OUP user menu

Identification and characterization of a gene required for α1,2-mannose extension in the O-linked glycan synthesis pathway in Schizosaccharomyces pombe

Yuka Ikeda, Takao Ohashi, Naotaka Tanaka, Kaoru Takegawa
DOI: http://dx.doi.org/10.1111/j.1567-1364.2008.00458.x 115-125 First published online: 1 February 2009


The KTRα1,2-mannosyltransferase gene family of Saccharomyces cerevisiae is responsible not only for outer-chain modifications of N-linked oligosaccharides but also for elongation of O-linked mannose residues. To identify genes involved in the elongation step of O-linked oligosaccharide chains in Schizosaccharomyces pombe, we characterized six genes, omh1+–omh6+, that share significant sequence similarity to the S. cerevisiae KTR family. Six deletion strains were constructed, each carrying a single disrupted omh allele. All strains were viable, indicating that none of the omh genes was essential. Heterologous expression of a chitinase from S. cerevisiae in the omh mutants revealed that O-glycosylation of chitinase had decreased in omh1Δ cells, but not in the other mutants, indicating that the other omh genes do not appear to be required for O-glycan synthesis. Addition of the second α1,2-linked mannose residue was blocked in omh1Δ cells. An Omh1–GFP fusion protein was found to be localized in the Golgi apparatus. These results indicate that Omh1p plays a major role in extending α1,2-linked mannose in the O-glycan pathway in S. pombe.

  • Schizosaccharomyces pombe
  • O-mannosyltransferase
  • glycosylation
  • Golgi apparatus


Glycosylation of secreted proteins in eukaryotic cells is the most common form of post-translational modification, and requires the participation of many different enzymes associated with the endoplasmic reticulum (ER) and Golgi apparatus (Spiro, 2002; Willer, 2003). Glycoproteins in Saccharomyces cerevisiae contain both N-linked and O-linked oligosaccharides. In S. cerevisiae, the synthesis of O-linked oligosaccharides is initiated in the lumen of the ER by an essential family of O-mannosyltransferases (Pmt1-7p). These enzymes catalyze the transfer of mannose from dolichyl phosphate-activated mannose (Dol–P–Man) to specific serine (Ser) or threonine (Thr) residues (Strahl-Bolsinger, 1999; Girrbach & Strahl, 2003). It has been reported that the PMT family consists of three subfamilies: PMT1, PMT2, and PMT4. Members of the PMT1 subfamily interact in pairs with members of the PMT2 subfamily, whereas the PMT4 subfamily forms homomeric complexes (Girrbach & Strahl, 2003). Addition of the first mannose onto Ser/Thr is not only important for yeast cell function in S. cerevisiae and in Schizosaccharomyces pombe (Tanaka, 2005; Willer, 2005) but also in mammals (Yoshida, 2001; Beltran-Valero de Bernabe, 2002; Willer, 2002).

Therefore, the initiation of protein O-mannosylation by the PMT family is highly conserved from yeasts to multicellular eukaryotes (Strahl-Bolsinger, 1999; Girrbach & Strahl, 2003; Willer, 2003). The strong dependence on proper interactions between the protein O-mannosyltransferases for transfer of the first mannose residue suggests that the process of O-mannosylation may place special demands on the stability, proper localization, and/or proper function of the individual proteins. The KTR and MNN1 gene families from S. cerevisiae encode α1,2- and α1,3-mannosyltransferases, respectively (Lussier, 1999). Each predicted sequence has a hydrophobic transmembrane region neighboring the N-terminus and a large luminal domain, indicating that they are type II transmembrane proteins. Their topology contains an active site termed the ‘C-terminal catalytic domain,’ which is responsible for glyosyltransferase-related events in the Golgi apparatus in both N-linked outer chain and O-linked oligosaccharide biosynthesis (Lussier, 1996, 1999; Rayner & Munro, 1998; Romero, 1999). This active site is composed of a highly conserved C-terminal catalytic domain of about 300 amino acids and a metal ion bound to the so-called ‘DXD’ motif. The motif sequence is not strictly conserved and variations have been observed in some glycosyltransferases. Although the crystal structure of the Kre2p α1,2-mannosyltransferase has been reported (Lobsanov, 2004), the detailed catalytic mechanism for retaining the function of the mannosyltransferase has not been resolved. The chemical features of the interactions between the GDP-mannose (donor) or oligosaccharides (acceptor) and amino acid residues within the active site of other glycosyltransferase remain to be clarified.

Current biochemical and genetic studies indicate that S. cerevisiae O-linked oligosaccharides have a linear structure Man1−5–Ser/Thr, in which the second and third mannose residues are elongated by Kre2p, Ktr1p, and Ktr3p (Lussier, 1997), while the fourth and fifth mannose residues are elongated by Mnn1p, Mnt2p, and Mnt3p (Romero, 1999). In Candida albicans and Pichia pastoris, α1,2-linked oligomannoses of varying length lacking decoration with α1,3-linked mannose are formed (Willer, 2003). In S. pombe, UDP-glucose and UDP-galactose can be interconverted (Huang & Snider, 1995), and a UDP-galactose transporter is known to transport UDP-galactose from the cytosol to the lumen of the Golgi apparatus (Tabuchi, 1997; Tanaka & Takegawa, 2001; Tanaka, 2001).

It has therefore been reported that S. pombe O-linked oligosaccharides have as many as two galactosyl residues that can be attached to the α1,2-linked mannose backbone, giving rise to a branched Gal0−2Man1−3–Ser/Thr structure (Gemmill & Trimble, 1999). However, the enzymes involved in the synthesis and processing of S. pombe O-linked oligosaccharides have not been characterized, except for the α1,2-galactosyltransferase, Gma12p (Chappell & Warren, 1989; Chappell, 1994; Gemmill & Trimble, 1999). The number of α1,2-mannosyltransferases required for elongation of O-linked oligosaccharides in S. pombe is currently unknown.

In order to clarify the specific functions of genes encoding putative α1,2-mannosyltransferases involved in the O-linked elongation step, we isolated and characterized six S. pombe genes omh1–omh6 (O-glycoside α1,2-mannosyltransferase homologs) that share significant amino acid similarity to the S. cerevisiae KTR family. Here we report the characterization of omh1+ encoding a mannosyltransferase responsible for elongation of O-linked mannosaccharides.

Materials and methods

Strains, media, and genetic methods

Escherichia coli XL-1 Blue (Stratagene) was used for all cloning procedures. The wild-type S. pombe strains ARC039 (h- leu1 ura4) were obtained from Yuko Giga-Hama (Asahi Glass, Japan). Standard rich medium (YES) and synthetic minimal medium (MM) for growing S. pombe were used as described (Moreno, 1991). Saccharomyces cerevisiae strains were grown in rich medium, 1% yeast extract, 2% peptone, and 2% glucose (YPD). Phosphate-free MM–P medium based on MM medium contains 14.6 mM sodium acetate instead of di-sodium hydrogen phosphate and potassium hydrogen phthalate. The S. cerevisiae chitinase expression medium based on MM contains 0.1% glucose and 2% fructose instead of 2% glucose. In this study, S. pombe cells were transformed by the lithium acetate method or by electroporation as described (Okazaki, 1990; Suga & Hatakeyama, 2001; Morita & Takegawa, 2004). Standard genetic methods have been described (Alfa, 1993).

Cloning and disruption of the S. pombe omh genes

The omh1+–omh6+ genes were disrupted using ura4 as a selective marker (Grimm, 1988). Approximately 1.3-, 1.6-, and 1.25-kb DNA fragments, including the whole or part of the omh1, omh2, and omh3 ORFs, respectively, were amplified from wild-type S. pombe genomic DNA and subcloned into pGEM-T Easy and pGEM-T vectors (Promega).

To disrupt the omh1+ locus (SPBC19C7.12c), a DNA fragment carrying the omh1+ gene was amplified by PCR, using the following oligonucleotides: sense, 5′-GGAAGCCATATTCGAAGTACTATAGTG-3,′ and antisense, 5′-TCAATGGATTCATAAGGATCACAGTCGC-3′. The KpnI–EcoRI sites within the cloned omh1+ ORF were digested and a 1.6-kb ura4+ cassette was inserted. To disrupt the omh2+ locus (SPBC16H5.09c), the following oligonucleotides were used: sense, 5′-GAAGTGGATAGCGTCTACTTTTAATGGTG-3,′ and antisense, 5′-TCCCTTTCTAGGGCAATGAAGAAGAGGAGC-3′. The EcoRV–HindIII sites within the cloned omh2+ ORF were digested and ura4+ was inserted. To disrupt the omh3+ locus (SPCC777.07), the following oligonucleotides were used: sense, 5′-CTGGCGTATCTTACACAATCGAACCATG-3′, and antisense, 5′-GCTGTCGCATGCGACTGTACATCCGTAGGGC-3′. The EcoRI–XhoI sites within the cloned omh3+ ORF were digested and ura4+ was inserted. Linearized DNA fragments carrying the disrupted omh1+, omh2+, and omh3+ genes were used to transform wild-type haploid ARC039 strains, and ura+ transformants were selected. To confirm that the omh1+, omh2+, and omh3+ genes had been disrupted, respectively, ura+ transformants were analyzed by Southern blotting and PCR to verify correct integration of the deletion constructs.

A loxP-flanked ura4 cassette was used to disrupt omh4+ (SPBC1773.08c), omh5+ (SPBC32H8.08c), and omh6+ (SPAC959.04c). DNA fragments for gene disruptions were constructed using the loxP cassette vector pBS loxP–ura4–loxP (Iwaki & Takegawa, 2004) as follows: two fragments of 0.6 kb containing the upstream or the downstream region of each ORF were amplified from wild-type S. pombe genomic DNA.

To disrupt the omh4+ locus, upstream fragments were amplified using a sense primer carrying an XhoI site and an antisense primer carrying a HindIII site. Downstream fragments were amplified by PCR using a sense primer carrying an EcoRI site and an antisense primer carrying a BamHI site using the flowing primers: upstream, sense, 5′-GTTTCTCGAGATCCCAGTGATTCAGCCACC-3,′ and antisense, 5′-GTTTAAGCTTGCACTATCAGTAGTAAGACC-3′; downstream, sense, 5′-GTTTGAATTCGCCAGAGTTAAAAGCTAGGG-3,′ and antisense, 5′-GTTTGGATCCTGTTTGGCCACACCGTCACC-3′. Amplified fragments were digested with the corresponding restriction enzymes, and the XhoI–HindIII upstream fragment and EcoRI–BamHI downstream fragment were inserted into the pBS loxP–ura4–loxP cassette. To disrupt the omh5+ locus, the following oligonucleotides were used: upstream, sense, 5′-GTTTGGATCCATAATGTGTTGACTCGCAGG-3,′ and antisense, 5′-GTTTGAATTCTCTTCTATTCTCTAACGACC-3′; downstream, sense, 5′-GTTTAAGCTTAATACGGGTACTTTACGCTC-3,′ and antisense, 5′-GTTTCTCGAGCATCACATCATTAACAGGCC-3′. Amplified fragments were digested with the corresponding restriction enzymes and the BamHI–EcoRI upstream fragment and the HindIII–XhoI downstream fragment were inserted into the pBS loxP–ura4–loxP cassette. To disrupt the omh6+ locus, the following oligonucleotides were used: upstream, sense, 5′-GTTTGGATCCTGTGCATTATGAACAACCAC-3,′ and antisense, 5′-GTTTGAATTCAGAAGCACTCAAATTGGAGC-3′; downstream, sense, 5′-GTTTAAGCTTTTGATGAATGAGCTTTACAG-3,′ and antisense, 5′-GTTTGGTACCCTTTTTTGACTGCTTGTTCC-3′. Amplified fragments were digested with the corresponding restriction enzymes and the BamHI–EcoRI upstream fragment and the HindIII–KpnI downstream fragment were inserted into the pBS loxP–ura4–loxP cassette. Linearized DNA fragments carrying the upstream and downstream region flanking the loxP–ura4–loxP cassette were used to transform wild-type haploid ARC039 strains, and ura+ transformants were selected. These deletion mutants were used in the following experiments without removal of ura4+ by Cre-mediated recombination (Iwaki & Takegawa, 2004).

Growth characterization assays

Cells cultured overnight in 5 mL YES were diluted with water to an OD600 nm of 0.5, corresponding to about 107 cells mL−1. Serial 10-fold dilutions were started with about 106 cells mL−1, and 7 μL aliquots were spotted onto YES plus 20 μg mL−1 hygromycin B or prewarmed YES plates at 30 or 37 °C. Plates were incubated for 3 days and photographed.

Analysis of acid phosphatase

Acid phosphatase from S. pombe cells was analyzed as described (Huang & Snider, 1995; Yoko-o, 1998), with the following modifications. Cells were grown in 5 mL YES medium to an OD600 nm of 1.5, and 1.5 × 108 cells were centrifuged, resuspended in 5 mL of MM-P medium, and incubated for 3 h at 30 °C to induce production of acid phosphatase. The cells were then collected by centrifugation, washed once with 62.5 mM Tris-HCl, pH 6.8, and suspended in 200 μL ice-cold lysis buffer (62.5 mM Tris-HCl, pH 6.8, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride, 0.1 mM dithiothreitol, and 10% glycerol). Cell lysates were prepared by vigorously mixing the cell suspension five times with 0.5-mm glass beads for 30 s at 4 °C using a Mini BeadBeater-8 (BioSpecProducts). The lysates were centrifuged at 15 000 g for 10 min, and the supernatants were recovered and centrifuged again at 15 000 g for 20 min. The supernatants were recovered again, and protein was determined using a BCA protein assay kit (Pierce).

Thirty micrograms of protein from each sample were mixed with 1/3 volume of 0.01% bromophenol blue, 15% glycerol, and 62.5 mM Tris-HCl, pH 6.8, and were immediately subjected to electrophoresis on a 6% polyacrylamide native gel. The electrophoresis buffer and the method for staining for acid phosphatase activity were used as described (Tanaka, 2001).

Plasmid constructs

To express the S. cerevisiae chitinase Cts1 in S. pombe, the CTS1 ORF was amplified by PCR and cloned into pFML1, the glucose-repressible expression vector (Tanaka, 2005). To tag the Omh1p C-terminus with green fluorescent protein (GFP), the omh1 ORF was amplified from genomic DNA, BglII and NotI sites were introduced at the 5′ and 3′ ends, respectively, and appropriate primers were used to amplify omh1 by PCR. The resultant PCR product was digested with BglII and NotI, and cloned into the corresponding sites of pTN197 derived from pREP41 (Nakamura, 2001). To tag Gms1p with red fluorescent protein (RFP), pAU–Gms1–RFP was constructed as described (Tanaka & Takegawa, 2001).

O-glycosylation analysis of chitinase

Heterologously expressed S. cerevisiae chitinase from the supernatant of saturated cultures was isolated, and the extent of O-glycosylation was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), as described (Tanaka, 2005). Fission yeast strains transformed with pFML1–CTS1 were grown in 5 mL of MM without leucine to the stationary phase, transferred to 10 mL of chitinase expression medium, and then incubated for 30 h at 30 °C. Saccharomyces cerevisiae strains were cultured in 10 mL of YPD at 30 °C. After adding 20 mg of chitin (Sigma), the tubes were rotated end over end at 4 °C for 12 h. The chitin was then pelleted by centrifugation and washed three times with cold sodium buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl). Chitinase was eluted from the chitin in a 50-μL SDS sample buffer by boiling for 5 min and then separated in 6% SDS-PAGE. The gel was stained with Coomassie Brilliant Blue R-250.

Preparation of 2-aminopyridine (PA)-labeled oligosaccharides from the S. pombe cell wall

Cells were grown in YES or MM medium at 30 °C, and harvested in the early stationary phase. Extraction and isolation of total cell surface galactomannan from the harvested cells were carried out as described (Ballou, 1990). O-linked oligosaccharides were released from glycoproteins in the lyophilized galactomannoproteins by hydrazinolysis as described (Natsuka & Hase, 1998). Briefly, 2 mg of the sample was heated at 60 °C for 6 h with 0.2 mL of anhydrous hydrazine. When the hydrazine had evaporated, N-acetylation was initiated with acetic anhydride in a saturated sodium bicarbonate solution. The glycans were passed through a cation exchanger, Dowex 50Wx2(H+), to remove sodium ions. The reducing ends of the released glycans were then tagged with a fluorophore, PA. Lyophilized samples were heated at 90 °C for 60 min with 20 μL of a pyridylamination reagent, and were heated at 80 °C for 35 min after the addition of 70 μL of reducing reagent, as described (Natsuka & Hase, 1998). Excess reagents were removed by phenol/chloroform extraction (Natsuka, 2002).

HPLC analysis of O-linked oligosaccharides

Size-fractionation HPLC was performed using an Asahipak NH2P-50 column (4.6 mm × 50 mm; Showa Denko) at a flow rate of 0.8 mL min−1 with solvent A (acetonitrile) and solvent B (200 mM acetic acid/triethylamine, pH 7.3) at 30 °C. After a sample had been injected, the proportion of solvent B was increased linearly from 7% to 35% for 35 min. The PA-oligosaccharides were detected with a fluorescence spectrophotometer at 310 nm excitation and 380 nm emission wavelengths. The molecular size of each PA-oligosaccharide is given in terms of glucose units based on the elution times of a PA-labeled standard marker (Takara Bio Inc.), PA-mannose or PA-isomalto-oligosaccharides.

Glycosidase treatments

Each PA-labeled O-linked oligosaccharide peak was collected and digested with 10 units mL−1 of jack bean α-mannosidase (EC; Sigma) in 50 mM sodium citrate buffer, pH 4.5, for 48 h at 37 °C, and 5 units mL−1 of green coffee bean α-galactosidase (EC; Sigma) in 150 mM sodium phosphate buffer, pH 6.5, for 48 h at 37 °C. The digestions were carried out in the presence of a trace amount of toluene and 10–100 pmol of substrate. Enzymatic reactions were stopped by heating samples at 100 °C for 5 min. A portion of the digests was analyzed by HPLC. Endoglycosidase H (Endo-β-N-acetylglucosaminidase H, EC was obtained from New England Biolabs.

Fluorescence microscopy

To confirm localization of Omh1p and Gms1p, cell were grown to an OD600 nm of 0.5 in MM without uracil and leucine, and were observed with an Olympus BX-60 fluorescence microscope using appropriate filter sets (Olympus, Tokyo, Japan). Images were captured with a Sensys Cooled CCD camera using MetaMorph (Roper Scientific, San Diego, CA), and were saved as adobe photoshop files on a Macintosh G4 computer.


Six α-mannosyltransferase homologs in S. pombe

In an effort to identify genes encoding α1,2-mannnosyltransferases that contribute to elongation of O-linked oligosaccharides in S. pombe, we found that the fission yeast genome contains six KTR-related genes, SPBC19C7.12c, SPBC16H5.09c, SPCC777.07, SPBC1773.08c, SPBC32H8.08c, and SPAC959.04c, designated omh1+omh6+. Sequence analysis revealed that Omh1p–Omh5p shared 33–53% overall identities with S. cerevisiae Kre2p, Ktr1p, and the Ktr3p subfamily (ScKtrp subfamily). On the other hand, Omh6p shared a relatively low sequence identity. Fission yeast Omh proteins possess a DXD-like motif that is well conserved in many glycosyltransferase families (Wiggins & Munro, 1998). In the case of S. cerevisiae Kre2p, unlike other glycosyltransferases, it has been reported that residues Glu-247, Pro-248, and Asp-249 (EPD) appear to form the putative ‘DXD’ motif. The first glutamate residue interacts directly with a divalent cation, and the Tyr-220 hydroxyl group plays an important role in the double-displacement mechanism between a GDP-mannose donor and the terminal α-mannoside of the acceptor (Lobsanov, 2004). The first conserved glutamate residues between Kre2p and the six Omh proteins are notable despite the fact that the last aspartate residue varies in the Omhs (EPD/S/G/E) (Fig. 1), suggesting that the six Omh proteins also catalyze a tyrosine-dependent mannosyl transfer reaction similar to Kre2p. In addition, the positions of seven cysteine residues are also conserved between Kre2 and the Omh proteins, suggesting a similarity in three-dimensional structures (Fig. 1). Omh6p has a shorter C-terminal region than the other Omh proteins, and five of seven cysteine residues in Omh6p are conserved relative to those of Kre2p (Fig. 1).


Comparison of Schizosaccharomyces pombe Omhp and the Saccharomyces cerevisiae Ktrp subfamily. Sequence alignments of the lumenal catalytic domain of the S. cerevisiae Ktrp subfamily (ScKre2p, ScKtr1p, and ScKtr3p) and S. pombe Omh1p–Omh6p are shown. Protein names are shown on the left, and numbered amino acid residues are indicated on the right. Identical residues and conserved amino acid substitutions in proteins are shaded with black or gray. The conserved seven Cys residues in ScKre2p, ScKtr1p, and ScKtr3p are indicated with dots; conserved regions with a large number of identical residues in all nine proteins are underlined. The putative DXD motif and presumed hydroxyl residue for the mannosyl transfer reaction are shown above the sequence as DXD or Y, respectively. Gaps (denoted by dashes) have been introduced to maximize the alignments.

An alignment of amino acid sequences revealed that active site residues are also conserved between the Omh proteins and Kre2p: YNLCHFWSNFEI. This region binds mannose from both the donor and the acceptor in mediating the tyrosine (Tyr-220)-dependent mannosyl transfer reaction (Fig. 1, underlined) (Høj, 1992; Lussier, 1999; Lobsanov, 2004).

Phenotypic characterization of omhΔ mutants

To investigate the effects of deletion of the omh genes on the cell growth and glycosylation of secreted glycoproteins, single-deletion strains were constructed and analyzed. All mutants harboring single gene disruptions of the omh genes were viable, indicating that these genes are not essential. Cells of the omh6Δ mutant were pear-shaped under normal growth conditions (Fig. 2a). The omh3Δ and omh6Δ mutants exhibited sensitivity to hygromycin B (Fig. 2b) or 0.005% SDS (data not shown), which are characteristic phenotypes of cell wall and N-linked glycosylation-defective mutants (Dean, 1995). Moreover, these two mutants were slow-growing and temperature-sensitive, while the other omh mutants displayed no obvious growth defects (Fig. 2b).


Cell morphology and growth defects of omh disruptants. (a) Wild-type (A), omh1Δ (B), omh2Δ (C), omh3Δ (D), omh4Δ (E), omh5Δ (F) and omh6Δ (G) strains were grown in YES liquid medium at 30°C and then observed by Nomarski optics. (b) Schizosaccharomyces pombe cells were grown on YES plates incubated at 30°C in the absence or presence of 20 μg mL−1 hygromycin B, or at 37°C on YES plates for 3 days. Yeast cultures harvested during early exponential growth were subjected to serial 10-fold dilutions (form left to right) and spotted onto YES plates.

The effect of deleting the omh1–omh6 genes on protein N-glycosylation was investigated by analyzing the electrophoretic mobility of acid phosphatase, a typical glycoprotein in S. pombe (Tanaka, 2001). Cell extracts from wild-type and each omh single mutant were subjected to native gel electrophoresis, and the position of the acid phosphatase was detected by activity staining (Fig. 3). Each omhΔ single mutant produced acid phosphatase with an electrophoretic mobility similar to that of the wild-type strain (Fig. 3). These results indicate that the Omh proteins do not appear to function in outer-chain elongation of N-linked galactomannan.


Acid phosphatase analysis of omh mutants. Wild-type (lane 1) and mutant strains, omh1Δ (2), omh2Δ (3), omh3Δ (4), omh4Δ (5), omh5Δ (6), and omh6Δ (7), were cultured in YES medium, and acid phosphatase was induced in YES-P medium. Lysates from induced cells were subjected to electrophoresis in a 6% polyacrylamide native gel and stained for acid phosphatase activity as described in Materials and methods.

O-glycosylation of secreted chitinase in omhΔ mutants

To test the involvement of omh1–omh6 in O-linked glycosylation, we analyzed the molecular weight of a recombinant S. cerevisiae chitinase, Cts1 (Fig. 4), expressed in S. pombe. Wild-type and omhΔ cells transformed with pREP41–ScCTS1 were grown to the stationary phase, and active secreted chitinase in the culture supernatants was bound by chitin beads. No chitinase band was detected in the culture supernatant from wild-type S. pombe cells (data not shown). The recombinant chitinase migrates with an apparent mass of 130 kDa (Fig. 4, lanes 1 and 2) in both wild-type S. pombe and S. cerevisiae. Interestingly, O-glycosylation of chitinase was unaffected in the omh2Δ–omh6Δ mutants, while chitinase migrated faster in the omh1Δ mutant as assessed by SDS-PAGE analysis (Fig. 4). These results suggest that O-glycosylation of extracellular chitinase is severely impaired in the absence of Omh1p, and therefore Omh1p is involved in the biosynthesis of O-linked oligosaccharides.


Detection of recombinant Saccharomyces cerevisiae chitinase in Schizosaccharomyces pombe cells. Chitinase was isolated from the culture medium of S. cerevisiae (lane 1), S. pombe wild type (2), omh1Δ (3), omh2Δ (4), omh3Δ (5), omh4Δ (6), omh5Δ (7), and omh6Δ (8) using chitin beads and was analyzed by SDS-PAGE. The gel was stained with Coomassie Brilliant Blue R-250.

Structural analysis of O-linked oligosaccharides in the omhΔ mutant

To gain more detailed information on cell-wall glycoproteins of wild-type and omh cells, O-linked oligosaccharides released by hydrazinolysis were pyridylaminated and analyzed by size-fractionation HPLC (Fig. 5). While similar types of O-linked oligosaccharides were found in most of the omh mutants, different sizes of the major oligosaccharides were found only in omh1Δ cells. In the case of oligosaccharides derived from wild-type cells, five major peaks were detected along with several unidentified peaks (Fig. 5a). To determine the structure of the oligosaccharide species in the wild-type strain, the five major peaks were collected and digested with jack bean α-mannosidase or coffee bean α-galactosidase. The position of peak 2-G among the five major peaks remained unchanged after α-mannosidase digestion, but shifted to the position of Man–PA after α-galactosidase digestion. On the other hand, peak 2-M shifted after α-mannosidase digestion, indicating that peak 2-G and peak 2-M consist of Gal–Man–PA and Man–Man–PA, respectively (Fig. 5a). We also investigated the structures of other oligosaccharide species by α-mannosidase and α-galactosidase digestion. Peaks 3-G and 3-M were identified as Gal–Man–Man–PA and Man–Man–Man–PA, respectively. However, peak 4-i shifted to the position of 2-M, 3-G, and 3-M after α-galactosidase digestion (data not shown), suggesting that peak 4-i still contains several different oligosaccharides (Fig. 5a, peaks 3-G, 3-M, and 4-i). omh1Δ cells had drastically reduced amounts of oligosaccharides larger than trisaccharides, and the O-linked oligosaccharides consisted of peak 2-G (Gal–Man–PA) (Fig. 5b). Peak 2-G from omh1Δ cells was collected and digested with α-mannosidase or α-galactosidase. While this material was found to be resistant to digestion by jack bean α-mannosidase, following digestion by coffee bean α-galactosidase, peak 2-G shifted completely to that corresponding to Man–PA (M1). These results strongly suggest that the α-1,2-linked mannose attached to the first mannose was completely blocked in omh1Δ cells, indicating that Omh1p plays an important role in elongation of α-linked mannose residues in O-linked oligosaccharides.


HPLC analysis of O-linked oligosaccharides prepared from galactomannoproteins. Galactomannoproteins were prepared from wild-type (a), omh1Δ (b), omh2Δ (c), omh3Δ (d), omh4Δ (e), omh5Δ (f), and omh6Δ (g) cells. O-linked oligosaccharides were released by hydrazinolysis, pyridylaminated, and analyzed by size-fractionation HPLC using an Asahipak NH2P-50 column. Monosaccharide (M1) at 7 min corresponds to standard PA–mannose [(a), indicated by an arrow)]. Arrows in (a), indicating peaks for 2-G, 2-M, 3-G, and 3-M, correspond to Gal–Man–PA, Man–Man–PA, Gal–Man–Man–PA, and Man–Man–Man–PA, respectively. The 4-i peak indicates PA–tetrasaccharides. The open arrowhead in (a) indicates unidentified peaks, detected in some samples. Black arrowheads in (b) indicate peaks for Gal–Man–PA (2G) and Man–Man–PA (2 M).

Localization of Omh1 protein

To determine localization of Omh1, an Omh1–GFP fusion protein was constructed and expressed in omh1Δ cells. The C-terminus was fused with GFP because a putative signal sequence is present at the N-terminus and is likely to be important for proper localization of this putative glycosyltransferase. Expression of Omh1–GFP complemented the variant O-linked oligosaccharide profiles in omh1Δ cells, indicating a functional protein (Fig. 6a). The GFP fusion protein exhibited a punctate pattern that fully overlapped with that of the Golgi-localized Gms1p/UDP galactose transporter (Fig. 6b). These results indicate that Omh1p is a type II-transmembrane protein localized to the Golgi apparatus in S. pombe.


Intracellular localization of Omh1p. (a) The pREP41–Omh1–GFP plasmid can complement the O-linked mannose elongation defect of the omh1Δ strain. omh1Δ cells containing plasmid (A) pREP41–GFP (control) or (B) pREP41–Omh1–GFP were grown in MM without leucine to the early stationary phase at 30°C. Galactomannoproteins were prepared and O-linked oligosaccharide profiles were analyzed by HPLC as described in Fig. 5. The black arrowhead indicates peaks recovered from the omh1mutant expressing the Omh1–GFP fusion protein. (b) Cells of omh1Δ strain transformed with plasmids pREP41–Omh1–GFP and pAU–Gms1–RFP were cultured in MM without uracil and leucine at 30°C. Cells harvested at OD600 nm=0.5 were then visualized using Nomarski optics and fluorescence microscopy.


We have reported the cloning and analyses of fission yeast omh genes homologous to the S. cerevisiae KTR gene family. Sequence analysis of the S. pombe omh genes revealed close homology to Kre2p, Ktr1p, and Ktr3p, which have α1,2-mannosyltransferase activity in S. cerevisiae (Lussier, 1997). Significant homology between Omh1p–Omh6p and S. cerevisiae Kre2p suggested the presence of a highly conserved C-terminal active site that interacts with a GDP-mannose donor, a terminal α-mannoside of the acceptor, and a divalent cation recognized by Omh proteins. All individual omhΔ mutants were found to be viable in S. pombe (Fig. 2a). Moreover, only the omh1 gene was found to be specifically responsible for the elongation step in O-glycan biosynthesis in fission yeast as deduced by HPLC analysis of O-linked oligosaccharides in omhΔ cells. O-linked oligosaccharides from wild-type cells contain a large amount of disaccharides, but also tetrasaccharides as described (Fig. 5; Gemmill & Trimble, 1999). The HPLC profile of O-linked oligosaccharides in omh1Δ cells was found to be quite different from that in the wild-type strain in terms of accumulated galactosyl-mannose disaccharides. We interpret this result as follows: Omh1p mainly transfers a second mannose to the O-linked mannose monomer and a galactosyltransferase(s) incorporates a galactose residue into the O-linked mannose monomer. This galactosyltransferase would be Gma12p, because Gma12p is responsible for the galactosylation of O-linked oligosaccharides (Yoko-o, 1998; Gemmill & Trimble, 1999). Interestingly, loss of omh1 did not lead to any obvious phenotypes, suggesting that at least a disaccharide unit (galactosyl-mannose) is sufficient for normal cell wall structure and function in fission yeast. Moreover, we were unable to construct an omh1Δgma12Δ double mutant, presumably due to a synthetic lethal interaction (Y. Ikeda & K. Takegawa, unpublished data). Thus, Omh1p and Gma12p can react with Man–Ser(Thr), and these enzymes may compete for sugar chains of this structure.

We previously reported that disruption of the gms1 gene led to a complete loss of protein galactosylation, due to a defect in transport of the galactosyltransferase substrate UDP–galactose from the cytosol into the lumen of the Golgi apparatus (Tabuchi, 1997; Tanaka, 2001). O-linked oligosaccharides in the gms1Δ mutant were found to consist exclusively of mannose residues, and no galactose-containing tetrasaccharides were detected by HPLC analysis of O-linked oligosaccharides. We were unable to construct an omh1Δgms1Δ double mutant, presumably due to a synthetic lethal interaction (Y. Ikeda & K. Takegawa, unpublished data). It has been reported that a defect in mannose elongation beyond two residues is not lethal in S. cerevisiae (Lussier, 1999). Therefore, we speculate that an O-linked mannose monomer is not sufficient for growth in fission yeast.

Although we have shown that Omh1p might be a major α1,2-mannosyltransferase, responsible for extending O-linked oligosaccharides in S. pombe, the functions of the other omh genes identified were not analyzed in this study. Overexpression of the omh2–omh6 genes in the omh1Δ mutant did not complement the O-glycosylation defect (data not shown), suggesting that Omh2p–Omh6p cannot substitute for Omh1p. In this study, we observed that the omh2–5 mutant had no apparent growth defects and produced a wild-type pattern of O-linked oligosaccharides. This observation suggests that these proteins are mannosyltransferases that add the third mannose residue onto O-linked oligosaccharides, or that Omh proteins other than Omh1p have redundant functions in O-glycan synthesis. No apparent defect in N-glycosylation was detected in the single omh disruptants, based on activity staining of acid phosphatase (Fig. 3). However, some of the omh genes might be involved in the synthesis of N-linked outer chains. To obtain more information about the functions of the Omh proteins, in vitro enzymatic studies are currently in progress.

Among the omh mutant cells, only omh6Δ cells exhibited sensitivity to the aminoglycoside antibiotic, hygromycin B, in addition to growing slowly, having an aberrant morphology, and being temperature-sensitive (Fig. 2b). However, the gel mobility of acid phosphatase produced in omh6Δ cells was indistinguishable from that in wild-type cells. Furthermore, omh6Δ and wild-type cells exhibited a similar HPLC profile of O-linked oligosaccharides. These results indicate that N- and O-glycosylation of secreted glycoproteins occurs independently of Omh6p. Omh6p has a short C-terminal region and shows a relatively low sequence identity to other Omh proteins (Fig. 1). A specific enzyme activity and the intracellular substrates for Omh6 protein are still unknown. We anticipate that our on-going analyses of Omh6 will clarify the catalytic activity of this protein.


We thank Drs Taro Nakamura, Takashi Toda, and Yuko Giga-Hama for providing S. pombe plasmids and strains. This work was partly supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture of Japan, and the Project for Development of a Technological Infrastructure for Industrial Bioprocesses on R&D of New Industrial Science and Technology Frontiers by the Ministry of Economy, Trade & Industry (METI), as supported by the New Energy and Industrial Technology Development Organization (NEDO).


  • Editor: Hyun Kang


View Abstract