Identification of Components of the SUMOylation Machinery in Candida glabrata

Regulation of protein function by reversible post-translational modification, SUMOylation, is widely conserved in the eukaryotic kingdom. SUMOylation is essential for cell growth, division, and adaptation to stress in most organisms, including fungi. As these are key factors in determination of fungal virulence, in this study, we have investigated the importance of SUMOylation in the human pathogen, Candida glabrata. We identified the enzymes involved in small ubiquitin-like modifier conjugation and show that there is strong conservation between Saccharomyces cerevisiae and C. glabrata. We demonstrate that SUMOylation is an essential process and that adaptation to stress involves changes in global SUMOylation in C. glabrata. Importantly, loss of the deSUMOylating enzyme CgUlp2 leads to highly reduced small ubiquitin-like modifier protein levels, and impaired growth, sensitivity to multiple stress conditions, reduced adherence to epithelial cells, and poor colonization of specific tissues in mice. Our study thus demonstrates a key role for protein SUMOylation in the life cycle and pathobiology of C. glabrata.

SUMOylation, the covalent reversible conjugation of small ubiquitin-like modifier (SUMO) 5 polypeptide to lysine residues, often within the canonical consensus motif ⌿KXE (⌿ and X represent a hydrophobic amino acid and any amino acid, respectively) in target proteins, is a post-translational modification that plays a key regulatory role in several cellular processes, including transcription, protein homeostasis, stress response, and development (1,2). The process of SUMO attachment consists of the following four steps: (i) processing of the ϳ10-kDa precursor SUMO peptide by SUMO-specific proteases to reveal a C-terminal diglycine motif in the mature SUMO; (ii) ATP-dependent activation of the processed SUMO through the thioester bond formation between the C-terminal glycine of SUMO and the catalytic cysteine of the E1-activating enzyme; (iii) transfer of the SUMO polypeptide from the E1 enzyme to a conserved cysteine in the E2-conjugating enzyme via a thioester linkage; and (iv) E3 ligase-mediated formation of an isopeptide bond between the C-terminal glycine of the SUMO and the ⑀-amino group of the lysine residue within the conserved sequence on the target protein (2,3). Besides the precursor SUMO maturation, the SUMO-specific peptidases are also able to hydrolyze the isopeptide bond between SUMO and SUMO-modified proteins thereby rendering the SUMOylation process reversible.
The SUMO polypeptide is ubiquitously present in all eukaryotes and highly conserved from yeast to mammals (1)(2)(3). SUMO modification of protein substrates has diverse functional consequences and range from increased protein stability to altered subcellular localization (1)(2)(3). Furthermore, deregulated expression of SUMOylation components has been implicated in several human diseases, including neurodegeneration, heart failure, cancer, diabetes, and infections by bacterial and viral pathogens (4,5).
Candida bloodstream infections, also known as candidemia, are a common occurrence in patients with immune dysfunctions and undergoing transplantation and radiation therapy (23,24). Candidemia often results in prolonged hospitalization in intensive care unit, high healthcare costs, and considerable morbidity and mortality (23). During last 2 decades, the incidence rate of candidemia has increased significantly with C. albicans being the most prevalent species followed by Candida glabrata (23,(25)(26)(27). C. glabrata accounts for up to 29% of total Candida bloodstream infections with a crude mortality rate of 40 -45% (26 -29).
Among the known virulence factors of C. glabrata, glycosylphosphatidylinositol-linked adhesins, cell surface-associated proteases, robust oxidative stress response, and the ability to form biofilms and evade immune response occupy central position (30). Although SUMOylation is known to be involved in the regulation of fungal development, differentiation, and virulence (22,31), nothing is known about its role in the pathogenesis of C. glabrata. Here, using the reverse genetics approach, we have identified components of the SUMOylation machinery in C. glabrata and show that the deSUMOylation peptidase CgUlp2 is required for biofilm formation, adhesion, and virulence of C. glabrata. We also report essentiality of C. glabrata Smt3 for cell growth and viability. Furthermore, we demonstrate for the first time a functional conservation of key SUMOylation components between S. cerevisiae and C. glabrata.

Results
Identification of Components of the SUMOylation Pathway in C. glabrata-To determine components of the SUMOylation pathway in C. glabrata, we performed whole proteome sequence and BLAST analyses, and identified C. glabrata orthologues of the proteins that are involved in SUMOylation in S. cerevisiae ( Table 1). Orthologues of all the components could be identified, and their percent similarity across the complete sequence is shown in Fig. 1 and Table 1. SUMO and the E2 ligase Ubc9 protein are the most conserved between C. glabrata and S. cerevisiae with both showing over 85% identity (Table 1 and data not shown). Other SUMOylation components in both yeasts also showed significant sequence similarities as well as conserved architecture of various characterized domains (see Fig. 1 and data not shown). One striking difference was the absence of the SAP domain in the C. glabrata Siz1 ligase (see Fig. 1). The SAP domain, found in SIZ/PIAS family (Sap and Miz/protein inhibitors of activated STAT) of SUMO ligases, is implicated in DNA binding and nuclear retention (32).
In S. cerevisiae, deletion of the SAP domain results in reduced nuclear localization of the Siz1 ligase, whereas the fulllength protein has a cell cycle-dependent localization. Although Siz1 is nuclear through most of the cell cycle stages, it relocates to the bud neck during cytokinesis, as SUMOylation of septins and other substrates at this stage is essential for completion of cell division (11,14,33,34). Siz1 relocation is dependent on the C-terminal domain, whereas its nuclear retention is contingent on the SAP domain (32). Therefore, loss of this well conserved domain in CgSiz1 ligase raises the possibility of additional C. glabrata-specific SUMO substrates outside the nucleus.
Functional Conservation between C. glabrata and S. cerevisiae SUMO Components-Given the strong conservation of sequence and architecture of all components of the SUMOylation pathway in C. glabrata, we next examined whether some of the key components would functionally complement the S. cerevisiae mutants. For the two essential genes tested, SMT3 and ULP1, we performed plasmid shuffling experiments in S. cerevisiae knock-outs. First, we transformed the S. cerevisiae smt3⌬ strain carrying the ScSMT3 gene on a URA3-based vector either with a LEU2-based plasmid expressing the CgSMT3 gene from the ADH1 promoter (CKM 379) or with the LEU2based vector alone. The double transformants were streaked on plates containing 5-fluoroorotic acid (5-FOA) to select for uracil auxotrophs. We found that the S. cerevisiae smt3⌬ strain expressing CgSMT3 survived, but the strain carrying the empty vector did not ( Fig. 2A). This establishes that the Scsmt3⌬ mutant could be complemented by CgSMT3, and therefore it could lose the ScSMT3-expressing plasmid. Through similar plasmid shuffle experiments, we next showed that CgULP1 could also complement the Sculp1⌬ mutant (Fig. 2B). The ULP2 gene is not essential for cell viability in S. cerevisiae, but its deletion results in growth retardation (35). As shown in Fig.  2C, slow growth of the S. cerevisiae ulp2⌬ mutant was rescued by ectopic expression of the CgULP2 gene indicative of a functional conservation between S. cerevisiae and C. glabrata Ulp2 deSUMOylase. However, CgUlp2 could not complement the telomeric silencing defect of the Sculp2⌬ mutant as CgULP2expressing Sculp2⌬ cells did not grow on medium containing 5-fluoroorotic acid (see Fig. 2C). These cells carry URA3 at the subtelomeric region of chromosome VII and would exhibit growth on FOA medium only if URA3 expression was suppressed due to the telomere position effect. As expected, S. cerevisiae wild-type strain harboring the URA3 gene at the subtelomeric locus exhibited robust growth on plates containing 5-FOA (see Fig. 2C). These data suggest that CgULP2 cannot complement all functions of ScULP2. Of note, we could not perform complementation studies with C. glabrata UBA2, AOS1, UBC9, and MMS21 genes because the corresponding heterozygous S. cerevisiae mutants sporulated extremely poorly (data not shown). Subcellular Localization of C. glabrata SUMO Enzymes-In S. cerevisiae, SUMOylated proteins are mostly nuclear with very few proteins SUMOylated in the cytosol or other subcellular compartments. This is because most of the SUMO ligases are found in the nucleus except Siz1, which localizes to the septin ring during cell division (13,36). To examine whether the location of SUMO enzymes was conserved between C. glabrata and S. cerevisiae, we introduced the dual His-FLAG tag sequence into C. glabrata SMT3, SIZ1, SIZ2, ULP1, and ULP2 gene sequences, and we transformed these plasmids into either wild-type strain (for essential genes) or the respective C. glabrata deletion strains. Immunofluorescence analysis with anti-FLAG antibody revealed that CgSmt3 was localized uniformly across the cell suggesting that SUMO and/or SUMOy-lated proteins are found both in the cytosol and the nucleus (Fig. 3). The two SUMO ligases, CgSiz1 and CgSiz2, and the desumoylating enzymes, CgUlp1 and CgUlp2, were all predominantly nuclear with only CgSiz1 and CgUlp1 showing few cytosolic spots (see Fig. 3). Interestingly, the characteristic nuclear pore complex localization of ScUlp1 was not observed in CgUlp1. Thus, despite uniform distribution of SUMOylated proteins across the cell, the enzymes involved in SUMOylation and deSUMOylation are predominantly nuclear in C. glabrata. These findings suggest that localization of most components of the SUMOylation pathway in C. glabrata is similar to that in S. cerevisiae.
Disruption of C. glabrata ORFs That Are Potentially Involved in SUMOylation in C. glabrata-To examine the effect of perturbation of SUMOylation machinery on physiology and pathogenesis of C. glabrata, we sought to generate strains lacking one or more SUMO components. Of the set of nine genes shown in Table 1, we were able to create deletion strains for CgSIZ1, CgSIZ2, and CgULP2 genes via a homologous recombination-based strategy. We also generated a double deletion strain for CgSIZ1 and CgSIZ2 genes to investigate the role of SUMO ligases in C. glabrata pathobiology. Despite several attempts, we could not delete CgAOS1, CgUBA2, CgUBA9, CgMMS21, CgULP1, and CgSMT3 ORFs. Notably, S. cerevisiae FIGURE 2. Complementation of S. cerevisiae mutants with C. glabrata orthologues. A, S. cerevisiae smt3⌬ carrying SMT3 on a plasmid with URA3 as selectable marker was transformed with CgSMT3 on a Leu-marked plasmid (3)(4)(5)(6) or empty Leu vector (2), or neither (1). Transformants were streaked on YPD (no selection), or Sc-Leu, Ura (selecting for both plasmids), or Sc-Leu FOA (for the absence of Ura plasmid). Only strains containing CgSMT3 could lose the ScSMT3 plasmid and grow on plates containing 5-FOA. B, S. cerevisiae ulp1⌬ carrying ScULP1 on a plasmid with URA3 as a selection marker were transformed with CgULP1 encoded on a plasmid with TRP1 as a selectable marker or empty Trp vector or neither. Only strains containing CgULP1 could lose the ScULP1 plasmid and grow on plates containing 5-FOA. C, Sculp2⌬ with telomeric URA3 (KRY744) was transformed with either plasmid encoding CgUlp2 or empty vector and tested for growth at 30°C and on 5-FOA plates for silencing phenotype. CgULP2 restores growth but not silencing in S. cerevisiae. orthologues of these six genes are essential for cell viability (5)(6)(7)(8)(9)(10). Furthermore, using the one-step disruption and plasmid loss methodologies, we could show that CgSMT3 is required for growth in vitro in C. glabrata (data not shown).
To investigate the role of SUMOylation in cell physiology, we conducted a comprehensive phenotypic characterization of the C. glabrata mutants generated. Growth analysis of Cgsiz1⌬, Cgsiz2⌬, Cgsiz1⌬siz2⌬, and Cgulp2⌬ mutants revealed that the Cgulp2⌬ mutant grew about 18% slower than the wild-type (WT) strain in rich medium (Fig. 4A). The Cgulp2⌬ mutant exhibited highly attenuated growth at 42°C and in medium containing caffeine and non-fermentable carbon sources (glycerol, oleic acid, and ethanol) (Fig. 4, B and C). Compared with WT cells, the Cgulp2⌬ mutant was more susceptible to the DNA-alkylating agent methyl methanesulfonate (MMS), replication fork staller hydroxyurea, thymine dimer-inducing ultraviolet (UV) radiation, and oxidative stress-inducing agent hydrogen peroxide (see Fig. 4B). Elevated sensitivity of the Cgulp2⌬ mutant to aforementioned stresses was complemented by ectopic expression of the CgULP2 gene, indicating that these effects were specifically induced by the absence of CgULP2 (see Fig. 4B). Contrary to the Cgulp2⌬ mutant, growth of the Cgsiz1⌬ mutant remained unaffected under diverse stressful conditions, viz. DNA damage (MMS, camptothecin, and hydroxyurea), oxidative (hydrogen peroxide and menadione), thermal (37 and 42°C), varied pH values (low (2.0) and neutral (7.0) pH), cell wall (caffeine), cell membrane SDS, nonfermentable carbon sources, and antifungal drug (fluconazole and caspofungin) stresses (see Fig. 4, B and C and data not shown). Interestingly, the Cgsiz2⌬ and Cgsiz1⌬siz2⌬ mutants displayed sensitivity to both UV and MMS, which could be rescued by ectopic expression of the CgSIZ2 gene (see Fig. 4B). Expression of CgSIZ1 partially restored slow growth of the Cgsiz1⌬siz2⌬ mutant in MMS-supplemented medium and upon UV treatment (see Fig. 4B). Altogether, these data implicate CgSiz2 and CgUlp2 in survival of DNA damage and thermal, oxidative, and DNA damage stresses, respectively.
Disruption of CgULP2 Rendered C. glabrata Cells Hypoadherent to Lec2 Ovary Epithelial Cells-During our phenotype profiling analysis, we noticed that about 5% of the Cgulp2⌬ mutant population displayed elongated pseudohypha-like structures after 48 h of growth in the YPD medium (Fig. 5A). To investigate whether this change in morphology is due to altered cell wall composition, we examined sensitivity of the Cgulp2⌬ mutant to digestion with zymolyase, which hydrolyzes ␤-glucan in the cell wall. The fungal cell wall is a complex and dynamic structure and consists of an inner layer of ␤-glucanchitin complex and an outer layer of heavily O-and N-glycosylated mannoproteins (37). Compared with WT cells, the Cgulp2⌬ mutant displayed resistance to zymolyase digestion, which was reversed upon ectopic expression of the CgULP2 gene in the mutant (Fig. 5B). Furthermore, cell wall chitin analysis revealed 1.5-fold elevated chitin content in the Cgulp2⌬ mutant compared with WT cells (Fig. 5C). Consistently, staining with calcofluor white, which binds to chitin in the cell wall, showed a diffused signal along the cell wall in the Cgulp2⌬ mutant compared with bud scar-limited staining of WT cells Wild-type C. glabrata cells were transformed with plasmid encoding the indicated FLAG-tagged genes. Immunofluorescence was performed using anti-FLAG antibody (green), and DAPI was used to delineate the nucleus. The 4th panel shows the merged signals in DIC to visualize the cell outline. Bar ϭ 2 m.
( Fig. 5D). Overall, these findings are indicative of an altered cell wall architecture in the Cgulp2⌬ mutant.
Cell wall integrity in C. glabrata is maintained by protein kinase C (PKC)-mediated signal transduction pathway, and mutants with cell wall defects have previously been reported to exhibit constitutively activated PKC signaling cascade (38). To examine the status of the PKC-dependent signaling pathway in the Cgulp2⌬ mutant, we checked phosphorylation levels of CgSlt2, which is the terminal mitogen-activated protein kinase (MAPK) of the PKC pathway, in the mutant. As shown in Fig.   FIGURE 4. CgUlp2 is required for survival of thermal and DNA damage stress. A, growth curve analysis. Overnight-grown cultures of indicated C. glabrata strains were inoculated into the YPD medium to an initial A 600 of 0.1. Absorbance at 600 nm was recorded over a 48-h time course at indicated time intervals. Data represent mean Ϯ S.E. of three to five independent experiments. Doubling times were calculated during the exponential phase of growth and are presented on the bottom right side of the graph. Statistical analysis was performed using an unpaired, two-tailed, Student's t test (*, p Յ 0.05). B, serial dilution-spotting assay to assess the growth under indicated stress conditions. 3 l of 10-fold serial dilutions of overnight-grown and 1.0 A 600 -normalized cultures of indicated C. glabrata strains were spotted on different media. Plate images were captured after 2 days of incubation at 30°C. Methylmethanesulfonate (MMS), hydroxyurea (HU) and hydrogen peroxide (H 2 O 2 ) were used at the concentrations of 0.03%, 50 mM, and 50 mM, respectively. To check sensitivity to ultraviolet radiation, 3 l of 10-fold serial culture dilutions was spotted on the YPD medium, and the plate was exposed to 40 J/m 2 UV radiation at 256 nm before incubation at 30°C. C, heat map illustrating cell growth in the presence of diverse stress-causing agents. 3 l of 10-fold serial dilutions of overnight YPD medium-grown and 1.0 A 600 -normalized cultures of indicated C. glabrata strains were spotted on different media. Growth profiles, recorded after 1-8 days of incubation, are color-coded and indicated at the bottom. Columns correspond to various growth condition mutants and rows to mutants. Stressful conditions used were as follows: DNA damage stress (camptothecin; 10 mM); cell wall stress (caffeine; 10 mM); antifungal stress (fluconazole (FLC; 16 g/ml) and caspofungin (75 ng/ml)); membrane stress (SDS; 0.005%); varied pH stress (pH 2.0 and 7.0); and utilization of alternative carbon sources (glycerol (3%), ethanol (2%), and sodium acetate (2%)). Cgulp2⌬ mutant grown in the YPD medium for 72 h. B, zymolyase digestion assay. Log phase cultures of indicated C. glabrata strains were treated with 50 g/ml zymolyase, and absorbance at 600 nm was recorded at regular time intervals. Initial A 600 of the cultures was considered as 100%, and data (means of three (Cgsiz1⌬2⌬/CgSIZ2 and Cgulp2⌬/V strains), four (Cgsiz1⌬2⌬/V and Cgsiz1⌬2⌬/CgSIZ1 strains), and five (WT/V and Cgulp2⌬/CgULP2 strains) independent experiments) are plotted as percentage of the starting A 600 . Statistical analysis was performed using an unpaired, two-tailed, Student's t test (*, p Յ 0.05). Statistically significant differences between WT and the Cgulp2⌬ mutant are marked. C, cell wall chitin measurement. Log phase cultures of indicated C. glabrata strains were stained with 25 g/ml calcofluor white (CFW), and fluorescence intensity was measured by flow cytometry. Signal intensity mean values Ϯ S.E. represent data from three independent experiments. Unpaired, two-tailed, Student's t test is shown (***, p Յ 0.001). Statistically significant differences between WT and the Cgulp2⌬ mutant are marked. D, representative confocal images illustrating calcofluor white (CFW)-stained cell walls of log phase cultures of WT and Cgulp2⌬ mutant. Scale bar, 10 m; DIC, differential interference contrast. E, Western blotting of CgSlt2 phosphorylation in indicated C. glabrata strains grown in YPD medium, YPD medium containing 16 g/ml fluconazole (FLC), and 10 mM caffeine for 4 h at 30°C. Whole cell protein extracts were prepared by glass bead lysis and quantified using the Thermo Scientific Pierce BCA protein assay kit. 30 g of protein was separated on SDS-PAGE, and Western blots were developed with an anti-phospho-ERK1/2 MAPK (Thr 202 /Tyr 204 ) antibody. Individual band intensity was quantified using the ImageJ software. CgSlt2 phosphorylation signal was normalized to the corresponding CgGapdh signal in each lane, and results, representative of at least three independent experiments, are presented as fold change (Ϯ S.E.) in phosphorylation levels compared with CgSlt2 phosphorylation in YPD-grown wild-type cells (considered as 1). 5E, the Cgulp2⌬ mutant exhibited high levels of phosphorylated CgSlt2 compared with wild type and complemented strains indicating a constitutively active PKC-mediated cell wall integrity pathway. Furthermore, consistent with earlier studies, C. glabrata WT cells responded to the antifungal fluconazole and the cell wall stressor caffeine by activating the PKC cascade (see Fig. 5E). However, caffeine treatment, instead of activating the PKC pathway, resulted in down-regulation of CgSlt2 phosphorylation in the Cgulp2⌬ mutant (see Fig. 5E). Of note, the Cgulp2⌬ mutant was fully proficient in activation of the PKC pathway upon fluconazole exposure (see Fig. 5E). These data indicate that lack of CgULP2 adversely affects the ability of C. glabrata cells to respond to the cell wall stressor caffeine. Importantly, these findings are in accordance with regular and elevated sensitivity of the Cgulp2⌬ mutant to fluconazole and caffeine, respectively (see Fig. 4C).
Next, to investigate the effect of altered cell wall structure on the adhesion capacity of the Cgulp2⌬ mutant, we examined the ability of Cgulp2⌬ to adhere to Lec2 ovary epithelial cells. As a control, adherence assays were also carried out with Cgsiz1⌬, Cgsiz2⌬, and Cgsiz1⌬siz2⌬ mutants, all of which displayed no detectable cell wall abnormalities (see Fig. 5, B and C). As shown in Fig. 6A, the Cgulp2⌬ mutant displayed 2-fold less adherence to epithelial cells compared with that of the WT cells, which was again restored back to WT levels in the Cgulp2⌬-complemented strain. The hypo-adherence of the Cgulp2⌬ mutant was unexpected and found to be, in part, due to a 3-4-fold reduced expression of two epithelial adhesin-encoding genes EPA1 and EPA6 in the mutant (Fig. 6B). Notably, Epa1 and Epa6 belong to a family of at least 23 cell wall adhesins that mediate adherence of C. glabrata cells to host epithelial cells (39 -41). Epa6 has also been shown to be pivotal to biofilm formation in vitro (42). To examine the effect of EPA6 transcript levels on biofilm formation, we measured the ability of WT, Cgulp2⌬, and Cgulp2⌬complemented strains to make biofilm on polystyrene-coated plates. We observed that the CgULP2 disruption led to a 50% reduction in the biofilm formation capacity of C. glabrata cells (Fig. 6C). Collectively, these data indicate that CgUlp2 plays a role in regulated expression of adhesin-encoding genes, biofilm formation, and maintenance of cell wall architecture.
Perturbation of SUMOylation Affects Growth of C. glabrata-SUMOylation of protein targets is regulated in part by recruitment of SUMO ligases and isopeptidases to specific subcellular sites at specific instances (13,43). Therefore, we tested the effect of an additional 1-2 copies of E3 SUMO ligases and isopeptidases on the physiology of C. glabrata by transforming WT cells with the pRK74 plasmid expressing CgSIZ1, CgSIZ2, CgMMS21, CgULP1, or CgULP2 genes from the PGK1 promoter. We found that although additional copies of CgSIZ1, CgMMS21, CgULP1, or CgULP2 genes had no measurable effect on growth (data not shown), an additional copy of the CgSIZ2 gene resulted in perturbed growth (Fig. 7). This effect was specific to the minimal medium as C. glabrata WT cells carrying CgSIZ2-expressing plasmid exhibited normal growth on the YPD medium (see Fig. 7). Intriguingly, additional copy of CgSIZ2 also rendered cells sensitive to MMS, again in the minimal medium (see Fig. 7). Because the CgUlp2 deletion also caused similar phenotypes, albeit in all media, including YPD (see Figs. 4B and 7), we speculate that this may either be due to the increased SUMOylation of a critical substrate by CgSiz2 or absence of deSUMOylation of that substrate in the Cgulp2⌬ mutant.
Global SUMOylation Pattern Is Altered in C. glabrata Mutants Lacking SUMO Ligases and deSUMOylase-To detect the SUMO proteome of C. glabrata, we tagged the CgSmt3 protein with His 6 and FLAG epitopes at the N terminus (hereafter referred to as "dual tagged Smt3") and performed Western blots on whole cell extracts of the Cgsmt3⌬ mutant complemented either with CgSmt3 or dual tagged CgSmt3 using anti-FLAG antibody. The dual tagged CgSmt3 was able to complement the viability defect of the Cgsmt3⌬ mutant. As seen in Fig. 8A, we could detect multiple proteins only in the presence of FLAGlabeled CgSmt3 confirming that the tagged CgSmt3 was conjugated to cellular proteins. Although global protein SUMOylation was reduced in Cgsiz1⌬ and Cgsiz2⌬ mutants, the Cgsiz1⌬siz2⌬ double mutant had hardly any detectable SUMOylated proteins (see Fig. 8A). Furthermore, the presence of differentially SUMOylated proteins in Cgsiz1⌬ and Cgsiz2⌬ mutants is indicative of substrate specificity of CgSiz1 and CgSiz2 ligases (see Fig. 8A). Unexpectedly, in the Cgulp2⌬ mutant, we could detect neither SUMOylated proteins nor free SUMO. Next, we examined the SUMOylation pattern in C. glabrata WT cells expressing additional copies of CgSiz1, CgSiz2, CgUlp1, and CgUlp2 enzymes. CgSiz1 or CgSiz2 hyperexpression led to increased protein SUMOylation, with CgSiz2 being more effective than CgSiz1 with several high molecular weight SUMOylated proteins (Fig. 8B). Interestingly, elevating the dosage of the SUMO peptidases, CgUlp1 and CgUlp2, did not reduce the SUMOylated proteins (see Fig. 8B) suggesting that deSUMOylation is very well regulated.
The result that the Cgulp2⌬ mutant contains neither any SUMOylated proteins nor free SUMO is intriguing. We expected to see accumulation of SUMO-conjugated proteins in the absence of a deSUMOylase. There are two possible expla-nations for this observation as follows: first, as in Aspergillus nidulans (31), CgUlp2 is the SUMO-processing enzyme, and therefore, in its absence, no SUMOylation can be detected. Alternatively, increased accumulation of polySUMOylated proteins in the Cgulp2⌬ mutant leads to their degradation by the SUMO-dependent ubiquitination pathway. To test whether CgUlp2 is required for SUMO-processing, we made a CgSMT3 construct that encodes SMT3 without the last four amino acids and terminating with the diglycine motif that could be directly used for conjugation to substrates. This construct was expressed in Cgulp2⌬ mutant cells. We found that expressing mature SUMO improved the growth  of Cgulp2⌬ measurably but not up to wild-type levels (Fig.  9A). Similarly, it also conferred improved resistance to MMS; in both cases, merely adding additional copies of fulllength CgSMT3 did not improve growth (see Fig. 9A). We confirmed that the mature SUMO complemented the Cgsmt3⌬ phenotype using the plasmid shuffle assay described earlier (Fig. 9B). Western blot analyses revealed that now we could detect free SUMO and a very slight increase in the SUMO proteome (Fig. 9C). These observations together suggest that either providing mature SUMO improves the stability of critical SUMOylated proteins or having mature free SUMO improves survival by non-covalent associations. . Severity of the Cgulp2⌬ mutant phenotype is partly reduced by mature SUMO. A, copy of either empty vector or full-length CgSMT3 (SMT3) or CgSMT3 encoding mature SUMO (mSMT3) was introduced in the Cgulp2⌬ mutant and WT cells. Cells were tested for growth at non-permissive temperatures and plates containing MMS as indicated by spotting 5 l of 10-fold serially diluted cultures. Improved survival of the Cgulp2⌬ mutant could be observed in both conditions. B, Cgsmt3⌬ (YRK1022) was transformed with either dual-tagged SMT3 (1 and 2; pCKM405) or dual-tagged m-SMT3 (3 and 4; pCKM469) and tested for complementation of SMT3 function by loss of WT-CgSMT3. YPD-Nat growth indicates presence of the pCN-PDC1 plasmid, and growth on 5-FOA plates indicates ability to lose the WT-CgSMT3 carried on pGRB2.2 plasmid in the Cgsmt3⌬ strain. C, total protein extracts were made from indicated C. glabrata strains, and the SUMOylation pattern was analyzed as described above. The mature SUMO but not the full-length SUMO could be detected clearly in the Cgulp2⌬ mutant. Ponceau S-stained membrane is displayed as a loading control. D, indicated C. glabrata strains were grown overnight in the YNB medium containing 0.1% proline as a sole nitrogen source at 30°C and 200 rpm. Cells were harvested; A 600 was adjusted to 0.5 with fresh media containing 0.003% SDS and incubated for 4 h. Cultures were transferred into a 96-well plate containing 100 l of YNB-proline media and 300 M MG132 and incubated at 30°C at 175 rpm. After a 2 h incubation, MMS (0.045%) was added to the media and grown for 3 h. Cultures were 10-fold serially diluted, and a 3 l volume was spotted on YPD medium. Plate images were captured after 48 h of incubation at 30°C.
To address the other possibility that increased accumulation of polySUMOylated proteins in the Cgulp2⌬ mutant leads to their degradation by the ubiquitin-proteasome pathway, we checked the SUMOylation pattern in Cgulp2⌬ cells treated with the proteasome inhibitor MG132, but we did not find any increase in SUMOylation levels of proteins (data not shown). Consistently, MG132 treatment could only marginally rescue the MMS sensitivity of the Cgulp2⌬ mutant (Fig. 9D). Of note, no further rescue of MMS sensitivity was observed upon expression of full-length or mature SUMO protein in the Cgulp2⌬ mutant (see Fig. 9D). These data suggest that the hyperactivated ubiquitin-proteasome proteolytic pathway is unlikely to be the sole cause of overall reduction in SUMOylation levels in the Cgulp2⌬ mutant.
Ethanol Stress and Macrophage Internalization Induce Accumulation of SUMO-conjugated Proteins in C. glabrata-SUMOylation has previously been implicated in the cellular response to diverse stresses (42,43). We therefore checked alterations in the SUMO proteome of C. glabrata cells upon a 4-h exposure to DNA damage (0.075% MMS), ethanol (10%), oxidative (150 mM H 2 O 2 ) stresses, and antifungal (fluconazole 64 -100 g/ml) stress. Western blot analysis using anti-FLAG antibody revealed that treatment with alcohol led to increased SUMOylation of several proteins, whereas MMS and H 2 O 2 exposure showed increased SUMOylation for a few proteins (Fig. 10A) suggesting that SUMO conjugation of multiple proteins is a part of cellular stress response in C. glabrata. Notably, contrary to expectations, C. glabrata SUMOylation-defective mutants, Cgsiz1⌬, Cgsiz2⌬, Cgsiz1⌬siz2⌬, and Cgulp2⌬ mutants, did not exhibit elevated susceptibility to ethanol stress (data not shown), which could be reflective of functional redun-dancy among components of the SUMOylation machinery in C. glabrata.
C. glabrata cells encounter several stressful conditions in the internal milieu of macrophages, including the nutrient-limiting reactive oxygen species-generating environment (30). To gain insights into whether intracellular survival/proliferation of C. glabrata cells in macrophages involves SUMO proteome modifications, we performed co-incubation assays. Macrophages derived from the human monocytic cell line THP-1 were co-incubated for 8 h with dual tagged SMT3-expressing C. glabrata cells, and the internalized C. glabrata were harvested. Total protein from these yeast cells was compared with RPMI 1640 medium-grown C. glabrata cells for any changes in SUMO proteome. As seen in Fig. 10B, in comparison with RPMI 1640 medium-grown C. glabrata cells, macrophage-internalized C. glabrata cells showed several prominently SUMOylated proteins. This establishes that to survive and multiply within macrophages, several C. glabrata proteins are SUMOylated.
Next, to investigate whether components of the SUMOylation machinery are required for virulence of C. glabrata, we examined fungal burden in BALB/c mice infected intravenously either with the wild-type or the Cgsiz1⌬siz2⌬ and Cgulp2⌬ mutant strains. Approximately, 10-and 8-fold lower yeast CFUs were recovered from the kidneys and liver, respectively, of the mice infected with the Cgulp2⌬ mutant compared with CFUs retrieved from corresponding organs of the WTinfected mice (Fig. 11, C and D). Ectopic expression of the CgULP2 gene restored the organ fungal burden in the Cgulp2⌬infected mice (see Fig. 11, C and D). Of note, no statistically significant differences in the fungal burden were seen between the spleen of WT-and Cgulp2⌬-infected mice (Fig. 11E).
Importantly, statistically similar yeast CFUs were obtained from all three target organs of WT-and Cgsiz1⌬siz2⌬-infected mice (see Fig. 11, C-E). Together, these data indicate an organspecific role for the CgUlp2 deSUMOylase and dispensability of CgSiz1 and CgSiz2 SUMO ligases in survival of C. glabrata in the murine model of disseminated candidiasis.

Discussion
In this work, we have initiated studies to understand the importance of protein SUMOylation in the pathobiology of C. glabrata. First, we identified components of the SUMOylation machinery in C. glabrata based on homology with S. cerevisiae. Second, we performed complementation studies in S. cerevisiae to confirm the predicted activities. Third, we generated C. glabrata deletion strains for non-essential SUMOylation genes and examined the effects of perturbed SUMOylation on stress response and survival in vivo. Finally, we demonstrated the essentiality of CgSmt3 for cell growth in C. glabrata. Our studies firmly establish that SUMOylation, like in S. cerevisiae, is essential in C. glabrata. This is different FIGURE 11. CgUlp2 desumoylase is required for virulence in the murine model of disseminated candidiasis. A, 1 ϫ 10 5 cells of indicated C. glabrata strains were added to phorbol 12-myristate 13-acetate-differentiated THP1 cells (1 ϫ 10 6 ) in a 24-well plate followed by removal of non-phagocytosed C. glabrata cells after 2 h. At indicated time intervals, intracellular yeast cells were recovered by lysing THP-1 cells in water and plating appropriate dilutions of lysates on the YPD medium. Yeast colonies were counted after 24 -48 h of incubation at 30°C. Data represent the mean of three to four independent analyses (Ϯ S.E.). B, cell proliferation of the indicated C. glabrata strains in the RPMI 1640 medium containing 10% FBS (fetal bovine serum) medium was assessed by CFU assay. Data represent the means of three to four independent analyses (ϮS.E.). C-E, 6 -8-week-old female BALB/c mice were infected intravenously with 4 ϫ 10 7 cells of indicated C. glabrata strains and sacrificed 7 days post-infection. Diamonds represent the CFUs recovered from target organs kidney (C), liver (D), and spleen (E) for individual mice. Horizontal line represents the geometric mean (n ϭ 8 -14) of CFUs per organ for all mice in one group. Statistically significant differences in CFUs between WT and mutant strains are indicated (**, p Յ 0.01; Mann-Whitney test). from other yeast/fungi like Schizosaccharomyces pombe, A. nidulans, and C. albicans, where SMT3 is not essential for survival (22,31,44). We also speculate that the essential SUMO conjugation in C. glabrata may be carried out by the CgMms21 because the Cgsiz1siz2⌬ mutant is viable (see Fig. 4A) and the Cgmms21⌬ mutant could not be generated. In addition, we could not create knock-outs for genes encoding SUMO-processing enzyme CgULP1, SUMO-activating enzymes CgAOS1 and CgUBA2, and SUMO-conjugating enzyme CgUBC9, suggesting that these, like in S. cerevisiae, could also be essential in C. glabrata.
An initial examination of the SUMO proteome in C. glabrata revealed several proteins to be SUMOylated. Lack of SUMO ligases, CgSiz1 and CgSiz2, led to loss of the majority of SUMOylation (see Fig. 8A), although their increased dosage resulted in accumulation of additional SUMO-conjugated proteins (see Fig. 8B). Interestingly, we observed several high molecular weight SUMOylated proteins with additional copies of CgSiz2 (see Fig. 8B) suggesting that, as in S. cerevisiae (45), CgSiz2 may also be involved in polySUMOylation in C. glabrata. SUMOylation pattern studies in the Cgulp2⌬ mutant revealed intriguing results as, surprisingly, we could neither detect SUMOylated proteins nor the SUMO protein in the mutant (see Fig. 8A). We reasoned that there maybe two possible explanations for this observation as follows. First, CgUlp2 is the SUMO-processing enzyme and therefore, in the absence of mature SUMO, no sumoylation takes place. Second, in the Cgulp2⌬ mutant, SUMOylated proteins accumulate and are targeted for degradation by the SUMO-dependent ubiquitination pathway. Our experiments to test these possibilities suggest that CgUlp2 is unlikely to be the processing enzyme because introduction of the processed SUMO does not completely alleviate Cgulp2⌬ phenotypes. Similarly, inhibition of proteasome also does not completely rescue the phenotype. However, both treatments improve Cgulp2⌬ growth perceptibly suggesting that tilting the balance toward retaining SUMOylated proteins can improve Cgulp2⌬ phenotypes. It further suggests that lack of a SUMOylated protein rather than deSUMOylating a critical protein(s) as a cause for Cgulp2⌬ mutant phenotypes cannot be ruled out.
Three lines of evidence indicate an important role for SUMOylation in cellular stress response in C. glabrata. First, perturbing SUMOylation by deleting SUMO ligases CgSiz1 and CgSiz2 or deSUMOylating enzyme CgUlp2 resulted in diminished survival under different stress conditions (see Fig. 4, B and C). Second, C. glabrata cells respond to a changing environment, including ethanol stress and macrophage internal milieu by SUMOylating multiple cellular proteins (see Fig. 10). Third, the C. glabrata mutant disrupted for CgULP2 exhibited altered cell morphology and cell wall architecture and the constitutively active terminal MAPK (CgSlt2) of the cell wall integrity signaling pathway (see Fig. 5). These findings along with strong survival defects of the Cgulp2⌬ mutant in specific tissues in the disseminated candidiasis model (see Fig. 11, C-E) indicate that SUMOylation is pivotal to stress response in C. glabrata.
The fungal cell wall is a dynamic organelle, and any alterations to its core constituents (␤-glucan, chitin, and mannoproteins) result in the activation of compensatory mechanism(s) (37). The Cgulp2⌬ mutant contained elevated levels of chitin in the cell wall (Fig. 5C). Importantly, an increase in the cell wall chitin content has previously been associated with reduced susceptibility to the ␤-1,3-glucan synthesis-targeting echinocandin antifungals and diminished virulence in C. albicans (46,47). However, despite elevated chitin levels, the Cgulp2⌬ mutant was found to not be resistant to caspofungin (see Fig. 4C). Of note, C. glabrata mutants with perturbed SUMOylation (Cgsiz1⌬, Cgsiz2⌬, Cgsiz1⌬Cgsiz2⌬, and Cgulp2⌬) did not exhibit altered sensitivity to the ergosterol biosynthesis inhibitor, fluconazole, either (see Fig. 4C). Consistently, cellular SUMOylation status remained unaffected upon exposure of C. glabrata cells to fluconazole (see Fig. 10A) suggesting that SUMOylation is dispensable for response of C. glabrata cells to azole and echinocandin antifungal drugs.
An important feature of this work is the strong conservation of the SUMOylation pathway between S. cerevisiae and C. glabrata. Similar to S. cerevisiae (48,49), SUMOylation appears to be involved in the regulation of the telomere position effect in C. glabrata. Adherence to the host tissue, an important virulence attribute of C. glabrata, is mediated by a large family of cell wall proteins, including Epa adhesins (30,40). The majority of the adhesin-encoding EPA genes in C. glabrata are transcriptionally silenced due to their close proximity to telomeres (40 -42). SUMOylation is a reversible process, and the tight regulation of SUMO conjugation and SUMO removal is probably necessary to fine-tune substrate functions (1,2). Our finding that a lack of the CgUlp2 deSUMOylase resulted in further repression of EPA1 and EPA6 expression and hypoadherence (see Fig. 6, A and B) indicates that removal of the SUMO modification from component(s) of the subtelomeric silencing machinery is probably pivotal to the expression of subtelomeric genes. This is particularly relevant for the virulence of C. glabrata due to the subtelomere-based localization of EPA genes that are likely to be relieved of the telomere position effect in response to host environmental cues. Notably, transcriptional activation of EPA6, which is required for biofilm formation, has been shown to be regulated by niacin levels in the mouse model of urinary tract infection (47,50).
Taken together, our findings demonstrate for the first time a pivotal role for SUMOylation in the life cycle of C. glabrata.

Experimental Procedures
Strains and Culture Conditions-Bacterial and C. glabrata strains were routinely maintained in the LB medium (1% tryptone, 0.5% yeast extract, and 1% NaCl) at 37°C and YPD (1% yeast extract, 2% peptone, and 2% dextrose) medium at 30°C, respectively. For alternative carbon source utilization, the synthetically defined YNB (0.67% YNB and 2% dextrose) medium was used. C. glabrata cultures were grown in the CAA medium (0.67% yeast nitrogen base (YNB) without amino acids, 0.6% casamino acids, and 2% dextrose) for adherence analysis. Logarithmic (log) phase C. glabrata cells were obtained after incubation of overnight cultures for 4 h in the fresh medium at 30°C with shaking at 200 rpm. S. cerevisiae strains were routinely grown either in YPD medium or minimal medium lacking the indicated nutrient. Bacterial, S. cerevisiae, and C. glabrata strains and plasmids used in this study are listed in Table 2.

TABLE 2 List of strains and plasmids used in the study
Construction of C. glabrata Deletion Strains and Plasmids-C. glabrata siz1⌬, siz2⌬, and ulp2⌬ strains were created using the homologous recombination-based strategy as described previously (51). For generation of the Cgsiz1⌬siz2⌬ strain, 433 and 406 bp of 5ЈUTR and 3ЈUTR regions, respectively, of the CgSIZ2 ORF were cloned in the pAP599 plasmid in such a way that CgSIZ2 UTRs flank each end of the hph1 gene. The 2.96-kb fragment containing 5ЈUTR-CgSIZ2, hph1 gene, and 3ЈUTR-CgSIZ2 was obtained by digestion of the plasmid pRK990 with KpnI and SacI restriction enzymes and transformed into the Cgsiz1⌬::nat1 strain via the lithium acetate method. Transformants were selected on the YPD medium containing hygromycin and confirmed for disruption of the CgSIZ2 ORF with the hph1 gene via PCR. To generate the Cgsmt3⌬/CgSMT3 strain, C. glabrata wild-type strain (YRK19) was first transformed with the plasmid pRK1069 containing full-length CgSMT3 gene under the PGK1 promoter. Transformants were selected for uracil prototrophy and colony purified followed by replacement of the genomic CgSMT3 locus with the hph1 gene. Disruption of the genomic CgSMT3 locus in the YRK1020 strain was achieved by transforming with a linear DNA fragment (3.91 kb), carrying 5ЈUTR-CgSMT3, hph1 gene, and 3ЈUTR-Cg-SMT3, obtained from the XhoI and SacI-digested pRK985 plasmid. Hygromycin-resistant colonies were checked for disruption of the CgSMT3 ORF with the hph1 gene via PCR.
All plasmids encoding C. glabrata genes were constructed by amplifying sequence-encoding full-length protein from genomic DNA and placed downstream of the PGK1 promoter in pRK74 plasmid. The His 6 -3ϫFLAG tags were constructed by first ligating an oligonucleotide encoding three copies of the FLAG sequence in the NheI/BamHI site of pRSETa plasmid (Invitrogen). Then the cassette containing His-FLAG was isolated as an XbaI/BamHI fragment and placed downstream of the PGK1 promoter in the pRK74 plasmid. For tagging of C. glabrata proteins at the N terminus, C. glabrata genes were placed in-frame with the His-FLAG sequence in the pRK74 plasmid. Construct encoding mature SUMO was made by amplifying the coding region of SMT3 until the diglycine motif and including a stop codon in the reverse primer. Sequence of primers used and further details are available upon request.
Microscopy Analysis-For differential interference contrast (DIC) microscopy, YPD medium-grown overnight cultures of wild-type (WT) and Cgulp2⌬ strains were inoculated in the fresh YPD medium to an A 600 of 0.1. After 48 -72 h of growth at 30°C, cells were washed twice with sterile PBS, suspended in PBS, and visualized using the Nikon eclipse 80i microscope (ϫ100 oil immersion objective). For calcofluor white staining, 1 ml of the above-prepared cultures were fixed with 4% formaldehyde for 2 h at room temperature, washed three times, and suspended in PBS. A 4-l cell suspension was stained with 1 l of calcofluor white (1 mg/ml solution) and imaged using the confocal microscope (Carl Zeiss LSM 700).
Immunofluorescence was performed with a slight modification to the procedure described earlier (48). Briefly, overnight cultures were subcultured to 0.5-0.6 A 600 and spun down at 3000 ϫg for 5 min. After washing with sterile water, cells were suspended in 200 l of 10 mM DTT and 0.1 M EDTA-KOH and incubated at 30°C for 10 min. Cells were collected by centrifu-gation and resuspended in 500 l of YPD containing 1.2 M sorbitol. 50 l of 2.5 mg/ml zymolyase and a pinch of lyticase was added to the cell suspension and incubated at 30°C for 45 min. Spheroplasting was monitored under a light microscope, and when complete, the cells were harvested by spinning at 2000 ϫg for 10 min and washed three times with YPD medium containing 1.2 M sorbitol. Finally, the spheroplasts were resuspended in 100 l of YPD supplemented with 1.2 M sorbitol. Immunofluorescence was performed as described earlier (48). Imaging was done in Zeiss Axio Scope A1 microscope equipped with an AxioCam camera and processed using the Zen software.
Quantitative Real Time PCR-C. glabrata strains were grown in the YPD medium for 48 h followed by incubation in fresh YPD medium for 1 h. Cells were collected, and RNA was extracted using the acid phenol extraction method. First-strand cDNA synthesis was done using the SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen), and quantitative PCR was performed using the SYBR Green Master Mix (Eurogentec). CgGAPDH (glyceraldehyde-3-phosphate dehydrogenase) mRNA was used to normalize RT-quantitative PCR data.
Western Blot-Whole cell extracts were made from overnight C. glabrata cultures using trichloroacetic acid (TCA) precipitation (49). An equal number of cells as measured by absorbance were used for protein extractions, and protein levels in samples were estimated by SDS-PAGE/Ponceau-stained blots and, in some cases, additionally by Pierce BCA kit. Protein samples were separated on 10% SDS-PAGE and transferred onto PVDF membrane. After blocking with 3% skimmed milk powder for 1 h, membrane was incubated with FLAG primary antibody (Sigma, 1:10,000) for 1 h, washed three times with TBST, and incubated with the mouse HRP-conjugated secondary antibody (Jackson ImmunoResearch, 1:15,000) for 1 h. Blots were washed three times with TBST and developed with chemiluminal developing solutions from G-Biosciences.
Chitin Estimation-Overnight cultures were inoculated in the YPD medium to an A 600 of 0.1 and incubated at 30°C for 6 h. Cells were washed twice with PBS, normalized to an A 600 of 2.0, and incubated with 2.5 l of calcofluor white solution (10 mg/ml) for 15 min at room temperature in the dark. After two PBS washes, 12.5 l of cell suspension (ϳ50,000 yeast cells) was diluted 24-fold in PBS and used to measure mean fluorescence intensity via flow cytometry (BD FACS ARIA III). Mean fluorescence intensity ratio was calculated by dividing the fluorescence intensity value of the mutant sample with that of the WT sample.
Biofilm Assay-The ability of C. glabrata cells to produce biofilms on polystyrene-coated plates was assessed as described previously (52). One ml of the YPD-grown logarithmic culture (1 ϫ 10 5 cells in PBS) was added to a well of the polystyrenecoated 24-well plate and incubated at 37°C for 90 min. After two PBS washes, 1 ml of the RPMI 1640 medium containing 10% FBS was added to each well, and the plate was incubated at 37°C with shaking (75 rpm). After 24 h, 500 l of the spent medium was replaced with the fresh RPMI 1640 medium, and incubation was continued for another 24 h. Unbound C. glabrata cells were removed with three PBS washes, and the plate was air-dried for 45 min and incubated with 250 l of crystal violet solution (0.4% in 20% ethanol) for 45 min. Well attached C. glabrata cells were washed four times with PBS to eliminate surplus crystal violet stain and incubated with 95% ethanol for 45 min. Absorbance of the 100-l destaining solution was recorded at 595 nm, which is reflective of the number of biofilm-forming C. glabrata cells. Absorbance values of wells without C. glabrata cells were subtracted from those of yeastcontaining wells, and the biofilm ratio was calculated by dividing the mutant absorbance units by those of WT cells.
THP-1 Macrophage Infection-THP-1 monocytes seeded at a density of 1 ϫ 10 6 per well of a 24-well tissue culture plate were treated with 16 nM phorbol 12-myristate 13-acetate for 12 h followed by infection with C. glabrata strains to a multiplicity of infection of 0.1. After 2 h, the wells were washed three times with PBS to remove extracellular yeast cells. THP-1 cells were lysed in water, and the number of intracellular C. glabrata cells was determined by plating appropriate dilutions of lysates on rich medium.
Mouse Infection Assay-Experiments involving mice were performed at the Centre for DNA Fingerprinting and Diagnostics animal facility, VIMTA Labs Ltd., Hyderabad, India (www. vimta.com), in strict accordance with the guidelines of The Committee for the Purpose of Control and Supervision of Experiments on Animals, Government of India. The protocol was approved by the Institutional Animal Ethics Committee of the Vimta Labs Ltd. (Institutional Animal Ethics Committee protocol approval number PCD/CDFD/05). YPD mediumgrown C. glabrata cells (4 ϫ 10 7 , 100 l of PBS cell suspension) were injected into 6 -8-week-old female BALB/c mice through the tail vein. At day 7 post-infection, mice were sacrificed, and three target organs, kidneys, liver, and spleen, were collected. Organs were homogenized in 1 ml of sterile PBS, and organ fungal load was determined by plating appropriate homogenate dilutions on the YPD medium containing penicillin and streptomycin.
Other Procedures-S. cerevisiae proteins were retrieved from the Saccharomyces Genome Database, and their orthologues in C. glabrata were searched for using Blastp. Percent similarity and identity were calculated using EMBOSS Stretcher (pairwise sequence alignment) tool (53). The protein sequences were scanned for annotated domains using Pfam and HMMER. Maps of proteins along with their domains were generated using DOG (Domain Graph) (54). Zymolyase digestion, CgSlt2 phosphorylation, and adherence analysis were performed as described previously (38,51).
Author Contributions-R. G., S. V., K. K., S. S. T., K. M., and R. K. conceived the experiments. R. G., S. V., and K. K. performed the experiments. R. G., S. V., K. M., and R. K. analyzed the data and wrote the manuscript.