J Biol Chem, Vol. 273, Issue 52, 34843-34849, December 25, 1998

Synthesis of Poly-N-acetyllactosamine in Core 2 Branched O-Glycans
THE REQUIREMENT OF NOVEL -1,4-GALACTOSYLTRANSFERASE IV AND -1,3-N-ACETYLGLUCOSAMINYLTRANSFERASE^*

Minoru Ujita, Joseph McAuliffe, Tilo Schwientek^§, Raquel Almeida^§, Ole Hindsgaul, Henrik Clausen^§, and Minoru Fukuda^¶

From the Glycobiology Program, Cancer Research Center, The Burnham Institute, La Jolla, California 92037 and ^§ School of Dentistry, University of Copenhagen, DK-2200 Copenhagen, Denmark

	ABSTRACT

Top Abstract Introduction Procedures Results Discussion References

Poly-N-acetyllactosamine is a unique carbohydrate composed of N-acetyllactosamine repeats and provides the backbone structure for additional modifications such as sialyl Le^x. Poly-N-acetyllactosamines in mucin-type O-glycans can be formed in core 2 branched oligosaccharides, which are synthesized by core 2 -1,6-N-acetylglucosaminyltransferase.

Using a -1,4-galactosyltransferase (4Gal-TI) present in milk and the recently cloned -1,3-N-acetylglucosaminyltransferase, the formation of poly-N-acetyllactosamine was found to be extremely inefficient starting from a core 2 branched oligosaccharide, GlcNAc16(Gal13)GalNAcR. Since the majority of synthesized oligosaccharides contained N-acetylglucosamine at the nonreducing ends, galactosylation was judged to be inefficient, prompting us to test novel members of the 4Gal-T gene family for this synthesis. Using various synthetic acceptors and recombinant 4Gal-Ts, 4Gal-TIV was found to be most efficient in the addition of a single galactose residue to GlcNAc16(Gal13)GalNAcR. Moreover, 4Gal-TIV, together with -1,3-N-acetylglucosaminyltransferase, was capable of synthesizing poly-N-acetyllactosamine in core 2 branched oligosaccharides. On the other hand, 4Gal-TI was found to be most efficient for poly-N-acetyllactosamine synthesis in N-glycans. In contrast to 4Gal-TI, the efficiency of 4Gal-TIV decreased dramatically as the acceptors contained more N-acetyllactosamine repeats, consistent with the fact that core 2 branched O-glycans contain fewer and shorter poly-N-acetyllactosamines than N-glycans in many cells. These results, as a whole, indicate that 4Gal-TIV is responsible for poly-N-acetyllactosamine synthesis in core 2 branched O-glycans.

	INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Mucin-type O-glycans are present in a wide variety of cells and play various roles in different cells. Mucin-type glycoproteins are also present in the plasma membrane, and they are often involved in cell-cell interaction (1). For example, O-glycans present in eggs were shown to be a receptor for both mouse and sea urchin (2, 3). In granulocytes, monocytes, and certain T lymphocytes, mucin-type O-glycans can carry sialyl Le^x, NeuNAc23Gal14(Fuc13)GlcNAcR, at their termini (4-6). Sialyl Le^x and its sulfated form are ligands for E-, P-, and L-selectin (7-11). Importantly, these selectins, in particular P- and L-selectin, preferentially bind to sialyl Le^x in a limited number of mucin-type glycoproteins such as PSGL-1 (for P-selectin) and GlyCAM-1 and CD34 (for L-selectin) (12-14). As shown previously, sialyl Le^x and its derivatives of O-glycans in blood cells can be only formed on core 2 branches, Gal14GlcNAc16(Gal13)GalNAcR (4, 5). Recent studies demonstrate that sialyl Le^x and sialyl Le^a in core 2 branches are highly correlated to tumor invasion and vessel invasion of colon carcinomas (15), probably because tumor cells utilize selectin-carbohydrate interaction for their adhesion.

In patients with immunodeficiency such as Wiskott-Aldrich syndrome, AIDS, and leukemia, leukocytes in the peripheral blood express a substantial amount of core 2 branched oligosaccharides, while leukocytes of normal individuals do not express them (16-19). Most recent studies employing transgenic mice demonstrated that such an overexpression of core 2 oligosaccharides weakens the interactions between T lymphocytes and antigen-presenting cells or B lymphocytes, resulting in reduced immune responses such as delayed type hypersensitivity and immunoglobulin isotype switching (20, 21). It has also been shown that AIDS patients produce antibodies against leukosialin expressing core 2 branched oligosaccharides, possibly causing T lymphocyte depletion in those patients (22, 23). Moreover, the overexpression of a mucin-type glycoprotein carrying those oligosaccharides was shown to interfere with cell adhesion (24), while knockout of the leukosialin gene in mice resulted in hyperimmune responses (25). It has been also shown that poly-N-acetyllactosamine can be extended from core 2 branches, forming poly-N-acetyllactosaminyl O-glycans (4-6, 26). Poly-N-acetyllactosamine is a unique carbohydrate composed of N-acetyllactosamine repeats (Gal14GlcNAc13)_n. Poly-N-acetyllactosamines are susceptible to endo--galactosidase and larger than typical N-glycans or O-glycans containing only one N-acetyllactosamine in a side chain. Poly-N-acetyllactosamines provide the backbone structure for additional modifications, which are often cell type-specific oligosaccharides, such as sialyl Le^x (1). These results, as a whole, indicate that core 2 branched oligosaccharides play critical roles in cell-cell interaction.

These results indicate that it is crucial to understand the synthesis of core 2 branched oligosaccharide and its further extension to poly-N-acetyllactosamines. To this end, we have cloned the cDNAs encoding core 2 -1,6-N-acetylglucosaminyltransferase (C2GnT)¹ that forms a core 2 branch (27) and -1,3-N-acetylglucosaminyltransferase (iGnT) that forms poly-N-acetyllactosamine together with -1,4-galactosyltransferase (4Gal-T) (28).

When we tried to synthesize poly-N-acetyllactosamine on core 2 branched oligosaccharides, iGnT and milk 4Gal-T (4Gal-TI) failed to form poly-N-acetyllactosamines. Since the majority of the products contained N-acetylglucosamine at the nonreducing ends, inefficient galactosylation by 4Gal-TI was a likely cause for the lack of poly-N-acetyllactosamine synthesis in core 2 branched oligosaccharides. These unexpected results prompted us to test if any of the new members of 4Gal-Ts, which have been identified recently (29-33), is responsible for poly-N-acetyllactosaminyl extension in core 2 branched oligosaccharides. In this report, we summarize these findings and demonstrate that 4Gal-TIV (33) is the enzyme involved in poly-N-acetyllactosamine extension of core 2 branched oligosaccharides. Moreover, we show that 4Gal-TIV is a rate-limiting factor and responsible for short poly-N-acetyllactosamine extension in core 2 branched O-glycans.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Isolation of cDNA Encoding iGnT and iGnTc-- cDNA encoding the iGnT was isolated, as described previously (28). The cDNA insert in the cloned pAMo vector was digested with HindIII and XmnI and cloned into the HindIII and EcoRV sites of pcDNA3.1 (Invitrogen), resulting in pcDNA3.1-iGnT as described previously (28). pcDNAI-A, harboring cDNA encoding a signal peptide sequence, and the IgG binding domain of Staphylococcus aureus protein A, was constructed as described before (28). The catalytic domain of iGnT was cloned into this vector, resulting in pcDNAI-A·iGnTc (28).

Expression of the Protein A-iGnT Fusion Vector-- pcDNAI-A·iGnTc and pcDNAI-A were separately transfected with Lipofectamine (Life Technologies, Inc.) into COS-1 cells as described previously (28), and 48 h after the transfection the medium was replaced with serum-free medium, Opti-MEM (Life Technologies) and cultured for an additional 24 h. The chimeric iGnT secreted into the Opti-MEM was adsorbed into IgG-Sepharose 6FF (Amersham Pharmacia Biotech), and the enzyme bound to the beads was used as an enzyme source (34). Alternatively, the culture medium was concentrated 100-fold (for iGnT alone) or 1000-fold (for poly-N-acetyllactosamine synthesis) by a Centricon 10 concentrator (Amicon) and directly used as an enzyme source. In most of the studies, the concentrated culture medium was used, since IgG-Sepharose-bound iGnT had a low activity as seen for other glycosyltransferases (35, 36). For poly-N-acetyllactosamine synthesis, the activity of iGnT in the incubation mixtures was 380 nmol/h/ml using 0.5 mM Gal14GlcpNP as an acceptor substrate. As described previously (28), the supernatant from mock-transfected COS-1 cells contained less than one-fifth of the activity compared with that derived from pcDNAI-A·iGnTc-transfected COS-1 cells.

Isolation and Expression of cDNAs Encoding 4Gal-TII, -TIII, -TIV, and -TV-- Isolation of the cDNAs encoding 4Gal-TII, -TIII, and -TIV has been described previously (29, 33). Based on the nucleotide sequences of these cDNAs, cDNAs encoding catalytic domains of 4Gal-TII and -TIII have been prepared using RT-PCR as described before (29). The catalytic domain of 4Gal-TIV was prepared by PCR using the expressed sequence tag sequence (expressed sequence tag 489768) as a template as described previously (33). These cDNAs were cloned into pAcGP67 (Pharmingen) and expressed in insect cells as described before (33). The supernatants from these transfected insect cells were used as an enzyme source.

4Gal-TV was cloned by PCR based on the published nucleotide sequence (31). The cDNA encoding a soluble form of 4Gal-TV was prepared by PCR using the obtained cDNA as a template. 5'- and 3'-primers for this PCR were 5'-CCGGATCCCCAAGGCATTCTGATCCGGGAC-3' (BamHI site is underlined) and 5'-CCCTCGAGTCAGTACTCGTTCACCTGAGCCAG-3' (XhoI site is underlined). The resultant cDNA encoding codon 49 to the stop codon was cloned into BamHI and XhoI sites of pcDNAI-A, resulting in pcDNAI-A·4Gal-TVc. pcDNAI-A·Gal-TVc was transiently expressed in COS-1 cells as described above for pcDNAI-A·iGnTc. The supernatant from the transfected COS-1 cells was then concentrated 100-fold as described above and used as an enzyme source.

Oligosaccharides-- GlcNAc16(Gal13)GalNAcp-nitrophenol was purchased from Toronto Research Chemicals. This oligosaccharide was enzymatically converted to Gal14GlcNAc16(Gal13)GalNAcpNP as described previously (37). Briefly, GlcNAc16(Gal13)GalNAcpNP (1.41 µmol) was incubated with 494 milliunits of bovine milk 4Gal-TI (Sigma), 5.8 units of calf intestine alkaline phosphatase (Boehringer Mannheim), 3.5 µmol of UDP-galactose in 300 µl of 50 mM HEPES buffer, pH 7.5, containing 14 mM MnCl₂. After incubation for 16 h at 37 °C, the oligosaccharide synthesized was purified by a C18 reverse-phase Sep-Pak cartridge (Waters). The purity of the oligosaccharide was ascertained by thin layer chromatography using silica gel and the solvent system of dichloromethane/methanol/water (12:7:1, v/v/v). The oligosaccharides were detected by charring after spraying with 10% H₂SO₄ in ethanol. The amount of 4Gal-TI necessary in this reaction was approximately 30 times more than that for galactosylation of GlcNAc16Man16ManO(CH₂)₇CH₃ (see below).

The acceptors (Gal14GlcNAc13)_nGal14GlcNAc16 Man16Man1O(CH₂)₇CH₃(octyl), where n = 0, 1, and 2, were synthesized from octyl 2,3,4-tri-O-benzyl--D-mannopyranosyl(16)-2,3,4-tri-O-benzyl--D-mannopyranoside (compound 1) and 2,6-di-O-acetyl-3,4-di-O-chloroacetyl--D-galactopyranosyl(14)-3,6-di-O-benzyl-2-deoxy-2-phthalimido--D-glucopyranosyl trichloroacetimidate (com- pound 2) (38). Thus, the acceptor compound 1 was glycosylated with the donor compound 2 (1.2 eq) in dichloromethane at 40 °C with catalytic triethylsilyl trifluoromethanesulfonate to give the derivative of tetrasaccharide (compound 3). Selective removal of the chloroacetyl groups of compound 3 was effected with thiourea and 2,6-lutidine in ethanol, the product of which was then glycosylated with compound 2 as before to give the hexasaccharide (compound 4). Repetition of this cycle of chloroacetyl group removal, followed by glycosylation with compound 2, gave the octasaccharide (compound 5). Deprotection of compounds 3-5 was achieved using published conditions (38). The crude products, compounds 6, 7, and 8, were isolated on C18 reverse-phase silica cartridges and purified on LH-20 Sephadex. The products were characterized by ¹H NMR spectroscopy and matrix-assisted laser desorption ionization-time of flight mass spectrometry. Compounds 6, 7, and 8 contain one, two, and three N-acetyllactosamines, corresponding to those where n = 0, 1, and 2, respectively, in the above structure.

Compounds 6, 7, and 8 (4 mM) were treated with E. coli -galactosidase (10 units) in 50 mM Tris buffer, pH 7.4, at 37 °C overnight, resulting in the GlcNAc terminated compounds 9, 10, and 11, respectively. Isolation, purification, and characterization of these products was performed as described above. Detailed procedures of the synthesis will be published elsewhere.²

Partial ¹H NMR (300 MHz, CD₃OD) (compounds 6 and 9); (500 MHz, D₂O) (compounds 7, 8, 10, and 11) are as follows. Compound 6:  4.80 (d, J_1',2' = 1.5 Hz, H-1'), 4.50 (d, J_1",2" = 8.1 Hz, H-1"), 4.49 (s, H-1), 4.38 (d, J_1,2 = 8.0 Hz, H-1), 1.98 (s, 3H, NHAc). Compound 7:  4.86 (s, H-1'), 4.68 (d, J = 8.2 Hz, 1H) 4.64 (s, H-1), 4.55 (d, J = 8.0 Hz, 1H), 4.42-4.47 (m, 2H), 2.02, 2.00 (2 s, 6H, NHAc). Compound 8:  4.84 (s, H-1'), 4.62-4.66 (m, 3H), 4.52 (d, J = 8.0 Hz, 1H), 4.39-4.45 (m, 3H), 2.02, 1.99 (2 s, 9H, NHAc). Compound 9:  4.79 (d, J_{1', 2'} = 1.5 Hz, H-1'), 4.49 (d, J_1",2" = 8.1 Hz, H-1"), 4.48 (s, H-1), 2.00 (s, 3H, NHAc). Compound 10:  4.84 (s, H-1'), 4.63 (s, H-1), 4.62 (d, J = 8.0 Hz, 1H), 4.52 (d, J = 8.2 Hz, 1H), 4.41 (d, J = 7.8 Hz, 1H), 2.00, 1.98 (2 s, 6H, NHAc). Compound 11:  4.84 (s, H-1'), 4.60-4.65 (m, 3H), 4.52 (d, J = 8.0 Hz, 1H), 4.39-4.44 (m, 2H), 2.01, 1.99 (2 s, 9H, NHAc).

The addition of N-Acetylglucosamine by iGnT-- To assay the transfer of N-acetylglucosamine residues, the reaction mixture contained 5 mM UDP-[³H]GlcNAc (2 × 10⁴ cpm/nmol; NEN Life Science Products), 20 mM MnCl₂, 20 µl of iGnT preparation as described above, 10 mM N-acetylglucosamine-1,5-lactone, and various concentrations of an acceptor in 50 µl (final volume) of 100 mM cacodylate buffer, pH 7.0. As acceptors, Gal14GlcNAc16Man16Manoctyl and Gal14GlcNAc16(Gal13)GalNAcpNP were used. After incubation for 1 h at 37 °C, the reaction mixture was diluted with 5 ml of water and applied to a Sep-Pak column and washed with water. The product was eluted with 2 ml of 30% acetonitrile in H₂O, and the radioactivity of the aliquots was determined by a scintillation counter (39).

Substrate Specificity of 4Gal-TI, -TII, -TIII, -TIV, and -TV-- Assays of 4Gal-TIV were performed in a 50-µl reaction mixture containing 25 mM Tris-HCl, pH 7.5, 4 mM MnCl₂, 0.1% Triton X-100, 5 mM UDP-[³H]Gal (2 × 10⁴ cpm/nmol) (NEN Life Science Products), and an appropriate acceptor (33). For 4Gal-TI, -TII, and -TIII, 25 mM Tris-HCl, pH 7.5, containing 10 mM MnCl₂ and 0.25% Triton X-100 (29) was used, while 25 mM Tris-HCl, pH 7.0, containing 20 mM MnCl₂ and 10 mM galactono-1,5-lactone was used for 4Gal-TV (31). As an enzyme source of 4Gal-TI, human milk 4Gal-T preparation (Sigma) was directly used.

As acceptors, the following oligosaccharides were used: GlcNAc16(Gal13)GalNAcpNP for galactosylation of core 2 oligosaccharide and GlcNAc16Man16Manoctyl, GlcNAc13Gal14GlcNAc16Man16Manoctyl, and GlcNAc13Gal14GlcNAc13Gal14GlcNAc16Man16Manoctyl for galactosylation of an antennary extended from GlcNAc16Man16 branch, which is formed by N-acetylglucosaminyltransferase V. After incubation for 1 h at 37 °C, the reaction mixture was applied to a Sep-Pak column, and the amount of the product was determined by measuring the radioactivity of the eluted product as described above. To compare the enzymatic activity of different 4Gal-T samples, each enzyme preparation was first calibrated using 0.5 mM GlcNAcp-nitrophenol (Sigma) as an acceptor. This assay showed that the samples of 4Gal-TI, -TII, -TIII, -TIV, and -TV had the activity of 1.90, 1.36, 1.90, 1.27, and 0.962 (all expressed in µmol/h/ml), respectively. The final concentration of each enzyme was adjusted to 38.0 nmol/h/ml in all experiments.

Poly-N-acetyllactosamine Formation in Core 2 Oligosaccharide and N-Glycan Oligosaccharide-- To assay poly-N-acetyllactosamine formation, 0.5 mM Gal14GlcNAc16(Gal13)GalNAcpNP or Gal14GlcNAc16Man16Manoctyl was incubated with different 4Gal-Ts (760 nmol/h/ml), iGnT (380 nmol/h/ml), 5 mM UDP-[³H]GlcNAc, and 5 mM UDP-[³H]Gal as described above. There is a slight difference in the optimum incubation conditions between iGnT and various 4Gal-Ts. The optimal conditions for 4Gal-T were used, since 4Gal-T was found to be a rate-limiting enzyme.

After incubation for 10 h at 37 °C, the product was purified by a Sep-Pak column as described above. The sample was then lyophilized and subjected to HPLC using a column (4 × 300 mm) of NH₂-bonded silica (Varian Micropak AX-5) using Gilson 306. The column was eluted for 60 min with a linear gradient from a mixture of solvent A (80%) and solvent B (20%) to 100% of solvent B; solvent A is composed of 90% acetonitrile and 10% H₂O, while solvent B is composed of 40% acetonitrile and 60% 15 mM KH₂PO₄ in H₂O (40).

Galactosylation of [³H]GlcNAc16(Gal13)GalNAc with 4Gal-TI and 4Gal-TIV-- [³H]GlcNAc16(Gal13)GalNAc was synthesized by incubating 5 mM Gal13GalNAc (Sigma) with C2GnTc and 5 mM UDP-[³H]GlcNAc under the same conditions described (27). As a source for C2GnTc, pcDNAI-A·C2GnTc (27) was transiently expressed in COS-1 cells, and the spent medium obtained was used. After incubation, unreactive UDP-[³H]GlcNAc was removed by QAE-Sephadex gel as described (41). [³H]GlcNAc16(Gal13)GalNAc was then isolated after applying the sample to Bio-Gel P-4 gel filtration as described (28). The synthesized [³H]GlcNAc16(Gal13)GalNAc (0.5 mM) was incubated with 4Gal-TI or 4Gal-TIV and nonradioactive UDP-Gal under the same incubation mixture as described above, except that the final concentration of 4Gal-TI or 4Gal-TIV was 80.0 nmol/h/ml as assayed using 0.5 mM GlcNAc as an acceptor. After the incubation, UDP-Gal was removed by QAE-Sephadex gel (41), and the products were analyzed by HPLC using the same conditions described above.

Analysis of Products by Endo--Galactosidase Digestion-- Radioactively labeled products were digested with Escherichia freundii endo--galactosidase for 18 h at 37 °C (42). The digests were applied to a column (1.0 × 120 cm) of Bio-Gel P-2 (400 mesh) equilibrated with 0.1 M NH₄HCO₃.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Poly-N-acetyllactosamine Formation in Core 2 Branched O-Glycan and N-Glycan Acceptor-- Recently, we have cloned a cDNA encoding iGnT that forms poly-N-acetyllactosamine together with 4Gal-T in lacto-N-neotetraose (28). To determine if the cloned iGnT and milk 4Gal-T, 4Gal-TI, can form poly-N-acetyllactosamines on N- and O-glycans, Gal14GlcNAc16Man16Manoctyl (octyltetrasaccharide) and galactosylated core 2 oligosaccharide were used as acceptors. As shown in Fig. 1A, one to four N-acetyllactosamine repeats were added to the octyltetrasaccharide. After endo--galactosidase digestion, peak 1 produced Gal14GlcNAc13Gal, while peaks 2-4 produced Gal14GlcNAc13Gal and GlcNAc13Gal in a ratio expected from the side chains containing 2-4 N-acetyllactosamine repeats (Fig. 1, B and C). Notably, all of the products contained galactose at nonreducing termini, as seen in many cells (39, 43-46). From galactosylated core 2 branched oligosaccharide, in contrast, GlcNAc13 Gal14GlcNAc16(Gal13)GalNAcpNP (peak 1) and Gal14GlcNAc13Gal14GlcNAc16(Gal13) GalNAcpNP (peak 2) were produced (Fig. 1D). Peaks 1 and 2 yielded GlcNAc13Gal and Gal14GlcNAc13Gal after endo--galactosidase treatment, respectively, as expected from their structures (Fig. 1, E and F). In nature, core 2 branched oligosaccharides rarely contain N-acetylglucosamine at nonreducing termini and are almost exclusively terminated with galactose or sialylated galactose (4-6). Moreover, the addition of N-acetyllactosamine repeats in the core 2 oligosaccharide was only 14.3% of that in the octyltetrasaccharide, indicating that N-acetyllactosamine formation was inefficient in the core 2 oligosaccharides. To determine whether 4Gal-T or iGnT is responsible for this unexpected, inefficient synthesis of poly-N-acetyllactosamine on core 2 branches, iGnT or 4Gal-TI was incubated with appropriate acceptors. The results demonstrated that iGnT can add N-acetylglucosamine with almost equal efficiency to N-glycan acceptor and core 2 branches (Fig. 2, A and B). In contrast, 4Gal-TI exhibited substrate inhibition when the core 2 branched oligosaccharide was used as an acceptor (Fig. 2D) but not toward an N-glycan acceptor (Fig. 2C) or N-acetylglucosamine pNP (Fig. 2E). It should be noted, however, that substrate inhibition for 4Gal-TI was reported when higher than 5 mM benzyl--GlcNAc or chitobiose or chitotriose was used (29, 47). When the CHO cell lysate was used as an enzyme source, the core 2 branched oligosaccharide was galactosylated even at its high concentrations (Fig. 2F), consistent with the fact that core 2 branched O-glycans in CHO cells transfected with C2GnT are fully galactosylated (48). These results, shown in Figs. 1 and 2, combined together, indicate that another 4Gal-T, which is apparently present in CHO cells, is responsible for galactosylation of core 2 branches.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Analysis of the products after incubation of core 2 and N-glycan acceptors with iGnT and 4Gal-TI. A and D, HPLC analysis of the products derived from 0.5 mM Gal14GlcNAc16Man16Manoctyl (A) and 0.5 mM Gal14GlcNAc16(Gal13)GalNAcpNP (D), respectively. B, C, E, and F, Bio-Gel P-2 gel filtration of endo--galactosidase-digested peak 1 in A (B), peak 4 in A (C), peak 1 in D (E), and peak 2 in D (F). Peaks  and  denote the elution positions of GlcNAc13Gal and Gal14GlcNAc13Gal, respectively.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   iGnT and 4Gal-T activity on core 2 O-glycan or N-glycan acceptors with various concentrations. A and B, varying concentrations of Gal14GlcNAc16Man16Manoctyl (A) and Gal14GlcNAc16(Gal13)GalNAcpNP (B) were incubated with iGnT and 5 mM UDP-[3H]GlcNAc for 1 h. C, D, and E, varying concentrations of GlcNAc16Man16Manoctyl (C), GlcNAc16(Gal13)GalNAcpNP (D), and GlcNAcpNP (E) were incubated with 4Gal-TI and 5 mM UDP-[3H]Gal for 1 h. F, varying concentrations of GlcNAc16(Gal13)GalNAcpNP were incubated with the lysates of CHO cells and 5 mM UDP-[3H]Gal for 1 h. All products were isolated by Sep-Pak columns, and the radioactivity eluted was determined. CHO cell lysates were prepared and assayed as described before (48).

4Gal-TIV Is Responsible for Galactosylation of Core 2 Branches-- Recently, it has been demonstrated from several laboratories that there exist at least five 4Gal-Ts in addition to 4Gal-TI (29-33, 49). The above results suggest that one of these newly identified 4Gal-Ts might be involved in galactosylation of core 2 branches. To determine if any of these new members of the 4Gal-T family can form galactosylated core 2 branch, GlcNAc16(Gal13)GalNAcpNP was incubated with each of these new members of the 4Gal-T family. As shown in Fig. 3A, 4Gal-TII, -TIII, and -TV exhibited substrate inhibition toward this substrate as did 4Gal-TI. Similar results were obtained when the concentration of these enzymes was increased 5-fold or decreased 5-fold. This substrate inhibition is most likely the reason why the amount of the final product was very low in 4Gal-TI-directed reaction (see also Fig. 1D). This substrate inhibition occurs probably because -galactose in the acceptor competes with UDP-Gal.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Dependence of 4Gal-T activity on the concentration of GlcNAc16(Gal13)GalNAcpNP. GlcNAc16(Gal13)GalNAcpNP with varying concentrations was incubated with 4Gal-TI (closed circle), -TII (open triangle), -TIII (open square), or -TV (closed square) (A) or with 4Gal-TIV (open circle) (B) and 5 mM UDP-[3H]Gal for 1 h. The same amount of the enzyme, 38.0 nmol/h/ml determined using 0.5 mM GlcNAcpNP, was present in these experiments.

In contrast, the core 2 branch was efficiently galactosylated by 4Gal-TIV with a K_m of 0.29 mM (Fig. 3B; also see Table I). The identical K_m was obtained when the concentration of 4Gal-TIV was decreased 5-fold as expected. These results indicate that 4Gal-TIV is most likely responsible for galactosylation of core 2 branched oligosaccharides. 4Gal-TVI was not included in the studies, since this enzyme is the least related to 4Gal-TI and shown to synthesize Gal14Glcceramide from Glcceramide (49).

View this table: [in this window] [in a new window]   Table I Kinetic properties of 4Gal-TI and 4Gal-TIV

We then tested if 4Gal-TIV works efficiently on an N-glycan acceptor, GlcNAc16Man16Manoctyl. The results shown in Fig. 4B showed that 4Gal-TIV added galactose much less efficiently to the N-glycan acceptor than the core 2 branch acceptor. Among other 4Gal-Ts, 4Gal-TI most efficiently galactosylated the N-glycan acceptor (Fig. 4A), indicating that 4Gal-TI is most likely involved in poly-N-acetyllactosamine synthesis in N-glycans.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   4Gal-T activity on core 2 or N-glycan acceptor. A, GlcNAc16Man16Manoctyl with varying concentrations was incubated with 4Gal-TI (closed circle), -TII (open triangle), -TIII (open square), and -TV (closed square). B, various concentrations of GlcNAc16(Gal13)GalNAcpNP (solid line) and GlcNAc16Man16Manoctyl (dotted line) were incubated with 4Gal-TIV. 5 mM UDP-[3H]Gal was used in all experiments. C and D, [3H]GlcNAc16(Gal13)GalNAc was incubated with 4Gal-TI (C) and 4Gal-TIV (D), and products were analyzed by HPLC as described in Fig. 1. Peaks at fraction 33 and 38 correspond to [3H]GlcNAc16(Gal13)GalNAc and Gal14[3H]GlcNAc16(Gal13)GalNAc, respectively.

To address the question of whether the nature of the aglycon attached to the core 2 branched acceptor influences the galactosylation, [³H]GlcNAc16(Gal13)GalNAc was enzymatically synthesized using recombinant soluble C2GnT (27), UDP-[³H]GlcNAc, and Gal13GalNAc. 4Gal-TI or 4Gal-TIV was then incubated with this synthesized [³H]GlcNAc16(Gal13)GalNAc. The results, shown in Fig. 4, C and D, clearly demonstrated that the reaction by 4Gal-TI was still incomplete, since the majority of the acceptor was not galactosylated (see the peak at fraction 33). In contrast, 4Gal-TIV completely galactosylated the acceptor. These results indicate that 4Gal-TIV is the enzyme responsible for galactosylation of core 2 branches and that such a synthesis is not influenced by the nature of aglycons.

Poly-N-acetyllactosamine Synthesis by iGnT and 4Gal-TIV-- We then tested if iGnT and 4Gal-TIV together can form poly-N-acetyllactosamine on core 2 branched oligosaccharide or N-glycan oligosaccharide acceptor. As shown in Fig. 5B, the products obtained from the core 2 branched acceptor consisted of (Gal14GlcNAc13)_1 or 2Gal14GlcNAc16(Gal13)GalNAcpNP, of which structures were elucidated by endo--galactosidase digestion (data not shown). The results are also consistent with our previous finding that iGnT cannot add N-acetylglucosamine to a Gal13GalNAc side chain (28). In contrast, only Gal14GlcNAc13Gal14GlcNAc16Man16Manoctyl was produced from Gal14GlcNAc16Man16Manoctyl (Fig. 5A). These results indicate that 4Gal-TIV, together with iGnT, efficiently formed N-acetyllactosamine repeats on core 2 branched oligosaccharides but not on N-glycans.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   HPLC analysis of the products after incubation with iGnT and 4Gal-TIV. Gal14GlcNAc16Man16Manoctyl (A) or Gal14GlcNAc16(Gal13)GalNAcpNP (B) was incubated with iGnT and 4Gal-TIV and analyzed by HPLC. The product in A was eluted at the position of [3H]Gal14[3H]GlcNAc13Gal14GlcNAc16Man16Manoctyl, while the products in B contained one (eluted at fraction 34) and two (eluted at fraction 40) additional [3H]Gal14[3H]GlcNAc13 to Gal14GlcNAc16 (Gal13)GalNAcpNP.

It is noteworthy that the size of poly-N-acetyllactosamine in core 2 branched oligosaccharides shown in Fig. 5B was much smaller than that in N-glycan acceptor shown in Fig. 1A, when the same amount of the enzymes was used in both experiments. While four and possibly five N-acetyllactosamine units were added to Gal14GlcNAc16Man16Manoctyl (Fig. 1A), only two N-acetyllactosamine units were added as a maximum to Gal14GlcNAc16(Gal13)GalNAcpNP (Fig. 5B).

These results are consistent with the facts that poly-N-acetyllactosamines on core 2 branched oligosaccharides are shorter than those in N-glycans and that the majority of poly-N-acetyllactosaminyl core 2 O-glycans contains only two N-acetyllactosamine repeats in many cells (4-6, 48).

4Gal-TIV Is Responsible for Short Poly-N-acetyllactosamines in Core 2 Branched O-Glycans-- The above results suggest that 4Gal-TI and 4Gal-TIV may differ in the efficiency for adding a galactose to acceptors containing poly-N-acetyllactosamines. We were particularly interested in determining if the inefficiency in galactosylation by 4Gal-TIV can be universally observed in various acceptors containing poly-N-acetyllactosamines. To determine whether this is the case, (GlcNAc13Gal14)_nGlcNAc16Man16Manoctyl, where n = 0, 1, or 2, were used as acceptors for 4Gal-TI or 4Gal-TIV. Fig. 6B demonstrates that 4Gal-TIV substantially decreased the efficiency of the enzymatic reaction as the acceptor became longer (Table I). Such a dramatic decrease was not, however, observed for galactosylation by 4Gal-TI (Fig. 6A and Table I). These results, combined together, indicate that the intrinsic nature of 4Gal-TIV is a likely cause for shorter poly-N-acetyllactosamines in core 2 branched oligosaccharides.

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Dependence of 4Gal-TI (A) and 4Gal-TIV (B) on the concentration of acceptors containing different sizes of N-acetyllactosamine repeats. A GlcNAc16Man16Manoctyl (closed circle), GlcNAc13Gal14GlcNAc16Man16Manoctyl (open circle), and GlcNAc13Gal14GlcNAc13Gal14GlcNAc16Man16Manoctyl (closed square) were incubated with 4Gal-TI (A) or 4Gal-TIV (B) and 5 mM UDP-[3H]Gal.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In the present study, we found that 4Gal-TI, abundantly present in milk, cannot efficiently add a galactose residue to a core 2 branched oligosaccharide, GlcNAc16(Gal13)GalNAc (Figs. 1 and 2). This unexpected finding led us to discover that a novel member of the 4Gal-T family, 4Gal-TIV, is responsible for galactosylation of core 2 branched oligosaccharides and, together with iGnT, can form poly-N-acetyllactosaminyl core 2 branched O-glycans (Figs. 3-5).

After cloning 4Gal-TI from bovine and human milk (50-52), 4Gal-TI has been the only 4Gal-T recognized until recently. The presence of additional members of 4Gal-TI was, however, suggested from the knockout of the 4Gal-TI gene in mouse, since the mice deficient in Gal-TI survived during embryonic development (53). More recent accumulation in the expressed sequence tag data base, PCR homology cloning, and purification and cloning of lactosylceramide synthase allowed several laboratories to isolate novel members of the 4Gal-T family, 4Gal-TII to 4Gal-TVI (29, 31-33, 49). There appear to be overlapping but different acceptor specificities in 4Gal-TI, -TII, and -TIII, and all of them are capable of adding a galactose residue to form N-acetyllactosamine in both glycoproteins and glycolipids. Except for 4Gal-TVI, which is mainly responsible for lactosylceramide synthesis from glucosylceramide, however, the roles of the novel 4Gal-Ts have been elusive. For example, 4Gal-TV, which is remotely related to 4Gal-TI, was reported to be inactive toward a glycolipid or glycoprotein acceptor (31).

To our knowledge, the present study is the first report among novel members of 4Gal-Ts that a particular 4Gal-T, 4Gal-TIV, is exclusively responsible for forming specific structures. 4Gal-TIV is the most recently added member of the 4Gal-T gene family (33). In the present study, we demonstrated that 4Gal-TIV is very efficient in adding galactose to the core 2 branched oligosaccharide but inefficient in adding to N-glycan-related acceptors such as GlcNAc16Man16Manoctyl. This finding is consistent with the previous report that 4Gal-TIV adds very little galactose to asialoagalactotransferrin, which contains only N-glycans (33). Our results are also consistent with the report that 4Gal-TIV acts inefficiently on asialoagalactofetuin, since O-glycans in fetuin lack core 2 branches (54). It was shown in the previous report that 4Gal-TIV can act with reasonable efficiency on glycolipid acceptors (33). We were thus concerned about whether the hydrophobic nature of the aglycon, p-nitrophenol in acceptors used, might affect the enzymatic reactions. We thus used a core 2 branched acceptor without an aglycon and found that 4Gal-TIV acts on this acceptor much better than 4Gal-TI, (Fig. 4, C and D), demonstrating that 4Gal-TIV efficiently acts on core 2 branched oligosaccharide regardless of whether or not a hydrophobic aglycon is attached. These results, taken together, support our conclusion that 4Gal-TIV and iGnT cooperatively form poly-N-acetyllactosamines on core 2 branched oligosaccharides initiated by C2GnT.

The present study also demonstrated that poly-N-acetyllactosamines formed in core 2 branched oligosaccharides were shorter than those formed in N-glycan acceptors when both were synthesized under the same conditions (see Fig. 1A and Fig. 5B). This finding is consistent with the fact that core 2 branched O-glycans are mostly composed of those containing one or two N-acetyllactosamine units, while poly-N-acetyllactosaminyl N-glycans contain three or more N-acetyllactosamine units (4-6, 42-46). Moreover, the amount of poly-N-acetyllactosaminylated O-glycans is much less than that in N-glycans when N-glycans and O-glycans were analyzed in the same lamp molecules (55) or the same CHO cells (46, 48, 55). The present study also demonstrated that 4Gal-TIV, but not 4Gal-TI, drastically reduces its efficiency as acceptors become longer (Fig. 6, Table I). 4Gal-TIV was also shown to act less efficiently on longer lacto-series glycolipids than shorter ones (33). These results, combined together, indicate that