A family of glycosylphosphatidylinositol-linked aspartyl proteases is required for virulence of Candida glabrata

Abstract

Candida glabrata is a yeast pathogen of humans. We have established a tissue culture model to analyze the interaction of C. glabrata with macrophages. Transcript profiling of yeast ingested by macrophages reveals global changes in metabolism as well as increased expression of a gene family (YPS genes) encoding extracellular glycosylphosphatidylinositol-linked aspartyl proteases. Eight of these YPS genes are found in a cluster that is unique to C. glabrata. Genetic analysis shows that the C. glabrata YPS genes are required for cell wall integrity, adherence to mammalian cells, survival in macrophages and virulence. By monitoring the processing of a cell wall adhesin, Epa1, we also show that Yps proteases play an important role in cell wall re-modeling by removal and release of glycosylphosphatidylinositol-anchored cell wall proteins.

Candida species are opportunistic pathogens of humans causing both mucosal and disseminated infections. Candida glabrata and Candida albicans are responsible for ≈15% and 60% of candidiasis, respectively (1). In C. albicans, important virulence attributes include the ability to grow in both yeast and hyphal forms and the production of secreted proteinase activity (2). C. glabrata's ability to cause disease is independent of both of these, because it does not secrete proteinase activity and apparently cannot make true hyphae (1).

Phagocytic cells, including neutrophils and macrophages, are important elements in the host defense against Candida infection. Morphologically, phagocytosed C. albicans yeast cells rapidly differentiate into hyphae that grow out of the macrophage, eventually killing it. C. albicans mutants unable to switch to hyphal growth are avirulent (3). Analysis of the C. albicans transcriptional response to macrophage internalization shows a remodeling of carbon metabolism, including repression of genes in the glycolytic pathway and induction of genes in gluconeogenic pathways that are required for growth on C2 carbon sources, probably derived from β-oxidation of fatty acids (4, 5).

Because C. glabrata does not make true hyphae, does it have an alternative response to phagocytosis? In this study, we analyze the interaction of C. glabrata with the mouse macrophage-like cell line J774A.1. We show that macrophage-internalized C. glabrata exhibit transcriptional induction of a C. glabrata-specific cluster of eight genes encoding a family of putative aspartyl proteases. These genes are closely related to the YPS (Yapsin) genes of S. cerevisiae (6). The S. cerevisiae yapsins are a family of five glycosylphosphatidylinositol (GPI)-linked aspartyl proteases (Yps1–3, Yps6, and Yps7) that have been shown to cleave peptides C-terminal to basic residues both in vitro and in vivo. The S. cerevisiae YPS genes are induced during cell wall remodeling, and strains deleted for YPS genes are sensitive to cell wall disrupting agents and have reduced amounts of β-1,3 and β-1,6 glucans in their cell walls (7). In C. albicans, the Yps-related proteases Sap9 and Sap10 have been implicated in C. albicans virulence: deletion of the SAP9 and SAP10 genes alters adherence of yeast to epithelial cells and reduces virulence in an in vitro model of oral candidiasis (8).

In this study, we show that the C. glabrata YPS genes have important roles in activation of, and survival within, macrophages and they are required for virulence. Moreover, we demonstrate a physiological role of C. glabrata Yps-family proteases in processing the GPI-linked adhesin, Epa1 (9).

Results

C. glabrata Survives and Replicates in J774A.1 Macrophages.

To study the interaction of C. glabrata with macrophages, we infected the mouse macrophage-like cell line J774A.1 with C. glabrata wild-type cells (strain BG2) at a MOI of either 1:1 or 1:10. Survival and growth of C. glabrata was monitored as a function of time. Within 1 h after infection, microscopic examination showed that essentially all yeast were internalized (data not shown). After 24 h, we observed a consistent increase (4- to 6-fold) in yeast colony-forming units (CFUs) (Table 1 and data not shown). Similar replication (4.5- to 5-fold) was observed for four additional clinical isolates of C. glabrata (data not shown). Microscopic analysis showed no evidence of extracellular yeast during the time course; moreover, we found that the number of yeast per individual macrophage increased over time (data not shown), indicating that the observed increase in CFUs is due to intracellular replication. As a control, we found that for S. cerevisiae (strains BY4742 and five additional clinical isolates), there was no replication over 24 h (Table 1 and data not shown).

Table 1.

Survival of C. glabrata ypsΔ strains in macrophages and nitrite production in infected macrophages

Strain	Mφ-associated yeast (24 h/2 h)*	Nitrite, μM†
IFN-γ alone	Not applicable	3.8 ± 0.2
BG2	4.4 ± 0.6	3.7 ± 0.4
yps1Δ	1.1 ± 0.4	5.5 ± 0.6
yps7Δ	3.9 ± 0.5	5.0 ± 0.4
yps1Δ yps7Δ	0.4 ± 0.05	7.4 ± 0.6
ypsCΔ	5.4 ± 1.9	3.2 ± 0.4
yps1Δ ypsCΔ	0.3 ± 0.1	9.8 ± 0.7
yps7Δ ypsCΔ	2.4 ± 1.4	5.3 ± 0.4
yps(1–11)Δ	0.03 ± 0.007	16.6 ± 0.7
S. cerevisiae‡	1.07 ± 0.09	7.4 ± 0.8

*The ratio of yeast CFUs recovered from macrophages at 24 hr versus 2 hr after infection. Values shown are means ± SD from 3 biological replicates.

^†The nitrite concentration measured from the culture medium of 5 × 10⁵ macrophages at 24 hr after infection. The experiments were repeated three times with biological replicates. Values shown are means ± SD from one representative experiment (done in quadruplicate).

^‡The strain used is BY4742.

To analyze the transcriptional response of C. glabrata following macrophage internalization, we used whole genome oligonucleotide microarrays to compare the transcript profiles of yeast recovered 2 h and 6 h postinfection versus yeast grown in the same tissue culture media in the absence of macrophages. A total of 131 and 288 genes were significantly induced (>2-fold) and 485 and 453 genes were repressed (>2-fold) at 2 h and 6 h postinfection, respectively [complete data sets can be found in the Gene Expression Omnibus (GEO) database (accession no. GPL3922)]. Our transcription profiling shows a response that mirrors very closely the response previously described for C. albicans coincubated with macrophages or other phagocytic cells (4, 5). In particular, phagocytosed C. glabrata represses genes involved in glycolysis (CDC19) whereas up-regulating genes involved in gluconeogenesis (e.g., FBP1 and PCK1), β-oxidation of fatty acids (e.g., FAA2, FOX2, POT1, POX1), glyoxylate cycle (e.g., ICL1, ACO1, MLS1) and methylcitrate cycle (PDH1, CIT3, ICL2). This response is most striking at the 2 h time point, but these pathways remain significantly up-regulated at the 6 h time point as well. Notably, there is an induction of transporters for amino acid (GAP1, CAN1), and acetate (ADY2) as well as concerted up-regulation of the Arg and Lys biosynthetic pathways. Also, as observed for C. albicans, there is a concerted down-regulation of the translational apparatus (ribosomal protein genes, tRNA synthetases, translation initiation, and elongation factors) after phagocytosis. Our data suggest that the response of C. glabrata to macrophage internalization is highly similar to that of C. albicans, and includes a wholesale shift in carbon metabolism as well as down-regulation of the translational apparatus.

Up-Regulation of a Family of GPI-Linked Aspartyl Proteases upon Macrophage Internalization.

Among the C. glabrata genes transcriptionally induced by phagocytosis were genes encoding putative GPI-linked aspartyl proteases [Fig. 1 and supporting information (SI) Table 3]. C. glabrata encodes 11 predicted GPI-linked aspartyl proteases (Fig. 1A and http://cbi.labri.fr/Genolevures/elt/CAGL). These genes show structural similarity to the S. cerevisiae YPS genes and we have assigned them the names YPS(1-11). CgYPS1 and CgYPS7 are the closest orthologues of ScYPS1 and ScYPS7, respectively, and are encoded at syntenic loci; CgYPS2 is syntenic with ScYPS2. The remaining eight C. glabrata genes (CgYPS3-6, 8-11) are encoded in a cluster (called hereafter the YPS cluster) on chromosome E, 36 kb from the YPS2 locus. Flanking the YPS cluster are the C. glabrata orthologues of ScMDH2 (YOL126c) and ScYOL125w, genes that form an adjacent pair in both S. cerevisiae and Ashbya gossypii, suggesting recent acquisition of the YPS cluster in the C. glabrata phylogenetic lineage. YPS4 and YPS11 may not encode bona fide GPI-proteins, because, in the published genomic sequence, YPS4 lacks a C-terminal hydrophobic GPI-addition signal, whereas YPS11 lacks a signal sequence.

Fig. 1.

Macrophage-induced transcriptions of C. glabrata YPS genes. (A) Schematic representation of C. glabrata YPS gene loci. (B) Relative mRNA abundance of YPS genes in C. glabrata coincubated with J774A.1 macrophages for 6 h (filled bars) and in cells grown in DMEM alone (open bars), as measured by quantitative real-time PCR. Results are the means ± SD of two biological duplicate experiments, each performed in triplicate.

To confirm our microarray results, we used reverse transcription followed by quantitative real-time PCR to assess the transcription of the C. glabrata YPS genes after phagocytosis (Fig. 1B). Of the YPS genes encoded outside the cluster, transcription of YPS1 and YPS7 is not affected by coincubation with macrophages; transcription of YPS2, which is expressed at very low levels, increases 2.5-fold in the presence of macrophages. Within the YPS cluster, YPS3 and YPS6 are expressed in yeast coincubated with macrophages or grown in medium alone. The remaining six YPS genes in the cluster are induced transcriptionally in the presence of macrophages (3.5- to 30-fold).

Characterization of Aspartyl Protease Mutants.

To study the role of the Yps proteases in C. glabrata virulence, we made a panel of strains disrupting combinations of YPS genes (SI Table 4). These include yps1Δ, yps7Δ, and yps1Δ yps7Δ strains. In addition, we deleted YPS2 and the YPS cluster as a group, (abbreviated ypsCΔ, where “C” stands for cluster). Other strains constructed are yps1Δ ypsCΔ, yps7Δ ypsCΔ, and yps(1-11)Δ (deleted for all 11 YPS genes). The yps(1-11)Δ and yps1Δyps7Δ mutants showed modest growth defects in liquid yeast extract/peptone/dextrose (YPD), whereas all other mutants grew as well as wild-type (SI Fig. 5 and data not shown).

Because yapsins are required for cell wall integrity in S. cerevisiae (7), we tested C. glabrata ypsΔ strains for cell wall-related phenotypes. Unlike the S. cerevisiae yps mutants, the C. glabrata yps(1-11)Δ strain showed no significant growth defect at 37°C. None of the ypsΔ strains showed sensitivity when grown in the presence of CaCl₂, fluconazole, or H₂O₂ (data not shown). As shown in Fig. 2A, yps1Δ strains were sensitive to NaCl and caffeine, but displayed wild-type growth in the presence of Congo red (CR) (which disrupts chitin and β-glucan fiber formation) and calcofluor white (CW) (an inhibitor of chitin polymer assembly). yps7Δ strains were sensitive to CR and CW. The yps1Δyps7Δ and yps(1-11)Δ strains were sensitive to NaCl, caffeine, CW, and CR, whereas the ypsCΔ strain displayed wild-type growth under all conditions tested. Strains lacking YPS1 or YPS7 alone or in combination with other YPS genes showed elevated resistance to zymolyase treatment as compared with the wild-type or ypsCΔ strains (SI Fig. 6). C. glabrata strains lacking YPS1 alone or in combination with other YPS genes show a dramatic loss in viability in stationary phase, with the yps1Δyps7Δ or yps(1-11)Δ strain displaying ≈0.1% survival after 96 h in culture (Fig. 2B). Taken together, these data suggest that YPS1 and YPS7 have critical roles in the maintenance of cell wall integrity as well as in stationary phase survival, whereas the macrophage-induced genes (YPS2 and the YPS cluster) had no apparent role in either cell wall integrity or stationary phase survival under the conditions tested.

Fig. 2.

Phenotypic characterization of C. glabrata ypsΔ strains. (A) C. glabrata ypsΔ strains display sensitivity to drugs causing cell wall stress. Equal number of cells were spotted in 10-fold serial dilutions onto YPD plates alone or supplemented with different compounds. Plates were photographed after 2 days at 30°C. (B) C. glabrata ypsΔ strains lose viability in stationary phase. Cells were grown in liquid YPD at 30°C, and viable cells as a percentage of total cells were determined as a function of time over 96 h.

Role of the YPS Cluster in C. glabrata Survival in Macrophages.

Transcription of YPS2 and the YPS cluster is induced upon macrophage internalization. To determine whether the YPS genes are required for C. glabrata survival or replication in macrophages, we infected J774A.1 cells with either the wild-type or ypsΔ strains at an MOI of 1:10 and monitored intracellular survival. As shown in Table 1, the ypsCΔ strain behaved like wild-type, showing a 4- to 6-fold increase in CFUs over 24 h. The yps7Δ strain also had no significant phenotype, whereas the yps7Δ ypsCΔ strain showed slightly reduced replication (2-fold). Over 24 h, the yps1Δ strain showed no increase in CFUs, the yps1Δ yps7Δ strain showed a 2-fold reduction, and the yps1Δ ypsCΔ strain showed a 3-fold reduction, whereas the yps(1-11)Δ strain showed a 33-fold reduction in CFUs. These data make clear that the C. glabrata YPS genes are required for survival in the macrophage intracellular environment. Importantly, the macrophage-induced YPS genes do have some role in intracellular survival and replication, which can best be seen by comparing the intracellular survival of the yps1Δ yps7Δ with that of the yps(1-11)Δ strain in which the macrophage-induced genes are additionally deleted.

To test whether macrophages are activated by exposure to C. glabrata, we monitored the production of nitric oxide (NO), a reaction catalyzed by inducible nitric oxide synthase (iNOS) (10, 11). NO is converted to nitrite, which can be measured spectro-photometrically. We pretreated J774A.1 macrophages with Interferon-γ (IFN-γ) for 5 h, followed by infection with either wild-type C. glabrata or S. cerevisiae (MOI 1:1). Macrophages infected with strain BG2 or any of the four C. glabrata clinical isolates for 24 h showed no increase in nitrite production over background levels (treated with IFN-γ alone). In contrast, a 2.5- to 3-fold increase in nitrite levels was seen after infection with BY4742 or additional S. cerevisiae clinical isolates (Table 1 and data not shown). When macrophages were infected with the yps(1-11)Δ strain, a 4.5-fold increase in total nitrite production was seen; modest increases of <2-fold were seen for the yps1Δ and yps7Δ strains, whereas 2-fold and 2.7-fold increases were observed for the yps1Δ yps7Δ and yps1Δ ypsCΔ strains, respectively (Table 1). We conclude that the C. glabrata ypsΔ strains stimulate macrophages to produce significantly more NO than is stimulated by wild-type strains. A role for the macrophage-induced YPS genes is indicated by the fact that the maximal induction of NO occurs in the strain lacking all of the YPS genes [compare the yps1Δ yps7Δ and yps(1-11)Δ strains].

The YPS Genes Are Required for C. glabrata Virulence.

We next examined the role of YPS genes during infection. We used a mouse model of disseminated candidiasis and carried out competitive infections with mixes of wild-type C. glabrata and different ypsΔ strains. Mice were infected via tail vein and killed after 7 days. Yeast CFUs were recovered from three target organs, kidney, liver and spleen. The ypsΔ strains all carry a hygromycin resistance (Hyg^R) cassette; therefore, colonies arising from the deletion strains can be distinguished from wild-type ones by their Hyg^R phenotype. As a control, we carried out competitive infections with a mix of the wild-type and a C. glabrata tnr1Δ strain (disrupting ORF CAGL0L13354g, which encodes a nicotinamide transporter). C. glabrata has two nearly identical TNR genes that are functionally redundant (B.M. and B.C. unpublished data), and we expected, therefore, that the tnr1Δ strain would have no virulence phenotype. As anticipated, infection with a mix of the wild-type and tnr1Δ strains, yielded a competitive index (CI) (see Materials and Methods) of ≈1 (Table 2). The ypsCΔ mutant also had a CI close to 1, as did the yps7Δ mutant; the yps1Δ mutant was modestly attenuated. By contrast, the yps1Δyps7Δ or yps(1-11)Δ strains were significantly attenuated, having aggregate CI of 0.07 and 0.04, respectively.

Table 2.

CI of C. glabrata ypsΔ strains in murine disseminated infection

Strains	CI in kidneys	CI in liver	CI in spleen
yps1Δ	0.37 ± 0.34	0.52 ± 0.07	0.41 ± 0.10
yps7Δ	1.44 ± 0.80	0.95 ± 0.24	0.99 ± 0.11
yps1Δ yps7Δ	0.02 ± 0.02	0.05 ± 0.02	0.13 ± 0.08
ypsCΔ	1.69 ± 1.43	0.80 ± 0.11	1.08 ± 0.12
yps1Δ ypsCΔ	0.09 ± 0.11	0.17 ± 0.07	0.28 ± 0.03
yps7Δ ypsCΔ	0.53 ± 0.35	0.48 ± 0.13	0.92 ± 0.19
yps(1–11)Δ	0.02 ± 0.03	0.03 ± 0.03	0.08 ± 0.04
tnr1Δ	1.31 ± 0.73	1.46 ± 0.23	1.33 ± 0.27

CI in each column is the ratio of mutant CFUs versus wild-type CFUs recovered from mouse organs 7 days after infection divided by the ratio of mutant CFUs versus wild-type CFUs in the strain mix for infection. Values shown are means ± SD from group of eight mice. A CI = 1 indicates the equal fitness, whereas a lower CI indicates reduced fitness of mutant versus wild-type strains.

To corroborate these findings, we carried out single infections with the yps(1-11)Δ and yps1Δyps7Δ strains. As shown in Fig. 3, the yps1Δyps7Δ strain is attenuated ≈2-logs in kidneys, and 1-log in liver and spleen, whereas the yps(1-11)Δ strain is attenuated ≈3-logs in kidney and liver, and 2-logs in spleen. These data implicate the YPS genes in C. glabrata survival in the host. Although ypsCΔ mutants have no virulence phenotype, the relative virulence of the yps(1-11)Δ and yps1Δyps7Δ strains indicates that the macrophage-induced YPS genes do play a role during infection.

Fig. 3.

C. glabrata ypsΔ mutants are compromised for virulence. Groups of 10 mice were infected with each C. glabrata strain via tail vein injection and killed 7 days after infection. Recovered CFUs from three target organs are indicated for individual mice as a diamond, and the geometric mean is shown as a bar.

Potential Substrates of C. glabrata YPS Proteases.

The Yps proteases are predicted to be GPI-anchored proteins. In yeast, GPI proteins are localized to the plasma membrane or, following a processing event, to the cell wall. We considered whether GPI-anchored cell wall proteins (GPI-CWPs) might be potential substrates of the Yps proteases. In C. glabrata, the GPI-CWPs include a family of cell wall-localized adhesins encoded by the EPA genes. To test whether Epa proteins might be substrates for the Yps proteases, we examined the stability of Epa1 protein (12) at the cell surface in wild-type and ypsΔ strains, using fluorescence activated cell sorting (FACS) and Western blot analysis. EPA1 transcription is normally induced specifically in lag phase after cells are diluted into fresh media, and then repressed to background levels within 2 h (S. Pan, A. de Las Peñas, and B.C. unpublished data). Thus, any Epa1 protein is derived primarily from transcript present during the first 2 h of growth, and protein stability can be monitored as cells continue to grow in log phase.

We used FACS to follow the fate of Epa1 at the cell surface (Fig. 4A and data not shown). In the wild-type and ypsCΔ strains, we found that maximal surface expression of Epa1, measured with a polyclonal antibody raised against the N-terminal domain of Epa1 (amino acid 30–336), occurs 2 h after dilution of stationary cells into fresh media. Levels of Epa1 remained constant at the cell surface for an additional 3 h before declining, reaching background levels after 10 h. By contrast, in the yps(1-11)Δ and yps1Δyps7Δ strains, which also exhibited maximal Epa1 surface expression by 2 h, Epa1 remained at maximal levels on the cell surface throughout the 10-h time course. In both the yps1Δ and yps7Δ strains, Epa1 levels fell by 2-fold over 10 h. We carried out Western blot analysis on isolated cell wall and culture media fractions from the wild-type and ypsΔ strains, using the antibody described above (Fig. 4B). For wild-type, Epa1 proteolytic products encompassing the N-terminal domain were detected in the culture media 4 h after dilution into fresh media and reaching maximal levels after 10 h. This was accompanied by a decrease in the amount of Epa1 protein in the cell wall fraction at hr 8 and 10. In the yps(1-11)Δ and yps1Δyps7Δ strains, we observed higher levels of Epa1 in the cell wall fraction compared with the wild-type strain, as well as much lower levels of the Epa1 proteolytic fragment released into the media. Even in the absence of all YPS genes, however, a small amount of Epa1 was cleaved and released into the media at 8 and 10 h (Fig. 4B). Not surprisingly, because Epa1 is a major adhesin in C. glabrata, the YPS deletion mutants are hyperadherent to epithelial cells (SI Table 5). These data implicate the YPS genes in proteolytic processing of Epa1 to remove it from the cell wall.

Fig. 4.

Epal is stabilized on the cell surface in ypsΔ strains. (A) FACS analysis of surface expressed Epa1 in C. glabrata wild-type and ypsΔ strains. Epa1 surface expression during a 10-h time course in YPD was assessed by an anti-Epa1 antibody and a FITC-conjugated secondary antibody. The geometric mean of fluorescence is indicated. (B) Western blot analysis of Epa1 protein in the cell wall and media fractions of the wild-type and ypsΔ strains grown in YPD. The samples were resolved on a 3–8% SDS/PAGE gel and labeled with anti-Epa1 antibody. The locations of molecular weight markers are indicated.

Discussion

C. glabrata is a successful pathogen of humans, suggesting that it has evolved mechanisms for colonization of and survival within the host. In this study, we have analyzed the interaction of C. glabrata with J774A.1 macrophage-like cells. Our data suggest that C. glabrata can replicate within the macrophage, a characteristic not shared by its closely related but nonpathogenic species S. cerevisiae. The overall transcriptional response of C. glabrata exposed to J774A.1 cells is highly similar to that described for C. albicans. This response is characterized primarily by a remodeling of carbon metabolism, which includes induction of genes encoding enzymes involved in β-oxidation, glyoxylate cycle, and gluconeogenesis. This is consistent with earlier proposals that phagocytosed C. albicans uses β-oxidation of lipids to generate acetyl CoA, which can be used for energy and, via the glyoxylate and gluconeogenesis pathway, to generate 5- and 6-carbon sugar metabolic building blocks (4, 5). In C. glabrata, we also found induction of the PDH1, CIT3, ICL2, and ACO1 genes. In S. cerevisiae, proteins encoded by the orthologues of these genes are proposed to function in the methylcitrate cycle, which converts propionate or propionyl-CoA to pyruvate and succinate (13). Induction of these genes in C. glabrata suggests that phagocytosed C. glabrata may face increased metabolic flux of propionyl-CoA, possibly from degradation of branched chain amino acids. Lastly, as was seen with C. albicans, there is a concerted down-regulation of the C. glabrata translation machinery following phagocytosis, which persists for at least 6 h. Interestingly, this down-regulation of translation, common to both pathogens, was not reported for phagocytosed S. cerevisiae (4).

We found that contact with macrophages up-regulates members of the C. glabrata YPS gene family encoding putative GPI-linked aspartyl proteases. Our genetic analysis, consistent with earlier work on the S. cerevisiae YPS genes (7), demonstrates a role for the C. glabrata YPS genes in cell wall metabolism. We show that the YPS genes, in particular YPS1 and YPS7, are required in vitro for C. glabrata survival during stationary phase or under conditions of cell wall stress. In terms of the cell wall integrity phenotypes, C. glabrata yps1Δ and yps7Δ strains are phenotypically similar to S. cerevisiae yps1 and yps7 strains (7), suggesting a broad conservation between C. glabrata and S. cerevisiae. Consistent with this, previous results show that CgYPS1 can functionally complement a S. cerevisiae yps1 mutant (7).

The YPS genes are also required for survival within macrophages and for virulence in a murine model of disseminated candidiasis. YPS1 has a primary role in both functions, with some role for YPS7. What is the role of the macrophage-induced YPS genes? Whereas the YPS cluster genes are apparently totally unimportant for in vitro growth (Fig. 2), they do function in the context of the mammalian host. In both the disseminated infection model and the macrophage assays, we observed no phenotype associated with deletion of the YPS cluster alone. However, in strains lacking YPS1, or lacking both YPS1 and YPS7, deletion of the YPS cluster had a strong additive phenotype (Fig. 3 and Tables 1 and 2). These data suggest that the YPS cluster functions primarily in the interactions with the host, where its function may overlap with other YPS genes, particularly YPS1. Consistent with this functional overlap, we found that during in vitro growth in YPD, transcript levels of YPS3, YPS10, and YPS11 increased 2- to 12-fold in the yps1Δ yps7Δ strain compared with the wild-type strain (data not shown).

The physiological substrates for yapsin-like proteases are not known in either C. albicans or S. cerevisiae. It has been proposed that substrates might include GPI-CWPs, a hypothesis consistent with the altered adherence phenotype of C. albicans sap9 and sap10 mutants (8). In this regard, our data show that C. glabrata ypsΔ mutants exhibit a profound defect in the processing of Epa1 from the cell surface. In the ypsΔ strains, Epa1 is stabilized at the cell surface, and the amount of proteolyzed fragment released into the culture media is greatly reduced relative to that seen in the wild-type strain. This strongly suggests that the substrates of Yps proteases might include GPI-CWPs, like Epa1. We favor a model in which Epa1 is proteolyzed directly by the Yps proteases, but we cannot exclude a more indirect role. The Yps proteases might, for example, serve to activate a different protease, which itself processes Epa1. Indeed, in the yps(1-11)Δ strains, there is still residual processing of Epa1. Potential proteases responsible for this residual activity include orthologues of S. cerevisiae Bar1 (which functions extracellularly) or perhaps Kex2 (although Kex2 is thought to function primarily in the Golgi). We have found that the C. glabrata bar1Δ strains have neither cell wall phenotypes, nor defects in removal of Epa1 from the cell surface (data not shown). We have not constructed a yps(1-11)Δ bar1Δ mutant.

More generally, our data implies an important role, either direct or indirect, for the Yps proteases in remodeling the yeast cell wall by removal of GPI-CWPs. This remodeling is apparently essential for virulence. We propose that one important role for the Yps proteases is to remodel the cell surface by removal of certain GPI-CWPs in response to different host environments. This might in principle be necessary for subsequent incorporation of other cell wall proteins more suited to a given environment. Alternatively, the Yps proteases might protect Candida from immune recognition by acting to remove GPI-CWP targets of the innate or adaptive immune responses. In this regard, it is interesting to note that infecting macrophages with wild-type S. cerevisiae activates them to produce increased levels of NO, but infecting with wild-type C. glabrata does not. Notably, the C. glabrata ypsΔ mutants strongly stimulate macrophage production of NO, suggesting that YPS-mediated cell wall remodeling may play a role in altering or suppressing macrophage activation.

Materials and Methods

Strains and Growth Conditions.

C. glabrata mutant strains are derivatives of a clinical isolate BG2 (16); additional clinical isolates (strains 4405, 4452, 4566, 4787) were a gift of M. Pfaller (Department of Pathology, University of Iowa School of Medicine, Iowa City, IA). All mutant C. glabrata strains (SI Table 4) are derived from BG14 (BG2 ura3Δ) (14). S. cerevisiae strain BY4742 MATa his3Δ leu2Δ lys2Δ ura3Δ and four clinical S. cerevisiae strains (YJM128, YJM264, YJM309, YJM336, YJM436) were used (15). Strains were cultured in YPD at 30°C. Yeast transformations were carried out as described in refs. 16 and 17.

Plasmid and Strain Construction.

Plasmids used to make targeted disruptions are detailed in SI Table 6. Gene fragments for targeting deletion constructs were generated by PCR (SI Table 7) and verified by sequencing. Yeast genes were disrupted by two-step (18) or one-step methods, using the hph gene (conferring Hyg^R) as a selectable marker (19). Disruptants were confirmed by PCR. For each target gene, two independent transformants were generated and tested with essentially identical results. Whereas it was not practicable to restore each YPS in each deletion strain, we verified that the restoration of YPS1 in the yps(1-11)Δ strain reverted the hypersensitivity to NaCl and caffeine, as well as the Epa1 processing defect (data not shown).

Cell Wall Assays.

Sensitivity of C. glabrata ypsΔ mutants to cell wall stress was tested by growth on YPD containing calcofluor white (250 μg/ml), congo red (1 mg/ml), caffeine (7.5 mM), or NaCl (0.5 M) (Sigma, St. Louis, MO). For the Zymolyase sensitivity assay, log phase cells were resuspended in 10 mM Tris·HCl (pH 7.5) containing 50 μg/ml Zymolyase (ICN Biomedicals, Costa Mesa, CA) and the OD₆₀₀ was measured as a function of time. To calculate survival in stationary phase, the total number of cells was assessed by hemocytometer, and viable CFUs were assessed by plating on YPD.

Cell Culture and Macrophage Survival Assay.

For the macrophage infection assays, cells of the murine macrophage-like cell line J774A.1 (ATCC) were seeded in 24-well plates. After incubation at 37°C for 16–18 h, log phase yeast cells were added and the plates were centrifuged at 200 × g for 1 min. After 1 h incubation at 37°C, the non-cell-associated yeast were removed by washing with DMEM. To measure yeast survival/replication in macrophages, lysates of infected macrophages at various time points were plated on YPD plates to determine CFUs.

Transcriptional Profiling by Microarray and RT-PCR Confirmation.

The printing conditions for C. glabrata microarray can be found in the Gene Expression Omnibus (GEO) database (accession no. GPL3922). To isolate RNA from macrophage-ingested C. glabrata, the infected macrophages were lysed in ice-cold H₂O containing RNase (Ambion, Austin, TX) to digest mammalian RNA. The yeast cells were washed in H₂O with protectRNA RNase inhibitor (Sigma), frozen on dry ice, and disrupted with glass beads in guanidium-isothiocyanate. Yeast RNA was isolated by acid phenol extraction. The protocols for synthesis of Cyanine 5- or Cyanine 3-labeled cDNA probes and microarray hybridization are in the Gene Expression Omnibus (GEO) database (accession no. GSE6058). Dye-swap controlled experiments were performed from three biological replicates for each time point. For each feature on microarray, the average Log₂(treated/control) value from the dye-swap experiment was calculated. Six such values (from three biological repeats with duplicate features on the microarray) were imported into SAM software (http://www-stat.stanford.edu/∼tibs/SAM/) for statistical analysis. In identifying significantly induced or repressed genes, the median false discovery rate equals 0, whereas the 90% false discovery rate was <0.001. For quantitative real time PCR, 10 μg of total RNA was used to synthesize first strand cDNA with Oligo(dT₂₀) in a final volume of 40 μl. 1 μl of cDNA was used as the template in individual PCR with primer pairs specific for each YPS gene or for ACT1 (SI Table 7), using a QuantiTect SYBR Green PCR kit (Qiagen, Valencia, CA). Real-time PCR was performed on an Applied Biosystems (Foster City, CA) 7500 Real-Time PCR system in a 96-well plate format. Dilutions of C. glabrata genomic DNA was used to generate a standard curve for each gene. mRNA abundance was normalized to ACT1 transcript levels. The quantitative real time PCR was done in triplicate from each of two independent biological samples.

Nitrite Determination.

J777A.1 cells were seeded in 24-well plates of 2 × 10⁵ cells per well. After 16–18 h, 10 ng/ml IFN-γ was added, and cells were incubated for another 5 h. Yeast were added at MOI 1:1, and after 1 h incubation, non-cell-associated yeast were washed away with DMEM. Subsequently, fresh DMEM containing 10 ng/ml IFN-γ (Sigma) was added. After 24 h incubation, the culture medium was collected and nitrite was measured by the Griess reaction with NaNO₃ as a standard as described in ref. 20.

FACS Assay and Western Analysis.

Surface Epa1 was detected by FACS as described in ref. 13, using a rabbit polyclonal anti-Epa1 antibody raised against recombinant Epa1 fragment (amino acid 30–336). For Western blot analysis, the cell wall and membrane fractions were prepared as described in ref. 21. Proteins in the medium were precipitated at −20°C after addition of three volumes of acetone. Protein was detected with the antibody described above and visualized by using an ECL-Plus kit (Amersham Pharmacia, Piscataway, NJ).

Adherence Assay.

Adherence assays were carried out as described in ref. 9.

Animal Studies.

Yeast was grown for 16 h in YPD at 30°C. Cells were collected and resuspended in PBS to 4 × 10⁸ cells/ml. Groups of 8–10 Balb/C mice (6- to 8-wk-old, Taconic, Rockville, MD) were infected with 100 μl of cell suspension by tail vein injection. Mice were killed after 7 days, and organs were harvested. Appropriate dilutions of homogenates were plated on YPD to assess CFUs. For competitive infections, a mix of wild-type and mutant cells (at a ratio of ≈1:1) were used. Because the mutant strain is Hyg^R, the mutant CFUs could be differentiated from wild-type ones by growth on YPD supplemented with 500 μg/ml Hygromycin (Calbiochem, San Diego, CA). The CI is the ratio of mutant CFUs versus wild-type CFUs recovered from organs divided by the ratio of mutant CFUs versus wild-type CFUs in the strain mix for infection (22).

Abbreviations

CI

competitive index

CWP

cell wall proteins

GPI

glycosylphosphatidylinositol

YPD

yeast extract/peptone/dextrose.