IE913180A1

IE913180A1 - Glycoprotein sialylation by gene manipulation

Info

Publication number: IE913180A1
Application number: IE318091A
Authority: IE
Original assignee: Behringwerke Ag
Priority date: 1990-09-11
Filing date: 1991-09-10
Publication date: 1992-03-11
Also published as: EP0475354A3; JPH06105692A; KR920006501A; DE4028800A1; AU661824B2; AU8376091A; CA2051047A1; PT98917A; EP0475354A2

Abstract

The invention relates to the sialylation of glycoproteins by genetic manipulation, and to sialylated glycoproteins ("oversialylated glycoproteins"). The sialylation by genetic manipulation is effected by bringing about the expression of a DNA sequence which codes for a sialyltransferase together with a DNA sequence which codes for the glycoprotein to be sialylated in a eukaryotic cell, and isolating the sialylated glycoprotein which has been formed.

Description

BEHRINGWERKE AKTIENGESELLSCHAFT HOE 90/B 030 - Ma 858 Dr. Lp./Wr.

Glycoprotein sialvlation bv gene manipulation The present invention relates to glycoprotein sialylation 5 by gene manipulation, and to novel slalylated glycoproteins (hypersialylated glycoproteins).

A number of glycoproteins which can be employed for therapeutic purposes, such as, for example, factor VIII.C, factor IX, erythropoietin, antithrom10 bin III and monoclonal antibodies, can be synthesized with the aid of known processes in mammalian cells which are able to construct complex carbohydrate side chains (Kaufman et al., J. Biol. Chem. Vol. 263 (1988), 63526362; Kaufman et al., J. Biol. Chem. Vol. 261 (1986), 9622-9628; Lin et al., Proc. Natl. Acad. Sci. USA, Vol. 82 (1985), 7580-7584; ZettlmeiBl et al., Bio/Technology, Vol. 5, (1987) 720-725, Kdhler and Milstein, Nature 256 (1975) 495-497). The biochemical characterization of the carbohydrate side chains of recombinant glycoproteins prepared in this way has revealed that, depending on the expression system used, a recombinant glycoprotein displays differences in the glycosylation pattern compared with the naturally occurring glycoprotein (Pfeiffer et al., Eur. J. Biochem., Vol. 186 (1989), 273-286; ZettlmeiBl et al., J. Biol. Chem., Vol. 264 (1989), 21153-21159). Thus, on synthesis of recombinant glycoproteins in Chinese hamster ovary (CHO) cells there is often attachment of terminal sialic acid exclusively in a-2,3 linkage to the carbohydrate side chains of, for example, recombinant erythropoietin (Takeuchi et al., J. Biol. Chem., Vol. 263 (1988), 3657-3663) or recombinant antithrombin III (ZettlmeiBl et al., J. Biol. Chem., Vol. 264 (1989), 21153-21159). By contrast, a-2,6-linked sialic acid is predominantly found in natural forms.

Furthermore, Incomplete sialylation of the carbohydrate - 2 side chains has been observed in complex glycosylated proteins which have been isolated from natural sources (for example human urine or human plasma) (Franzen et al., J. Biol. Chem. Vol. 255, 5090-5093) or have been obtained by expression in mammalian cells (Sasaki et al., J. Biol. Chem., Vol. 262 (1987), 12059-12076). It is known that various glycoproteins, such as, for example, erythropoietin or antithrombin III, have a shorter halflife in vivo when slalylatlon is incomplete. It is furthermore known that so-called asialoglycoproteins, which have terminal galactose in place of terminal sialic acid on their sugar chains, have a particularly short half-life in vivo. This is because incompletely sialylated glycostructures of this type are recognized by receptors in the liver and are more rapidly removed from the bloodstream.

It is therefore very desirable and advantageous for in vivo use of the glycoproteins to have available more extensively or completely sialylated glycoproteins, because the latter have a longer half-life in vivo and thus better activity than the natural glycoproteins.

Known processes for the enzymatic sialylation of purified glycoproteins in vitro by soluble or immobilized highpurity sialyltransferases have technical and economic limitations owing to the low availability of the relevant enzymes and to the elaborate synthesis of the substrate (CMP-neuraminic acid) for sialyltransferase.

The present invention is based on the object of providing more extensively or completely sialylated glycoproteins for therapeutic purposes, the intention being to prepare them by genetic engineering methods. This object is achieved by bringing about the expression of a DNA sequence coding for a sialyltransferase, together with a DNA sequence coding for the glycoprotein to be sialyl35 ated, in a eukaryotic cell, and isolating the sialylated glycoprotein which is formed. - 3 The cDNA sequences for the ^-galactoside a-2,6sialyltransferase (Gal a-2,6-ST) from rat liver (Weinstein et al., J. Biol. Chem., Vol. 262 (1987), 17735-17743) and from human placenta (Grundmann et al., Nucl. Acids Res., Vol. 18 (1990), 667) have been isolated and characterized. It has been possible to express the Gal a-2,6-ST from rat liver in functional form in CHO cells. The result of this was that the CHO cell proteins synthesized by the CHO cells had sialic acid in a-2,6 linkage besides the a-2,3-linked sialic acid usual for this cell line (Lee et al., J. Biol. Chem., Vol. 264 (1989), 13848-13855).

It is now possible, surprisingly, according to the invention to modify a heterologous glycoprotein of therapeutic interest, both in the linkage pattern (for example a-2,3 vs. a-2,6) of the terminal sialic acids and, especially, in respect of a higher degree of sialylation, i.e. maximum completeness of occupation of terminal galactose residues with sialic acid, by coexpres20 sion with a heterologous sialyl transferase in a eukaryotic cell. The highly sialylated glycoproteins obtained according to this invention are particularly distinguished by long-lasting serum levels and thus by an improved therapeutic efficacy. Another advantage of the process is its industrial applicability, which results in more homogeneous products (homogeneous sialylation). To carry out the process of the invention, the coding DNA sequence for a sialyl transferase is cloned in a known manner downstream from a strong eukaryotic promoter (for example SV40 early promoter, compare Fig. 2) and inserted together with the transcription unit for a selectable marker gene (for example dihydrofolate reductase, aminoglycoside 3'-phosphotransferase, puromycin N-acetyltransferase, Wirth et al., Gene Vol. 73 (1988), 419-426; glutamine synthetase, Bebbington and Hentschel, DNA Cloning Vol. Ill (1987), Glover, ed., IRL press) into a eukaryotic cell suitable for the preparation of the glycoprotein to be sialylated, and is expressed in this - 4 producer cell. Selected eukaryotic cells which express the sialyltransferase together with the glycoprotein to be sialylated can be employed for the industrial synthesis of the glycoprotein in highly sialylated form.

Particularly preferred marker genes are those which permit coamplification of the DMA sequence coding for sialyl transferase, such as, for example, the cDNAs for dihydrofolate reductase or glutamine synthetase.

Typical examples of DNA sequences which code for sequences for galactoside galactoside a-2,6N-acetylgalactosamide coding DNA sequence for sialyltransferases are the a-2 ,3-sialyltransferase, sialyltransferase and a-2,6-sialyltransferase. The galactoside a-2,6-sialyltransferase is particularly preferred (Weinstein et al., loc. cit.; Grundmann et al., loc. cit.).

Typical and preferred examples of the recombinant glycoprotein to be sialylated ares human antithrombin III (Bock et al., Nucl. Acids Res., Vol. 10 (1982), 811320 8125), human erythropoietin (Jacobs et al., Nature, Vol. 313 (1985), 806-809), human Factor VII (Hagen et al., Proc. Natl. Acad. Sci. USA, Vol. 83 (1986), 24122416), human Factor VIII.C (Toole et al., Nature, Vol. 312 (1984), 342-347), human Factor IX (Choo et al., Nucl. Acids Res., Vol. 15 (1987), 871-884), glycosylated human receptor molecules such as, for example, tissue factor (Fisher et al., Thromb. Res., Vol. 48 (1987), 8999), interleukin-1 receptor (Sims et al., Proc. Natl. Acad. Sci. USA, Vol. 86 (1989), 8946-8950), interleukin30 4 receptor (EP-A 0 367 566), interleukin-7 receptor (Goodwin et al., Cell, Vol. 60 (1990), 941-951), tumor necrosis factor receptor (Schall et al., Cell, Vol. 61 (1990), 361-370), CD4 (Maddon et al., Cell, Vol. 42 (1985), 93-104).

The claimed process can be carried out in a variety of variants: (1) First the DNA sequence coding for - 5 sialyltransferase is inserted together with a marker gene and then the DNA sequence coding for the glycoprotein to be sialylated is inserted together with another marker gene into the eukaryotic cell. (2) First the DNA sequence for the glycoprotein to be sialylated is inserted together with a marker gene and then the DNA sequence coding for the sialyltransferase Is inserted with another marker gene into the eukaryotic cell. (3) The DNA sequence coding for the sialyltransferase and the DNA sequence coding for the glycoprotein to be sialylated are inserted at the same time together with a marker gene into the eukaryotic cell.

Typical examples of eukaryotic cells are yeasts, fungi, plant cells or animal cells. Permanent mammalian cells are preferred for the process to which the present invention relates. Mammalian cells such as BHK, CHO, C127 which are commercially available are particularly preferred.

In another embodiment of the invention, the DNA sequence coding for sialyltransferase can be inserted with a marker gene into a cell line which naturally forms glycoproteins. Preferred in this connection are hybridoma cells which form monoclonal antibodies (Kdhler and Milstein, loc. cit.) or cells which are employed for growing viruses.

The sialylated proteins obtained according to the invention can be purified and analyzed for the sialic acid content by standard methods (Hermentin and Seidat, GBF Workshop Protein Glycosylation: Cellular, Biotechno30 logical and Analytical Aspects, June 28-30, 1990, Braunschweig, Abstracts of Papers, p. 49).

The invention is explained further in the figure and the tables: Table 1 shows the complete cDNA sequence of the clone coding 14 for human galactoside a-2,6-sialyltransferase with its flanking EcoRI linker portions. The figure shows the expression plasmid for human galactoside a-2,6sialyltransferase pAB Sial.

Table 2 summarizes the results on the sialic acid linkage of human antithrombin III purified from cell culture supernatants of BHK cells before and after transfection with the plasmid pAB Sial.

The invention is explained by the examples. These ex10 amples should not be regarded as limiting. Further information on the molecular biological processes employed here are to be found in Sambrook et al., Molecular Cloning, 2nd edition, 1989, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., USA.

Example 1: Cloning of the cDNA for human ^-galactoside a—2,6—sialyltransferase (Gal a-2,6-ST) A cDNA bank, described in EP-A 0 236 978, from human placenta in the phage vector lambda gtlO was screened, by hybridization with synthetic oligonucleotide probes, for clones which harbor a cDNA coding for Gal a-2,6-ST. The two oligonucleotide probes used for this purpose were derived from the cDNA sequence of rat Gal a-2,6-ST (Weinstein et al., loc. cit.) taking account of the codon usage in human protein-encoding sequences (Lathe, J. Mol.

Biol., Vol. 183 (1985), 1-12) and had the following sequences: A) 5'CTC CTG GCT CTT GGG CAT CTG GAA CTC CTT GGC CTG CAG GGT CAG GGC CTC ATA GTC 3' (57mer) B) 5'ATG ATG ACC CTG TGT GAC CAG GTG GAC ATC TAT GAG TTC CTG CCA TCC AAG 3' (51mer) Both oligonucleotide sequences were labeled at the 5' end with T4 polynucleotide kinase using (gamma-32P)-ATP (equimolar amounts of DNA and (gamma-32P)-ATP: 3000 Ci/mmol, 10 pCi/^l, employing 5 pl/50 pi per reaction mixture). The probes had a specific activity of 2 x 10® Bq/pg or 4 x 104 Bq/pmol.

For the screening of the placental cDNA bank with the oligonucleotide probes described above, 3 x 104 pfu (plague forming units) were plated out with cells of the E. coli K 12 strain C 600 hfl in soft agar on 13.5 cm Petri dishes and incubated at 37 C for 6 h. Lysis was still incomplete after this period. The plates were incubated in a refrigerator overnight and the phages were transferred to nitrocellulose filters (Schleicher & Schull, BA 85, Ref. No. 401124) (duplicates). Nitrocel15 lulose filters and Petri dishes were marked with an injection needle to permit subsequent assignment. The Petri dishes were stored in a cold room during the subsequent treatment of the nitrocellulose filters. The DNA present on the nitrocellulose filters was denatured by placing the filters for 5 rain on a filter paper (Whatman 3M) impregnated in 1.5 M NaCl, 0.5 M NaOH. The filters were then renatured in the same way with 1.5 M NaCl, 0.5 M tris (pH 8.0) and washed with 2 x SSPE (0.36 M NaCl, 16 mM NaOH, 20 mM NaH2PO4, 2 mM EDTA). The filters were then baked at 80 °C under reduced pressure for 2 h. The filters were prehybridized at 65eC for 4 h (prehybridization solution: 0.6 M NaCl, 0.06 M tris (pH 8.3), 6 mM EDTA, 0.2 % nonionic synthetic sucrose polymer (FicoH*), 0.2 % polyvinylpyrrolidone 40, 0.2 % BSA, 0.1 % SDS, 50 pg/ml denatured herring sperm DNA).

The filters were finally incubated in beakers or in sealed polyethylene films, shaking gently, with the addition of 100 000 - 200 000 Bg of the labeled oligonucleotide/ml of hybridization solution (as prehybridiza35 tion solution but without herring sperm DNA) overnight.

The hybridization temperature was 65’C. The nitrocellulose filters were washed with 6 x SSC, 0.05 M sodium pyrophosphate at room temperature for one hour and at - 8 hybridization temperature for another hour. The filters were dried and autoradiographed overnight. Signals which occurred in identical position on the X-ray film with both duplicates were assigned to the Petri dish, and the region (about 30 plagues) was punched out with the wide end of a Pasteur pipette, and the phages were resuspended in 1 ml of SM buffer. Phage mixtures which contained phage cDNAs hybridizing with the probe were singled out over three cycles until it was possible to identify a single positive clone.

In total, 106 pfu of the placental cDNA bank were screened, and two phages which gave positive signals were analyzed in more detail (clones Nos. 14 and 15). The two phage clones 14 and 15 were grown and their DNAs were extracted. The DNAs were partially cleaved with EcoRI, isolated and ligated into the EcoRI restriction cleavage site of the Bluescript M13 vector (Stratagene, San Diego, CA, USA) for restriction and sequence analyses. The resulting recombinant plasmids were called pSiall4 (L or R) and pSiallS (L or R) depending on the orientation of the EcoRI fragment. They all contained the entire coding sequence. Tab. 1 showed the complete cDNA sequence of clone 14 (2188 bp) with its flanking EcoRI linker portions (5'AG 'AATTCT). The sequence has an open reading frame and codes for a protein of 406 amino acids.

Example 2: Coexpression of human Gal a-2,6-ST in a baby hamster kidney (BHK) 21 cell which expresses human antithrombin III (AT III) a) Preparation of a BHK21 cell line which expresses human AT III The preparation of the BHK mixed clone 3MK1 which secretes human AT III into the culture medium and which was used for the subsequent experiments is described in Example lb of EP-A 0 330 977. Singling out of clones was carried out by limiting dilution with this mixed clone 3MK1 (400 pg/ml G418; 10 pM methotrexate (MTX); see Tab. 2 of EP-A 0 330 977). A single clone obtained in this way, 3MK1-3-B11, was grown in culture medium (DME medium which contains 10 % fetal calf serum (FCS), 400 pg/ml G418 and 10 pM MTX) (cell ine 3MK1-3-B11). b) Construction of an expression vector for Gal a-2,6-ST A cDNA fragment (2.2 kb), coding for Gal a-2,6-ST, from the vector pSiall4L (Example 1) was obtained by cleavage with the restriction enzymes Hind III and Xbal and was cloned into the vector pAB3-l (ZettlmelBl et al., Behring Inst. Mitt. Vol. 82 (1988), 127-143) opened with Hindlll/Xbal. The resulting plasmid which harbors the Gal a-2,6-ST cDNA in the required orientation downstream of the SV40 early promoter was called pABSial (Fig. ). c) Cotransfection of the cell line 3MK1-3-B11 with an expression plasmid for Gal a-2,6-ST (pABSial) and an expression plasmid for glutamine synthetase (pSVLGSl) The cell line 3MK1-B11 obtained as in (a) was adapted for 8 passages to growth in GMEM medium (Bebbington and Hentschel (1987), loc. cit.) which contained 10 % FCS, 400 pg/ml G418 and 10 pM MTX (GS medium). The cells obtained in this way expressed 11 ± 2 pg of AT 40 III/106 cells/24 h and were transfected with the plasmids pAB Sial (20 pg) and pSVLGSl (5 pg; Bebbington and Hentschel (1987), loc. cit.) using the calcium phosphate technique described in EP-A 0 330 977. Transfected cells were selected In GS medium which contained 15 pM methionine sulfoximine (MSX). After about 4 weeks, 9 MSX-resistant single clones were isolated. RNA was isolated from three of these clones (42-A3, 42-A11, 42-C1) by standard processes and hybridized in a Northern blot with a radiolabeled fragment of human Galx a-2,6-ST cDNA as probe.

In contrast to the initial clone (3MK1-3-B11), all 3 clones contained a Gal a-2,6-ST specific mRNA. The strength of the Gal a-2,6-ST specific signal in the clones 42-A3 and 42-C1 was comparable with that of the signal of the AT III specific mRNA, which signal remained unaffected, compared with the initial clone 3MK1-3-B11, after transfection with the vectors pABSial/pSVLGSl in all 3 analyzed clones 42-A3, 42-All and 42-C1.

The expression rates for antithrombin III for all 3 ana10 lyzed Gal a-2,6-ST expressing cell lines were, moreover, identical to that for the initial clone (11 ± 2 ^g/106 cells/24 h).

Example 3: Purification and characterization of human AT III from a BHK cell line which coexpresses human Gal a-2,6-ST a) Growth of cells and purification of AT III The BHK cell lines 3MK1-3-B11 (AT III) and 42-C1 (AT III + Gal a-2,6-ST) were cultured to confluence in GS medium in 1750 cm2 roller bottles. After removal of the GS medium and a single wash (4 hours) with 100 ml of serum-free Iskove's medium (Behringwerke AG, Marburg, FRG) the cells were incubated in 500 ml of serum-free Iskove's medium four times for 48-72 h each. This resulted in AT III concentrations above 12 mg/1 in the medium for both cell lines. The recombinant human AT III was purified from the culture supernatant by a standard process (ZettlmeiBl et al. (1989), loc. cit.). b) Sialic acid determination by HPAE-PAD The sialic acid content of the various r-AT III samples was determined in a manner known per se (Hermentin and Seidat, loc. cit.) by hydrolytic liberation (0.1 N H2SO4, 1 h, 80eC), fractionation (high-pH anion-exchange chromatography , HPAE) and determination of the sialic acid content by pulsed amperometric detection (PAD). The sialic acid content proved in the case of the cell line 42-C1 transfected with Gal a-2,6-ST to be Increased by about 40 % compared with the untransfected cell line 3MK1-3-B11. c) Glycan differentiation The cell line 42-C1 transfected with Gal a-2,6-ST, and the untransfected cell line 3MK1-3-B11 were examined for the linkage of sialic acid to terminal galactose residues of the N-glycans present In AT III in a manner known per se using the glycan differentiation kit supplied by Boehringer Mannheim (Ref. No. 1142 372). This test makes use of a so-called dot blot assay to recognize a-2,315 linked sialic acid using the digoxigenin-carrying lectin MAA (Maackia amurensis agglutinin) and a-2,6-linked sialic acid using the digoxigenin-carrying lectin SNA (Sambucus nigra agglutinin) and detection thereof via a color reaction in a sandwich assay with sheep anti20 digoxigenin which is conjugated to the enzyme alkaline phosphatase.

The following standard glycoproteins in the glycan differentiation kit were used as controls: Fetuin: positive for the lectins MAA, SNA and DSA; asialofetuin: negative for the lectin SNA; transferrin: (negative control) 1.0, 0.5 and 0.2 pg of glycoprotein, in each case in serial dilutions, was applied to each well.

Exclusively a-2,3-linked sialic acid was detected with AT III isolated from the untransfected cell line 3MK1-3-B11. a-2,6-linked sialic acid was detected, in addition to normal a-2,3-linked sialic acid, with the AT III isolated from the cell line 42-C1 transfected with Gal a-2,6-ST (Table 2). - 12 The result demonstrates that human a-2,6-sialyltransferase is active after transfection in the DNA sequence which is to be converted in the 42-C1 cell line and makes possible an a-2,6-sialylation, in addition to the normal a-2,3-sialylation, of human antithrombin III.

Example 4: It is possible in a manner analogous to Examples 1-3 to achieve improved sialylation of recombinant erythropoietin from various eukaryotic cell lines. The following cell lines are particularly preferred for this: BHK, C127, CHO - 13 Table 1 50 aattCtGCCCGGCGTTAACAAAGGGAGCCGATACCGACCGGCGTGGGCGCGGAGCGGGCG 90 110 GCCGCCACCGAGCGTGCTGAGCAACCGCAGCCTCCGCGGCCGAGAGTGCAGCGAGCAAGG 130 150 170 GGAGAGCCAGTTGCGCAGAGCCCTGCAACCAGCAGTCCAGGGAGAAGTGGTGAATGTCAT 190 210 230 GGAGCCCAGCTGAAATGGACTGGCCCCCTTGAGCCTGTCCCAAGCCCTGGTGCCAGGTGT 250 270 290 CCATCCCCGTGCTGAGATGAGTTTTGATCATCCTGAGAAAAATGGGCCTTGGCCTGCAGA 310 330 350 CCCAATAAACCTTCCCTCCCATGGATAATAGTGCTAATTCCTGAGGACCTGAAGGCCTGC 370 390 410 CGCCCCTGGGGGATTAGCCAGAAGCAGGCTTGTTTTCCTGCTCAGAACAAAGTGACTTCC 430 450 470 CTGAACACATCTTCATTATGATTCACACCAACCTGAAGAAAAAGTTCAGCTGCTGCGTCC MIHTNLKKKFSCCVL 490 510 530 TGGTCTTTCTTCTGTTTGCAGTCATCTGTGTGTGGAAGGAAAAGAAGAAAGGGAGTTACT VFLLFAVICVWKEKKKGSYY 550 570 590 ATGATTCCTTTAAATTGCAAACCAAGGAATTCCAGGTGTTAAAGAGTCTGGGGAAATTGG DS FKLQTKEFQVLKSLGKLA 610 630 650 CCATGGGGTCTGATTCCCAGTCTGTATCCTCAAGCAGCACCCAGGACCCCCACAGGGGCC MGSDSQSVSSSSTQDPHRGR 670 690 710 GCCAGACCCTCGGCAGTCTCAGAGGCCTAGCCAAGGCCAAACCAGAGGCCTCCTTCCAGG QTLGSLRGLAKAKPEASFQV 730 750 770 TGTGGAACAAGGACAGCTCTTCCAAAAACCTTATCCCTAGGCTGCAAAAGATCTGGAAGA WNKDSSSKNLIPRLQKIWKN 790 810 830 ATTACCTAAGCATGAACAAGTACAAAGTGTCCTACAAGGGGCCAGGACCAGGCATCAAGT YLS MNKYKVSYKGPGPGIKF 850 870 890 TCAGTGCAGAGGCCCTGCGCTGCCACCTCCGGGACCATGTGAATGTATCCATGGTAGAGG SAEALRCHLRDHVNVSMVEV 910 930 950 TCACAGATTTTCCCTTCAATACCTCTGAATGGGAGGGTTATCTGCCCAAGGAGAGCATTA TDFPFNTSEWEGYLPKESIR Table 1 (continuation) 970 990 1010 GGACCAAGGCTGGGCCTTGGGGCAGGTGTGCTGTTGTGTCGTCAGCGGGATCTCTGAAGT TKAGPWGRCAVVSSAGSLKS 1030 1050 1070 CCTCCCAACTAGGCAGAGAAATCGATGATCATGACGCAGTCCTGAGGTTTAATGGGGCAC SQLGREIDDHDAVLRFNGAP 1090 1110 1130 CCACAGCCAACTTCCAACAAGATGTGGGCACAAAAACTACCATTCGCCTGATGAACTCTC TAN FQQDVGTK T T I RLMNSQ 1150 1170 1190 AGTTGGTTACCACAGAGAAGCGCTTCCTCAAAGACAGTTTGTACAATGAAGGAATCCTAA LVTTEKRFLKDSLYNEGILI 1210 1230 1250 TTGTATGGGACCCATCTGTATACCACTCAGATATCCCAAAGTGGTACCAGAATCCGGATT VWDPSVYHSDI PKWYQNPDY 1270 1290 1310 ATAATTTCTTTAACAACTACAAGACTTATCGTAAGCTGCACCCCAATCAGCCCTTTTACA NFFNNYKTYRKLHPNQPFYI 1330 1350 1370 TCCTCAAGCCCCAGATGCCTTGGGAGCTATGGGACATTCTTCAAGAAATCTCCCCAGAAG LKPQMPWELWDILQEISPEE 1390 1410 1430 AGATTCAGCCAAACCCCCCATCCTCTGGGATGCTTGGTATCATCATCATGATGACGCTGT IQPNPPSSGMLGI IIMMTLC 1450 1470 1490 GTGACCAGGTGGATATTTATGAGTTCCTCCCATCCAAGCGCAAGACTGACGTGTGCTACT DQVDIYEFLPSKRKTDVCYY 1510 1530 1550 ACTACCAGAAGTTCTTCGATAGTGCCTGCACGATGGGTGCCTACCACCCGCTGCTCTATG YQKFFDSACTMGAYHPLLYE 1570 1590 1610 AGAAGAATTTGGTGAAGCATCTCAACCAGGGCACAGATGAGGACATCTACCTGCTTGGAA KNLVKHLNQGTDEDIYLLGK 1630 1650 1670 AAGCCACACTGCCTGGCTTCCGGACCATTCACTGCTAAGCACAGGCTCCTCACTCTTCTC ATLPGFRTIHC 1690 1710 1730 CAT CAGG CATTAAATGAATGGTCTCTTGGCCAC CC CAGC CTGGGAAGAACATTTTCCTGA 1750 1770 1790 ACAATTCCAGCCTGCTCCTTTTACTCTAGGGGCCTCTGTCAGCAAGACCATGGGACTTCA 1810 1830 1850 AGAGCCTGTGGTCAGGAAATCAGGTCCAGCCTTCCCTGTAGCCAGACAGTTTATGAGCCC 1870 1890 1910 AGAGCCTCCTGCCACACACATGCACACATATCTAGCATTCTTTCCAAGACAGCATCCTCC - 15 Table 1 (continuation) 1930 1950 1970 CCGCCTTCCACCTTGTAGATGCAAGGTCTATCTCTCCCATCAGGGCTGCCAAAGCTGGGC 1990 2010 2030 TTTGTTTTTCCCAGCAGAATGATGCCATTCTCACAAACCAATGCTCTATATTGCTTGAAG 2050 2070 2090 TCTGCATCTAAATATTGATTTCACGTTTTAAAGAAATTCTCTTAAATTACAATTGTGCCC 2110 2130 2150 AATGCAGGGTGGCTCTGGGGGGCAAGTAGGTGGTACAGGGGATTGGAAACAATCGTCCGC 2170 2188 GCCTCCAGAGAAAAGTTGCTCCCGAGag - 16 Table 2: Determination of the sialic acid linkage using the glycan differentiation kit U 1 CN !> •r| P P rtf Ο» ω 3 P Φ 3 co 3 e tH •H Ο to Ο β •H e •H CO <4-1 Q 3 P •r| P Φ 0 0 m 0 •r| 3 o P P P 1 ft r-l Ό P r-l << P P 0 p <4-1 m 0 3 P 0 co Φ P 1 P (0 Φ P Q P ι-1 ft w H «-4 c P <4-1 P 0 o 0 H H •H 3 CD c Ό i-l P Ο t—1 H s 3 0 2 3 0 3 >0 P S cn P o to <0 u <0 •r, 3 rl H Ej Φ P n O 0 rt! ft E4 O HOE 90/B 030

Claims

1. Patent claims:

1. A process for the sialylation of glycoproteins, which comprises expressing a DNA sequence coding for a 5 sialyltransferase together with a DNA sequence coding for the glycoprotein to be sialylated in a eukaryotic cell, and isolating the sialylated glycoprotein which is formed.

2. The process as claimed in claim 1, wherein the DNA 10 sequence coding for a galactoside a-2,3-sialyltransferase is used.

3. The process as claimed in claim 1, wherein the DNA sequence coding for a galactoside a-2,6-sialyl transferase is used. 15

4. The process as claimed in claim 1, wherein the DNA sequence coding for N-acetylgalactosaminoside a-2,6sialyltransferase is used.

5. The process as claimed in claim 1, wherein the DNA sequence coding for a human galactoside a-2,320 sialyltransferase is used.

6. The process as claimed in claim 1, wherein the DNA sequence coding for a human galactoside a-2,6sialyltransferase is used.

7. The process as claimed in claim 1, wherein the DNA 25 coding for a human N-acetylgalactosaminoside a-2,6sialyltransferase is used.

8. The process as claimed in claim 1, wherein first the DNA sequence coding for the sialyltransferase is inserted together with a marker gene and then the DNA sequence 30 coding for the glycoprotein to be sialylated is inserted together with another marker gene into the eukaryotic cell. - 18

9. The process as claimed in claim 1, wherein first the DNA sequence coding for the glycoprotein to be sialylated is inserted together with a marker gene and then the DNA sequence coding for the sialyltransferase is inserted 5 together with another marker gene into the eukaryotic cell.

10. The process as claimed in claim 1, wherein the DNA sequence coding for the sialyltransferase and the DNA sequence coding for the glycoprotein to be sialylated are 10 inserted at the same time together with a marker gene into the eukaryotic cell.

11. The process as claimed in any of claims 8-10, wherein the coding sequences for dihydrofolate reductase, glutamine synthetase, adenosine deaminase, thymidine 15 kinase, puromycin N-acetyltransferase, aminoglycoside 3'phosphotransferase are used as marker genes.

12. The process as claimed in claim 1, wherein mammalian cells are used as eukaryotic cells.

13. The process as claimed in claim 1, wherein BHK, C127, 20 CHO, mouse myeloma, vero or HeLa cells are used as eukaryotic cells.

14. A process for the sialylation of glycoproteins, which comprises inserting a DNA sequence coding for sialyltransferase into a cell which normally forms a 25 glycoprotein, and isolating the sialylated glycoprotein which is formed.

15. The process as claimed in claim 14, wherein a monoclonal antibody producing hybridoma cell is used as eukaryotic cell. 30

16. A sialylated glycoprotein obtainable as claimed in any of claims 1 to 15.

17. A process according to claim 1 or 14 for the sialylation of a glycoprotein, substantially as hereinbefore described and exemplified.

18. A sialylated glycoprotein whenever obtained by a process claimed in claim 17.