. 36, 1977, 59-77, ATCC (CRL1573) per 80 mm petri dish 24 hours before

-22 incubates MEM / F12 broth in DUlbecco's medium

1: 1 with 15 mM Hepes, Gibco with 10% heat inactivated fetal. calf serum at 37 ° C in an atmosphere of 5% carbon dioxide. Cells were transferred to fresh medium for 3 hours prior to transfection and further incubated at 37 ° C. 10 µg of plasmid DNA (purified by two centrifugation using CsCl gradient) of plasmid pAD19B-IFN is dissolved in 0.5 ml of 250 mM CaCl ₂ and then added dropwise to 0.5 ml of 2x HBS (16.36 g / L).

NaCl, 11.9 g / 1 Hepes, 0.40 g / 1 Na _2? 0 ₄ pH 7.12). The precipitate formed is added to a Petri dish and the cells are incubated at 37 ° C for 4 hours. The cells were washed with PBS, 15% glycerol in 1 x HBS (shock) was added for 30 seconds, then the cells were washed once more with PBS and incubated in 10 ml fresh medium with 10% calf serum at 37 ° C. After 72 hours, the supernatant is collected and the secreted IFN is determined.

Example 4

Construction of plasmid pAD-CMVL and pAD-cmv2 for expression

From the parts of the plasmid for expression of pCDM8 (Seed and Aruffo,

Proc. Natl. Acad. Sci. USA 84, 1987, 8573-8577, Seed, Nature 329, 1987, 840-842, Invitrogen Inc., San Diego, CA), plasmid pSV2gptDHFR20 (EP A1 0321 842) and Bluescript SK + (Short et al., Nucleic Acids Res) ., 11, 5521-5540, Stratagene, La Jolla, CA) was constructed a new plasmid containing, instead of the multiple cloning for controlled storage heterologous to s _# i niho string of DNA plasmid that can be replicated in E. coli's resistance against ampicillin with a high copy number. The M13 intergenic region allows single stranded plasmid DNA to be prepared by superinfecting transformed bacteria with helper phage (e.g., R408 or M13K07) to facilitate analysis of the plasmid DNA sequence and utagenesis. The T7 promoter, located upstream of the multiple cloning site, allows trans-i preparation

RNA in vitro. In mammalian cells, heterologous genes are expressed using the cytomegalovirus (CMV) promoter and enhancer (Boshart et al., Cell 41, 1985, 521-530). The origin of SV40 replication allows, in suitable cell lines (e.g., SV40 transformed cells such as COS-7, adenovirus-transformed cell line 293 (ATCC CRL1573)) to increase autonomous replication of plasmids for expression into high copy numbers and thereby achieve high levels of transient expression. The modified hamster minigene (promoter with the coding region and the first intron) for dihydrofolate reductase (DHFR) is used as a label for the next selection, followed by amplification of the methotrexate proexpression cassette.

a) Preparation of part of the vector and promoter by PCR

Plasmid Bluescript SK + is linearized by HindIII and 5 ng of DNA is used in 100 µl PCR mix (Reaction Buffer: 50 mM KCl, 10 mM Trs-Cl pH 8.3, 1.5 mM MgCl ₂ , 0.01% gelatin, 0.2 mM all four desoxynucleotide triphosphate (aATP, dGTP, dCTP, and dTTP), 2.5 units Taq-polymerase per 100 µl). As primer, 50 pmoles of the synthetic oligonucleotide EBI-1786 (5'-GGAATTCAGCCTGAATGGCGAATGGG-3 '), which binds the immediate outer M13-ori-region in the Bluescript plasmids, position 475, independently of the M13 orientation, and E3I-1729 (5) are used. '-CCTCCGAGCGTTGCTGGCGTTTTTCTC-3'), which binds to Bluescript at position 1195 before the ori and corresponds to the origin of the Bluescript chain in pCDM8, 6 and 5 is the XhoI site). After denaturation for 5 minutes at 94 ° C, PCR is performed for 20 cycles (40 sec at 94 ° C, 45 sec at 55 ° C and 5 minutes at 72 ° C, Perkin Elmers Cetus Thermal Cycler). Oligonucleotides flank the intergenic region of M13, i. origin of replication (ori) with the intermediate beta-lactamase gene. At the same time, at the end of the origin of replication, an XhoI site occurs and an EcoRI site at the other end. The reaction mixture is deprived of proteins by extraction of a mixture of phenol and chloroform a

24 DNA is precipitated with ethanol. The DNA obtained was digested with XhoI and EcoRI, and a 2.3 kb fragment was isolated after agarose gel electrophoresis.

ng of SacII-linearized plasmid pCMD8 was amplified using EBI-1733 (5 * -GGTCGACATTGATTATTGACTAG-3 * oligonucleotide), ligated to the CMV promoter region at position 1542 of plasmid pCDM8, corresponding to position 1 in pAD-CMV, cloning site for cloning ) and oligonucleotide EBI-I734 (S'-GGAATTCCCTAGGAATACAGCGG-3 ', binds to the polyoma origin of the 3'SV40 polyA region at position 3590 of plasmid pCDM8) using PCR under identical conditions as described above for plasmid Bluescript SK +. Oligonucleotides bind to the origin of the CMV promoter / transcription enhancer and contain a SalI site (EBI 1733) or bind at the end of the SV40 polyadenylation site and contain an EcoRI site (EBI-1734). The PCR product was digested with SalI and EcoRI, and a 1.8 kb fragment was isolated on an agarose gel.

Both PCR products were ligated with T4 DNA ligase and the obtained product was used to transform E. coli HB1O1 and plsamide DNA was amplified and isolated by standard procedures. The plasmid obtained was designated pCMV-M13 (Fig. 8).

The SV40 origin of replication (oriSV40) was isolated from plasmid pSV2gptDHFR2O (EP-A1 0321842). To this end, the plasmid was digested with HindIII and PvuII and the DNA ends filled with a large fragment of E. coli DNA polymerase (Klenow enzyme) in the presence of all four desoxynucleotide triphosphates. The obtained 0.36 kb DNA fragment was isolated from an agarose gel and was ligated with EcoRI linearized plasmid pCMV-M13. The plasmid obtained after transformation

E. coli HB101 and containing SV40 ori in the same orientation as the beta-lactamase gene and the CMV promoter was designated pCMV-SV40. The construction of this plasmid is shown in FIG. 8th

b) Mutagenesis of the DHFR gene

For the preparation of plasmids for expression with a multiple cloning site, two sites were removed from the DHFR minigene by mutagenesis and the deletion of the three restriction enzyme sites was deleted. BglII of plasmids pUC219 (iBI) and plasmid pUCDHFR was obtained. Plasmid pUCDHFR transformed E. coli JM109 cells (Stratagene) was infected with an approximately 40-fold excess of helper phage R408 (stratagene) and shaken at 37 ° C in LB for 16 hours. Single-stranded plasmid DNA was isolated from the bacterial supernatant.

Controlled mutagenesis was performed in two consecutive steps using the RPN1523 in vitro mutagenesis system (Amersham). The EcoRI site at the beginning of exon 2 was disrupted by one base change in GAATTC to form GAGTTC. This exchange of bases does not lead to any change in the encoded amino acid sequence and, moreover, corresponds to the nucleotide sequence in the natural mouse DHFR gene (McGrogan et al., J. Biol. Chem.

260, 1985, 2307-2314, Mitchell et al., Mol. Cell. Biol. 6.

1986, 425-440). The oligonucleotide EBI-1751 (5'-GTACTTGAACTOGTTCCTG-3 ') was used (upstream) for mutagenesis.

The plasmid with the desired mutation was prepared as a single stranded DNA as described above, and the PstI enzyme in the first intron was removed by mutagenesis using oligonucleotide EBI-1857 (5'-GGCAAGGGCAGCAGCCGG-3 ') and upstream. , the change was made from CTGCAG to CTGCTG. The mutations were verified by chain analysis and the plasmid obtained was designated pUCDHFR-Mut2. The 1.7 kb BglII frsgment was isolated from the plasmids pUCDHFRMut and ligated to the plasmid pSV2gptDHFR2O, digested with BglII and BamHI.

After transformation of E. coli, aplication ^? and recovering the DNA by:

a plasmid with the desired properties, which was designated t i

PSV2gprDHRF-Mut2. After BamHI digestion, a 0.12 kb DNA fragment containing a KpnI digestion site was removed from the 3 * non-DHFR gene region downstream of the BglII site. By combining with BglII and BamHI the resulting DNA ends, the sites of action for both enzymes are also removed.

Plasmid pCMV-SV4Q was digested with EcoRI and BamHI and the DNA ends were filled with Klenow enzyme. The DNA was then purified by extraction of the phenol / chloroform mixture and ethanol precipitation. Finally, the DNA was dephosphorylated by incubation with alkaline phosphatase and the DNA of the 4.4 kb vector was isolated from an agarose gel.

Plasmid pSV2gptDHFR-Mut2 (Fig. 9) was digested with EcoRI and PstI, and the DNA ends were filled with a 20 minute incubation at 11 ° C with 5 units of T4 DNA polymerase (50 mM tris-HCl pH 8.0, 5 mM MgC). (5 mM dithiothreitol, 0.1 mM each of the four desoxynucleotide triphosphates, 50 µg / ml bovine serum albumin). The 2.4 kb fragment containing the mutated DHFR gene was isolated from an agarose gel and ligated to the above-prepared plasmid pCMV-SV40. The plasmid obtained by transformation of E. coli containing the DHFR gene in the same orientation as the CMV promoter was designated pCMV-SV40DHFR. In the final step, the 0.4 kb fragment following the CMV promoter, which originated from the starting plasmid pCDM8, was exchanged for site t

for multiple cloning. For this purpose, plasmid pCMV-;

The SV40DHFR was digested with HindIII and XbaI, and the vector portion was:

isolated from an agarose gel. Multiple cloning site;

formed from both oligonucleotides, EBI-1823 (5'-AGCTTCTGCAGGTCGACATCGATGGATCCGGTACCTCGAGCGGCCGCGAATTCT-3 '), (5'-CTAGAGAATTCGCGGCCGCTCGAGGTACCGGATCCACCGATGTCA-3').

EBI-1829, including a clone compatible for cloning at the HindIII-XbaI site as well as PstI, SalI, ClaI sites

BamHI, KpnI, XhoI, Noti and EcoRI. !

t »

í f

- 27 • ... rsntmMíTw.r.,

1 µg of both oligonucleotides was incubated in 20 µl reaction buffer (70 mM Tris-Cl pH 7.6, 10 mM MgCl 2, 5 mM dithiothreitol, 0.1 mM ATP) with 5 units of T4-polynucleotide kinase for 1 hour at at 37 ° C to phosphorylate the 5-termini. The reaction was stopped by heating to 70 ° C for 10 minutes and the complementary oligonucleotides hybridized to each other such that the sample was incubated for an additional 10 minutes at 56 ° C and then slowly cooled to room temperature. 4 µl of hybridized oligonucleotides (100 ng) were ligated to 100 ng of plasmid vector and used to transform E. coli HB101. The plasmid that could be linearized with the enzymes of the multiple cloning site (except Nôti) was designated pAD-CMV1.

Of the large number of clones examined, no one could be identified that could be cleaved by Nôti. In the chain analysis, it has always been possible to detect the deletion of some of the bases in the chain in which the enzyme Nôti is located. In the same manner, using the oligonucleotide pair EBI-1820 (5 -AGCTCTAGAGAATTCGCGGCCGCTCGAGGTACCGGATCCATCGATGTCGACCTGCAGAAGCTTG-3 ') and EBI-1821 (5'- CTAGCAAGCTTCTGCAGGTCGACATCGATGGATCCGGTACCTCGAGCGGCCGCGAATTCTCTAG-3) Construction of expression plasmid pAD-CMV2, comprising restriction sites in the multicloning site in the reverse respectively. This gave a plasmid pAD-CMV2 which can be linearized with all restriction enzymes including Nôti.

The nucleotide sequence of the plasmid pAD-CMV1 of velikosti size

6414 bp (FIG. 10) is discussed in more detail in FIG. 11.>

l

I

The individual sections of this plasmid (based on the numbering of the individual bases) correspond to the following sequence of chains:

i>

and

I i

Baze:

1-21 EBI-1733, origin of transcription enhancer and promoter from CDM8 632-649 the T7 promoter 658-713 site for multiple cloning (HindIII to XbaI from EBI-1823, EBI-1829) 714-1412 Intron SV40 and polyadenylation site of CDM8 1413-2310 S'-non-coding region and hamster DHFR gene promoter (from plasmid pSV2gptDHFR2O) 2311-2396 DHFR hamsters: exon 1 2516 A-T: mutation-disrupted PstI site in DHFR intron 1 2701-3178 DHFR exons 2 to 6 (coding region) 2707 A - G: mutation of the EcoRI site 3272-3273 deletions between BglII and BamHI sites in the 3'-non-DHFR region 3831 end of DHFR gene (from plasmid pSV2gptDHFR20) 3832-4169 SV40 ori (from pSV2gptDHFR2O) 4170-4648 M13 ori (from pBluescript Sk +) 4780-5640 beta-lactamase (coding region) 6395-6414 EBI-1729, the end of the pBluescript vector string Preparation of plasmids pAD-CMV1 and pAD-CMV2 is shown

FIG. 10

Example 5

Development of recombinant CHO (Chinese hamster ovary) cell lines producing glycosylated human interferon-alpha2

a) Transfection of CHO cells and selection of stable transformed cell lines

Parental cell lines, CHO-DXB11 and CHO-DG44 (Proc. Natl. Acad. Sci. USA 77, 4216-4220, 1980, Som. Cell.

Mol. Genet. 12, 555-666, 1986) were grown in RPMI 1640 (Roswell Park Memorial Institute) supplemented with 10% fetal calf serum, hypoxanthine (10 yumol), thymidine (16 / umol), penicillin G sodium (100 units / ml) ) and streptomycin (50 units / ml). Two days before transfection, the cells were transferred to 25 cm flasks, at the time of transfection, the cells formed an almost continuous layer.

Transfection was performed as follows:

microliters of pAD19B-IFN plasmid solution (1 µg / ml) was diluted with 125 µl of calcium chloride and 855 µl of sterile deionized water. These solutions were added dropwise to 1 ml of 2 x HSB (1 x HSB contains the following components in 1 liter of solution:

8.18 g sodium chloride, 5.94 g Hepes, 0.2 g sodium hydrogen phosphate, pH 7.1). The CHO cell culture medium was removed and 0.25 ml of suspension was added to each flask. The cultures were then incubated for 4 hours at 37 ° C. Then she was sus-!

the cells were released from the surface of the bottle by a solution of trypsin and EDTA and suspended in a selection medium (the selection medium consisted of minimal basal medium, ribonucleotide-free alpha-modification and dsoxyribonucleotides and supplemented with 10% dialyzed fetal calf serum, sodium penicillin G - 100 units / ml, streptomy-;

i i

* í

T

I

I

I

With tin 50 units / ml and amphotericin B 2.5 µg / ml, 40 ml per bottle was used). The cell suspension was then transferred to the wells of two cell culture microtiter plates (96 wells per plate, 0.2 ml per well) and the cells were incubated for two weeks at 37 ° C. The selection medium was used, but without amphotericin B, in all other experiments.

Cell cultures were examined for cell growth visually. The digestion culture medium in which growth was observed was examined for IFN-alpha2 content by an immunoassay using two monoclonal antibodies against IFN-alpha2 (Biochem J. 276, 511-518, 1991). one week again with fresh supernatant. cells from positive cultures were detached from the surface by trypsin and EDTA solution and transferred to 24-well cell culture plates. Cultures with good cell growth were then repeatedly tested for IFN production. The concentrations of IFN-alpha2 in the supernatants were typically in the range between 2000 and more than 10,000 units / ml (1 ng of IFN.alpha2 protein corresponds to 230 Units).

b) Amplification of the IFN-alpha2 gene by selection using methotrexate

Clones following transfection of both parent cell lines, CH0-DXB11 and CH0-DG44, which in a greater number of assays showed high concentrations of IFN, were selected for the next amplification. In addition, 3 to 5 additional clones were pooled. These cultures were then maintained in 25 cm flasks in selective culture medium (without amphotericin B) to which methotrexate was added at a concentration of 20 nM or 50 nM. Fresh culture was changed once a week for each culture. Surviving clones were observed for two to three weeks. Once the cells covered approximately 50% of the growth area in the flask, the supernatants in the individual flasks were again assayed for IFN-alpha2 content. The cells were then detached from the plate, diluted and transferred to new flasks. Now, the concentration of methotrexate has been increased e to 5-fold, for example from 20 nM to 50 and 100 nM or from 50 nM to 100 and 200 nM. After a greater number of selection cycles as described above in the presence of increasing amounts of methotrexate while selecting resistant cultures still secreting IFN, ultimately methotrexate resistant cell lines up to a concentration of 5000 nM could be obtained, these lines simultaneously excluding a relatively large amount of IFN-alpha2.

The following Table I summarizes the productivity gains of the CHO-DXB11-IFN-alpha2c-3 / 2D4 cell line.

The results obtained using another cell line, CHO-DG44IFN-alpha2c-poolS, are then summarized in Table II below.

Concentration of IFN-alpha2c methotrexate in culture supernatant

-u. o c

Ό

ABOUT

Ό

about 0 about about about about ABOUT about about about 0 about about about about about about about about about about about about about about about about 't σ » m about about about about about about rh WHAT OJ OJ D about about m about 1 I rJ | 1 fH 1 00 ABOUT | 00 1 σι 1 1 about 1 about 1 0 1 ABOUT 1 about 1 about 1 ABOUT 1 about 1 ABOUT about about 0 about about about about about about about about about about about about about about about about OJ σι 10 about about about about about OJ ABOUT •with d) about H T-H rh 00 OJ CM

o o 0 about ABOUT ABOUT about ABOUT ABOUT OJ m about about about about about ABOUT R-1 OJ LO about about about OJ d)

J t

t t

t i

I

d.

jj

-P

I

WHAT l

C about about about about about about ABOUT about ra about about about about about about about about -P about about about about about about about about ra from-\ C rh WHAT OJ what about about about about about at ε what rh σ> OJ OJ 0- if) OJ M <u rh OJ OJ lf) WHAT WHAT H Q. >> 3 44 1 1 1 1 1 1 1 1 W -P ABOUT about about about ABOUT about about about about > C about about about ABOUT about about about about Ό about about about about about about about about ABOUT υ WHAT pole • '-i <3 C3 <3 rh about about about about ra OJ m ω θ ' <3 0- about C3 • - rh rh OJ rh

--t CC

I, 2

U.

jO H + J

CC

X o

d.

After

x:

JJ

d)

CC

CP c

(U o

what

.x

z-a about ABOUT ABOUT about ABOUT ABOUT about ABOUT ε OJ lf) ABOUT about ABOUT about about ABOUT C rh OJ lf) about about about '- rh pole lf)

IFN-alpha2 was purified from recombinant CHO cell supernatants by affinity chromatography using monoclonal antibodies (e.g., EBI-1 or E3I-10) in a known manner (Nature 285, 446, 1980, J. Biol. Chem. 265, 9290-9295 1990, Biochem. J. 276: 511-518 (1991).

The process described in EP 203 382 is particularly suitable for the purification of proteins according to the invention.

Example 6

Characterization of recombinant, glycosylated human IFN-alpha2c from Chinese hamster ovary cells -CHO

(a) Reverse phase HPLC (RP-HPLC)

Recombinant glycosylated IFN-alpha2c from CHO cells, purified by affinity chromatography, was compared with recombinant IFN-alpha2c from E. coli which was not glycosylated by RP-HPLC. The analytical method was described by Adolf et al., J. Biol. Chem. 265, 9290-9295, 1990. The glycosylated CHO-IFN-alpha2C (top of Figure 21) is formed by two major peaks (1 and 2) and two smaller peaks for IFN (3 and 4). The non-glycosylated E. coli IFNalpha2c (lower part of FIG. 21), on the other hand, consists of a single major peak (disulfide musts) and one smaller peak which is likely to be a secret disulfide must form. From the comparison of the retention time, it is evident that both major peaks of CHO-IFN-alpha2c elute somewhat earlier than the major peak of E. coli-IFN-alpha2c. The reason for this reduced hydrophobicity is the presence of oligosaccharide in CHO-IFN-alpha2c.

The two smaller peaks in CHO-IFN-alpha2c have approximately the same retention time as the major peak of E. coli-IFN-alpha2c and are most likely composed of a smaller non-glycosylated fraction of CHO-IFN-alpha2C.

b) Analysis of the N-terminal part of the chain

Both major peaks for CHO-IFN-alpha2c were isolated from RP-KPLC and their chains analyzed. The assay conditions were described in Adolf et al., J. Biol. Chem. 265, 9290-9295, 1990. The first 15 amino acids were identified in accordance with the appropriate cDNA strand. No heterogeneity was observed at the N-terminus.

c) Analysis of the C-terminal part of the chain

The major peak of E. coli-IFN-alpha2c and the two major peaks of CHO-IFN-alpha2C were also isolated by RP-HPLC and digested with trypsin. The resulting peptides were separated by RP-HPLC. The procedure was carried out under the conditions of Adolf et al., Biochem. J. 276, 511-518 (1991). 22 shows a comparison of the peptides obtained. There is only one difference between the peptides from peak laz of peak 2 of CHO-IFN-alpha2c (upper and middle): peptide 18 from peak 1 after tryptic cleavage is almost not present in peak 2 (referred to as peptide 15). Instead, a new peptide (# 19) is observed in peak 2, which is absent at all in peak 1 or in E. coli-IFN-alpha2c (lower part of Fig. 22).

Peptides 12 (from E. coli-IFN-alpha2C), 18 (from peak 1 CHO-IFN-alpha2C), 15 and 19 (from peak 2 CHO-IFN-alpha2C) were analyzed using plasma desorption mass spectrometry (PD- MS). Details of this procedure are given in Adolf et al., Biochem. J. 276, 511-518 (1991). The molecular weight corresponding to amino acids 150 to 162 of the IFN-alpha2a chain was determined for the first three of these samples. Peptide 19 from peak 2 of CHO-IFN-alpha2c, on the other hand, had a lower molecular weight corresponding to the amino acid chain of 150-161. The chain analysis of a portion of peptide 19 was also performed, clearly demonstrating that the peptide begins with amino acid 150 and ends with LEU-161. From these results, it can be concluded that peak 1 for CHO-IFN-alpha2c contains the complete IFN molecule, whereas peak 2 lacks 4 C-terminal amino acids 162-165. Amino acid residues 163-165 are not positive after trypsin digestion. that the resulting dipeptide (163/164) and free amino acid 165 elute with the total RP-column volume. A small proportion of peptide 15 (a truncated tryptic peptide with amino acids 150 to 162), which also occurs in peak 2 of CHO-IFN-alpha2c, is likely to be present due to peak 1 contamination since peaks 1 and 2 are not possible with RP -HPLC completely separate from each other.

In further experiments, it has been shown that C-terminal truncation of O-glycosylated IFN-alpha2 from CHO cells can be prevented by using trypsin-free solution (e.g., disodium disodium) instead of the usual trypsin-containing EDTA solution to release the cells. sul EDTA,

200 mg / l with D (+) - glucosomonohydrate 200 mg / l in sodium chloride solution with phosphate buffer pH 7.4). IFN-alpha2 purified from cultured cell cultures as described above has only peak 1 corresponding to full protein but not peak 2 corresponding to truncated protein chain in reverse phase HPLC (analogous to Fig. 21a). In addition, it was possible to demonstrate, following analysis of the peptide after tryptic cleavage (analogous to FIG. 22), that the protein composition obtained without trypsin was identical to that of the protein from which trypsin digestion yielded the peptide from peak 1 RP-HPLC as is shown in FIG. 22a.

SDS-electrophoresis

Peaks 1 and 2 of CHO-IFN-alpha2c were isolated from RP-HPLC. Then, the materials from these peaks were analyzed individually and compared to E. coli-IFN-alpha2c, both under reducing conditions (after boiling with dithiothreitol) and under non-reducing conditions using SDS-alecophoresis. For details on these procedures, see Adolf et al., J. Biol. Ckemi. 265, 9290-9295, 1990. The results of these tests are shown in FIG. 23 (spots 2 to 4 were obtained under non-reducing conditions, spots 5 to 7 under reducing conditions. The upper part in each spot was obtained when 4 µg IFN was applied, the lower part when 1 µg IFN was applied. at the bottom, it is clear that both peak 1 and peak 2 of CHO-IFN-alpha2c have a higher molecular weight than non-glycosylated E. coli-IFN-alpha2c, due to incomplete separation of peaks 1 and 2 using RP-HPLC. contamination of peaks 1 and 2. For the same reason, peak 2 is also contaminated with a small amount of unglycosylated CHO-IFN-alpha2c, which originates from peak 3) as shown in FIG. 21). In view of the above-mentioned contamination, the major peak bands 1 and 2 for CHO-IFN-alpha2c may be considered homogeneous. Since peaks 1 and 2 differ in their C-terminal chains (as described above), it can be concluded from the results obtained using SDS gel electrophoresis that the oligosaccharide portions of peaks 1 and 2 for CHO-IFN -alpha2c are identical (as will be discussed further in paragraph f below).

e) Deglycosylation of CHO-IFN-alpha2c

Peaks 1 and 2 for CHO-IFN-alpha2c were isolated from RP-HPLC and dried in appropriate equipment (SpeedVac). These samples as well as E. coli-IFN-alpha2c samples were incubated in 10 µl of 0.1 M NaOH for a total of 20 hours at room temperature. These samples, which were deglycosylated by beta-elimination, were analyzed using SDS-gel electrophoresis as compared to untreated samples. The results shown in FIG. 24 shows that the molecular weight of the peaks 1 and 2 material for CHO-IFN-alpha2c decreases substantially after NaOH treatment and is then identical to the molecular weight of E. coli-IFN-alpha2c after NaOH treatment. The diffuse appearance of the band for all NaOH-treated samples is aptly due to changes occurring in the peptide chain under the reaction conditions used.

f) Identification of glycopeptides

A comparison of the peptide resulting from trypsin digestion of E. coli-IFnalph2c (FIG. 22, bottom) and the peptide from the top (complete C-terminal chain) for CHO-IFN-alpha2c (FIG. 22, top) shows that the two peptides have different retention times. Peptides 18 and 21 from E. coli-IFN-alpha2c, which contain amino acid residues 84-112 and 71 to 112, do not occur in the peak 1 material of CHO-IFN-alpha2c. Instead, there are two novel peptides (Nos. 26 and 31) which absorb in a similar manner at 280 and 214 nm as peptides 18 and 21 from E. coli-IFN-alpha2c.

It can be concluded that peptides 26 and 31 represent a glycosylated verse of amino acids 84 to 112 and 71-112.

Probably also for this reason, said peptides from the RP column elute substantially earlier than the analogous peptides from E. coli-IFN-alpha2c. For both possible peptide lengths after tryptic cleavage (amino acid residues 84 - ++ 2 and 71-112), one peak can be observed (peptide 26 and 31), so that it can be concluded that the oligosaccharide moiety is substantially homogeneous.

Comparing the peptides obtained from peaks 1 and 2 of CHOIFN-alpha2c (upper and middle portions of Fig. 22), it is evident that glycopeptides 26 and 31 from peak 1 and glycopeptides 24 and 30 from peak are identical. This also implies that the four missing amino acid residues at the C-teminal end of peak 2 constitute the only difference between the material of peak 1 and 2.

All major peptides in all three amino acid chain 84- ++ 2 and 71-112 samples were isolated from RP-HPLC and further digested with protease from Staphylococcus aureus V8 at the C-terminal end of glutamic acid. Reaction conditions are described in Adolf et al., Biochem. J., 276, 511-518, 1991. The results were compared, all different peaks were isolated and analyzed by N-terminal termination analysis and / or mass spectrometry.

One of the peptides found at peaks 1 and 2 of CHO-IFN-alpha2c, but not in E. coli-IFN-alpha2c after protease digestion with Staphylococcus aureus, contained amino acids 97-112 of the IFN-alpha2c chain. THR-10O was not detectable in this peptide by N-terminal analysis. It can be concluded that THR-106 is glycosylated in this peptide. In the Edman chain analysis method, the glycosylated amino acids are derivatized and cleaved the same as the non-glycosylated amino acids, but the glycosylated amino acids cannot be extracted from the reaction mixture by butyl chloride due to their increased hydrophilicity. This means that the amino acid cannot be detected at this stage, but the chain continues unhindered. Further evidence that the oligosaccharide is bound to THR-106 can be seen in that GLU-107 / Thr-108 binding is a protease from s. aureus split only partially. Obviously, the accessibility of this peptide bond is limited in the presence of an oligosaccharide. In an analogous peptide from E. coli-IFN-alpha2c, this peptide bond is almost completely cleaved.

Another peptide that can only be found in CHO-IFN-alpha2c after digestion with S. aureus protease was analyzed by PD mass spectrometry. The molecular weight obtained corresponded to the amino acid sequence 97-112, including the oligosaccharide, which consists of one molecule of N-acetyl40 galactosamine and galactose and two molecules of N-acetylneuraminic acid.

From these results it is apparent that both the glycosylation site and the oligosaccharide moiety in CHO-IFN-alpha2c are identical to those obtained for wild-type IFN-alpha2 from leukocyte stimulated virus.

Isolation of O-glycosylated interferon from virus-stimulated cells proceeds as follows:

techniques

Biological test for interferon

The antiviral effect of the IFN-containing preparation was demonstrated on microtiter plates by an assay that determines the cyto-pathic effect (CPE) of the encephalomyocarditis virus (EMCV). Human lung cancer cells were used as test cells. For details of the assay, see Adolf G.R. Virol. 68, 1669-1676, 1987. In each bioassay, all titrations were performed twice. Human IFN-slfs2c produced by E. coli was also assayed for each assay. The efficacy of these preparations was determined by comparison with international reference means for human IFNalpha2, Gxa 01-901-535. All observed IFN potencies were corrected for the defined effect of said reference preparations.

ELISA for interferon

An ELISA assay was performed using your neutralized murine IgG MAb anti-IFN antibodies and one IFN-alpha2c reference preparation (listed above) as standards. The production of antibodies and their properties have been reported in the publication

Adolf et al., J. Cell Physiol. Suppl. 2, 51-66 (1982) and Adolf G.R. Virol. 68, 1669-1576, 1987. Antibody E3I-1 was used to overlay the assay plates. The Ξ3Ι10 antibody, covalently bound to horseradish peroxidase, was added along with the test sample. O-phenylenediamine and sodium perborate were used as substrates for the enzyme. The reaction was terminated by the addition of sulfuric acid and then the absorption of the resulting product was measured (492 nm, reference 690 nm).

Purification of natural human IFN-alpha2

An affinity chromatography column was prepared by binding 12 mg of a monoclonal antibody such as MAb EBI10 (purified from mouse ascites by ammonium sulfate precipitation and protein G affinity chromatography according to standard procedures) to 1 g Sepharose 43, activated by CNBr according to the manufacturer's instructions (Pharmacia). The fill volume in the column was approximately 3 ml. Partially purified interferon from human leukocytes (Cantell et al., Methods Enzymol. 78, 29-38, 1981, Cantell et al., Ibid., 499-505), from which the proportion of IFN-omega was removed (Adolf et al., J. Biol. Chem., 265, 9290-9295, 1990) and containing approximately 2 to 3 x 10θ IU / ml at a total protein concentration of 2 mg / ml was applied to the column at a flow rate of 1 ml / min. (200 and 350 ml). Then the column was washed with 0.1 M buffer (sodium phosphate) pH 7.5 - Buffer A and then eluted using a linear gradient of Buffer A and Buffer B (0.1 M sodium citrate at pH 2.1) in the system. FPLC '(Pharmacia) at a flow rate of 1 ml / min. The fractions obtained were examined for IFn activity by ELISA. The appropriate fractions were pooled, neutralized with 1 M NaOH, and re-plated on the same column equilibrated in buffer A. The same elution program was then used at a flow rate of 0.25 ml / min. The corresponding fractions were again pooled, neutralized, and then individual portions of the material were lyophilized.

SDS-gel electrophoresis, HPLC technique and amino acid chain analysis

SDS-polyacrylamide gel electrophoresis and RP-HPLC were used to analyze purified IFN-alpha2. These procedures have already been described in detail (Adolf et al., J. Biol. Chem. 265, 9290-5295, 1990. The N-terminal chain determination was performed on an automated chain nanalyzer (Applied Biosystems, Models 477A), amino acid derivatives being analyzed by RP-HPLC system on-line (Adolf et al.,

J. Biol. Chem. 265, 9290-9295 (1990).

Mapping of peptides obtained by proteolysis

Affinity chromatography purified IFN-alpha2 was further purified by RP-HPLC, denatured and de-salted (Adolf et al.

J. Biol. Chem. 265, 9290-9295, 1990. The main fractions were separated and dried on a SpeedVac. 29 µg (peak 1) and 66 µg (peak 2) protein were dissolved in 0.1 ml of 1% sodium bicarbonate solution. Then 0.5 µg or 1 µg trypsin (Boehringer Mannheim) in 3 or 6 µl of 0.01% trifluoroacetic acid was added and the reaction mixture was incubated at 37 ° C. After 6 hours of incubation, the same amount of trypsin was added again and the mixture was incubated for an additional 18 hours. The reaction mixture was then reduced for analysis by adding 10 µl of 0.5 M dithiothreitol and 100 µl of 6 M urea for 2 hours at room temperature. RP-HPLC was performed on a Delta Pak C18 column (Waters, 3.9 x 150 mm, particle size 5 µm, pore diameter 1 x 10 cm) at 30 ° C using the following solvents: Solvent A: 0.1% trifluoroacetic acid in water, Solvent B: 0.1% trifluoroacetic acid in acetonitrile. The following gradient (flow rate 1 ml / min) was used: 0-55 minutes: 0-55% B (linear gradient), 55-70 minutes: 50% B. Peptide detection was performed by their absorption at 214 and 280 nm. . The resulting samples were compared43 with those obtained from Ξ. coli IFN-alpha2c. Peptides obtained from natural IFN-alpha2 that behaved differently than the corresponding peptides from the recombinant material during elution were separated and analyzed for their N-terminal termination, or the peptides were further degraded by a protease from Staphylococcus aureus V8 ( Endopeptidase Glu C, Boehringer, Mannheim 0.88 µg (peak 1 / Ϊ), 2.6 µg (peak 2 / Ia) and 1.5 µg (peak 2 / Ib) of the peptide were always dissolved at 0 ° C. 1 ml of 25 mM phosphate buffer pH 7.8. A protease solution in water (17.5 ng, 52.5 ng and 29 ng) was then added and the reaction mixture was incubated at 37 ° C. After 6 hours, the same amount of protease was added and the mixture was incubated for 18 hours. The samples were then analyzed as described above using RP HPLC. The corresponding fractions were isolated and N-terminal termination analysis was performed.

Deglycosylation of IFN-alpha2

Purified, denatured and salt-free IFN-alpha2 was treated with Vibrio cholerae neuraminidase (Boehringer Mannheim) (50 mU / ml, 18 hours at 37 ° C in 20 µl of 50 mM sodium acetate at pH 5.5, 4 mM CaCl ₂ ) and / or by treatment with endo-alpha-N-acetylgalactosaminidase, i. O-glycanase (Boehringer Mannheim) (100 mM / ml, 18 hours at 37 ° C in the same buffer). Chemical elimination was achieved by incubation with 0.1 M NaOH for 20 hours at room temperature.

PD-mass spectrum

The trypsin digest mass spectrum of the peptides was measured on a BIO-ION 20 time-of-flight mass spectrometer (BI0-INN Nordic AB, Uppsala, Sweden). The samples were dissolved in 0.1% trifluoroacetic acid and loaded onto nitrocellulose-coated target materials (BIO-ION). Spectral accumulation times ranged from 0.5 to 12 hours depending on yield. The spectrum for individual samples was checked at a voltage of 17 kV.

Example 7

Purification of natural human IFN-alpha2

Human leukocyte interferon, obtained from human peripheral blood leukocytes after inoculation with Sendai virus, and partially purified by the method of Cantell et al., Methods Enzymol. 78, 29-38 and 78, 499-512 (1981) was used as a starting material for the isolation and purification of IFN-alpha2. Selective affinity chromatography using an anti IFN-omega monoclonal antibody, for example OMG-4, OMG-5 or OMG-7, has removed a proportion of IFN-omega (Adolf et al., Virology 175, 410-417, 1990, EP No. 262,571) ).

e

The specific antiviral activity was 1 to 2 x 10 IU / mg, which is

Q means that IFN-alpha with a specific potency of 2 x 10 IU / mg accounted for only about 1% of the total amount of protein. Highly selective anti-IFN-alpha2 monoclonal antibodies were used to purify IFN-alpha2 from contaminating foreign proteins as well as other IFN-alpha types. These antibodies are highly specific to IFN-alpha2 in a standard neutralization bioassay (Adolf G.R., J. Gen. Virol. 68, 1669-1676, 1987).

An affinity chromatography column was prepared in which such monoclonal antibodies, such as the EBI-10 antibody, were obtained, for example, according to J. Gen. et al. Virol. 68, 1669-1676, 1987 or according to DE 33 06 060.6 bound to Sepharose 4B, activated by CNBr. The antibody was purified from mouse ascites by ammonium sulfate precipitation followed by protein B affinity chromatography according to standard procedures. For example, 12 mg of monoclonal antibody EBI-10 bound to 1 g of CNBr activated Sepharose 4B is used, following the manufacturer's instructions. The column fill volume is approximately 3 ml.

Leukocyte interferon formulations were loaded onto the column. Approximately 20% of antiviral activity was bound. Elution was performed using a linear buffer gradient of 0.1 M sodium phosphate pH 7.5 and 0.1 M sodium citrate pH 2.1. Two peaks for protein were detected in the eluate (Fig. 12): Fraction A and Fraction B. Both fractions were analyzed for IFN-alpha content using a two-tailed ELISA using both EBI-10 and EBI-1. Both antibodies had high specificity for IFN-alpha2 (Adolf et al., J. Cell. Physiol. Suppl. 2, 61-68, 1982). Recorabinant IFN-alpha2c was used as a standard. Fractions eluted at low ph (Fig. 12, peak A) and sample fractions yielded titration curves that ran parallel to the titration curve of recombinant IFN-alpha2c. Curves with different steepness were obtained from peak 1 and its fraction (peak B). These materials could not be analyzed by quantitative ELISA (Fig. 13), but the materials were tested for their biological activity (Table III).

The lower pH required for peak A elution, as well as ELISA results, demonstrate that IFN-alpha2 constitutes the major portion of peak A. To ensure that the entire amount of immunoreactive IFN-alpha was bound by the antibody, the material was resuspended column and a second elution as described above. The obtained material contained less than 10% IFN activity bound at the first pass.

Fractions A and 3 were separated, neutralized and chromatographically purified on the same affinity chromatography column. In both cases, more than 95% of IFN activity was bound. Elution was performed using the same gradient as above. The starting product and fractions were examined in Coomassie Blue for protein content and bioassay for IFN content. The results are summarized in Table III below.

cu> N ku

CO o

x e (U x

W (U o c

C P • H H> O

about CM and ABOUT about r4 't Ό cn about T

O v

d)

CM or r *

CM r-l n o τ> Cl If)

Purification of natural IFN-alpha2

o ί

iH>

jj ε o

-p O. M

WHAT

C • H r4> o o oo

-h ε

Ή

P ε

<u • '' S

P

ABOUT

WHAT CM what ω » « * * CM CM σ> CM Π

o in d>

OCOC ^ CO ^ if if) <—I — 4 m

"what

C <u u

<0

WITH

CM Ci. I-I,

I

Q.

'Cfl o < and < C -i o u a 4> + j 44 44 and> □ * <Ϋ́ 'CC 'WHAT 'Cfl 44 o □ 3 □ 3 (Ú —i γ · 4 r-) rh ε w about <U <u CD - □ • • night what tn tn WHAT 3 o 3 3 3 3 —1 44 · r -1 S «3 in dc in > U > >> >> About Q. and about about about • ABOUT . . . . f-i a R? R-4 CM CM

ABOUT 4-3 w > » about > C0 0 3 0 3 in (U N ^ 4 £ p ABOUT s about >> -. ABOUT • r! w ABOUT) P 0 • r4 R '4 rh γ · 4 1 ABOUT WHAT • H N oo P cu 2 ε Lf) U. 0 M 1 2 '>> Ci. WHAT C 1-1 44 KU 0 44 Ή C «n C Ό • H KU ABOUT > the C P WHAT > cu with- R4 C 44 C > the n Ό cu Ό ABOUT 44 ABOUT > C what 4J 'Cfl 0 > and Q. r ~ '* CM

Example 8

Identification of protein purified by affinity methods such as

IFN- ~ 2

The affinity method of IFN-alpha purification was first analyzed for reverse phase KPLC and purification. Peak A contained two further incompletely separated peaks 1 and 2 at a weight ratio of approximately 1: 2 (Fig. 14 below). Peak 1 represents the more hydrophilic fraction of the protein. Both peak fractions were separated, rechromatographed and N-terminal chain analysis was performed. The following chain was obtained from both fractions (the Cys residues indicated in parentheses were not identified but were derived from conserved IFN chains):

( ¹ CYS) -ASP-LEU-PRO- ⁵ GLN-TKR-HIS-SER-LEU- ¹ ° GLY-SER-ARG-ARGthr- ¹⁵ leu-met-leu-leu-ala- ²⁰ gln-met-arg- By comparison with the published chains, both chains were identified as IFN-alpha2 chains. ²³ arg-ile- ²⁵ serLEU-PHE-SER- (CYS) - ³⁰ LEUS.

In both fractions of peak 1 and 2, the amino acid at position 23 was uniquely identified as arginine. Thus, a variant referred to as LeIFA which contains lysine at position 23 (Goeddel et al., Nature 290, 20-26, 1981) was not present in a detectable amount in the leukocyte preparation used. The amino acid at position 34 was identified as histidine. IFN-alpha2 belonged to the IFN-alpha2b variant.

Specific antiviral activity of annual IFN-alpha2 based on IFN-alpha2 international reference preparation,

Gxa01-901-535 based on the determination of protein content in the sample also by absorption at 214 nm according to Adolf et al.

and

h .... »

48 Virology 175, 410-417, 1990 was determined to be 1.5 x 8

IU / mg (averaged from five independent bioassays).

When comparing the retention time of native IFN-alpha2 for KPLC in reverse phase with the same value for recombinant E. coli-IFN-alpha2, it can be shown that the elution of the recombinant protein occurs much later, as shown in FIG. 14. The increased hydrophilicity of the natural protein as well as its heterogeneity must therefore be related in some way to post-translational modifications.

When peak 3 is subjected to reverse phase HPLC, a mixture of five incomplete separated peaks is obtained. The analysis of the chain materials from these peaks showed that although these peaks contained IFN-alpha material, none of them showed IFN-alpha2.

IFN-alpha2, which was purified by HPLC, was further analyzed using SDS-PAGE after reduction with dithiothreitol as shown in FIG. 20. An apparent molecular weight of 17,500 (molecular weight derived from the amino acid chain: 19,287) could be detected for the recombinant E. coli IFN-alpha2c recombinant conditions under selected conditions. Analysis of fraction 1 obtained by HPLC yielded a single wide band (apparent molecular weight of 20,000), whereas fraction 2 of the peak 2 yielded two major components (20,000 and 19,000 molecular weight, respectively) and one minor component of 21,000 molecular weight. These differences in molecular weight compared to recombinant E. coli protein, heterogeneity in size of individual molecules, as well as increased hydrophilicity, indicate that native IFN-alpha2 is glycosylated. Since there is no site for N-glycosylation in the structure of the molecule: it must be O-glycosylation.

and

Example 9

Reaction of natural IFN-alpha2 with endo- and exoglycosidases

The following experiments were performed with both peaks after separation by RP-HPLC (peaks 1 and 2 in Fig. 14b). Both samples were incubated with neuraminidase and then with O-glycanase. After each enzymatic reaction, a portion of the material was analyzed by SDS-PAGE.

As shown in FIG. 16, the peak 1 material did not react with either neuraminidase or O-glycanase. The apparent molecular weight remained constant at 20,000. In contrast, the three bands of peak 2 reacted with both neuraminidase and O-glycanase. Neuraminidase reactions resulted in a decrease in apparent molecular weight in both larger bands (21,000 and 20,000) to 19,000. However, traces of material with a molecular weight of 20,000 remained present. Subsequent incubations with O-glycanase lead to a further reduction of the apparent molecular weight from 19,000 to 17,500 (i.e. to the molecular weight of E. coli-IFN-alpha2c). The 19,000 molecular weight component was thus completely degraded. However, there still remains a trace amount of 20,000 molecular weight material. Since the separation of both peaks 1 and 2 of FIG. 14b by RP-HPLC remained incomplete, the cause of the non-cleavable fraction in peak 2 with a molecular weight of 20,000 is probably due to the contamination of this peak of materials from peak 1.

In another experiment, the peak 2 material was incubated with O-glycanase without prior neuraminidase treatment (O-glycanase reads Gal (beta 1 to 3) GalNAc disaccharide from the protein only if it is not substituted with any other substances). The reaction product was again analyzed by SDS-PAGE (Fig. 17). It is evident that only the lightest fraction of peak 2 (from 19,000 to 17,500) decreased molecular weight. The apparent molar weight of both other constituents (with molecular weights of 21,000 and 20,000) was retained.

Example 10

Reaction of native IFN-alpha2 with 0.1 M NaOH

Since O-glycosylated materials can be degraded under mildly alkaline conditions, an attempt was made to deglycosylate the O-glycanase-resistant component (peak 1 in Fig. 14b) by incubation with 0.1 M NaOH. The reaction was carried out as described above. Simultaneously, E.coli-IFN-alpha2c and peak 2 were incubated as controls. The resulting reaction products were analyzed by SDS-PAGE. As shown in FIG. 13, the molecular weight of all components of native IFN-alpha2 was reduced to the apparent molecular weight of E. coli-IFN-alpha2. Blurred protein bands were seen to be due to little disruption of the protein structure under the conditions used. Also, the formation of a higher molecular weight band (above 30,000) should be attributed to alkaline conditions.

Example 11

Identification of glycopeptides by peptide mapping

Both peaks, which were obtained from native IFNalpha2 (Fig. 14b) and also E. coli-IFN-alpha2c tulle digested with trypsin, were reduced and then resolved with RP-KPLC. In FIG. 19 shows sections of the chromatogram obtained. Obviously, the two E. coli-IFN-alpha2c peaks, due to their hydrophobicity and later elution from the column, are somewhat delayed compared to analogous peaks from native IFN-alpha2: peak I and peak II (for E. coli-IFNalpha2c). ) are eluted substantially later than the corresponding peaks 1/1 and 1 / II of peak 1 of FIG. 14b or peaks 2 / Ia, 2 / Ib, 2 / IIa and 2 / IIb of peak 2 of FIG. 14b.

I

By analyzing the chain from the N-terminal end for the natural IFN-alpha2 peak and for both E.coli peaks, an amino acid chain was obtained for the peaks I, 1/1, 2 / Ia, 2 / Ib (Figure 1S). AS) 84-112 and for peaks II, 2 / IIa, 1 / II and 2 / IIb, the chain AS 71 to 112 (the amino acid chain IFN-alpha2c is shown in Fig. 15). Thus, different retention times must be due to glycosylation of the peptide from natural IFN-alpha2.

Example 12

PD-mass spectroscopy of glycopeptides from natural IFN-alpha2

Peaks 1 / II, 2 / IIa, 2 / Ia and 2 / IIb were characterized using the PD-mass spectrum. The results of the measurements are summarized in Table IV below. The difference in molecular weight calculated for the individual peptides from the amino acid chain and the actual demonstrated molecular weight of these glycopeptides can be explained by the different glycan structure: the molecular weight of the peptide (AS 71 to 112), which is substituted with a tetrasaccharide consisting of two N-acetylhexosamine units and two hexose units. Analogous to the structures of such O-glycans already described, this could be an oligosaccharide with the following structure: Gal 1 to 3 (Gal 1 to 4 GlcNAc 1-6) GalNAc-.

Peptide 2 / Ia has a molecular weight of 3,975, which can be explained by substitution of this peptide with a trisaccharide with the structure NeuAc-Gal-GalNAc. The same glycan structure can be derived from the molecular weight of peptide 2 / IIa, 5,448.

A molecular weight of 5,132 was measured for peptide 2 / IIb, which corresponds to the glycosylation of this peptide with a disaccharide with a Gal-GalNAc structure. This paragraph always refers to atomic mass units, amu.

Essentially, a molecular weight increased by approximately 23 units can be demonstrated for all peaks analyzed. This can be explained by binding sodium ions to the peptide. These impurities could have been removed by intensive washing of the material prior to measurement, but in this case the washing was not intentionally performed to keep the glycopeptide losses as low as possible. From the glycosidase degradation results described above, and by mass spectrometer measurements, the glycan structures shown in Table IV could be derived. The small peaks observed in the glycopeptide map may be due to other variants of the glycosylation of peptides.

co □ j->

c cn

t .. cc o x c o

tfl E r

CC

d.

□

d.

4-i «

-ce c

CC

CC 'CC CC ^ 4

21i 4-1 and 21i 3 4) C 1-4i and n>

• H 1 • 3 N CO o 4J C- 2í W O O 1 C C N 4-> y 1y χ ~ - WHAT and □ 0 -U 4-> (JI = kj y < £ 2 Z O cheek o. y 1 CC> »N O 'CO 3' H> E 0 CC 0, -1 C 3 y 2í N y y r— * »—f 0 co s c

I >> o c

TH rl TH • H £, —i

4-1 CC y

Q. · W

0) HJ>

C-2I

ABOUT

JZ o

d.

>

r ·;

d) and what f *>> what c c what

CC

about 1 ϋ < y y from < < y -i □ l- < y y ABOUT T from 1 from 1 1 1 1 1-F (-T «· * what WHAT WHAT C | ABOUT 1 about 1 ABOUT I l y 1 y 1 y 1 y < < < < from from from from 1-4 '- ( WHAT CC what y C C U about

I

σ> OJ WHAT ^ 4 C3 WITH what WHAT

WHAT WHAT WHAT (0 WHAT about WHAT WHAT WHAT c- Ρ » IN WHAT in in

uo m what OJ WHAT c * 'C WHAT in σι 'C CH uo what m d)

cm 04 OJ CJ rM 1 "t rf RW R-4

I I I

rM T 1-4 t} r- what θ ' θ '

CC Ώ M CC H-F M H M h-1 M \ CM CM CM

I »rt> c_a

• 3 o • 3 c

-. CC 4J C C. <3 y r \ IN" M 0 about jjj and >> Ό R 4 y CA sz > £. and y ^ 4 '> » y > 'c4 • H Ό CC 4-1 Ό- 0 Ή C > Ό 0 y and n 4-1 TO 0 M 0 OO C CJ 4J CC 0 and E SZ WHAT I Ή 1 from C- and C 'CC w 0 3 sz 'rj y C nH 3 0 ABOUT with y

>

glycan structure. (1) the peak numbers of Fig. 19.

(2) amu, atomic mass unit, (3) calculated mass, including sodium ions.

CC

XJ

Example 13

Identification of O-glycosylated amino acids by gas phase chain analysis

Since the glycopeptides obtained after trypsin digestion were too long to be able to determine their entire chain, these peptides were further digested with protease V8 from Staphylococcus aureus and then separated by RP-HPLC. Similarly, peptides from E. coli-IFNalpha2c were also processed. After comparing the peptide maps, all peptides with different retention times were isolated and their chains analyzed. All glycopeptides from wild-type IFN-alpha2 contained amino acids 97-112. In E. coli-IFN-alpha2c it was possible to detect the THR peptide, and in peptides from wild-type IFN-alpha2 this amino acid could not be detected. Thus, THR can be considered as a site where glycosylation occurs.

- 55 P-2

Claims (15)

PATENT CLAIMS Interferon alpha, which is O-glycosylated and has substantial biological and / or immunological properties of IFN-alpha2, preferably is O-glycosylated human IFN-alpha2a, IFN-alpha2b or IFN-alpha2c Interferon alpha according to claim 1, wherein the threonine residue at position 106, THR-106, O-glycosylated. Interferon alpha according to claims 1 or 2, wherein the oligosaccharide is preferably a neutral disaccharide Gal-GalNAc, a mono- or disialylated variant thereof or a neutral tetrasaccharide Gal- (Gal-GlcNAc-) GalNAc. 4. A method according to claim 1, wherein the interferon alpha is produced a) leukocytes, preferably human leukocytes induce a virus, (b) the induced interferon lafa is purified by a series of gentle purification steps in which the protein precipitates and dissolves, the pH being always below 8.0 (Cantelluv procedure); (c) the interferon mixture obtained in step (a) or (a) and (b) is bound to a column for affinity chromatography with a monoclonal antibody against IFN-alpha2, (d) the bound protein is eluted in an appropriate manner; (e) separating the protein from the eluate and, if necessary, purifying the affinity chromatography compound several times. 5. The method of claim 4, wherein the monoclonal antibody to IFNalpha2 is EBI-10 or analogues thereof. 6. A process according to claim 4 for the production of interferon alpha, characterized in that a further chromatographic purification is carried out after step e). 7. A process for the production of interferon alpha according to claim 1, which comprises: (a) insert DNA encoding IFN-alpha into an expression plasmid suitable for transfection of cells of multicellular organisms; b) transfecting cells of a multicellular organism, preferably an orphan, using a plasmid thus obtained for expression; (c) the eyelids of this organism are grown after transfection in a suitable nutrient medium; d) separating the cell culture supernatant, e) O-glycosylated IFN-alpha is isolated and purified in the usual manner. 8. The method of claim 7, wherein in step a) the plasmid pAD-CMV13, 15 or 19, preferably pAD-CMV19, is used and the DNA stored encodes a protein which essentially has biological and / or immunological properties of IFN-alpha2, preferably human IFN-alpha2a, IFN-alpha2b or IFNalpha2c, especially human IFN-alpha2c. 9. The method of claims 7 or 8, wherein the plasmid for expression is plasmid pAD19B-IFN. 10. A method according to claim 7, 8 or 9, wherein CHO cells are used from the cells of multicellular organisms. Expression plasmid for transfection of multicellular organisms, namely pAD-CMV13, pAD-CMV-15 or pAD-CMV-19. The O-glycosylated interferon alpha obtained by the method of claims 4 to 10. Interferon alfa according to claims 1 to 3 or 12 for use as a medicament. 14. A pharmaceutical composition for the treatment of viral or cancer diseases, characterized in that it contains interferon alpha as active ingredient according to claims 1 to 3 or 12. 15. A pharmaceutical composition according to claim 14 comprising a mixture of at least two glycosylated proteins IFN-alpha2A, IFN-alpha2b or IFNalpha2c.