Classifications

• • C—CHEMISTRY; METALLURGY
  • C07—ORGANIC CHEMISTRY
  • C07K—PEPTIDES
  • C07K7/00—Peptides having 5 to 20 amino acids in a fully defined sequence; Derivatives thereof
  • C07K7/04—Linear peptides containing only normal peptide links
  • C07K7/06—Linear peptides containing only normal peptide links having 5 to 11 amino acids
• • C—CHEMISTRY; METALLURGY
  • C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
  • C12N—MICROORGANISMS OR ENZYMES; COMPOSITIONS THEREOF; PROPAGATING, PRESERVING, OR MAINTAINING MICROORGANISMS; MUTATION OR GENETIC ENGINEERING; CULTURE MEDIA
  • C12N9/00—Enzymes; Proenzymes; Compositions thereof; Processes for preparing, activating, inhibiting, separating or purifying enzymes
  • C12N9/10—Transferases (2.)
  • C12N9/1048—Glycosyltransferases (2.4)
  • C12N9/1051—Hexosyltransferases (2.4.1)
• • A—HUMAN NECESSITIES
  • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
  • A61K—PREPARATIONS FOR MEDICAL, DENTAL OR TOILETRY PURPOSES
  • A61K38/00—Medicinal preparations containing peptides

Definitions

• the present invention relates to glycosyltransferase enzymes and the genes corresponding to such enzymes.
• the present invention relates to the enzyme N-acetylgalactosaminyltransferase.
• the invention relates to the isolation and sequencing of the enzyme N-acetylgalactosaminyltransferase.
• the invention also relates to the construction of proteins capable of expressing the acceptor peptide for the enzyme N-acetylgalactosaminyltransferase.
• Carbohydrates are an important class of biological compounds.
• carbohydrates function as structural components where they regulate viscosity, store energy, or are key components of cell surfaces. Nearly all site specific intercellular interactions involve cell surface carbohydrates. For example, union of sperm and egg as well as the implantation of fertilized egg are both mediated by cell surface carbohydrates.
• a number of proteins that function as cell adhesion molecules including GMP-140, ELAM-1, and lymphocyte adhesion molecules like Mel- 14, exhibit structural features that mimic lectins, and are thought to bind specific cell surface carbohydrate structures (Stoolman, Cell (1989) 56:907-910). Glycosylated proteins as tumor-associated antigens are now being used to identify the presence of numerous carcinomas. Even isolated oligosaccharides have been found to exhibit biological activity on their own.
• oligosaccharides have an influence on the protein or lipid to which they are conjugated (Rademacher et al., Ann. Rev..).
• oligosaccharides have been shown to influence proteins, stability, rate of proteolysis, rate of in vivo clearance from the bloodstream, thermal stability and solubility. Changes in the oligosaccharide portion of cell surface carbohydrates have been noted in cells which have become cancerous. Other oligosaccharide changes have been detected during cell differentiation (Toone et al., Tetrahedron Report (1989) 45(17):5365-5422). As such, the significance of oligosaccharides to biological function cannot be understated.
• O-glycosidically linked (mucin type) oligosaccharides have been reported on a number of different types of glycoproteins (Sadler, (1984) Biology of Carbohydrates. (Ginsburg and Robbins, eds.) pp. 199-213, Vol. 2, John Wiley and Sons, New York). These structures have been assigned a diverse array of functions, ranging from quite specific such as being involved in cell-cell recognition and host-pathogen interaction, to more general such as providing protection from proteolytic degradation or supplying the appropriate charge and water binding properties to mucous secretions (Sadler (1984) Biology of Carbohydrates (supra); Paulson (1989) Trends Biochem. Sc , 14:272-275; and Jentoft (1990) Trends Biochem. Sci.. 15:291-294).
• the initial reaction in O-linked oligosaccharide biosynthesis is the transfer of an N-acetylgalactosamine residue from the nucleotide sugar UDP-N- acetylgalactosamine to a serine or threonine residue on the protein acceptor.
• This reaction which can occur post-translationally, is catalyzed by UDP- GalNAc olypeptide, N-acetylgalactosaminyltransferase (hereinafter referred to as GalNAc-transferase or GalNAcT) an intracellular membrane bound enzyme believed to be localized in the secretory pathway.
• GalNAc-transferase The exact location(s) of GalNAc-transferase is still controversial. It has been reported that the initial addition of N-acetylgalactosamine to the acceptor protein can take place early (even co-translationally) in the rough endoplasmic reticulum (ER). Other authors have suggested that this reaction is a post-translational event occurring in later ER compartments and/or in the cis region of the Golgi complex (e.g. Hanover et al. (1982) J. Biol. Chem. 257:10172-10177; Roth (1984) J. Cell Biol. 98:399-406; Elhammer and Kornfeld (1984) J. Cell Biol. 98:327-331; Tooze et al. (1988) J. Cell Biol. 106:1475-1487; Deschuyteneer et al. (1988) J. Biol. Chem.
• Enzyme-mediated catalytic synthesis would offer dramatic advantages over the classical synthetic organic pathways, producing very high yields of carbohydrates (e.g., oligosaccharides and/or polysaccharides) economically, under mild conditions in aqueous solutions, and without generating notable amounts of undesired side products.
• carbohydrates e.g., oligosaccharides and/or polysaccharides
• Such enzymes which include glycosyltransferase, are however difficult to isolate, especially from eukaryotic, e.g., mammalian sources, because these proteins are only found in low concentrations, and tend to be membrane- bound.
• the acceptor (peptide) specificity of GalNAc-transferase is poorly understood.
• acceptor site for this enzyme consists of acidic amino acids closely followed by the tetrapeptide Ser-Gly-Xaa-Gly, where Xaa may be any amino acid (Bourdon et al. 1987).
• the present invention is based upon the discoveries of the gene coding for the enzyme N-acetylgalactosaminyltransferase, the amino acid sequence of the enzyme N-acetylgalactosaminyltransferase, and the polypeptide sequence ofthe acceptor peptide for the enzyme N-acetylgalactosaminyltransferase. These discoveries allow for the control of glycosylation of a protein.
• the present invention involves controlling the glycosylation of a protein, either within a cell or in vitro, by introducing into the DNA sequence encoding the protein at least one gene which is capable of expressing the acceptor peptide for the enzyme N-acetylgalactosaminyltransferase, expressing a protein having an acceptor cite for that enzyme, and exposing the expressed protein to that enzyme.
• the present invention involves introducing into the DNA sequence encoding the protein a DNA sequence encoding an N- acetylgalactosaminyltransferase enzyme acceptor peptide having an amino acid sequence as follows: PPDAATAAPL [SEQ ID NO:20] wherein Proline is P, Aspartic Acid is D, Alanine is A, Threonine is T, and Leucine is L.
• the present invention also involves expressing a protein having a PPDAATAAPL [SEQ ID NO:20] acceptor cite for that enzyme, and exposing the expressed protein to that enzyme.
• the present invention also provides a process for altering the glycosylation of a protein produced by a cell where the process involves introducing into the cell at least one gene which is capable of expressing the enzyme N- acetylgalactosaminyltransferase followed by expressing a sufficient amount of the enzyme in the cell to thereby alter the glycosylation of the protein in the cell.
• N-terminal Requence of bovine colostrum GalNAc-transferase sequence of oligonucleotide nrimers. restriction map for cDNA clones (pCRl000-91B and PCR1000-52A) containing the GalNAc-transferase and the sequencing strategy.
• A N-terminal amino acid sequence (34 amino acids) [SEQ ID NO:l] obtained from purified bovine colostrum GalNAc-transferase.
• oligonucleotides A, B and C are 512, 64 and 64, respectively.
• B Nucleotide sequence of the region surrounding the EcoRI cloning site of the ⁇ gt 10 vector. Oligonucleotides F and G [SEQ ID NOS: 7 and 8, respectively] were synthesized and used in PCR reactions with the bovine small intestine cDNA library cloned in ⁇ gtlO [SEQ. ID NO: 16] (see text).
• C Restriction map of cDNA clones pCR1000-91B and pCR1000-52A.
• the protein coding region of the GalNAc- transferase protein is represented by the open box, the noncoding regions by the straight solid line and vector sequences by a solid box.
• the arrows beneath the 9 IB clone and above the 52A clone indicate the direction and extent of sequencing of the clones.
• Figure 3 Amino acid sequence [SEQ ID NO:91 of the cloned GalNAc-transferase inferred from the nucleotide sequence of cDNA clones 91B ISEQ ID NO:101 and 52A.
• the proposed transmembrane sequence is indicated by the solid boxed residues.
• N-linked glycosylation Potential sites for N-linked glycosylation are indicated by the dashed boxed residues and predicted sites for O-linked glycosylation are marked with a dot under the appropriate amino acid.
• the N-terminus of the soluble bovine GalNAc-transferase (determined by N-terminal sequencing) is indicated by the arrow.
• the consensus poly A+ sequence (AATAAA) is indicated with a solid box and the sequence of the 93 bp insert of pCR1000-93I and the 621 bp insert of pCRl000-600 are indicated by the dashed underline (931) or solid underline (600).
• the numbering of the nucleotide (upper) [SEQ ID NO: 10] or amino acid sequence (lower) [SEQ ID NO:9] is indicated to the right of the sequence.
• the first ATG codon obtained from the 9 IB clone [SEQ ID NO: 10] represents the beginning of the 1680 base pair nucleotide sequence for GalNAc-transferase [SEQ ID NO:ll].
• Figure 4 Predicted transmembrane domain and O-linked glvcosvlation sites for the cloned GalNAc-transferase.
• FIG. 6 Immunoprecipitation of in vivo 35 S-methionine labeled GalNAc-transferase expressed in baculovirus infected Sf9 cells.
• the cloned GalNAc-transferase DNA was expressed in Sf9 cells using a baculovirus vector.
• the infected cells were switched to culture medium containing 35 S-methionine 24 hours post-infection and harvested after another 24 hours.
• the cells were lysed in a detergent containing buffer and the labeled transferase was immunoprecipitated from the cell lysates and the corresponding culture media.
• the washed irnmunoprecipitates were separated by SDS-PAGE on a 10% polyacrylamide gel.
• Lanes 1, 3 and 5 contain radioactivity precipitated from cell lysates of cells infected with virus containing the constructs GalNAcT 2-l.A, GalNAcT 2-l.B and CMV Pol-1, respectively.
• Lanes 2, 4 and 6 contains radioactivity immunoprecipitated from the corresponding culture media. The two molecular mass forms of the immunoprecipitated protein is indicated by the arrow heads. The migration of molecular weight markers is indicated to the right.
• A Human granulocyte-macrophage colony-stimulating factor.
• B Human choriogonadotropin ⁇ -chain.
• C Subtilisin BPN'.
• D Bovine cytochrome C
• A Bovine rhodanese.
• B Chimeric protein constructed from the first two domains of human CD4 and the last three domains oi Pseudomonas exotoxin.
• C Human LDL receptor protein.
• D Human Alzheimer amyloid protein precursor.
• Figure 9. Lineweaver-Burk plots of GalNAc-transferase reaction velocities.
• FIG. 12 The domain structure of bovine UDP-GalNAc:polvpeptide. N- acetvlgalactosaminvltransferase: construction of the secreted, soluble enzvme.
• GalNAcT denotes the full-length transferase; the domain structure of the molecule is high-lighted by the symbols described in the key.
• GalNAcTs denotes the soluble fusion molecule; the melittin signal sequence and 5 amino acids forming the linkage between the signal sequence and the GalNAc-transferase sequence, are represented by the solid bar. The arrow indicates the signal peptidase cleavage site.
• Figure 13 The domain structure of bovine UDP-GalNAc:polvpeptide. N- acetvlgalactosaminvltransferase: construction of the secreted, soluble enzvme.
• GalNAcT denotes the full-length transferase; the domain structure of the molecule
• sequences coding for the cytoplasmic and membrane spanning domains of the full-length cDNA were replaced with sequences that code for the honeybee melittin signal peptide and five linker amino acids (78 nucleotides) [SEQ ID NO: 18].
• the honeybee melittin signal sequence was chosen since the intended expression system for the construct was baculovirus/Sf9 cells.
• Figure 14 Separation of soluble GalNAc-transferase on SDS-polvacrvlamide electrophoresis. Silver staining detected only one protein band on the 10% polyacrylamide gel. A molecular mass of approximately 61 kDa could be detected by Coomassie Blue staining.
• Figure 15 The nucleotide seouence of UDP-GalNAc:polvpeptide. N- acetylgalactosaminyl-transferase.
• the depicted nucleotide sequence [SEQ ID NO: 11] codes for the enzyme N-acetylgalactosaminyltransferase.
• Figure 16 The aminp acid sequence pf UDP-Ga_NAc;pQl ⁇ peptide.
• N- acetylgalactosaminyl-transferase The amino acid sequence of the enzyme N- acetylgalactosaminyltransferase [SEQ ID NO:9] is depicted.
• Figure 17. An amino acid sequence of a soluble form UDP-GalNAc:polvpeptide. N- acetylgalactosaminyl-transferase. The amino acid sequence of a secreted form of the enzyme N-acetylgalactosaminyltransferase [SEQ ID NO: 19] is depicted.
• Figure 18 GalNAc-transferase reaction velocity plot. The transfer of 3 H-acetylgalactosamine to the acceptor peptides by soluble GalNAc-transferase was assayed as outlined in Materials and Methods.
• the synthetic acceptor peptide of the instant experiment was Pro-Pro-Asp-Ala-Ala- Thr-Ala-Ala-Pro-Leu (PPDAATAAPL) [SEQ ID NO:20].
• N-acetylgalactosaminyl transferase GalNAcT
• GalNAcT N-acetylgalactosaminyl transferase
• Fig. 15 [SEQ ID NO: 11] and the amino acid sequence depicted in Fig. 16 [SEQ ID NO:9]. This definition is intended to encompass natural allelic variations in the
• GalNAct sequence and all references to GalNAcT, and nucleotide and amino acid sequences thereof are intended to encompass such allelic variations, both naturally- occurring and man-made.
• Cloned genes of the present invention may code for the GalNAcT enzyme of any species of origin, but preferably code for enzymes of mammalian, most preferably bovine, origin.
• Probes may be labeled with a detectable group such as a fluorescent group, a radioactive atom or a chemiluminescent group in accordance with known procedures and used in conventional hybridization assays.
• a detectable group such as a fluorescent group, a radioactive atom or a chemiluminescent group in accordance with known procedures and used in conventional hybridization assays.
• GalNAcT gene sequences may be obtained by use of the polymerase chain reaction (PCR) procedure, with the PCR oligonucleotide primers being produced from the GalNAcT gene sequence provided herein. See U.S. Patent Nos. 4,683,195 to Mullis et al. and 4,683,202 to Mullis.
• the GalNAcT enzyme may be synthesized in host cells transformed with vectors containing DNA encoding the GalNAcT enzyme.
• a vector is a replicable DNA construct. Vectors are used herein either to amplify DNA encoding the GalNAcT enzyme and/or to express DNA which encodes the GalNAcT enzyme.
• An expression vector is a replicable DNA construct in which a DNA sequence encoding the GalNAcT enzyme is operably linked to suitable control sequences capable of effecting the expression of the GalNAcT enzyme in a suitable host. The need for such control sequences will vary depending upon the host selected and the transformation method chosen.
• control sequences include a transcriptional promoter, an optional operator sequence to control transcription, a sequence encoding suitable mRNA ribosomal binding sites, and sequences which control the termination of transcription and translation.
• Amplification vectors do not require expression control domains. All that is needed is the ability to replicate in a host, usually conferred by an origin of replication, and a selection gene to facilitate recognition of transformants.
• Vectors useful for practicing the present invention include plasmids, viruses
• a useful vector is a baculovirus expression vector.
• the vector replicates and functions independently of the host genome, or may, in some instances, integrate into the genome itself. Suitable vectors will contain replicon and control sequences which are derived from species compatible with the intended expression host.
• Transformed host cells are cells which have been transformed or transfected with the GalNAcT enzyme constructed using recombinant DNA techniques. Transformed host cells ordinarily express the GalNAcT enzyme, but host cells transformed for purposes of cloning or amplifying the GalNAcT enzyme DNA need not express the GalNAcT enzyme. When expressed, the GalNAcT enzyme will typically be located in the host cell membrane.
• DNA regions are operably linked when they are functionally related to each other.
• a promoter is operably linked to a coding sequence if it controls the transcription of the sequence.
• a ribosome binding site is operably linked to a coding sequence if it is positioned so as to permit translation.
• operably linked means contiguous and, in the case of leader sequences, contiguous and in the same translational reading frame.
• Expression vectors for such cells ordinarily include (if necessary) an origin of replication, a promoter located upstream from the gene to be expressed, along with a ribosome binding site, RNA splice site (if intron-containing genomic DNA is used), a polyadenylation site, and a transcriptional termination sequence.
• the transcriptional and translation control sequences in expression vectors to be used in transforming vertebrate cells are often provided by viral sources.
• promoters are derived from polyoma, Adenovirus 2, and Simian Virus 40 (SV40).
• SV40 Simian Virus 40
• the early and late promoters of SV40 are useful because both are obtained easily from the virus as a fragment which also contains the SV40 viral origin of replication.
• the GalNAcT enzyme promoter, control and/or signal sequences may also be used, provided such control sequences are compatible with the host cell chosen.
• An crigin of replication may be provided either by construction of the vector to include an exogenous origin, such as may be derived from SV40 or other viral source, or may be provided by the host cell chromosomal replication mechanism. If the vector is integrated into the host cell chromosome, the latter may be sufficient.
• GalNAcT enzyme made from cloned genes in accordance with the present invention may be used for designing new compounds containing oligosaccharides for a variety of healthcare and industrial applications.
• host cells may be transformed with a vector of the present invention, GalNAcT enzyme expressed in that host, the cells lysed, and the enzyme isolated from the lyzed cells.
• the enzyme can then be used in vitro to begin the initial reaction in the O-linked oligosaccharide biosynthesis of the transfer of an N-acetylgalactosamine residue from the nucleotide sugar UDP-N-acetylgalactosamine to a serine or threonine residue on the protein acceptor.
• Cloned genes and vectors of the present invention are useful in molecular biology to transform cells which do not ordinarily express the GalNAcT enzyme to thereafter express this enzyme. Such cells are useful as intermediates for producing the enzyme. Such cells are also useful for the in vivo biosynthesis of an O-linked oligosaccharide to a protein acceptor.
• Milk (and colostrum) contains a number of glycosyltransferase activities (e.g. Prieels et al, 1975; Paulson et al, 1977; Bushway et al, 1979; Parodi et al, 1984).
• bovine colostrum contains what appears to be a soluble form of N-acetylgalactosaminyl transferase (GalNAcT) (Elhammer and Kornfeld, 1986) but did not provide a procedure for the purification of sufficient amounts of GalNAcT for N-terminal sequencing. The following procedure describes the purification of GalNAcT from bovine colostrum.
• the amino acid sequence of the enzyme is determined by N-terminal sequencing. This information is then used to isolate a cDNA clone encoding a full- length (membrane bound) transferase which upon expression in the insect cell line Sf9 resulted in the synthesis of a fully active enzyme.
• the acceptor specificity of the enzyme is then determined using a semiquantitative analysis of the amino acids surrounding known glycosylation sites in 16 different proteins followed by in vitro glycosylation studies of synthetic peptides.
• [ ⁇ - 32 P]dATP 300 Ci/mmol
• UDP-[lJH]N-acetylgalactosamine 8.3 Ci/mmol
• Na[ 125 I] (15.2 mCi/ ⁇ g)
• [ ⁇ - 33 P]dATP is from NEN/Dupont and 35 S- methionine is from ICN (Trans S-35 label, 1 mCi ml),.
• Bovine colostrum is obtained from a local farmer.
• UDP-N-acetylgalactosamine UDP, PMSF, chymostatin, leupeptin, antipain, pepstatin, aprotinin, bovine submaxillary mucin, Nonidet P-40 (NP-40), Triton X-100, taurodeoxycholate, Sephadex G-100 Superfine, rabbit anti- chicken IgG antibodies, ATP, myelin basic protein, subtilisin, rhodanese and cytochrome C (reduced and carboxymethylated as described by Heinrikson, R.L., 1973) are from Sigma. DEAE-Sephacel, Sepharose 6B and Protein A-Sepharose are from Pharmacia. IODOGEN is from Pierce.
• N-glycosidase F is from Oxford Glycosystems. Geneamp Kit (for PCR) is obtained from Perkin Elmer/Cetus. A bovine small intestine cDNA library cloned in a ⁇ gt 10 vector is purchased from Clontech (catalog # BLIOlOa). The TA cloning vector pCRlOOO is from Invitrogen. Sequenase version 2.0 is from U.S. Biochemical Corp. The baculoGold transfection kit is from PharMingen. 1 cc Bond Elut C lg columns were from Varian. Serum-free Grace's insect medium, Insect Express, was from BioWhitaker.
• the vector pVt-Bac was a gift from Dr. Thierry Vernet at the Biotechnology Research Institute, National Research Council of Canada. Patella vulgata ⁇ -N-acetylgalactosaminidase is from V-Labs, Inc. Restriction enzymes and all other reagents are from standard sources. In addition, the following buffers are used. Buffer A: 25 mM Imidazole, pH 8.
• buffer B 25 mM imidazole, pH 7.2, 1 M NaCl, 1% Triton X-100, 20 mM EDTA; buffer C: 25 mM Imidazole, pH 7.2, 30 mM MnCl 2 , 20 - mM NaCl; buffer D: 25 mM Imidazole pH 7.2, 0.5 M NaCl, 20 mM EDTA; buffer E: 25 mM Ir- ' ..
• Example 1 Isolation of N-acetvlgalactosaminvltransferase from Bovine Colostrum
• the first four steps in the purification of the transferase are identical to the procedure described by Elhammer and Kornfeld (1986) (which is herein incorporated by reference) except that the samples loaded on the affinity columns are adjusted to 1 mM ATP (in addition to the reported buffer, salt and UDP concentrations) to compensate for an apparently higher pyrophosphatase activitydes) in the colostrum used. Equilibration, loading, washing and elution buffer volumes are adjusted (scaled up) for the larger columns used. All steps in the purification procedure are performed at +4°C and enzyme activity is assayed with the following standard assay throughout the purification.
• the standard assay for UDP-GalNAc:polypeptide, N-acetylgalactosaminyl- transferase activity during purification contained the following components in a final volume of 80 ⁇ l: 50 mM Imidazole pH 7.2, 10 mM MnCl 2 , 0.5% Triton X-100, 15 ⁇ M UDP-GalNAc, UDP[1- 3 H-] GalNAc (27,000 cpm/assay), 0.15 mg/ml apomucin and varying amounts of enzyme (see individual experiments).
• the reaction mixture is incubated at 37°C for 5-10 minutes (see individual experiments) and the reaction product is TCA precipitated and radioactivity measured as described.
• the supernatant from the 100,000 g centrifugation is loaded directly on a DEAE- Sephacel column equilibrated in buffer A.
• the bed volume of this column should be approximately equal to the amount of 100,000 g supernatant loaded
• Apomucin deglycosylated mucin
• bovine submaxillary mucin by the method of Hagopian and Eylar with minor modifications.
• the carbohydrate content ofthe apomucin preparation is determined by the method of Reinhold.
• CNBr-activated Sepharose is prepared from Sepharose 6B essentially as described by Cautrecasas.
• the apomucin is coupled to the activated Sepharose in 0.1 M sodium carbonate buffer pH 9.2 at 4°C overnight. The protein concentration during the reaction is 2.5 mg/ml.
• the columns are run by gravity at a pressure of -30 cm H 2 0 during loading and -60 cm H 2 O during washing, elution and regeneration.
• the column Before loading, the column is washed with 400 ml buffer B (regeneration buffer) followed by 500 ml buffer C and 150 ml buffer C containing 0.25 mM UDP.
• the sample Prior to loading the column the sample ( -200U enzyme activity per 50 ml column in the first affinity step) is supplemented with MnCl 2 and UDP to final concentrations of 30 mM and 1.25 mM, respectively.
• the column is washed with 4 column volumes of buffer C containing 0.25 mM UDP and six 40 ml fractions are collected. The column is then eluted with buffer D.
• fractions 1 and 2 25 ml each, normally contains no, or very little activity
• fractions 3 and 4 50 ml each contains the bulk of the activity
• fractions 3 and 4 50 ml each, contains the bulk of the activity
• fractions 5 through 7 25 ml each, contains in some cases smaller amounts of activity.
• the individual fractions are dialyzed against 4 liters of buffer E (2 changes) immediately after elution, and assayed for enzyme activity. Typically only fractions 3 and 4 are used in the subsequent purification.
• Step 4 Apomucin affinity chromatography II
• the same type column is used as in the previous one.
• the column is first washed with 400 ml buffer B followed by 500 ml buffer F and 150 ml buffer F, containing 0.25 mM UDP.
• dialyzed fractions 3 and 4 from step 3 are supplemented with 1 M MnCl 2 , 4 M NaCl and UDP to achieve final concentrations of 30 mM, 100 mM and 1.25 mM respectively.
• Step 5 Gel filtration chromatography on Sephadex G-100 superfine
• the dialyzed fractions from three step 4 runs are pooled, 1/50 volume 5% taurodeoxycholate is added, and the material is concentrated to 2.5 ml on an Amicon YM-10 filter under 40 psi pressure.
• Half of this material, 1.25 ml is loaded on a Sephadex G-100 Superfine column (20-50 ⁇ m bead size; 1.5 x 100 cm) equilibrated in buffer G having 300 mM NaCl.
• the column is run at a pressure of 30 cm H 2 O, which resulted in a flow of approx. 2.3 ml/hour and fractions (100 total) are collected at 40 min.
• the fractions comprising the activity peak are pooled and concentrated as described above but without any further addition of detergent.
• Analytical gel filtration to determine the molecular weight ofthe transferase is carried out using the same procedure but with a smaller column (0.9 x 100 cm) and collecting 1.06 ml fractions. The recoveries from this step using the conditions described above typically ranged from 80-90%.
• the purified GalNAcT preparation contains only one polypeptide, with a molecular mass of approximately 70 kDa, detectable with silver staining (Figure LA). A portion ofthe purified preparation is labeled in vitro with 125 I and separated on SDS- PAGE before and after digestion with peptide N-glycosidase F. Figure IB shows that this treatment results in an approximately 6 kDa shift in the apparent molecular mass of the protein.
• N-terminal sequencing of the purified bovine colostrum GalNAcT is done by automated Edman degradation in an Applied Biosystems Sequencer (Model 470) fitted with an on-line HPLC analyzer (Model 120-A) for phenylthiohydantoins. Quantitation of the latter is afforded by the Nelson Analytical Turbochrom chromatography data system connected in parallel with the recorder to the output from the HPLC system. The 34 amino acid sequence is shown in Figure 2A [SEQ ID NO:l].
• Example 3 Isolation and Characterization of cDNA Clones Encoding Bovine GalNAc- Transferase
• Oligonucleotide primers are synthesized based on the partial N-terminal amino acid sequence of the purified bovine colostrum enzyme with an Applied Biosystems
• oligonucleotide (oligos A-E) [SEQ ID NOS: 2-6, respectively,] sequence of the primers and probes used in the Polymerase Chain Reaction (hereinafter referred to as PCR) and later in a Southern Blot analysis are shown in Fig. 2A below the GalNAcT amino acid sequence.
• PCR Polymerase Chain Reaction
• Fig. 2A The degeneracy of oligonucleotides A, B and C are 512, 64 and 64, respectively.
• the PCR is carried out in 0.1 ml of solution containing 50 mM KCl, 10 mM Tris-HCL pH 8.3, 1.5 mM MgCl 2 , 0.2 mM each of the four dNTP's, 1 ⁇ M of each oligonucleotide, either 5 ⁇ l of the bovine intestine cDNA library or 10 ng of plasmid or ⁇ DNA and 2.5 units of Taq polymerase.
• the reaction is covered with 0.1 ml of mineral oil and subjected to a temperature step cycle. When degenerate oligonucleotides are used the steps are 94°C (1 min), 37°C (2 min), 72°C (3 min) for a total of 35 cycles.
• oligonucleotides For nondegenerate oligonucleotides the steps are 94°C (1 min), 55°C (2 min), 72°C (3 min) for a total of 25 cycles. Standard DNA manipulations are performed as described in Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecular Cloning.
• the cDNA encoding the GalNAcT gene is cloned using the following approach. Oligonucleotides A [SEQ ID NO:2] and C [SEQ ID NO:4] are used as opposing primers in a PCR reaction. A bovine small intestine cDNA library cloned into a ⁇ gtlO vector is used as the template for the reaction. On the basis of the amino acid sequence, the predicted size ofthe amplified PCR product is 93 bp. The products ofthe PCR reaction are analyzed by Southern blot analysis using ohgonucleotide B [SEQ ID NO:3] as a probe ( Figure 2A).
• oligonucleotide primers D-G [SEQ ID NOS: 5-8, respectively] are synthesized. Oligonucleotides D [SEQ ID NO:5] and E [SEQ ID NO:6] are derived from the sequence ofthe pCRlOOO-931 insert and F [SEQ ID NO:7] and G [SEQ ID NO:8] are primers that directly flank either side of the EcoRI cloning site of ⁇ gtlO ( Figure 2B).
• PCR reactions are run using the bovine cDNA library as template with oUgonucleotides D+F or D+G as primers.
• the resulting PCR products are analyzed by Southern blot analysis using oligonucleotide E [SEQ ID NO:6] as a probe.
• the 621 bp insert contains a 207 amino acid open reading frame with the first 23 amino acids of that open reading frame being a perfect match to amino acids 12-34 of the purified protein ( Figure 2A) [SEQ ID NO:l].
• the 621 bp fragment contains a portion ofthe GalNAcT gene
• this fragment is labeled with [ ⁇ - 32 P]dATP by nick translation (Goldin et al, 1981) and is used as a probe to screen the bovine cDNA library.
• the cDNA library (containing 2.5 x IO 6 independent clones) is screened by plaque hybridization using the above labeled DNA fragment as a probe. Seven positive plaques are obtained from the primary screen and each isolate is plaque purified three times. Five of the seven isolates are found to contain inserts of 600 bp or smaller while the two remaining isolates contain inserts of approximately 1600 and 2300 bp.
• the two larger inserts are PCR amplified and cloned (using oUgonucleotides F and G as primers) into the TA cloning vector to yield pCR1000-52A (1600 bp insert) and pCR1000-91B (2300 bp insert).
• the size ofthe ⁇ inserts are analyzed on 1% agarose gels following restriction digest with EcoRI or by PCR using oUgonucleotides F and G as primers.
• Example 4 DNA Sequence Analysis of PCR Inserts and Predicted Amino Acid Sequence
• the inserts in pCR1000-93I, pCR1000-600, pCR1000-91B (2294 bp) and pCRl000-52A (1582 bp) are sequenced by the dideoxy chain termination method (Sanger et al. , 1977) using Sequenase version 2.0 with [ ⁇ - 33 P]dATP. Double stranded DNA sequencing (Ausubel et al. 1987) is done with 20-mer ohgonucleotide primers, synthesized according to the sequence of the cDNA insert. The sequencing strategy is shown in Figure 2C. Sequence analysis is performed using the Sequence Analysis software package of the University of Wisconsin Genetics Computer Group (Devereux et al, 1984).
• the first ATG codon ofthe sequence obtained from the 9 IB clone is present at nucleotide 53.
• the sequence of the 52A clone demonstrated that it is a truncated version of the 9 IB clone in that the sequence of this clone starts at nucleotide 162 and ends at nucleotide
• the 52A insert covers nearly all of the open reading frame sequences (missing codons for the first 37 amino acids) found in the 9 IB clone.
• nucleotide sequence of the 52A clone is identical to the 9 IB clone with the exception that nucleotide 358 is a G in the 52A clone instead of an A. This base change is in the wobble position (AGA. to AGGJ of codon 102 so it does not alter the arginine at that position.
• the 3'-untranslated region of the 9 IB clone is 562 bp in length, contains a consensus polyadenylation signal (nucleotides 2176-2182) and a track of 25 A residues at the end ofthe clone (Fig. 3), indicating that the 91B clone contains aU the 3' terminal sequences of the GalNAcT mRNA [SEQ ID NO: 10].
• MDBK cells bovine mammary tissue and various human tissues is analyzed by Northern blot analysis using the 600 bp insert of pCR 1000-600 as a hybridization probe.
• Total RNA and poly A* RNA is prepared from bovine mammary tissue and from MDBK using the Invitrogen Fastrack kit, following the manufacturers procedure. Two ⁇ g of poly A * RNA are denatured by glyoxylation and Northern blot analysis is performed as previously described (Homa et al. , 1986).
• a human multiple tissue Northern blot (Clontech (Cat # 7760-1)) is prehybridized in 50% formamide, 5 x SSC, 1 x Denhart's, 1% SDS, 100 ⁇ g per ml denatured salmon testes DNA, at 42°C for 2 h and then hybridized overnight at 42°C with the 32 P-labeled 600 bp insert isolated from the pCRlOOO-600. Filters are washed three times for 15 min in 0.1 X SSC, 0.1% SDS at 55°C As shown in Figure 5, at least two different sized
• GalNAcT mRNA's are detected from all the samples.
• the size of the bovine messages are approximately 4.1 and 3.2 kb, while all the human tissues express messages of 4.8 and 3.9 kb.
• a third mRNA of approximately 1.5 kb is detected in the skeletal muscle sample.
• the putative GalNAcT coding region, pCR1000-9lB is digested with Sstll and Hindlll (both enzymes cut only in pCRlOOO sequences that flank the insert; Figure 2C) and these sites are blunted using T4 DNA polymerase so that it can be cloned into a baculovirus expression vector.
• BamHI linkers are then ligated onto the blunted ends and the resulting sample is ligated into the BamHI site of the baculovirus expression vector pAC373 (Summers and Smith, 1986).
• the resulting isolates are screened for proper orientation of the GalNAcT open reading frame with respect to the baculovirus polyhedron promoter, to yield pAC373-GalNAcT.
• Cotransfection of Sf9 cells with pAC373-GalNAcT and linearized baculovirus DNA from PharMingen's baculoGold transfection kit is performed using calcium phosphate precipitation (Summers & Smith, 1986).
• the baculovirus DNA provided in the PharMingen transfection kit contains a lethal mutation that can be corrected by homologous recombination with sequences contained in the pAC373 vector. Therefore, following transfection, only recombinant viruses wiU grow on Sf9 cells.
• GalNAcT 2-1A and GalNAcT 2-1B Transfections are done in duplicate and the resulting virus samples are referred to as GalNAcT 2-1A and GalNAcT 2-1B.
• Cells are harvested 48 hours post infection and lysed in a detergent containing buffer. Following sedimentation of undissolved material, the cleared lysates are assayed for GalNAcT activity. Lysates from uninfected cells or from cells infected with either a baculovirus containing an unrelated gene, CMV-POL (human cytomegalovirus DNA polymerase gene), or two separate baculovirus isolates of the GalNAcT gene, GalNAcT 2-1A and GalNAcT 2- IB, are assayed.
• CMV-POL human cytomegalovirus DNA polymerase gene
• the baculovirus expressed protein is further examined by immunoprecipitation and SDS-PAGE analysis.
• Baculovirus infected cells are labeled from 24 to 48 hours postinfection with [ 35 S]methionine.
• GalNAcT is immunoprecipitated from lysates and culture media ofthe labeled cells using a chicken polyclonal antibody raised against the purified bovine colostrum enzyme.
• a chicken is injected with 100 ⁇ g purified enzyme axillary, intramuscularly (with Freund's complete adjuvant).
• One month later the chicken is boosted with another 50 ⁇ g antigen subcutaneously (with Freund's incomplete adjuvant); a second booster, 50 ⁇ g enzyme axillary, intra-muscularly, is administered after an additional 21 days.
• Test bleeds are done two weeks after each booster. After the second test bleed (which upon analysis is found to contain anti- GalNAcT antibodies) eggs are coUected each day and used as a source for antibodies.
• IgG is isolated from egg yolk as described by Jensenius et al., 1981.
• Immunoprecipitation ofthe in vivo 35 S-methionine labeled enzyme is done from crude cell lysates. Infected cells are labeled between 24-48 hours postinfection with 50 ⁇ Ci/ml 35 S-methionine in medium that contains one tenth the normal methionine concentration. Approximately 1.5 X 10° labeled, infected cells are dissolved in 670 ⁇ l PBS containing 0.5% Triton X-100, 0.5% taurodeoxycholate, 0.05% SDS, 0.1 TlU/ml of Aprotinin and 10 ⁇ g/ml each of leupeptin, antipain, chymostatin and pepstatin.
• any undissolved debris is sedimented at 10,000 x g for 20 minutes and the supernatant is collected.
• Immunoprecipitation is carried out by the addition of 4 ⁇ l (approximately 20 ⁇ g chicken IgG) of chicken anti GalNacT antibodies; purified IgG isolated from egg yolk is used for all immunoprecipitation experiments.
• the antigen- antibody complexes are isolated by over night adsorption to 22 ⁇ l (volume of sedimented gel) of protein A-Sepharose coated with rabbit anti-chicken IgG antibodies.
• the coated protein A-Sepharose is prepared by incubating 330 ⁇ l sedimented protein A-Sepharose with 2.3 mg rabbit anti-chicken IgG antibodies (an affinity purified IgG fraction) in 1 ml of PBS over night; the coated protein A-Sepharose is washed three times with 1 ml PBS containing 0.5% Triton X-100, 0.5% taurodeoxycholate, 0.05% SDS. Following adsorption of the antigen, the immunosorbent is sedimented by centrifugation and washed extensively essentially as described by Dunphy et al. (1985).
• the washed antigen-antibody-immunosorbent complexes are suspended in 50 ⁇ l SDS-PAGE sample buffer (Laemmli, 1970) and heated for five minutes on a boiUng water bath to release the bound antigen. Following sedimentation of the protein A-Sepharose the antigen containing supernatants are aspirated and loaded on SDS-PAGE. SDS-PAGE, and fluorography of the dried gels is done as described previously (Davis et al. , 1986) (Fig. 6).
• the soluble bovine colostrum enzyme is the result of proteolytic cleavage of a membrane bound molecule or if it represents a bona fide secretory protein.
• Soluble, enzymaticaUy active forms of a ⁇ l-4 galactosyltransferase and a ⁇ 2-6 sialyltransferase have been reported, both of which appear to be the result of proteolytic cleavage of membrane bound proteins (Paulson and CoUey, 1989 and references therein).
• the translation products from the different mRNA species related to both these molecules appears in most tissues to be membrane bound molecules (Joziasse, 1992).
• the larger sizes ofthe two GalNAcT messages are presumably related to untranslated sequences larger than those recovered in the isolated clones, in the 5' and/or 3' ends of the native molecules.
• Messenger RNA molecules from previously characterized cloned glycosyltransferases frequently contain extensive 5' and 3' untranslated sequences (e.g. Weinstein et al, 1987; Larsen et al, 1989; Russo et al, 1990; Scocca et al, 1990; Sarkar et al, 1991; Nagata et al, 1992).
• these two proteins are not known at present; they may represent different glycoforms of the enzyme or, perhaps more likely, the lower molecular mass form may be a proteolytic fragment, similar to the enzyme purified from bovine colostrum.
• the latter possibiUty is supported by two observations: 1), the mass difference between the two molecules is roughly equal to that ofthe sequence (40 amino acids) missing in bovine colostrum enzyme and 2), while the irnmunoprecipitates from cell lysates contains predominantly the higher molecular mass form ofthe enzyme, the culture medium appears to be enriched in the lower mass form. High-speed centrifugation of the culture medium failed to sediment more than approximately 30% ofthe enzymatic activity (Data not shown).
• the smaller molecular mass of the insect cell produced molecule as compared to the predicted mass of a membrane bound bovine enzyme may be the result of differences in glycosylation of the two molecules.
• Insect cells typically synthesize truncated, non-sialylated N- and O-linked oUgosaccharides (e.g. Hsieh and Robbins, 1984; Domingo and Throwbridge, 1988; Kuroda et al, 1990; Thomsen et al, 1990; Wathen et al, 1991; Chen and Bahl, 1991); this results in a reduced molecular mass of insect cell produced glycoproteins on SDS- PAGE. The identity of higher molecular mass bands, approximately 120-180 kDa, on the gel is not clear.
• Example 8 Construction and Expression in Sf9 cells of a Soluble GalNAc-transferase (GalNAcTs)
• GalNAcTs Soluble GalNAc-transferase
• the sequences coding for the cytoplasmic and membrane spanning domains of the full- length cDNA were replaced with sequences that code for the honeybee melittin signal peptide (Fig. 12).
• the honeybee melittin signal sequence was chosen since the intended expression system for the construct was baculovirus/Sf9 cells.
• the plasmid pAC373-GalNAcT (Homa et al., 1993) which contains the full length GalNAc-transferase gene under the control ofthe baculovirus polyhedron promoter was digested with Xbal and Bglll, which generated a 150 bp fragment, and with Bglll and Xhol, which generated a 9700 bp vector fragment. Both fragments were gel purified.
• the Xbal site used is located 7 amino acids from the N-terminus of the soluble colostrum enzyme, in a portion of the molecule corresponding to what is referred to as the "stem region" in other glycosyltransferases (reviewed by Shaper & Shaper, 1992).
• 2100 bp fragment generated by this digest was gel purified.
• the three gel purified fragments were added to the same tube and ligated.
• the resulting plasmid contains a GalNAc-transferase gene under the control of the baculovirus polyhedron promoter in which the first 47 amino acids (141 nucleotides) have been replaced with 21 amino acids (63 nucleotides) ofthe honeybee meUttin signal peptide plus five (5) amino acids (15 nucleotides) that link the two domains together (Fig. 13) [SEQ ID NO: 18].
• GalNAcTs-Mel in Sf9 cells resulted in 130-fold increase in GalNAc-transferase activity in the culture medium, as compared to uninfected cells (Table 2) or cells infected with an unrelated molecule ( ⁇ 6-3). This is more than 35 times the amount recovered in the medium of cells expressing the full length molecule (Homa & Elhammer, unpublished observations). A significant portion (36%) ofthe total enzymatic activity resulting from expression of the soluble molecule was, however, retained inside the cells; the reason for this is not clear at present.
• Example 9 Isolation and Characterization of GalNAcTs
• the bound enzyme was eluted from the column with 720 ml Buffer D and collected in seven 80 ml fractions. Run-through, wash, and eluted fractions were all dialyzed against Buffer E containing 300 mM NaCl (three changes) prior to assay for GalNAc- transferase activity. The recovery of enzyme activity on the column was invariably over 90%. The following concentration of the eluted enzyme, however, led to significant losses in activity. In fact this step accounted for the largest losses in the preparation procedure. Dialyzing the enzyme into a buffer containing 300 mM NaCl prior to concentration was an absolute necessity to avoid even higher losses in this step. The enzyme isolated from bovine colostrum shows a similar behavior in this regard (Elhammer & Kornfeld, 1986). The purified preparation was concentrated by ultrafiltration on a YM-10 membrane at 45 psi pressure.
• the crude medium sample was precipitated as described by Wessel and Flugge ( 1984) prior to electrophoresis; precipitate corresponding to approximately 250 ⁇ l medium was loaded.
• NH 2 -terminal sequencing ofthe purified molecule was done as described in Example 2. Interestingly, this purification procedure yielded a homogenous preparation only if expression of the molecule was carried out in serum- free medium.
• the efficient production ofthe cloned molecule using the baculovirus expression system facilitated preparation of GalNAc-transferase in amounts sufficient for detailed biochemical and enzymatic studies to determine the acceptor substrate specificity of GalNAc-transferase from a database of in vivo substrates and from the in vitro glycosylation of proteins and peptides. These studies have been facilitated by the avaUabiUty of information regarding the presence of glycosylated serine and threonine residues in proteins obtained during protein sequencing. This information is registered in the NBRF protein sequence repository.
• a search of the NBRF protein database yields several hundred definite or probable Thr and Ser O-glycosylation sites. From these, only those with reasonably unambiguous assignments are chosen and all proteoglycans are excluded since they contain primarily glycosaminoglycan chains where the anchoring sugar is xylose and not GalNAc. Also included into the reference set are the O-glycosylation sites identified
• glycosylation sites themselves show no homology.
• the complete reference set consists of the 196 glycosylated peptide segments (shown in Table 1 in Elhammer et al., 1993).
• the glycosylated peptides are listed as enneapeptide (ennea Greek, nine) segments, with the reactive Ser or Thr in the central position, designated as PO. Accordingly, the amino acid side chains toward the N-terminus are designated as the subsites Pl to P4 and those toward the C-terminus as subsites Pl' to P4'.
• a length of nine residues is chosen as a starting point, with the option that, depending on the results on the selectivity of the subsites, the portion of the peptides subject to analysis may be extended or truncated.
• the sequences show that besides the obvious need for Ser or Thr in PO, no other subsite has an absolute requirement for any given amino acid. This then suggests that specificity of the enzyme may be the result of the cooperation of several subsites, none of them essential, but all of them contributing to catalytic efficiency.
• PS is the abundance of glycosylatable peptides in all proteins and RP is the cumulative probability calculated as the product of all relevant s ⁇ values:
• a Ser or Thr-containing peptide may be predicted to be a substrate for the enzyme if the probabUity, h, is higher than a certain cutoff value, h c .
• a given peptide is predicted to be a glycosyl acceptor, if
• the probability pattern of these two proteins is shown in Figure 7. It is interesting to note that the calculated probabilities for these two proteins are not distributed uniformly between the two extremes. Rather, a small number of residues are associated with very high probabilities whereas the rest of the sequence indicates uniformly low probabilities. Furthermore, the residues with high probabilities are clustered into one or two distinct segments where the clustering of Ser and Thr residues may perhaps be a necessary but certainly not a sufficient criterion for creating a highly glycosylated protein segment.
• subtilisin BPN' (subsn.aa, Figure 7C) which is produced by a microbial system incapable of O-linked oUgosaccharide biosynthesis, contains a number of randomly distributed potential glycosylation sites, while very few nonglycosylated mammalian proteins contain any potential glycosylation sites. It is perhaps more typical to find no potential glycosylation sites at all, as in the case of horse hemoglobin (hbho.aa) or that of bovine cytochrome C (ccpg.aa, Figure 7D).
• this region of the protein consists of a fully exposed segment linking the two homologous domains ofthe enzyme, exposure of native or mildly denatured rhodanese to GalNAc-transferase should result in glycosylation ofthe molecule.
• the chimeric protein, CD4PE40 constructed from two domains of the human CD4 protein and three domains of the Pseudomonas exotoxin, shows two prominent potential glycosylation sites, both at regions linking individual domains, see Figure 8B. In the same vein, one would predict that subtiUsin could also be extensively glycosylated, if not in the native form then at least after mild denaturation.
• the potential of the predictive method is perhaps best illustrated by its application to the LDL receptor (ldlrec.aa, Figure 8C) and the Alzheimer precursor protein (alz.aa, Figure 8D), which have both been shown to be extensively O-glycosylated, each in a known, narrow segment of the polypeptide.
• the present method not only correctly identifies these regions of glycosylation but also specifically predicts which Ser and Thr residues may be modified.
• the above analysis allows one to hypothesize about the saUent features of the enzyme active site responsible for the specificity of glycosylation. Table 4 indicates that high selectivity is expressed at aU subsites, but only toward Ser, Thr, and Pro.
• the selectivity of a given subsite depends on how many times more frequent are at that site the surabundant residues than all the other amino acids. Also, selectivity is higher when the surabundant residue is one which occurs with low frequency in globular proteins.
• S j a specificity parameter for the subsite i, S j , as the number of surabundant residues found at that site, divided by the number of these same residues expected at that site from random distribution. This ratio is then multiplied by the fraction of surabundant residues at that site.
• S j The values of S it reported in Table 3, suggest that the binding site extends at least from P3 to P4' and perhaps even P4 is included in the substrate-enzyme interactions.
• the preferred conformations appear to be a random coil, a sharp bend, or a ⁇ -strand from P4 to PO followed by a turn.
• the enzyme does not require a preformed secondary structure but imposes one upon binding of the substrate.
• the hydration index of the amino acids in the potential glycosylation sites also shown in Table 6, indicates that most peptides are reasonably exposed to the aqueous environment.
• reaction products are characterized using alkaline sodium borohydride treatment essentially as described by Carlson (1968).
• Digestion with Patella vulgata ⁇ -N-acetylgalactosaminidase (approximately 1 unit ml) is done in 25 mM citrate buffer pH 4.0 in a final volume of 30 ⁇ l for 24 hours.
• Released radioactive sugars are separated on descending paper chromatography in pyridine-ethyl acetate-glacial acetic acid-water (5:5:1:3; v:v:v:v).
• Table 7 shows that, as predicted, both bovine rhodanese and, to a lesser extent, the bacterial protein subtihsin do indeed function as acceptors for the enzyme, although neither of them reacts unless reduced and carboxymethylated prior to exposure to the enzyme.
• bovine cytochrome C which contains one Ser and eight Thr residues but no predicted potential sites, is not an acceptor for the enzyme, whether in the native, or in the reduced and carboxymethylated state.
• Myelin basic protein a molecule which previously has been shown to be an efficient acceptor for GalNAc-transferase (Hagopian et al., 1971) is included as a positive control in this experiment.
• Rhodanese contains two additional predicted acceptor sites, Ser 142 and Ser 6 , (Fig. 8). However, due to the low rates of transfer to serine residues under our standard assay conditions, transfer to these sites should not contribute significantly to the total transfer in the assay.
• the lower rate of transfer to reduced and carboxymethylated rhodanese compared to that of myelin basic protein, may be related to incomplete exposure of the acceptor sites even by the reduction and carboxymethylation procedure, and/or differences in rate constants between the acceptor sequences on the two molecules.
• Myelin basic protein contains one site predicted with high probability and three additional low probability sites. The molecule can reportedly be glycosylated with 1.2 to 1.5 N-acetylgalactosamines per molecule (Cruz and MoscareUo, 1983).
• the bacterial protein subtilisin contains four predicted serine sites with probabilities higher than 0.6 (Fig. 8). Three of the serines have a high exposure index in the native protein (Kabsch and Sander, 1983), but the three-dimensional structure of the protein indicates that the hydroxyls are located in a restrained environment. Again, this could account for the need for reduction and carboxymethylation for acceptor activity.
• the 35 times slower transfer rate to denatured subtUisin, as compared to myelin basic protein indicates again a slower transfer to serines than to threonines, under the conditions used. Factors such as those discussed for rhodanese may also contribute to the low levels of transfer.
• cytochrome C which does not contain any predicted acceptor site, is completely inactive as an acceptor, whether in the native or in the reduced and carboxymethylated form.
• the ability of GalNAc-transferase to glycosylate a series of synthetic acceptor peptides is shown in Figure 9 and Table 8.
• PPAdSTdSAPG Pro-Pro-Ala-D-Ser-Thr-D-Ser-Ala-Pro-Gly
• the t-Boc-amino acids and the PAM resin solid supports are supplied by AB
• the completed peptides are removed from the supporting resin, concurrently with the side chain-protecting groups, by a standard HF cleavage procedure using anisole as a cation scavenger (10% v/v).
• the crude peptides are purified by preparative reverse phase chromatography on a C18 Vydac column (2.5 x 30 cm) using a water/acetonitrile gradient, each phase containing 0.1% TFA.
• Each purified peptide is characterized by FAB MS and shows a single symmetrical peak on analytical HPLC.
• glycosylated amino acids in the peptides PPASTSAPG [SEQ ID NO: 14] and PPASSSAPG [SEQ ID NO : 15] are identified by sequencing ofthe reaction products from the corresponding assay.
• Fig. 11 shows that for both glycosylated peptides the majority of the sugar-Unked radioactivity is associated with residue #5, the central amino acid, be it threonine, as in PPASTSAPG [SEQ ID NO: 14], or serine, as in PPASSSAPG [SEQ ID NO: 15].
• the measurable amounts of radioactivity associated with the residues following residue 5 are presumably due to the large load of peptide in the sequencer necessitated by the low specific radioactivity of the sample. Nevertheless, since the radioactivities associated with residues 7 and 8 extrapolate smoothly to that of residue 6, it is most likely that, within our experimental error, residue 6 is not labelled.
• N-acetylgalactosamine to peptide acceptors is assayed by two different assays.
• concentration of UDP-GalNAc is saturating in all assays; a !__, of 8 ⁇ M is reported for bovine colostrum GalNAc-transferase (Elhammer and Kornfeld,
• reaction mixture contains 50
• reaction product glycosylated peptide
• UDP-GalNAc a product that is separated from unreacted UDP-GalNAc by chromatography on Dowex-2 columns (0.5 ml bed volume) equilibrated in water; the run-through fraction (2.5 ml) containing the glycosylated peptide is collected, supplemented with scintillation fluid and counted for radioactivity.
• the assay conditions are as follows: 50 mM Imidazole, pH 7.2, 10 mM MnCl 2 , 0.5% Triton X-100, 150 ⁇ M UDP-GalNAc, approximately 260,000 cpm UDP-[ 3 H]-GalNAc and 3.2 mM acceptor peptide (the concentration of RSPPP is 3.7 mM).
• the assays are incubated for 20 minutes (PPASTSAPG) [SEQ ID NO: 14] or 8 hours (PPASSSAPG [SEQ ID NO: 15], PPAdSSdSAPG and RSPPP [SEQ ID NO: 13]).
• the enzyme is inactivated by placing the samples on a boiUng water bath for 1.5 minutes. The samples are then allowed to cool and the reaction products are separated from unreacted UDP-GalNAc and free GalNAc by chromatography on a Biogel P-2 column (1 X 50 cm) equiUbrated in 7% isopropanol; thirty 1.3 ml fractions are coUected.
• the peptide PPASTSAPG [SEQ ID NO: 14] is designed to contain a single Thr, at PO.
• the proline residues at P4, P3 and P3 1 provide maximum probabiUties at those positions; serine residues at Pl and Pl' result in good probabilities without much steric constraint.
• FinaUy, the alanine residues at P2 and P2' and the glycine at P4 are indifferent as to the probabUity of glycosylation but allow for flexibUity of the peptide backbone.
• Tables 8 and 9 show that this peptide is the most efficient of the acceptors tested and comparative assays show that its reactivity is very close to that of bovine apomucin (data not shown). Furthermore, the kinetic parameters for the two peptides, determined under our conditions, are quite comparable to those of the purified porcine submaxiUary GalNAc-transferase-catalyzed glycosylation of peptides whose structure is derived from sites identified in porcine submaxiUary mucin (Wang et al., 1992).
• the peptide RTPPP [SEQ ID NO: 12] derived from the major acceptor sequence in myelin basic protein (Hagopian et al.,1971), has a ___, lower than that of PPASTSAPG [SEQ ID NO: 14] but also a much lower V majl and, hence, its catalytic efficiency is only half of that of PPASTSAPG [SEQ ID NO: 14].
• the activities of the two corresponding peptides containing serine instead of threonines are measurable but too low for
• bovine colostrum GalNAc-transferase is capable of transferring GalNAc to the serine of these peptides, albeit — under our in vitro conditions — approximately 35 times slower than to threonine (Table 9).
• bovine colostrum GalNAc-transferase In contrast to the enzyme recently purified from porcine submaxiUary gland (Wang et al., 1992), however, the bovine colostrum GalNAc-transferase is definitely capable of glycosylating both threonine and serine residues.
• bovine colostrum GalNAc-transferase in experiments reported by O'Connel et al. (1992), bovine colostrum GalNAc-transferase faUed to glycosylate serine in a peptide derived from human erythropoietin. This phenomenon may be related to the specific acceptor peptide used, even though a serine in this position is glycosylated in vivo.
• Example 13 Transfer of N-acetvlgalactosamine to Synthetic Acceptor Peptides bv Soluble GalNAc-Transferase
• the abiUty of GalNAcTs to glycosylate a series of synthetic acceptor peptides was also studied.
• Assays for the determination of kinetic parameters for peptide acceptors were carried out as described by Elhammer et al. (1993) but using a modification ofthe method described by O'Connel and Tabak (1993) for isolation ofthe acceptor peptides: One ml Bond Elut columns containing 100 mg packing material were used. Before loading the assay samples, the columns were washed with 2 ml methanol followed by 2 ml 0.1% TFA (in water).
• the assay samples (40 ⁇ l) were diluted to 1 ml with 0.1% TFA and loaded on the columns. Unbound radioactivity was than washed out with 4 ml 0.1% TFA, after which the glycosylated acceptor peptides were eluted with 1.5 ml 35% acetonitrile, 0.1% TFA (in water), directly into scintillation vials. Calculation of kinetic parameters was done from double reciprocal plots (Uv versus 1 S) using standard procedures.
• the Km for UDP-GalNAc is approximately 1.7 ⁇ M and the Km:s for the threonine containing acceptor peptide PPASTSAPG [SEQ ID NO: 14] and the serine containing acceptor peptide PPDAASAAPLR [SEQ ID NO: 17] are approximately 6.5 and 3.6 mM, respectively.
• Transfer by GalNAcTs to another serine containing acceptor peptide PPASSSAPG [SEQ ID NO: 15] is approximately 70 times slower than to PPASTSAPG [SEQ ID NO: 14] (Data not shown).
• the specific activity ofthe purified enzyme preparation, using bovine apomucin as acceptor, is approximately 2,160 U/mg protein (Table 4).
• the enzymatic properties of the purified GalNacTs appear to be similar to the those determined for the enzymes purified from bovine colostrum and porcine submaxiUary gland (Elhammer and Kornfeld, 1986; Elhammer et al., 1993; Wang et al., 1992; Wang et al., 1993).
• the Km for the acceptor peptide PPASTSAPG [SEQ ID NO: 14] is almost identical for the colostrum enzyme and the baculo expressed molecule, 6.0 vs. 6.5 mM.
• the serine containing acceptor peptide PPDAASAAPLR [SEQ ID NO: 17] (O'Connel et al.
• GalNAcTs and the bovine colostrum enzyme glycosylate the peptide PPASSSAGP [SEQ ID NO: 15] are at least 35 times slower than PPASTSAPG [SEQ ID NO: 14] (Elhammer et al., 1993).
• the Km for UDP-GalNAc in assays using GalNAcTs is lower than those determined for the bovine colostrum and the porcine submaxiUary gland enzymes, 1.7 uM vs. 8 ⁇ M and 6 ⁇ M, respectively (Table V; Elhammer & Kornfeld, 1986; Wang et al., 1992). The reason for this is not clear at present.
• the amino acid sequence of the bovine colostrum and the cloned molecules should be identical except for five amino acids in the NH 2 -terminal end of the molecule; the colostrum enzyme sequence also contains an additional two amino acids at the NH 2 -terminus.
• differences in post-translational processing, in particular glycosylation, of the Sf9 produced vs. the bovine molecule may, to some extent, influence the kinetic characteristics of the two molecules.
• the oUgosaccharide structures on the insect produced molecule are most likely of high mannose and/or truncated high mannose type, while results from endoglycosidase digestion experiments suggest that the colostrum enzyme contains complex type oligosaccharides (Elhammer & Kornfeld, 1986). It is likely that the N-linked oUgosaccharide structures on the porcine enzyme also would be of the types normally synthesized by mammalian cells; peptide: N-glycosidase F digestion experiments suggest that this molecule contains 9 kDa of N-linked oUgosaccharides (Wang et al., 1992). Further experiments will however be needed to clarify this question.
• the ability of soluble GalNAc-transferase to glycosylate another synthetic acceptor peptide is shown in Figure 18.
• the synthetic acceptor peptide Pro-Pro-Asp- Ala-Ala-Thr-Ala-Ala-Pro-Leu (PPDAATAAPL) [SEQ ID NO:20] was synthesized by solid phase methodology as described in Example 12. Sequence analysis for the identification of the glycosylated amino acid(s) in the acceptor peptide PPDAATAAPL [SEQ ID NO:20] was also performed as described in Example 12. Experiments (data not shown) demonstrated that the incorporated radioactivity in the acceptor peptide PPDAATAAPL [SEQ ID NO:20] is in the form of N-acetylgalactosamine. Further, the glycosylated amino acids in the peptide PPDAATAAPL [SEQ ID NO:20] were identified as decribed in Example 12.
• the peptide PPDAATAAPL [SEQ ID NO:20] has a lower K_, but a similar V mll _.
• the synthetic acceptor peptide PPDAATAAPL [SEQ ID NO:20] has a much higher catalytic efficiency than that of PPASTSAPG [SEQ ID NO: 14].
• CeUs (1X10 6 ) were infected with recombinant virus containing GalNAcTs-Mel (5 pfu/ceU). The ceUs were harvested 65 hours post infection, lysed in a detergent containing buffer and the GalNac-transferase activity was determined in the cell lysates and the corresponding culture media; lysate and culture medium from uninfected ceUs were assayed as control. The numbers have been adjusted for differences in protein content in the cell lysates; the volume ofthe culture media was 5 ml.
• *1 unit equals one mole N-acetylgalactosamine transferred to apomucin per minute, under assay conditions.
• Surabundant amino acids surrounding the reactive Ser or Thr Surabundance at a given subsite for a given amino acid is expressed as the number of that amino acid found at the site in excess to that expected from random distribution, divided by the S.D. of the expected distribution.
• the excess of surabundant residues is equal to or higher than twice the S.D. of the expected residue
• N-acetylgalactosamine to protein acceptors was assayed under standard conditions (see Material and Methods).
• the acceptor concentration was 65 ⁇ M
• the enzyme concentration approximately 65 mU/ml
• assay time was 60 minutes.
• the transfer to both native and reduced-carboxymethylated acceptors was assayed.
• Bovine cytochrome C ⁇ 0.1 ⁇ 0.1
• PPASSSAPG [SEQ n.a. n.a. *8.5 ID NO:15]
• RSPPP [SEQ ID n.a. n.a. »0.4 b NO: 13]
• Assays were done as described in Materials and Methods. The products were separated by Biogel P-2 chromatography. Assay times were 20 minutes for PPASTSAPG [SEQ ID NO: 14] and 8 hours for PPASSSAPG [SEQ ID NO: 15], PPAdSTdSAPG and RSPPP [SEQ ID NO: 13].
• Glu lie Gly Thr Tyr Asp Ala Gly Met Asp Ile Trp Gly Gly Glu Asn 305 310 315 320
• GCACAAACTC CAATGCAGAC CATTCTCTTG GTACCTAGAG AATATTTATC CTGATTCTCA 1320
• GATTCCTCGT CACTATTTCT CTTTGGGAGA GATACGAAAT GTGGAAACAA ATCAGTGTCT 1380
• AAAAAAAAAA AAAA 2294 INFORMATION FOR SEQ ID NO: 11:
• GATTACTTTC AGGAAATTGG AACATATGAT GCTGGAATGG ATATTTGGGG AGGAGAAAAC 960
• AAGCTCAACT TTCGCTGGTA TCCTGTTCCC CAAAGAGAAA TGGACAGAAG GAAAGGTGAT 780
• Cys Pro Ile lie Asp Val Ile Ser Asp Asp Thr Phe Glu Tyr Met Ala 195 200 205

Landscapes

• Chemical & Material Sciences (AREA)
• Life Sciences & Earth Sciences (AREA)
• Organic Chemistry (AREA)
• Health & Medical Sciences (AREA)
• Genetics & Genomics (AREA)
• General Health & Medical Sciences (AREA)
• Zoology (AREA)
• Medicinal Chemistry (AREA)
• Molecular Biology (AREA)
• Engineering & Computer Science (AREA)
• Biochemistry (AREA)
• Wood Science & Technology (AREA)
• Bioinformatics & Cheminformatics (AREA)
• Microbiology (AREA)
• General Engineering & Computer Science (AREA)
• Biotechnology (AREA)
• Biomedical Technology (AREA)
• Biophysics (AREA)
• Proteomics, Peptides & Aminoacids (AREA)
• Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
• Enzymes And Modification Thereof (AREA)

Applications Claiming Priority (3)

Publications (1)

Family

ID=21713649

Family Applications (1)

Country Status (4)

Families Citing this family (2)

Family Cites Families (1)

• 1996
  • 1996-09-09 EP EP96930677A patent/EP0854882A1/de active Pending
  • 1996-09-09 JP JP9515030A patent/JPH11514232A/ja active Pending
  • 1996-09-09 WO PCT/US1996/014136 patent/WO1997013783A1/en not_active Application Discontinuation
  • 1996-09-09 AU AU69641/96A patent/AU6964196A/en not_active Abandoned

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events