Process for Controlling Intracellular Glycosylation of Proteins

1. Field of the Invention.

The present invention relates generally to the cellular mechanisms and machinery involved in the glycosylation of proteins manufactured by the cell. More particularly, the present invention involves altering the glycosylation capabilities of a cell in order to control the structure of carbohydrate groups attached during glycosylation.

2. Description of Related Art.

The publications and other reference materials referred to herein to describe the background of the invention and to provide additional detail regarding its practice are hereby incorporated by reference. For convenience, the reference materials are numerically referenced and grouped in the appended bibliography.

During the last decade, numerous processes and procedures have been developed for genetically engineering cells in order to produce a wide variety of proteins and glycoproteins. These procedures involve utilizing recombinant DNA technology to prepare a vector which includes genetic material that codes for a specific protein or glycoprotein. Upon introduction of the vector into the host cell, the inserted genetic material instructs the host cell's biochemical machinery to manufacture a specific protein or glycoprotein.

Problems have been experienced with the production of glycoproteins by genetically engineering host cells. Glycoproteins are proteins having carbohydrate groups attached at various points along the protein's amino acid backbone. The carbohydrate groups are commonly attached to asparagine, serine or threonine. The genetic sequence introduced into the host cell usually includes instructions with respect to the amino acid sequence of the protein and the location and structure of the carbohydrate groups. Most of the cell lines which are commonly used as host cells are capable of following the vector's instructions with respect to preparing a protein having a specific amino acid sequence. However, many host cells are not capable of following instructions with respect to glycosylation of the protein. For example, E. coli is a common host cell used in producing a wide variety of proteins. However, E. coli does not contain the cellular glycosylation machinery required to attach carbohydrate groups to the proteins it manufactures.

Unlike E. coli, many other host cells do have varying capabilities with respect to protein glycosylation. However, even though these cells have glycosylation capabilities, the glycosylation machinery is not controlled by the recombinant DNA vector. Accordingly, the glycoprotein produced by such host cells may differ in carbohydrate structure from the natural glycoprotein coded for by the vector.(1, 2)

Chinese hamster ovary (CHO) cells are a standard cell line used commercially for the high yield expression of glycoproteins from vectors engineered through recombinant DNA technology. The protein sequence of the glycoprotein expressed by CHO comes from the DNA transinfected into the cell while the structure of the carbohydrate portion of the glycoprotein is determined by the cellular machinery of the CHO cells. While most glycoproteins normally contain a mixture of NeuAc-.alpha.-2,6Gal and NeuAc-.alpha.-2,3Gal linkages on their N-linked oligosaccharides, CHO cells only make asparagine linked carbohydrate chains with terminal sialic acids in the NeuAc-.alpha.-2,3Gal linkage.(1,2)

For example, erythropoietin is a glycoprotein naturally occurring in humans which has N-linked carbohydrate groups with both the NeuAc-.alpha.-2,6Gal and NeuAc-.alpha. -2,3Gal linkages. CHO cells which are genetically engineered to produce erythropoietin can only produce this protein with the NeuAc-.alpha.-2,3Gal linkages.(1) Although a number of mutant CHO cell lines have been developed which have altered capabilities for protein glycosylation,(3) they are not suitable for the production of glycoproteins intended for use in animals. Indeed, the carbohydrate groups produced by the cells are truncated, resulting in the rapid clearance of the recombinant glycoproteins from the blood followed by degradation. Thus, while the glycoproteins produced by these mutant CHO cell lines do display in vitro biological activity, they are inactive in vivo because of the rapid clearance from the blood stream.

From the above, it is apparent that there is a need to develop a process which can be used to alter the glycosylation machinery of host cells in order to control the structure of carbohydrates attached during glycosylation. Such a process for controlling host cell glycosylation would be useful not only in expressing glycoproteins which accurately mimic naturally occurring proteins, but would also be useful in preparing glycoproteins having selected altered carbohydrate structures for diagnostic and research uses.

In accordance with the present invention, a process is disclosed which provides for control of cellular glycosylation. The invention is based upon the discovery that the glycosylation machinery of host cells can be altered and controlled by introducing a gene into the host cell which codes for at least one enzyme which is capable of affecting glycosylation of a protein in the cell.

The present invention involves controlling the glycosylation of a protein in a host cell wherein attachment of the carbohydrate moiety to proteins during glycosylation is dependent upon a number of naturally occurring enzymes which are present in the cell. In accordance with the present invention, at least one gene is introduced into the cell which is capable of expressing at least one supplemental enzyme which is capable of affecting the glycosylation mechanism. The expression of the supplemental enzyme in the cell produces a cell having both naturally occurring and supplemental enzymes wherein the presence of the supplemental enzyme alters the cell's glycosylation mechanism.

As a feature of the present invention, it was discovered that transfection of CHO cells with cDNA coding for a glycosyltransferase resulted in production of the glycosyltransferase enzyme in the CHO cell and subsequent alteration of the carbohydrate structure of glycoproteins produced by the CHO cell.

A feature of the present invention is that the glycosylation machinery of a host cell can be controlled to produce glycoproteins wherein the location and structure of carbohydrates is equivalent to a given naturally occurring glycoprotein. As an additional feature of the present invention, the glycosylation process of the host cell can also be controlled to produce glycoproteins wherein the carbohydrate structure is changed slightly from the naturally occurring glycoprotein. Such purposely altered glycoproteins are useful as both diagnostic and research tools in studying the biochemistry of various naturally occurring glycoproteins.

The above discussed and many other features and attendant advantages of the present invention will become better understood by reference to the following detailed description when taken in conjunction with the accompanying drawings.

FIG. 1 is a diagram showing the plasmid construct used to transfect CHO cells to alter the terminal structures of the N-linked carbohydrate groups by expression of sialyltransferase. The vector was constructed by inserting a cDNA coding for the sialyltransferase into the cloning site of the pECE expression vector described by Ellis, et al.(4)

FIG. 2 is a diagrammatic representation outlining the mechanism by which applicant's process controls carbohydrate structure during glycosylation within CHO cells.

FIGS. 3a and 3b are the nucleotide sequence of the sialytransferase cDNA prior to mutagenisis at the EcoRI site.

The present invention involves controlling the glycosylation machinery of a cell by using genetic engineering to instruct the cell to produce various enzymes upon which glycosylation in the cell is dependent. The process for controlling glycosylation is based on the well-documented fact that a given cell will synthesize carbohydrate groups whose structures are determined by the specificities of the glycosyltransferases produced by that cell. (2) The most bioactive terminal sugars are attached to common core structures by "terminal" glycosyltransferases.(5) When two terminal enzymes compete with each other, the ultimate carbohydrate structure is determined by the specificity of the enzyme that acts first. The invention relies on the concept that the introduction and over expression of a terminal (or branching) glycosyltransferase, not normally produced by a cell, will result in the successful competition with the endogenous enzymes, and will produce carbohydrate groups with a structure specified by the new enzyme.

The basic procedure involves transfection of a host cell with a vector carrying a gene which expresses at least one enzyme upon which glycosylation in the host is dependent. The resultant enzyme(s) which is expressed in the cell provides internal control of the glycosylation machinery of the cell. Accordingly, the invention provides a useful procedure for controlling the structure of carbohydrates attached during glycosylation to more closely resemble naturally occurring glycolipids. In addition, one can use the present invention to alter the carbohydrate structure of a glycoprotein produced in a host cell for investigational purposes.

The invention has wide application to host cells which are naturally capable of glycosylation. Exemplary cell lines to which the present invention is amenable include Chinese hamster ovary (CHO) cells, mouse L cells, mouse A9 cells, baby hamster kidney cells, C127 cells, PC8 cells and other eukaryotic cell lines capable of the expression of recombinant glycoproteins.

The particular procedure used to introduce genetic material into the host cell for expression of the glycosyl transferase is not particularly critical. Any of the well-known procedures for introducing foreign nucleotide sequences into host cells may be used. These include the use of plasmid vectors, viral vectors and any of the other well-known methods for introducing cDNA or other foreign genetic material into a host cell. It is only necessary that the particular genetic engineering procedure utilized be capable of successfully introducing at least one gene into the host cell which is capable of expressing at least one enzyme which is known to be involved in glycosylation. Further, the genetic material must be introduced in such a way that the host cell expresses the enzyme coded for by the inserted genetic material so that, upon expression, the enzyme alters the glycosylation capabilities of the cell.

The preferred enzymes for production within a host cell are glycosyltransferases such as sialyltransferase. Other possible enzymes which may be coded for and produced within the cell to alter the glycosylation machinery include fucosyltransferases, galactosyltransferases, .beta.-acteylgalactosaminyltransferases, N-acetylglycosaminyltransferases, and sulfotransferases.(5)

The particular vector used is also not particularly critical. Any of the conventional vectors used for expression of recombinant glycoproteins in eukaryotic cells may be used. Exemplary vectors include pMSG, pAV009/A.sup.+, pMT010/A.sup.+ and any other vector allowing expression of glycoproteins under direction of the SV-40 early promotor, metallothionein promotor, murine mammary tumor virus promotor, Rous sarcoma virus promotor or other promotors shown effective for expression in eukaryotic cells. A suitable vector is the pECE vector which is described in Ellis, et al.(4)

The types of glycoprotein which would be expressed having modified carbohydrate structural forms include erythropoietin, insulin, plasminogen activator (TPA), interferon and various glycopeptide hormones. The nucleotide sequences for various cDNA coding for these proteins are known.

The following portion of this detailed description is limited to alteration of the glycosylation machinery of CHO cells. However, it is understood that the principles disclosed with respect to the CHO cell line also apply to the other various host cells previously mentioned.

In this specific example, a CHO cell line is produced which produces modified terminal sialic acid groupings on its N-linked oligosaccharides. Normally these cells make N-linked carbohydrate groups that contain sialic acid exclusively in the NeuAcc.alpha. 2,3Gal sequence. By expression of a .beta.-galactosideo.alpha. 2,6 sialyltransferase cDNA in these cells, the N-linked carbohydrate groups are directed to produce the NeuAc.alpha. 2,6Gal sequence commonly found on many glycoproteins.

The expression vector used to transfect the CHO cells is shown in FIG. 1. The vector was constructed as follows. ST3, a 1.6 kb cDNA encompassing the complete amino acid coding sequences for the .beta.-galactoside .alpha. 2,6 sialyltransferase, was shotgun subcloned from an EcoRI digest of ST3 into M13mp19 as described by Weinstein, et al.(6). Although an internal EcoRI site was present, the two fragments were correctly oriented as determined by dideoxy sequencing(7). Site directed mutagenesis by the procedure of Zoller and Smith(8) eliminated the internal EcoRI site using the primer GCCAAGGAGTTCCAGAT which binds to nucleotides 115-132. An A to G transition abolished the EcoRI recognition site, GGAATTC, but preserved the native amino acid coding sequence. The mutation in ST3 was confirmed by the dideoxy chain termination DNA sequencing method (7).

The nucleotide sequence of the sialyltransferase cDNA prior to mutagenesis at the EcoRI site is set forth in FIGS. 3a and 3b. Also shown is the complete amino acid sequence inferred from the nucleotide sequences. Peptide sequence overlaps (black boxes) include the NH.sub.2 -terminal sequence of the purified sialyltransferase (arrow). Stippled areas indicate residues that were not identified. Potential glycosylation sites with the sequence Asn-X-Thr/Ser are boxed. The proposed signal-anchor sequence is underscored with the cross-hatched box bordered at either end by open boxes highlighting charged lysine residues.

The altered cDNA was subcloned from M13mp19 into the EcoRI site of the polylinker of bluescript (bs-ST3) and subsequently into the EcoRI site of pECE (pECE-ST3) for expression of the sialyltransferase as shown in FIG. 1. In this vector the sialyltransferase cDNA is under the direction of SV-40 virus early promotor, allowing the sialyltransferase to be expressed in a wide variety of eukaryotic cell lines. The vector was obtained from William J. Rutter (University of California at San Francisco School of Medicine-Department of Biochemistry and Biophysics), and is described in Ellis, et al.(4)

CHO cells were transfected with pECE-ST3 according to the method of Graham and Van der Eb(9). Cells at 50% confluency were transfected with 20 ug supercoiled pECEST3 and 2 ug supercoiled pSV2neo per 100 mm dish. After 48 hours the cells were split 1:5 and replated in 75 cm.sup.2 flasks with selection medium containing the antibiotic G418. After six weeks in selection medium, resistant cells were presumed to have the transfected DNA stably integrated in the genome, and the cells were then maintained in the absence of G418.

To select clonal cell lines expressing the sialyltransferase, advantage was taken of a newly described plant lectin, Sambucus nigra agglutinin (SNA), which recognizes the product of the sialyltransferase, NeuAc.alpha. 2,6Gal, with 50-100 fold higher avidity than the NeuAc.alpha. 2,3Gal sequence normally produced by the CHO cells.(10) Accordingly, cells producing the sialyltransferase were found to bind the fluorescent labeled lectin (FITC-SNA), but not the wild type cells. Following automated fluorescence activated cell sorting (FACS), clonal cell lines expressing the .beta.-galactoside .alpha. 2,6 sialyltransferase were readily isolated and amplified.

Detailed analysis of one cell line showed that the .beta.-galactoside .alpha. 2,6 sialyltransferase is expressed at equivalent levels to the endogenous .beta.-galactoside .alpha. 2,3 sialyltransferase normally present in CHO cells. The result is that 20-25% of the total cell surface carbohydrate groups contain the NeuAc.alpha. 2,6Gal sequence instead of the NeuAc.alpha. 2,3Gal sequence. Thus, this cell line produced the mixture of sialic acid linkages found on many naturally produced glycoproteins, rather than only the NeuAc.alpha. 2,3Gal sequence produced by wild type CHO cells. This invention is therefore useful in the expression of recombinant glycoproteins such as erythropoietin, where the natural mixture of sialic acid linkages differs from that of the recombinant glycoprotein produced in CHO cells.

A comparison of the carbohydrate structures which result from addition of terminal sialyl acid groups in endogenous CHO cells and CHO cells transfected with the cDNA clone expressing the .alpha. 2,6 sialyltransferase is set forth in FIG. 2. In FIG. 2, GlcNAc=N-acetylglucosamine; Gal=galactose; Man=mannose; SA=sialic acid (N-acetyl neuraminic acid); and Asn=asparagine. As can be seen, the terminal SA groups in the endogenous CHO cells are only attached by .alpha. 2,3 linkages. However, the CHO cells which are modified in accordance with the present invention also produces .alpha. 2,6 terminal SA linkages.

The ratio of the NeuAc.alpha. 2,6Gal and NeuAc alpha 2,3Gal linkages can now be controlled by controlling the level of expression of the .beta.-galactoside .alpha. 2,6 sialyltransferase. To this end an expression vector has been constructed placing the sialyltransferase under the control of the metallothionein promotor in a plasmid (pMT010/A.sup.+) containing the DHFR gene.(11) When transfected into cells, this vector allows a twofold control on the level of expression. The first is by induction of the metallothionein promotor with metal ions, and the second by amplification of the gene by selection with methotrexate. Such vectors will allow amplification of expression 100 fold over that obtained in the transfected cell lines examined to date.(11) Thus, a simple alteration of the procedure described above, will allow additional control over the terminal sialic acid linkages of CHO cells.

The general applicability of the procedure of the present invention is limited only by the availability of cDNA's coding for glycosyltransferases making terminal structures not normally found on the target cells, and the availability of an appropriate lectin or carbohydrate specific antibody capable of recognizing the newly expressed carbohydrate structure on the surface of the transfected cells. Numerous specific carbohydrate lectins and monoclonal antibodies suitable for this purpose have been reported and are available.(12,13)

Having thus described exemplary embodiments of the present invention, it should be noted by those skilled in the art that the within disclosures are exemplary only and that various other alternatives, adaptations and modifications may be made within the scope of the present invention. Accordingly, the present invention is not limited to the specific embodiments as illustrated herein, but is only limited by the following claims.

(1) Takeuchi, M., et al. (1988) J. Biol. Chem 263, 3657-3663.

(2) Kagawa, Y., et al. (1988) J. Biol. Chem. 263, 17508-17515.

(3) Stanley, P. (1987) Meth. Enzymol. 138, 443-470.

(4) Ellis, L., et al. (1986) Cell 45, 721-732.

(5) Beyer, T., et al. (1981) Advances in Enzymol. 52, 24-175.

(6) Weinstein, et al., (1987) J. Biol. Chem. 17735-17743.

(7) Sanger, F. (1977) Proc. Nat. Acad. Sci. USA 74, 5463-5467.

(8) Zoller, M.J. and Smith, M. (1984) DNA 3, 479-488.

(9) Graham, F.L. and Van der Eb, A.J. (1973) Virol. 52, 456-467.

(10) Shibuya, N. (1987) J. Biol. Chem. 262, 1596-1601.

(11) Choo, K.H. et al. (1986) DNA 5, 529-537.

(12) Product catalogs of Sigma Chemical Co. and E.Y. Labs, San Mateo, Calif.

(13) Hakomori, S. (1984) In "Monoclonal antibodies and functional cell lines" (R.H. Kennett, K.B. Bechtol and T.J. McKearn) Plenum Pub. Corp. New York, pp. 67-100.