Outer Chain N-Glycans Are Required for Cell Wall Integrity and Virulence of Candida albicans*

The outer layer of the Candida albicans cell wall is enriched in highly glycosylated mannoproteins that are the immediate point of contact with the host and strongly influence the host-fungal interaction. N-Glycans are the major form of mannoprotein modification and consist of a core structure, common to all eukaryotes, that is further elaborated in the Golgi to form the highly branched outer chain that is characteristic of fungi. In yeasts, outer chain branching is initiated by the action of the α1,6-mannosyltransferase Och1p; therefore, we disrupted the C. albicans OCH1 homolog to determine the importance of outer chain N-glycans on the host-fungal interaction. Loss of CaOCH1 resulted in a temperature-sensitive growth defect and cellular aggregation. Outer chain elongation of N-glycans was absent in the null mutant, demonstrated by the lack of the α1,6-linked polymannose backbone and the underglycosylation of N-acetylglucosaminidase. A null mutant lacking OCH1 was hypersensitive to a range of cell wall perturbing agents and had a constitutively activated cell wall integrity pathway. These mutants had near normal growth rates in vitro but were attenuated in virulence in a murine model of systemic infection. However, tissue burdens for the Caoch1Δ null mutant were similar to control strains with normal N-glycosylation, suggesting the host-fungal interaction was altered such that high burdens were tolerated. This demonstrates the importance of N-glycan outer chain epitopes to the host-fungal interaction and virulence.

nine, consist of a linear chain of one to five ␣1,2-linked mannose residues (12)(13)(14) and are known to be required for full virulence (14). The process of N-glycosylation has been studied extensively in Saccharomyces cerevisiae. N-Linked glycosylation is initiated in the endoplasmic reticulum with the transfer of the Glc 3 Man 9 GlcNAc 2 oligosaccharide precursor to the protein target (15,16). The oligosaccharide precursor is then processed by endoplasmic reticulum-resident glucosidases and a mannosidase to yield the mature triantennary Man 8 GlcNAc 2 core (17). Outer chain branching is initiated through the addition of a single ␣1,6linked mannose residue to the Man 8 GlcNAc 2 core by Och1p as the protein passes through the Golgi. Subsequently the ␣1,6-mannose backbone is extended by the sequential action of the mannan polymerase I and mannan polymerase II enzyme complexes. The linear ␣1,6linked polymannose backbone is then branched by the action of further mannosyltransferases (18,19). These side chains in C. albicans consist of ␣1,2and ␣1,3-linked mannose residues (20 -22) and in serotype A strains ␤1,2-linked mannose residues (23). The side chains and positions within the core are also modified by the addition of phosphomannan, consisting of a chain of ␤1,2-linked mannose residues attached via a phosphodiester bond (24). Previous studies in C. albicans have demonstrated the overall importance of glycosylation for cellular viability and virulence (25)(26)(27). Analysis of the PMT and MNT gene families required for O-glycosylation has demonstrated its relative importance in the host-fungal interaction and virulence (12, 14, 28 -31). Phosphomannan has been implicated in the interaction with phagocytic leukocytes; however, deletion of MNN4, required for phosphomannan production, demonstrated that it was not required for macrophage interaction or virulence (32).
In S. cerevisiae OCH1 encodes the specific ␣1,6-mannosyltransferase that initiates outer chain branching through its action on the Man 8 GlcNAc 2 core (33)(34)(35). ScOch1p is unusual among mannosyltransferases because it displays a narrow acceptor specificity requiring the whole Man 8 GlcNAc 2 core for efficient recognition (34,36). Mutants lacking the OCH1 gene in S. cerevisiae are viable but display temperature-sensitive growth (33). The Scoch1⌬ mutant has a severe N-glycosylation defect, with a complete loss of the ␣1,6-linked polymannose backbone. N-Glycans isolated from the Scoch1⌬ mutant consist of only Man 8 -10 GlcNAc 2 , where the Man 8 GlcNAc 2 core is modified by the addition of ␣1,3-linked mannose residues to antennae of the core (35). N-Glycosylation och1⌬ mutants have also been characterized for Schizosaccharomyces pombe and Pichia pastoris (37)(38)(39).
To assess the importance of outer chain N-glycosylation in virulence and host-fungal interactions, we disrupted the C. albicans OCH1 homolog. The Caoch1⌬ null mutant had a severe defect in N-glycosylation, displaying loss of the ␣1,6-linked polymannose backbone, although the core was further modified by the addition of ␣1,2-mannose residues. The Caoch1⌬ null mutant displayed cell wall defects and was attenuated in virulence in a murine model of systemic infection despite being able to colonize organs to almost wild type levels. This work demonstrates the importance of epitopes carried on branched outer chain N-glycans for the host-fungal interaction and for virulence.
Construction of Caoch1 Null Mutant, Re-integrant, and Conditional Mutant Strains-The CaOCH1 gene was disrupted by the ura-blaster method (42). The 5Ј and 3Ј regions of homology were amplified by PCR (5Ј primer pair 5Ј-CTAGAGCTCATGCTACAACTAAGAGAACCC-C-3Ј and 5Ј-GATGATGATAGATCTCGATTTCGGCATGATTCT-ATAAGCTTTAGC-3Ј; the SacI and BglII restriction sites are underlined, respectively; 3Ј primer pair 5Ј-CTACTAGTCGACCGATCT-GGACTTGTAGTTGGT-3Ј and 5Ј-GATGCATGCGGGTGTGTAC-ACAAGGGGTT-3Ј; the SalI and SphI restriction sites are underlined, respectively) and cloned into the relevant sites in pMB-7. The ura-blaster cassette was released by digestion with SacI and SphI and was flanked by 505-bp upstream and 521-bp downstream sequences complementary to CaOCH1. CaOCH1 was disrupted in strain CAI-4 by sequential gene replacement and recycling of the URA3 marker by selection on SD plus 5-fluoroorotic acid (1 mg/ml) and uridine (50 g/ml). The Ura Ϫ Caoch1⌬ null mutant was transformed with the StuI-digested CIp10 plasmid so that URA3 was expressed from the neutral RPS1 locus (43,44). As a control, a re-integrant strain was also constructed in which a wild type copy of CaOCH1 was transformed into the null mutant. The CaOCH1 open reading frame plus 969 bp of its own promoter and 425 bp of its terminator sequences were amplified by PCR (primer pair 5Ј-CAGACTGTGGGCCCGATTTAATTTGGATTTCAAG-3Ј and 5Ј-TAATGCGTGGTTCTCATGTC-3Ј), and the product was cloned into pGEM-T Easy (Promega Ltd., Southampton, UK). The plasmid insert was subcloned into the NotI site of CIp10. The resulting plasmid was then linearized with StuI before transformation into the Caoch1⌬ null mutant.
A conditional mutant was also constructed in which CaOCH1 was placed under the control of the tetracycline-regulated promoter (45) in strain THE1 that expresses the tetracycline transactivator (a fusion protein of Escherichia coli TetR and the activation domain of S. cerevisiae HAP4). The first allele of CaOCH1 was disrupted by means of the URA3 recyclable PCR-directed gene disruption system (46). The recyclable URA3 cassette was amplified from pDDB57 (primer pair 5Ј-GAAACA-GACACGAGACCTTTACTATAACTGATACTTTTGTTTCCATT-TTCTTTTCCAATTATTGGGAATAAATATTTCATGTGAAATT-GTGAGCGGATA-3Ј and 5Ј-GGTAATTTAGTTTACAAAATCGG-TTATAAATAGAAAATATGTGCATTTAAATATATATGGGTG-TGTACTCAAGGGGTTAGTTTTCCCAGTCACGACGTT-3Ј; the region of homology to pDDB57 is underlined) and transformed into THE1 to generate a heterozygous strain. The URA3 marker was then recycled by selection on SD plus 5-fluoroorotic acid (1 mg/ml) and uridine (50 g/ml). The promoter replacement cassette was produced by a modification of the published system in order to utilize PCR-directed targeting. The URA3-TR promoter cassette was amplified from p99-CAU1 (primer pair 5Ј-GAAACAGACACGAGACCTTTACTATAA-CTGATACTTTTGTTTCCATTTTCTTTTCCAATTATTGGGAA-TAAATATTTCGTAATACGACTCACTATAGGG-3Ј and 5Ј-TAT-GACTACTATTCCTAAAACGGCTAGTTTTAGATGTTTATGA-ACCATTTGGGGTTCTCTTAGTTGTAGCATGGGGTCTAGTT-TTCTGAGATAAAGCTG-3Ј; the region of homology to p99CAU1 is underlined) and transformed into the heterozygous mutant in the THE1 strain background.
Labeling of Glycans and TLC-For the analysis of acid-labile and O-linked glycans, the strains were labeled with D-[2-3 H]mannose. Cells As CAI-4 but pmr1⌬::hisG/pmr1⌬::hisG, RPS1/rps1⌬:: CIp10 27 growing in 2 ml of SC ϩ GlcNAc were incubated with 1.85 MBq of D- [2-3 H]mannose (555 GBq mmol Ϫ1 ; PerkinElmer Life Sciences) for 16 h at 30°C. Cells were then collected and washed twice with water before O-linked glycans were released by ␤-elimination with 100 mM NaOH for 24 h at room temperature. Cells were then pelleted, and the supernatant containing O-glycans was retained for TLC analysis. The cells were washed twice in water and then boiled in 10 mM HCl for 1 h to release acid-labile glycans. The remaining cellular material was pelleted, and the supernatant was retained for TLC analysis.
For TLC the samples were spotted and dried onto Silica Gel 60 TLC plates (Whatman). The plates were eluted twice in the solvent (3:4:2.5:4 ethyl acetate/butan-1-ol/acetic acid/water). For detection of labeled carbohydrates, the plates were sprayed with En 3 Hance (PerkinElmer Life Sciences) and visualized by autofluorography (Kodak BioMax XLS).
In Situ N-Acetylglucosaminidase Activity Staining-Strains were grown for 16 h in SC ϩ GlcNAc to induce HexNAcase expression. Cells were washed and resuspended in 10 mM Tris-HCl, pH 8, containing protease inhibitor mixture (Roche Applied Science) and then disrupted with glass beads in a FastPrep machine (Qbiogene, Cambridge, UK). The lysate was clarified by centrifugation at 21,500 ϫ g for 10 min. For endoglycosidase H (Endo H) treatment, the native sample was treated with 25 milliunits of Endo H (Roche Applied Science) for 16 h at 37°C in 50 mM sodium acetate, pH 5.2. Samples were mixed with native loading dye (62.5 mM Tris-HCl, pH 6.8, 0.01% bromphenol blue, and 15% glycerol) and run on a 3-8% Tris acetate-polyacrylamide gel (Invitrogen) for 1 h at 150 V under nondenaturing conditions. The gel was washed in 0.1 M citrate/KOH buffer, pH 4, for 10 min at room temperature and then incubated in the substrate solution (0.18 mM naphthyl-GlcNAc (Glycosynth Ltd., Warrington, UK) in 0.1 M citrate/KOH buffer, pH 4) for 30 min at 37°C. Finally the reaction was visualized by incubation in the substrate solution plus 0.7 mM Fast Blue at 60°C until the color developed.
Cell Wall Analysis and Sensitivity Testing-Alcian blue binding was assayed essentially as described previously (27,32). A suspension of ϳ1 ϫ 10 7 stationary phase yeast cells was washed and resuspended in 1 ml of 30 g/ml Alcian blue in 0.02 M HCl. Cells were incubated in Alcian blue at room temperature for 10 min and then spun down, and the A 620 of the supernatant was determined. Alcian blue concentration was calculated by reference to a standard curve. From these data the level of Alcian blue bound to cells was determined (g of Alcian blue bound per A 600 unit of cell suspension). Cell wall carbohydrate composition was analyzed by high performance anion exchange chromatography as described previously (48,49).
Strains were tested for sensitivity to cell wall stressing agents by microdilution testing as described previously (27). Standardized inocula were prepared from 24-h YEPD cultures. The cells were washed with water and resuspended at an A 600 ϭ 1. These cells were then inoculated into YEPD at an A 600 ϭ 0.01 and 95 l dispensed into 96-well plates. Cell wall stressing agents were added in a 5-l volume across a range of doubling dilutions. Plates were incubated at 30°C for 16 h, and the A 600 was determined. The agents tested were as follows: Calcofluor White (100 g/ml), Congo Red (100 g/ml), SDS (0.1%), hygromycin B (500 g/ml), NaCl 2 (1 M), KCl (1 M), caffeine (50 mM), vanadate (80 mM), and tunicamycin (100 g/ml). The concentrations listed are the maximum concentration tested for each agent.
Preparation of Total N-Glycan-Total N-linked glycans were released from cell walls by enzymatic digest with peptide-N-glycosidase F. Cell walls were prepared by a method modified from de Groot et al. (50). Cells from a stationary phase culture (100 ml) were washed three times in 50 mM Tris-HCl, pH 6.8, resuspended in 50 mM Tris-HCl, pH 6.8, containing protease inhibitor mixture (Roche Applied Science) to a total volume of 5 ml, and disrupted by three passes through a French press at 15.5 MPa. Cell walls were washed extensively in 1 M NaCl followed by water and extracted twice with 2% (w/v) SDS, 100 mM EDTA, 40 mM ␤-mercaptoethanol, 50 mM Tris-Cl, pH 7.8, at 100°C for 5 min to remove noncovalently bound proteins. SDS-extracted cell walls were then washed extensively in water.
Phosphodiester-linked glycans were removed from the cell wall preparations by mild hydrolysis with 40 mM trifluoroacetic acid at 100°C for 10 min. The cell walls were washed three times with water, resuspended in 250 l of 1% (w/v) SDS, 50 mM Tris-HCl, pH 6.8, and incubated at 100°C for 10 min to denature cell wall proteins. The suspension was adjusted to 2.5 ml of 0.1% (w/v) SDS, 1% (v/v) Triton X-100, 50 mM Tris-Cl, pH 6.8, and N-glycans were released by treatment with 3750 units of peptide-N-glycosidase F (New England Biolabs, Hitchin, UK) at 37°C and shaken for 16 h (51). Cell wall material was pelleted, and the supernatants containing N-glycans were retained and adjusted to pH 8.8 with 1 M NaOH. Phosphate groups present on the N-glycans were removed by treatment with 400 units of bovine intestinal alkaline phosphatase (Sigma) at 37°C for 16 h. SDS was removed by precipitation with potassium acetate (final concentration 20 mM) overnight on ice, and Triton X-100 was removed by extensive extraction with toluene (52). The N-glycans were desalted on columns of 2.5 ml of AG-50-Hϩ over 1 ml AG-4 OH Ϫ , eluted fully in water, and lyophilized. The N-glycans were further desalted by gel filtration on a Bio-Gel P2 column (25 ϫ 1.5 cm) and eluted with water at 12 ml h Ϫ1 . Fractions were collected and glycans detected by spotting onto Silica Gel 60 TLC plates and staining with orcinol. Fractions containing glycans were pooled, lyophilized, and dissolved in water.
Methylation Linkage Analysis-Methylation linkage analysis was performed on N-glycans isolated from the wild type and Caoch1⌬ null mutant. Glycans were converted to their component monosaccharides in the form of partially methylated alditol acetates and analyzed by gas chromatography-mass spectrometry on a Supelco SP2380 column as described previously (53).
Mass Spectrometric Analysis of Permethylated N-Glycans-Samples of permethylated glycans in 80% acetonitrile containing 0.5 mM sodium acetate were loaded into nanospray tips (Micromass type F) for ES-MS and ES-MS/MS. Samples were analyzed in positive ion mode with capillary and cone voltages of 0.9 kV and 40 V using a Micromass Q-ToF2 orthogonal quadrupole-time of flight mass spectrometer (Micromass, Manchester, UK). MS/MS spectra were recorded at 3 ϫ 10 Ϫ3 torr with argon as the collision gas and with collision voltages of 45-70 V. All spectra were collected and processed using MassLynx software.
Protein Extracts and Western Analysis-Activation of the cell integrity pathway was assessed by Western blot analysis utilizing the Phos-phoPlus p44/42 MAPK antibody kit (New England Biolabs) that crossreacts with CaMkc1p in its phosphorylated form. Strains were grown in YEPD at 30°C to mid-exponential phase, as positive control strains were also treated with Calcofluor White (100 g/ml) 2 h before collection. Protein extracts were prepared in 100 mM Tris-HCl, pH 7.5, 0.01% SDS, 1 mM dithiothreitol, 10% glycerol containing protease inhibitor mixture (Roche Applied Science) by means of glass bead disruption in a FastPrep machine (Qbiogene, Cambridge, UK). The resulting lysate was clarified by centrifugation at 21,500 ϫ g for 10 min. Protein samples (50 g) were separated on a 4 -14% NuPAGE gel (Invitrogen) and blotted onto a polyvinylidene difluoride membrane. The membrane was blocked in phosphate-buffered saline plus 0.1% Tween 20 and 5% bovine serum albumin for 2 h at room temperature. The PhosphoPlus p44/42 MAPK antibody kit (New England Biolabs) was then used to develop Western blots according to the manufacturer's instructions.
Virulence Tests-Female, immunocompentent BALB/c mice (Harlan Sera-Lab Ltd., Loughborough, UK) were challenged intravenously with yeasts grown for 18 -24 h in NGY medium at 30°C. The cells were collected, washed twice with water, and resuspended in physiological saline to give a challenge inoculum of 1.8 ϫ 10 4 cfu/g mouse body weight in a 100-l volume. The challenge inoculum concentration was confirmed by the determination of cellular ATP content, hemocytometer counting, and viable counting. Groups of five or six mice were intravenously inoculated and monitored over 28 days. Mice showing signs of illness were humanely terminated, and their deaths recorded as occurring the following day. Mice surviving the course of the experiment were humanely terminated on day 28. Kidneys and brain were removed aseptically postmortem, homogenized in 0.5 ml of water, and tissue burdens determined by viable counting.

RESULTS
Isolation and Deletion of CaOCH1-CaOCH1 was initially isolated by screening a cDNA library with PCR primers based on an unpublished sequence displaying homology to ScOCH1. The mannan polymerase II complex in S. cerevisiae contains ScHoc1p that displays homology to ScOch1p. Completion of the C. albicans genome-sequencing project confirmed the CaOCH1 open reading frame (GenBank TM accession number AY064420; orf19.7391) and identified a CaHOC1 homolog (orf19.3445). The degree of overall sequence identity confirms that CaOch1p reported here is the true homolog of ScOch1p. CaOch1p displays 37.5% identity to ScOch1p and 33.8% identity to ScHoc1p, whereas CaHoc1p displays 25% identity to ScOch1p but 41.4% identify to ScHoc1p. A multiple sequence alignment (ClustalW) also supports the direct sequence comparisons with the C. albicans and S. cerevisiae Och1p homologs and the Hoc1p homologs clustering together (not shown). It also identified regions that were only present in either the Och1p or Hoc1p homologs. The CaOCH1 open reading frame of 1158 bp is predicted to encode a 385-amino acid type II membrane protein with a 17-amino acid cytosolic tail, followed by an 18-amino acid membrane region before the catalytic domain. It also displays a classic DXD motif (residues 165-214) known to be required for binding the donor nucleotide sugar.
The CaOCH1 open reading frame was disrupted in strain CAI-4 by sequential gene replacement following the ura-blaster method (42). The resulting Caoch1⌬ null mutant had URA3 introduced into the neutral RSP1 locus to avoid problems with the level of URA3 expression (43,44). CaOCH1 was reintroduced into the null mutant under the control of its own promoter at the RPS1 locus as a re-integrant control. The CaOCH1 open reading frame was also placed under the control of a tetracyclineregulatable promoter in the THE1 strain that expresses the tetracycline transactivator (45). The first allele of CaOCH1 was disrupted in THE1, and the second allele was regulated by promoter replacement. In the resulting strain, TET-OCH1, expression of CaOCH1 was repressed by the addition of doxycycline to the medium. Under inducing conditions, the strain is referred to as TET-OCH1 ON and in repressing conditions, through the addition of doxycycline, as TET-OCH1 OFF.
Growth and Morphology of the Caoch1⌬ Null Mutant-The Caoch1 null mutant had a specific growth rate of 75% of the wild type control in YEPD at 30 and 37°C. The mutant also displayed an altered cellular morphology with cells forming tight aggregates and the presence of swollen cells (Fig. 1A). After growth in medium containing chitinase, cellular aggregates were dispersed, suggesting they were the result of a failure in cell separation. The clumping of mutant cells resulted in a change in colony morphology from the smooth white appearance of wild type to a highly crenulated appearance in the Caoch1⌬ null mutant (Fig. 1B). The Caoch1⌬ null mutant colonies invaded the agar surface more than the wild type (Fig. 1C), and the invading cells were predominantly hyphal. The mutant also displayed a temperature-sensitive defect with a failure to grow at 42°C (Fig. 2). In terms of hyphal development, the Caoch1⌬ null mutant responded normally to the presence of 20% (v/v) serum, although germ tubes were shorter, indicating either a decrease in extension rate or a structural defect limiting cell extension. In response to the weaker hypha-inducing signal of pH and temperature shift, in Lee's medium or RPMI 1640, the null mutant generated mixed morphologies consisting of true hyphal and pseudohyphal forms (data not shown). The null mutant also failed to induce filamentous growth on solid Spider medium (data not shown). In all cases the growth defects were restored to wild type in the re-integrant strain and the tetracyclineregulated strain displayed the same phenotype as the Caoch1⌬ null mutant after growth in repressing conditions. Reduced Mannosyltransferase Activity toward N-Linked Core Glycan in the Caoch1⌬ Mutant-In S. cerevisiae Och1p initiates outer chain elongation of N-linked core oligosaccharides. The ␣1,6-mannosyltransferase activity of ScOch1p is novel in that it displays a tight substrate specificity requiring Man 8 GlcNAc 2 for efficient acceptor recognition (36). We carried out in vitro mannosyltransferase activity assays on mixed membrane preparations of the Caoch1⌬ null mutant and the TET-OCH1 strain under inducing and repressing conditions utilizing Man 8 GlcNAc 2 as the acceptor ( Table 2). The Caoch1⌬ null mutant had 59% of the wild type activity. The remaining mannosyltransferase activity suggests that other mannosyltransferases can act on the complex Man 8 GlcNAc 2 acceptor structure in vitro. The TET-OCH1 strain also displayed a significant drop in transferase activity when grown under repressing conditions (73% of wild type). When growing exponentially under inducing conditions, the TET-OCH1 strain appeared to overexpress Och1p with detected mannosyltransferase activity rising 2-fold (207% of wild type).  Glycosylation Defects in the Caoch1⌬ Null Mutant-Previously, the electrophoretic mobility of secreted acid phosphatase on native gels has been used as a maker of N-glycosylation status (13,25,27). However, no acid phosphatase activity was detected in zymograms of the Caoch1⌬ null mutant. Additionally, in in vitro assays, only low levels of acid phosphatase activity were detected for the Caoch1⌬ null mutant, in cellassociated assays, and in secreted fractions (data not shown). We therefore developed an in situ activity assay for the well characterized hydrolytic enzyme HexNAcase as an alternative marker for N-glycosylation status in C. albicans. HexNAcase, encoded by CaHEX1, is induced after growth in medium containing GlcNAc as the sole carbon source. It has seven potential N-glycosylation sites and has been demonstrated to be highly N-glycosylated (54 -56). The in situ activity assay used a naphthyl derivative of GlcNAc as the substrate and the tetrazolium salt Fast Blue for visualization after the separation of protein extracts by native gel electrophoresis. The HexNAcase of the Caoch1⌬ null mutant had an increased electrophoretic mobility, indicating an N-glycosylation defect (Fig. 3A). The TET-OCH1 strain also showed a similar increase in the electrophoretic mobility of HexNAcase after growth under repressing conditions. The electrophoretic mobility of HexNAcase of the re-integrant control and TET-OCH1 strain grown under inducing conditions was similar to wild type. After Endo H treatment to remove N-glycans, the HexNAcase migrated faster through the gel as two distinct bands. It has been reported previously that HexNAcase contains three N-glycosylation sites that receive only the mature core N-glycan and lack outer chain elongation. These sites were partially resistant to Endo H cleavage when HexNAcase was in its native nondenatured form (55). The N-glycosylation defect in the Caoch1⌬ null mutant was more severe than in the Capmr1⌬ null mutant (27) that has a general glycosylation defect (Fig. 3B), indicating that CaOCH1 acts at an early step in N-glycan outer chain elaboration.
The outer chain N-glycans contain the majority of the acid-labile phosphomannan fraction consisting of ␤1,2-linked mannose residues attached via a phosphodiester linkage. The extent of mannosylphosphorylation can be detected by the extent of binding of the cationic dye Alcian blue. The Caoch1⌬ null mutant and TET-OCH1 OFF strains had reduced Alcian blue binding, with 16.7 and 24.6% of wild type levels, respectively ( Table 2). The re-integrant control and TET-OCH1 ON had wild type levels of Alcian blue binding. The acid-labile phosphomannan fraction was also directly resolved by TLC analysis. This had greatly reduced levels of mannosylphosphate with only small amounts of Man 3 and Man 4 ␤1,2-linked mannose residues present in the Caoch1⌬ null mutant and TET-OCH1 OFF strains (Fig. 4A). The wild type, re-integrant control, and TET-OCH1 ON had a normal phosphomannan structure. The structure of O-mannan was also assessed by TLC analysis and displayed no change resulting from the loss of CaOCH1 (Fig. 4B).
Analysis of N-Linked Glycans-Cell walls were treated with SDS/␤mercaptoethanol followed by mild acid hydrolysis to remove noncovalently linked mannoproteins and phosphomannan, respectively. The purified cell walls were then treated with peptide-N-glycosidase F to release the N-linked glycans. These were further treated with alkaline phosphatase to render the glycans neutral by removing the 6-linked phosphate groups that remain after removal of phosphomannan. Desalted N-linked glycans were permethylated and analyzed by ES-MS and gas chromatography-mass spectrometry methylation linkage analysis.
The methylation analysis of the wild type glycans (Table 3) was in accord with the detailed analysis of Shibata et al. (57). The material contained substantial amounts of nonreducing terminal Man, 2-O-substituted Man, 6-O-substituted Man, and 2,6-di-O-substituted Man, as well as the 3,6-di-O-substituted Man of the core glycan and occasional outer arm branch points (Fig. 5). Methylation analysis of the Caoch1⌬ null mutant glycans (Table 3) showed an almost complete loss of 6-Osubstituted Man and 2,6-di-O-substituted Man and a significant decrease in nonreducing terminal Man residues. This result is consistent with the loss of the 1,6-linked arm of the wild type N-glycan structures. Methylation analysis of the Caoch1⌬ null mutant glycans was consistent with the residual N-linked glycans being based on the conventional Man 8 GlcNAc 2 core glycan because they contained almost exclusively terminal Man, 2-O-substituted Man and 3,6-di-O-substituted Man. However, ES-MS analysis revealed that they were larger than the conventional core ranging from Hex 11 HexNAc 2 to Hex 16 HexNAc 2 (Fig. 6). The methylation analysis suggests these additional hexoses were in the form of 1,2-linked Man. The original doubly charged [M ϩ 2Na] 2ϩ ions that were deconvoled (Fig. 6) were analyzed

Och1p Required for Outer Chain N-Glycosylation and Virulence
by ES-MS/MS, and the product ion spectra confirmed that these contained a GlcNAc 2 core and were of the triantennary oligomannose type (data not shown). Analysis of cross-ring cleavage ions (58) suggests that the additional 1,2-linked Man residues can be attached to one or more antennae of the conventional Man 8 GlcNAc 2 core (Fig. 5). To determine the effect of the Caoch1⌬ mutation on the integrity of the cell wall, we tested the null mutant for sensitivity to a range of cell wall perturbing agents and other compounds associated with glycosylation defects. The Caoch1⌬ null mutant was hypersensitive to the cell wall perturbing agents Calcofluor White and Congo Red and to SDS that would affect both cell wall proteins and the plasma membrane (Fig.  7). The null mutant was also hypersensitive to hygromycin B and tunicamycin (Fig. 7) and slightly more resistant to vanadate (data not shown), characteristics that are common to N-glycosylation mutants in other fungi (59,60). There was no change in sensitivity toward other agents and stress such as caffeine, NaCl, or KCl (data not shown). A similar sensitivity profile was seen with the TET-OCH1 strain when the initial inoculum was pre-grown under repressing conditions. Antifungal susceptibility testing demonstrated no change in sensitivity to itraconazole, flucytosine, amphotericin B, or caspofungin (data not shown).

Cell Wall Composition, Sensitivity, and Cell Integrity Pathway
Because the Caoch1⌬ null mutant had a cell wall defect, we carried out Western blotting to test whether the protein kinase C cell integrity pathway was activated by using an antibody that recognizes the phosphorylated form of the MAPK Mkc1p. Mkc1p was activated in the Caoch1⌬ null mutant and TET-OCH1 strain under repressing conditions but was not activated in the wild type and TET-OCH1 strain under inducing conditions (Fig. 8). As a positive control, the strains were also treated with Calcofluor White, which is known to activate the pathway. An additional band was also detected only in the Caoch1⌬ null mutant and TET-OCH1 strain under repressing conditions. This band migrated at 49 kDa and most likely corresponds to the phosphorylated form of CaCek1p. It has been shown recently that this phospho-specific antibody can recognize activated CaCek1p (61). Therefore, CaCek1p may act in an alternative cell integrity pathway activated by glycosylation defects. A similar role has been reported for Kss1p in S. cerevisiae (62,63). The increased sensitivity to cell wall perturbing agents and constitutive activation of the cell integrity pathway confirm a cell wall defect resulting from the loss of CaOCH1.
Attenuation of Virulence in Caoch1⌬-The effect of the loss of CaOCH1 on virulence was assessed in a mouse model of systemic infection. The Caoch1⌬ null mutant was significantly attenuated in virulence with a mean survival time of 22 days, compared with 6 days for the wild type control (Fig. 9, log rank test; p Ͻ 0.01). The re-integrant control was completely restored in virulence with a mean survival time of 7 days. Mice infected with the wild type, re-integrant control, or Caoch1⌬ null mutant all had similar tissue burdens in both brain and kidney (Table 4). Therefore, the host-fungal interaction is altered in the Caoch1⌬ null mutant that lacks N-glycan outer chains, such that significant tissue

Och1p Required for Outer Chain N-Glycosylation and Virulence
burdens are generated but evidently cause less damage resulting in an increased survival time. This work demonstrates the importance of N-glycan outer chain epitopes in the host-fungal interaction and virulence.

DISCUSSION
Here we describe the importance of outer chain elongation of N-glycans in the structure of the C. albicans cell wall and the interaction of these epitopes with the host. The addition of N-glycans is the major form of modification of mannoproteins. The carbohydrate structure consists of the Man 8 GlcNAc 2 core that is extensively elaborated in the Golgi by the addition of the highly branched outer chain. Previous stud-ies have demonstrated the overall importance of glycosylation to the cell wall structure and virulence (25)(26)(27). Loss of O-mannan has been shown to result in reduced adherence and the attenuation of virulence, suggesting that O-mannan may act as a ligand in the interaction with host surfaces (12, 14, 28 -31). Conversely, phosphomannan has been shown not to be required for normal host-fungal interactions or virulence (32). We have extended these studies by analyzing the importance of the N-glycan outer chain epitopes. This was achieved by disrupting the C. albicans homolog of OCH1, which initiates outer chain elongation. The Caoch1⌬ null mutant had a clear defect in the addition of the N-glycan outer chain and correspondingly displayed a weakened cell wall. The N-glycan outer chain was shown to be necessary for normal host-fungal interactions and virulence.
The Caoch1⌬ null mutant had slightly reduced growth rate and formed cellular aggregates. These aggregates could be dispersed by treatment with chitinase suggesting that they are the result of a cell separation defect. After chitinase treatment, it was clear that the null mutant was growing in chains of short pseudohyphae. These aggregates of short pseudohyphae resulted in the formation of a crenulated colony morphology. Most notably, the null mutant cells also invaded the agar surface, and the invading cells were predominantly hyphal in morphology under conditions that normally support growth by budding. Western analysis has demonstrated that both the Cek1p and Mkc1p MAPKs are constitutively activated in the null mutant. Both Cek1p and Mkc1p have been shown to be important in invasive growth, signaling through nutrient limiting conditions or colonial growth and physical contact, respectively (64,65). Therefore, the constitutive activation of these pathways in the Caoch1⌬ null mutant may explain why it so readily invades the agar surface under nonhyphal inducing conditions.
Previously we have measured the electrophoretic mobility of secreted acid phosphatase as a marker for changes in N-glycosylation. However, we were not able to detect acid phosphatase in soluble protein extracts of the Caoch1⌬ null mutant, and only low levels of cell-associated or secreted activity were seen. In crude cellular extracts, low activity was detected after native PAGE, but it failed to migrate into the polyacrylamide gel. This would suggest that the glycosylation defect in Caoch1⌬ resulted in acid phosphatase becoming partially insoluble with low activity. As an alternative we developed an in-gel activity assay for Hex-NAcase. HexNAcase is exclusively N-glycosylated and is readily detected in nondenaturing polyacrylamide gels. Therefore, it can be used as a more sensitive marker of N-glycosylation defects.
Loss of CaOCH1 resulted in clear defects in N-glycosylation demonstrated by the increased mobility of HexNAcase and a reduction in phosphomannan. The remaining phosphomannan present in the null mutant, low levels of Man 3 and Man 4 , could be attached to the N-linked core structure or potentially O-mannan. As expected, the O-mannan in the Caoch1⌬ null had a normal structure. The null mutant also displayed a significant reduction in the in vitro mannosyltransferase activ-

Och1p Required for Outer Chain N-Glycosylation and Virulence
ity toward the Man 8 GlcNAc 2 core structure. However, activity toward this acceptor was not completely abolished in the null mutant. The Man 8 GlcNAc 2 core is a complex structure and could be modified in vivo by additions to the terminal antennae of the core by other mannosyltransferases. Mannosylphosphate is also added to the N-glycan core, although in this case the resulting charged product would not be detected in the assay. Additionally, in in vitro assays of activity many mannosyltransferases display low substrate specificity. Indeed, heterologously expressed CaMnt1p has been shown to display activity toward a Man 5 GlcNAc core-like structure in vitro (66), although in vivo it specifically acts in the extension of O-linked glycans (14). Therefore, background activity toward the Man 8 GlcNAc 2 core in the Caoch1⌬ null mutant is not unexpected.
Purified N-glycans from the wild type and Caoch1⌬ null mutant were analyzed by methylation linkage analysis and ES-MS. Wild type N-glycans had no ions detectable by ES-MS because of their extreme size. Methylation linkage analysis revealed substantial amounts of terminal mannose and 2-O-substituted, 6-O-substituted, and 2,6-di-O-substituted mannose, consistent with the current structural model for the outer chain (57) (Fig. 5). The two unique 3,6-di-O-substituted mannose linkages in the N-glycan core were used to determine the stoichiometry of the various linkage groups in the outer chain. However, the outer chain has also been reported to contain some branched side chains (57), and these structures would also contribute to the level to 3,6-di-Osubstituted mannose. Hence, the proportions of linkages in the wild type N-glycans may be a minimum estimate of the extent of outer chain elongation.
Methylation linkage analysis of the N-glycans purified from the Caoch1⌬ null mutant demonstrated an almost complete loss of 6-Osubstituted and 2,6-di-O-substituted mannose. This confirms the absence of the outer chain glycan in the Caoch1⌬ null mutant. Therefore, CaOCH1 functions in initiating outer chain elongation in C. albicans. Consistent with the loss of the outer chain, there was a reduction in the proportion of terminal and 2-O-substituted mannose residues. However, 2-O-substituted mannose residues were not reduced to the level expected for a simple Man 8 GlcNAc 2 core. Indeed, the core appears to undergo further modification in the Caoch1⌬ null mutant, with elaborations from Man 11 GlcNAc 2 to Man 16 GlcNAc 2 . This is in contrast to the situation in S. cerevisiae where N-glycans of the Scoch1⌬ null mutant consist of only Man 8 -10 GlcNAc 2 (35). These further modifications to the core in the Caoch1⌬ null mutant appear to consist of chains of 1,2-linked mannose residues added to one or more of the antenna residues of the core (Fig. 5). It is not yet clear whether such modifications to the core also occur in wild type N-glycans or are the result of the lack of normal acceptor substrates for mannosyltransferases in the absence of the outer chain. The presence of these modifications is in accord with the mannosyltransferase activity detected toward the Man 8 GlcNAc 2 core in the null mutant. The migration pattern of HexNAcase from the null mutant is also fairly diffuse, which would suggest that heterogeneity exists in the core N-glycan structures. Low levels of phosphomannan may also contribute to this heterogeneity.
The Caoch1⌬ null mutant had a weakened cell wall, as demonstrated by its hypersensitivity to a range of cell wall perturbing agents and other agents whose action is indicative of glycosylation defects. Commensurate with the cell wall sensitivity phenotype, we also showed constitutive activation of the cell wall integrity pathway by assessing the phosphorylation status of Mkc1p. It was also noticed that Cek1p is also activated constitutively in the null mutant, although it was not activated in response to Calcofluor White. In S. cerevisiae Kss1p, a homolog of CaCek1p, was also found to be activated constitutively in the Scoch1⌬ null mutant and was shown to increase the expression of the ␤1,3glucan synthase Fks2p (62). This required components from the high osmolarity glycerol, pheromone response, and invasive growth pathways and has been suggested to act as a separate pathway in vegetative cells responding to changes in protein glycosylation and acting as a further cell wall integrity pathway (62,63). Finding that Cek1p is constitutively active in the Caoch1⌬ null mutant suggests a similar pathway may exist in C. albicans that responds to severe glycosylation defects.
The Caoch1⌬ null mutant had significantly attenuated virulence in the mouse model of systemic infection. However, tissue burdens in the kidney and brain were still found to be high in animals infected with the null mutant, even when no symptoms of disease were seen. This indicates that the null mutant has no defect in targeting and colonizing the organs normally associated with disease progression. However, high burdens were tolerated by the host, suggesting that the nature of the host-fungal interaction is altered. One possibility is that the null mutant was able to colonize the affected tissues but could not express factors needed for invasive growth. An alternative hypothesis is that the immune response is tolerant of the Caoch1⌬ null mutant. The balance of pro-and anti-inflammatory response components is known to be important for a number of fungal pathogens, where a strong pro-inflammatory response results in tissue damage and disease progression. In the case of the Caoch1⌬ null mutant, the changes to the cell surface, resulting from the lack of N-glycan outer chain epitopes, may alter the recognition of the invading organism by cells of the innate immune system and result in a shift of the host reaction away from Th2 toward a Th1 response. In any case it is clear that outer chain N-glycans play vital roles in the interaction between C. albicans and the host during invasion and pathogenesis.