Previous studies have demonstrated an important role for the vacuole in the virulence of the fungus Cryptococcus and studies in yeast have implicated the vacuolar protein Vps41 in copper loading of proteins such as iron transporters. However, our studies found that a cryptococcal vps41Δ strain displayed wild-type growth on media containing iron and copper chelators and normal activity of the copper-containing virulence factor laccase as well as almost normal growth at 37°C and wild-type production of the virulence factor capsule. Despite these attributes, the vps41Δ mutant strain showed a dramatic attenuation of virulence in mice and co-incubation of mutant cells with the macrophage cell line, J774.16, resulted in a dramatic loss in viability of the vps41Δ mutant strain at 10 h compared with wild-type and complemented strains. Closer examination revealed that the vps41Δ mutant displayed a dramatic loss in viability after nutrient starvation which was traced to a failure to undergo G2 arrest, but there was no defect in the formation of autophagic or proteolytic vesicles. Our results indicate that VPS41 plays a key role in regulating starvation response in this pathogenic organism and that defects in cell cycle arrest are associated with attenuated pathogenic fitness in mammalian hosts.
Cryptococcus neoformans is a major fungal pathogen primarily infecting immunocompromised hosts, especially patients with impaired cellular immunity related to HIV infection (Perfect et al., 1998). It is generally believed that inhalation of desiccated yeasts or basidiospores from the environment is the route for pulmonary C. neoformans infection. The alveolar macrophages (AMs) therefore serve as one of the first lines of anticryptococcal defence through phagocytosis and subsequent killing of the ingested yeast cells. An effective immune response of the immunocompetent host normally results in clearance of intracellular yeast cells or containment of the fungus in a latent state (Feldmesser et al., 2001a). However, cryptococcal infection of hosts with impaired cellular immunity such as AIDS patients results in intracellular growth of the fungus following phagocytosis by AMs and eventually leads to the lysis of macrophages and release of replicated infectious yeast cells (Feldmesser et al., 2001b). Dissemination of cryptococcal cells to the brain causes life-threatening cryptococcal meningitis, the predominant manifestations in the clinical setting associated with cryptococcal infection of AIDS patients (Casadevall and Perfect, 1998).
The ability of C. neoformans to elaborate a variety of virulence factors forms the foundation of a rich repertoire of pathogenic strategies as an opportunistic pathogen and enables it to survive in both extracellular and intracellular environments. For example, production of laccase, a key enzyme involved in lignin degradation and in vitro pigment formation, was shown in vitro to interfere with hydroxyl radical production by host macrophages and thus likely contributes to cryptococcal survival inside macrophage (Liu et al., 1999a). Moreover, melanized cryptococcal cells exhibit more resistance to antibody-mediated phagocytosis and the killing effects of murine macrophages than non-melanized cells (Wang et al., 1995), although the formation of significant amounts of melanin during in vivo pathogenesis has been called into question (Liu et al., 1999b). A second virulence factor of C. neoformans is the production of an extensive polysaccharide capsule which inhibits phagocytosis and promotes survival after phagocytosis (Bulmer and Sans, 1967; Fromtling and Shadomy, 1982). Accumulation of polysaccharide-containing vesicles in the cytoplasm of infected macrophages has been hypothesized to augment intracellular survival by the encapsulated cryptococcal cells (Tucker and Casadevall, 2002). Furthermore, the ability to grow at host temperatures allows C. neoformans either to proliferate or to establish latent infection in mammalian hosts, serving as afundamental component of the pathogenic pathway of this organism (Odom et al., 1997). Additional genes that have been shown to contribute to the intracellular and extracellular survival of cryptoccal cells, by inhibiting phagocytosis or by counteracting the oxidative and nitrosative stresses of macrophages, include AOX1 (Akhter et al., 2003), URE1, PLB1, SOD1 (Cox et al., 2000; 2001; 2003), APP1 (Luberto et al., 2001), FHB1 (de Jesus-Berrios et al., 2003), TSA1 and TRX1 (Missall et al., 2004; Missall and Lodge, 2005).
Previously, a vph1Δ mutant was isolated from a screening for laccase-deficient mutations using an insertional mutagenesis approach (Erickson et al., 2001). The VPH1 gene encodes a component of the vesicular proton pump required for post-translational modification of proteins. Importantly, loss of VPH1 function resulted in defects in expression of four virulence traits: laccase, capsule, urease and growth at 37°C, suggesting an important role for normal vesicular/vacuolar functions in the expression of virulence factors in C. neoformans. In addition, studies of vacuolar mutants of C. neoformans has shed light on the role of these proteins in the biosynthesis of metalo-enzymes. For example, defective vesicular acidification resulted in defects in metalation of the copper-containing virulence factor, laccase, which could be reconstituted with exogenous copper (Zhu et al., 2002). Similarly, the vesicular chloride channel, CLC1, was found to be important in laccase activity and has also been implicated in efficient metalation of copper-containing factors, such as Fet3 in Saccharomyces cerevisiae (Zhu and Williamson, 2003). To further test the potential role of cryptococcal vesicular function in virulence expression, we deleted a cryptococcal VPS41 homologue, which functions in both the alkaline phosphate (ALP) and carboxypeptidase Y (CPY) sorting pathways and has been implicated in copper loading of the iron transporter, Fet3 in yeast (Radisky et al., 1997).
In this study, we report that although VPS41 is dispensable for normal growth of C. neoformans even under conditions of low iron availability and plays no role in expression of several important metal-dependent virulence factors such as laccase, capsule and urease, it is essential for virulence and survival in macrophages. Furthermore, the VPS41 mutation was associated with a defect in tolerance to nutrient starvation, related to an inability to arrest in G2. These studies thus provide genetic support for a role for adaptation to nutrient starvation as a component of the virulence composite of pathogenic fungi.
Isolation of the VPS41 gene and creation of a vps41 deletion mutant
A blast search of the Duke serotype A genome database with the protein sequence of the S. cerevisiae VPS41 gene identified a single cryptococcal VPS41 homologue (E = 3.9 e-56; http://cneo.genetics.duke.edu/) and a ∼6 kb genomic fragment encompassing the whole VPS41 gene was PCR-amplified, cloned and sequenced (as detailed in Experimental procedures). To accurately predict the protein sequence of VPS41, a 3245 bp cDNA fragment containing the full open reading frame (ORF) of VPS41 was isolated from total cryptococcal RNA by reverse transcription polymerase chain reaction (RT-PCR) amplification. Sequence comparison of the genomic DNA of VPS41 and its cDNA revealed that the full length of the cryptococcal VPS41 gene contains 13 introns. Sequence analysis indicated that the full ORF of VPS41 is predicted to encode a hydrophilic protein of 1038 amino acids. Alignment with the Saccharomyces Vps41 protein showed that the two share distinct conserved domains, namely, an N-terminal Domain I, a domain required for binding of S. cerevisiae Vps41p to the C-terminal coiled coil containing domain of Apl5p (Darsow et al., 2001), and two clathrin heavy-chain repeat (CHCR) domains involving in homo-oligomerization and a RING-H2 motif present in higher eukaryotes but not Vps41 from S. cerevisiae (Ward et al., 2001). However, the overall shared identity of the two proteins is moderate [27.58% (246/892)]. Southern blot analysis indicated that, as in S. cerevisiae, the C. neoformans genome contains only one copy of the VPS41 gene (data not shown).
To determine the potential role of the VPS41 gene in the virulence of C. neoformans, we disrupted the VPS41 locus using a 1.6 kb URA5 fragment to replace the 447 bp EcoRV/BglII fragment of the wild-type locus via a biolistic transformation approach. Of 125 transformants obtained, three vps41Δ deletion mutants were detected based on initial PCR screening. Southern blot analysis showed that all three mutants contained a specific integration of the URA5 marker into the targeted genomic region of the VPS41 gene. One of the deletion mutants, designed as vps41Δ, was used for experiments in this study to evaluate the functional roles of the VPS41 gene in C. neoformans.
To complement the vps41Δ mutant, a cloned ∼6 kb genomic fragment containing the full ORF of VPS41 plus ∼800 bp of the 5′ promoter region was transformed into the vps41Δ mutant cells by electroporation using the 2 kb HgR gene, which confers hygromycin-B resistance, as a selection marker and genomic insertion verified by Southern blot hybridization using uncut genomic DNA.
VPS41-related virulence phenotypes in C. neoformans
As VPS41 has been implicated in intact function of the copper-dependent iron transporter, Fet3 in Saccharomyces, we investigated whether VPS41 deletion resulted in a reduction in activity of the virulence factor, laccase, which is also a multicopper oxidase of the same family as Fet3 (Liu et al., 1999a). Surprisingly, the vps41 mutation showed no effect on laccase expression, as well as other metal-dependent virulence factors of C. neoformans such as capsule production or urease expression, suggesting that metal sensing and trafficking processes are not dependent on this vesicular factor in this fungus (Fig. 1B). Growth rate of the vps41Δ mutant was identical to wild type in YPD broth at 30°C (doubling time, H99 wild type: 2.8 ± 0.2 h; vps41Δ: 2.9 ± 0.2 h; vps41Δ+VPS41: 2.7 ± 0.2 h) and nearly so at 37°C (doubling time, H99 wild type: 2.9 ± 0.2 h; vps41Δ: 3.4 ± 0.2 h; vps41Δ+VPS41: 2.8 ± 0.2 h).
To evaluate the potential role of VPS41 in the pathogenesis of C. neoformans in vivo, virulence of the yeast cells of H99 wild-type, vps41Δ mutant and vps41Δ+VPS41-reconstituted strain was tested in a mouse tail vein model. After injection of 1 × 106 yeast cells, the median survival times of mice infected independently with H99 wild type, and the vps41Δ+VPS41-reconstituted strain were 7 and 9 days, respectively (Fig. 1C), whereas all mice infected with the vps41Δ mutant survived at least 1 month (P < 0.001) and cultures of brain at 1 month failed to yield live vps41Δ mutant cells. The complete loss of virulence of the vps41 mutant in a mouse model despite a high inoculum was surprising, given the lack of large defects in known virulence attributes and suggested further studies to determine the predominant defect leading to such a dramatic reduction in virulence in the mutant strain.
Effects of VPS41 disruption on macrophage killing and stress responses of C. neoformans
As C. neoformans is a facultative intracellular pathogen which can reside and replicate within macrophages (Feldmesser et al., 2001a), we assessed for survival of the vps41Δ mutant within the J774.16 macrophage-like cell line. Normally, incubation of C. neoformans cells within this cell line results in long-term survival and even replication of yeast cells within the phagolysosome (Feldmesser et al., 2001a). However, incubation and phagocytosis of the vps41Δ mutant resulted in rapid reduction in fungal viability as measured by colony-forming units (cfu), compared with the viability of the wild-type or the VPS41-reconstituted strain (Fig. 2A). In contrast, incubation of the vps41Δ mutant strain in tissue culture media containing either 10% serum or anti-capsular antibody resulted in recovery of greater than 95% of cells during the same time period (data not shown). In separate experiments, phagocytosis of mutants of capsule biosynthesis (cap64) or laccase (lac1Δ) with J774.16 cells under the same conditions resulted in over 90% viability of these strains in the same time period. These experiments demonstrate a requirement for intact VPS41-dependent processes for intracellular survival. As an additional control, the phagocytotic index using J774.16 cells was determined and was found to be similar between the three strains (wild type: 18 ± 3%; vps41Δ: 16 ± 3%; vps41Δ+VPS41: 18 ± 3%).
In the yeast, S. cerevisiae, vacuolar processes are essential for a wide range of physiological processes, including pH homeostasis and osmoregulation, protein degradation, and storage of amino acids, ions and polyphosphates (Bryant, 1998). To examine the effects of VPS41 disruption on cryptococcal growth under stress conditions that might be encountered in macrophages, we tested the ability of the vps41Δ mutant to grow on media containing hydrogen peroxide or the redox cycling agent menadione, the latter of which tests for the sensititivy of a cell to superoxide radicals (Narasipura et al., 2003). Compared with H99 wild type, and the reconstituted vps41Δ+VPS41 strain, no observable differences in growth were observed for the vps41Δ mutant (Fig. 2C), suggesting that the function of VPS41 is dispensable for survival and growth in the presence of oxidative stresses. The vps41Δ mutant strain was also found to be insensitive to osomotic and ionic stress indicated by wild-type growth on media containing 1.2 M NaCl or 1.8 M sorbitol (data not shown). In addition, as VPS41 deletion in S. cerevisiae produces growth defects on low-iron medium (Radisky et al., 1997), conditions that are believed to exist within the intracellular phagolysosome (Fang et al., 2000; Chung et al., 2004), we next investigated whether VSP41 is required for normal cryptococcal growth in iron- or copper-deficient medium at 30°C and 37°C. H99 wild type, vps41Δ mutant and the reconstituted vps41Δ+VPS41 strain were allowed to grow on YPD agar with addition of either the iron chelator 3-(2-Pyridyl)-5,6-diphenyl-1,2,4-triazine-p,p′-disulfonic acid, monosodium (500 μM) or the copper chelator Bathocuproinedisulphonic acid, disodium (1 mM). Again, no defects in growth were detected for the vps41Δ mutant and its growth on both conditions was comparable to that of wild-type and reconstituted vps41Δ+VPS41 strains (Fig. 2C). Thus, unlike that of its counterpart in S. cerevisiae, VPS41 function is not required for C. neoformans to grow on iron-deficient medium, suggesting that cryptococal cells may use a different iron transport pathway or that there are other gene(s) possessing overlapping function with VPS41 in this pathogenic fungus. Finally, we assessed the functional role of VPS41 in C. neoformans in response to nutrient starvation conditions as previous studies showing a role for the glyoxylate pathway and gluconeogenesis in fungal pathogenesis suggest an importance for nutrient deprivation in protection from fungal pathogens (Lorenz and Fink, 2001; Panepinto et al., 2005). The wild-type, vps41Δ mutant and reconstituted vps41Δ+VPS41 strains were cultured at 37°C with shaking (250 r.p.m. min−1) in sterile water supplemented with YNB without amino acids and ammonium sulphate for various time periods. As shown in Fig. 2B, the vps41Δ mutant strain exhibited severe defects in response to nutrient starvation characterized by an 80% decline in viability at the 10 h point, compared with a viability of 84.5% and 97.4% for the wild-type and reconstituted vps41Δ+VPS41 strains, respectively, over the same time period (P < 0.001). Interestingly, the cryptococcal vps41Δ mutant survival defect could not be complemented by expression of a VPS41 open reading frame from S. cerevisiae under the C. neoformans promoter, suggesting a requirement for unique cryptococcal sequences such as the C-terminus RING-H2 motif described above. In summary, the above results indicate that the function of VPS41 is essential for cryptococcal survival under starvation conditions and represents a new phenotypic association for this gene in eukaryotes.
VPS41-associated cellular function under nutrient starvation conditions
Previous studies in Saccharomyces have implicated VPS41 in intact vacuolar morphology (Radisky et al., 1997), and recently the Vps41 protein has been localized to the yeast vacuole (LaGrassa TJ, 2005), a phenotype shared by the cryptococcal homologue as demonstrated by colocalization of a Vps41-green fluorescent protein (GFP) construct with the vacuolar marker, FM4-64 (Fig. 3A). In addition, smaller vesicular structures have been observed in yeast vps41Δ mutants, but the role of nutrient starvation in this phenomenon has not been studied. Thus, to investigate the role of VPS41 during nutrient-replete as well as under starvation conditions, and to assess this role in C. neoformans, we used a coumarin dye, 7-amino-4-chloromethyl-coumarin (CMAC), which localizes to acidic vesicular compartments (Roberts et al., 1991; Shoji et al., 2006). As shown in Fig. 3B, during log phase, nutrient-replete conditions, both wild-type and vps41Δ mutant C. neoformans cells were observed to form numerous small acidic vesicles observed by both CMAC and DIC imaging. However, under nutrient starvation conditions, wild-type cells formed larger, more typical vacuolar structures, whereas the vps41Δ mutant cells retained the small vesicular structures characteristic of log phase. Reconstitution with wild-type VPS41 or a Vps41-GFP construct restored the ability to form large acidic vacuoles during nutrient starvation (Fig. 3A and B). This failure to form appropriately sized vacuole compartments in the mutant strain could thus be due to a primary defect in vacuole formation or a failure to signal events leading to starvation-appropriate phenotypes.
During formation of large vacuolar structures during nutrient starvation, studies in yeast have shown that cells form smaller cytosolic autophagic vesicules which engulf cellular elements, transporting them into the vacuole for programmatic digestion, thus allowing survival under these conditions (Klionsky, 2004). In contrast, cells defective in autophagic processes show defects in tolerance to nutrient starvation (Klionsky, 2004) and genes involved in autophagy have recently been shown to be upregulated by phagocytosed C. neoformans (Fan et al., 2005). Thus, we investigated autophagic properties of the vps41Δ mutant to identify processes that may be involved in cryptococcal macrophage survival. As shown in Fig. 4A, panel a, wild-type cryptococcal cells starved by incubation in YNB without amino acids and ammonium sulphate for 3 h at 37°C formed typical small granular bodies that can be labelled with the specific autophagic marker, Atg8. Atg8 is highly conserved among eukaryotes (Lang et al., 1998; Wu et al., 2006) and includes a single cryptococcal homologue (http://www.tigr.org) that is 93% identical to Atg8 from S. cerevisiae. Similarly, during starvation conditions, the vps41Δ mutant cells also formed Atg8-labelled autophagic digestive vesicles (Fig. 4A, panel d). In contrast, log-phase, nutrient-replete growth conditions resulted in repressed production of Atg8-labelled autophagic vesicles in wild-type C. neoformans (data not shown) as described in yeast (Lang et al., 1998). To further confirm that vesicles observed during starvation showed intact proteolytic function required for autophagy, the fluorescent dye markers, 7-amino-4-chloromethylcoumarin, l-arginine amide (CMAC-Arg) and 7-amino-4-chloromethylcoumarin, l-alanyl-l-proline amide (CMAC-Ala-Pro), were used that require proteolytic cleavage by amino-peptidase or dipeptidyl-peptidase enzymes, respectively, to activate fluorescence (Suarez-Rendueles et al., 1981; Achstetter et al., 1983; Roberts et al., 1991). As shown in Fig. 4B, in log-phase cells and under starvation conditions, both wild-type and vps41Δ mutant cells form small proteolytic vesicles which colocalize to the vesicular membrane marker, MDY-64. Similar results were observed using the dipeptidyl peptidase substrate, corroborating the presence of intact proteolytic vesicles in the vps41Δ mutant strains (data not shown).
Deletion of VPS41 results in failure of G2 arrest during nutrient starvation
Cell cycle arrest is a prominent response to nutrient deprivation of higher eukaryotes and has been observed by laser scanning cytometry in C. neoformans cells arrested in stationary-phase and oxygen deprivation (Ohkusu et al., 2004). Initial studies with the vps41Δ mutant showed that upon incubation in nutrient-deprived conditions in YNB without amino acids and ammonium sulphate for 24 h at 37°C, the budding index of the mutant was elevated (31.6%) compared with that of wild-type cells (15.8%) incubated under the same conditions (P < 0.0001). This suggested that VPS41 may have a role in cell cycle control under starvation conditions, as abnormalities in the budding index have been shown to correlate with abnormalities in cell cycle (Zettel et al., 2003). In contrast, the budding indexes of the two strains in log phase were found to be similar (H99 wild type: 69.7% versus vps41Δ: 74.4; P > 0.10). As cell cycle arrest is an important requirement for survival under nutrient-deprived conditions (Werner-Washburne et al., 1996), we used laser scanning cytometry to assess for successful arrest in vps41Δ cells. As shown in Fig. 5A, wild-type C. neoformans cells in log phase are distributed almost equally between G1 and G2, indicated by a similar proportion of cells with 1C DNA and 2C DNA. Cell size of G2 cells were also greater than that of G1 cells as indicated by a ratio of forward scatter of G2/G1 of approximately 1.4 for wild-type cells. (Peaks greater than 2C are most likely due to adherent, budding cells.) However, in response to nutrient starvation, wild-type cells arrested in G2, resulting in almost all cells found in this phase of the cell cycle as described previously for C. neoformans (Ohkusu et al., 2004; Fig. 5A). As cells with a 2C DNA content can also be due to cells arrested in mitosis, we visualized nuclei by 4′,6-diaminidino-2-phenylindole dilactate (DAPI) staining to distinguish the stages. As shown in Fig. 5B, wild-type cells were found to arrest with a well-formed, large nuclei typical for G2 arrest. In contrast, we failed to identify any mitotic cells in over 200 cells counted. Similar to the wild-type cells, the vps41Δ mutant cells were also found to be distributed between G1 and G2 during log phase. (Ratio of forward scatter G2/G1 was again approximately 1.4. More tailing of the G2 peak is most likely due to slower release of buds in the vps41Δ cells.) However, after 24 h of nutrient starvation, the vps41Δ cells continued to be distributed between G1 and G2 in amounts similar to that of log phase (Fig. 5A). DAPI staining also showed two populations of cells, one containing larger nuclei typical of G2 arrest, and a population with typical early postmitotic nuclei, which are small and condensed (Straube et al., 2003). However, mutant cells were still metabolically active after incubation in starvation conditions, as shown by metabolic processing of FUN 1 dye to a fluorescent vesicular product as previously described (Martinez and Casadevall, 2005). Flow cytometry-derived DNA distributions of starved vps41Δ mutant were not altered by addition of 1 μg ml−1 rapamycin, an inhibitor of TOR in C. neoformans (Cruz et al., 2001), suggesting that the VPS41 gene acts either downstream of or is independent of TOR under these starvation conditions. Overall, these data suggest that VPS41 may play a role in effective cell cycle control during nutrient deprivation and provide an explanation for the mutant's inability to tolerate nutrient starvation conditions encountered in the intracellular environment and the mammalian host.
Targeted deletion of VPS41 was initiated as an effort to further evaluate the potential role of vacuolar functions in laccase expression and virulence of C. neoformans. Interestingly, many of the functions of proteins localized to vacuoles also play roles in biosynthetic machinery within secretory vesicles. For example, in both Saccharomyces and Cryptococcus, deletion of genes encoding components of the vacuolar chloride channel and proton pump have impacts on the bio-incorporation of copper cofactors into the iron transporter Fet3 and the virulence factor laccase respectively (Davis-Kaplan et al., 1998; Zhu and Williamson, 2003; Zhu et al., 2003). As Vps41 has also been implicated to play a role in the metalation of Fet3 in Saccharomyces with growth defects on iron or copper-limiting media (Radisky et al., 1997), we investigated its role in the expression of laccase. Structurally, the cryptococcal Vps41 protein is highly similar to the Saccharomyces homologue, but the cryptococcal homologue also contains a vesicle-binding RING-H2 motif more typical of higher eukaryotes such as mammalian cells (Ward et al., 2001), suggesting that there could be divergent functions of VPS41 between ascomycete and basidiomycete fungi. Such a structural divergence in VPS41 may explain the lack of a role in copper homeostasis in C. neoformans; alternatively, there may be redundant pathways for copper insertion in basidiomycete fungus.
Despite the normal expression of virulence factors in the vps41Δ mutant and a minimal effect on growth rate at 37°C, a dramatic attenuation in virulence was observed using a mouse model with survival of all inoculated mice and clearance of organisms despite using a high inoculum (106 cfu) of fungi. To investigate this in more detail, fungal cells were incubated with a J774.16 macrophage cell line to study phagocytosis and killing. While phagocytosis was similar to wild type for the mutant strain, a dramatic rate of killing was observed within hours of phagocyosis that was more pronounced than that observed for mutants of two principal virulence factors of C. neoformans, capsule and laccase. This dramatic rate of killing is unusual for a viable cryptococcal mutant and could not be accounted for by the slight reduction in growth at 37°C or the slower bud release observed during flow cytometry. Indeed, previous studies of wild-type C. neoformans have shown successful replication within J774.16 cells and even acapsular strains show significant survival within a 24 or 48 h period (Mukherjee et al., 1996). Thus, rapid killing of the vps41Δ cells over a period of a few hours suggests that VPS41 is involved in expression of an additional phenotype essential for intracellular survival, a property that has recently been implicated in both acute and chronic cryptococcal infections (Feldmesser et al., 2001a). Expression of such critical phenotypes appears to be key to the ability of a pathogen to survive in the host, a property recently described as pathogenic fitness (Panepinto and Williamson, 2006).
Previous studies have suggested that macrophage cell lines derived from BALB/c mice such as J774 contain defects in the membrane transporter NRamp, which is required for wild-type oxidative burst (Barton et al., 1995) and is involved in anticryptococcal killing (Blasi et al., 2001). Thus, it may be less likely for an oxidative killing mechanism to be primarily responsible for the rapid demise of the vps41Δ strain in this cell line, and, indeed, the vps41Δ mutant in vitro did not appear to show altered survival in response to oxidative stresses such as hydrogen peroxide or menadione, previously shown to be fungicidal for C. neoformans (Narasipura et al., 2003; de Jesus-Berrios et al., 2003). Instead, the cryptococcal vps41Δ mutant cells displayed a defect in tolerance to nutrient starvation with a rapid rate of demise similar to that observed after phagocytosis by the macrophage cell line. This represents a new role for Vps41 proteins, and may explain previous data showing that class C complex vacuolar genes such as VPS41 and VPS39 are hypersensitive to the TOR inhibitor rapamycin whose pathway is involved in tolerance to nutrient deprivation by activation of cell cycle arrest and autophagy (Xie et al., 2005). Indeed, these data may shine light on a recent finding that VPS41 was identified in a haplo-insufficient screen for mutants defective in the filamentation, suggesting the relationship between nutrient-induced cell cycle arrest and filamentation (Uhl et al., 2003).
These data led to further investigation of starvation-related phenotypes in the cryptococcal vps41Δ mutant which suggested several defects in phenotypes important to survival under nutrient-deprived conditions. For example, in wild-type C. neoformans, log-phase cells demonstrated numerous small staining vesicles which accumulated the acidic CMAC fluorescent dye, and, under starvation conditions, developed larger, more typical vacuoles similar to other yeasts. However, the vps41Δ mutant failed to demonstrate enlarged vacuoles upon nutrient starvation, typical for the aberrant morphology described in class C vacuolar mutants of S. cerevisiae (Radisky et al., 1997). This defect could be due to an inherent defect in enlargement of vacuoles during the transition to nutrient starvation conditions, or it could be due to a defect in either nutrient sensing or execution of processes associated with cell cycle arrest. A second starvation-associated pathway, the formation of Atg8-labelled autophagic vesicles (Noda and Ohsumi, 1998; Kamada et al., 2004), appeared to be intact during nutrient starvation, and formation of proteolytic competent vesicles required for autophagic digestion also appeared intact. However, we were not able to study proteolytic maturation of specific VPS41-associated proteases such as CPY due to a lack of appropriate homologues in C. neoformans.
Surprisingly, investigation of cell cycle in the VPS41 mutants demonstrated that the VPS41 gene may have a role in cell cycle control mechanisms during starvation that may help to explain the poor survival of the mutant strain both in vitro and after macrophage phagocytosis. G2 arrest upon nutrient deprivation has been previously described under stress conditions and stationary phase (Ohkusu et al., 2001), and thus may play an important role in nutrient conservation in the intracellular habitat. Failure of successful cell cycle arrest results in rapid depletion of cellular nutrient stores and has been associated with poor survival under nutrient-depleted conditions in yeast (Iida and Yahara, 1984). An alternative explanation may be that the mutant cells simply shut down and die during nutrient starvation, but the retention of starvation-specific metabolic events such as expression of enzymatically active laccase, production of autophagic vesicles, as well as metabolic conversion of the FUN 1 fluoresecent dye, make this explanation less likely. However, a specific role of VPS41 in cell cycle control will require the identification of the specific pathways and target genes involved and is the subject of active investigation. Interestingly, addition of rapamycin to mutants during starvation failed to restore wild-type cell cycle arrest as reported previously in yeast (Zaragoza et al., 1998), suggesting that VPS41 in C. neoformans may act downstream of or act independently of TOR under these conditions.
These data thus suggest a role of VPS41 in tolerance to nutrient stress expected to occur within the macrophage intracellular environment. Despite the need for host cells to obtain basic nutrients in its own growth and survival, pathogens appear to be denied these basic tools for life during intracellular phagocytosis. Thus, in addition to a need for the pathogen to obtain nutrients under the adverse host environment, it appears that an essential tool for pathogen survival is also to prevent growth by mechanisms such as cell cycle arrest under conditions that may temporarily deprive the fungal pathogen of necessary substrates required for growth.
Fungal strains, plasmids and media
Cryptococcus neoformans ATCC 208821 (H99) was a generous gift of J. Perfect. Strain H99FOA19, a uracil auxotrophic mutant derived from H99 (Erickson et al., 2001), was employed as a recipient strain for creating targeted gene deletion. The strains were grown in YPD media (2% glucose, 1% yeast extract, 2% Bacto-peptone) or YPD agar medium (adding 2% agar) or incubated in yeast nitrogen base (1.7 g l−1 YNB, without amino acids and ammonium sulphate, Becton, Dickinson, MD) media for starvation experiments. Asparagine minimum selective media for transformant selection and for detection of laccase production were previously described (Zhu and Williamson, 2003). Plasmids pPM8 and pCIP containing the URA5 gene are a kind gift of K.J. Kwon-Chung. Plasmids containing the hygromycin resistance gene was a gift of G. Cox.
Cell lines and culture media.
J774.16 (ATCC) is a murine macrophage-like cell line derived from a reticulum sarcoma. Cells were maintained at 37°C in 10% CO2 in DMEM (Invitrogen, Rockville, MD) that was supplemented with 10% heat-killed fetal calf serum (FCS) (Harlan, Indianapolis), Cellgrow 1× non-essential amino acids (Mediatech, Haerndon, VA), Cellgrow 100 mg ml−1 penicillin/streptomycin and 10% NCTC-109 medium (Invitrogen). Cell lines were used between 5 and 15 passages.
Disruption and complementation of VPS41
To make the deletion construct, a 2.5 kb genomic fragment encompassing the major part of VPS41 ORF was PCR-amplified using primer pairs of VPS41F1 (5′-GCTCCAGAAGAGCACGGATG-3′) VPS41R1 (5′-GTGTGACATGTTTGCGTATG-3′) and subsequently cloned into TOPO TA vector (Invitrogen). The resultant plasmid was digested with EcoRV and BglII and the 447 bp EcoRV/BglII fragment was replaced by a 1.6 kb fragment of a URA5 transformation maker described previously (Varma et al., 1992) to generate the deletion construct plasmid. The final disruption allele with a 1.6 kb URA5 marker flanked on either side by a 500 bp DNA sequence homologous to genomic regions of the VPS41 gene was PCR-amplified and purified by spin column (Invitrogen) and introduced into H99FOA cells via a biolistic approach (Cox et al., 1996). Transformants were screened for potential VPS41 deletion mutant by a PCR approach using one set of primers, designed to detect disruption events, including a primer just outside of the transforming vector and an opposing primer based on the URA5 sequence. The specific disruption of the VPS41 gene in candidate mutants was verified by using a set of primers outside the sequences of the knockout construct and by Southern blot analysis. To complement the vps41Δ mutant strain, a 6 kb genomic fragment encompassing the full ORF of VPS41 plus ∼800 bp of the 5′ promoter region was PCR-amplified using a primer set of VPS41FC (5′-TCGAGCTTGGACAGCAGAG-3′) and VPS41RC (5′-ATTCACTGCGCACACGTCTGC-3′) and then cloned into a modified Bluescript SK vector (Stratagene) containing the 2 kb hygromycin-B resistance gene under the control of a cryptococcal actin promoter (Cox et al., 1996) to generate the complementation construct, which was introduced into vps41Δ mutant cells by electroporation as previously described (Erickson et al., 2001) and transformants were selected on hygromycin-containing YEPD agar plates (200 μ ml−1).
Macrophage killing assay
The J774.16 macrophage-like cell line (ATCC) was used to evaluate the ability of the vps41Δ mutant to grow inside macrophages by a previously described method (Tucker and Casadevall, 2002). Cells were allowed to grow for 5–7 days in Dulbecco's modified Eagle medium (Cellgro, Herndon, VA) supplemented 10% FCS, 100 μg ml−1 Cellgro penicillin-streptomycin at 37°C in the presence of 5% CO2, and then harvested from monolayers using 0.25% trypsin and the number of cells were counted with a hemocytometer. The macrophage concentration was adjusted to 105 cells ml−1 and 100 μl of the macrophage suspension was added to each well of a 96-well plate. The cells were primed with murine gamma interferon (IFNγ) (Sigma) at a concentration of 50 U ml−1 and were incubated at 37°C, 5% CO2 overnight. Yeast cell suspensions (107 ml−1) of H99, vps41Δ and the complement were prepared from fresh cultures and antibody 18BF (Y1) (a generous gift of A. Casadevall) added (10 μg ml−1) and incubated at 37°C for 1 h. To each well in the 96-well plate, 104 antibody-treated cryptococcal cells were added plus 50 units of IFNγ (Sigma) and 1 μg of LPS and incubated at 37°C 5% CO2. The macrophage and yeast mixtures were harvested at various time points (1, 3, 10 h) and extracellular yeast cells were removed by washing with phosphate-buffered saline (PBS) three times and lysed with 0.1% SDS in water. The collected yeast cells were finally washed with and suspended in PBS and plated on chloramphenicol-containing YPD agar for colony counts. All experiments were performed in triplicate.
Macrophage phagocytosis assay
Macrophage phagocytosis assays were conducted according to the method of Tucker and Casadevall (2002). Briefly, 18–24 h before assays, macrophages were plated at a density of 1–3 × 105 cells per ml in culture media as above. Log-phase cryptococcal cells were washed three times in PBS and counted in a haemacytometer. Yeast cells were added to cultured cells at a multiplicity of infection of 1:1 in the presence of 10 mg ml−1 monoclonal antibody 18B7 (a generous gift of A. Casadevall) and IFNγ and LPS added as described above. Phagocytosis was allowed to proceed at 37°C for 1 h in 10% CO2, and extracellular yeast cells were removed by three successive washes with regular culture feeding medium without LPS or IFNγ, then stained with Diff Quick and phagocytosis assessed by counting 100 macrophages. Five randomly selected fields were counted. The phagocytosis index was expressed as a mean percentage of AMs attaching to or ingesting one or more organisms per 100 Ams ± standard error (SE) as previously defined (Liu et al., 1999a).
Starvation stress response assay
To assess the role of VPS41 in starvation stress response, yeast cells of overnight cultures in YPD liquid medium of H99 wild-type, vps41Δ mutant and the reconstituted vps41Δ+VPS41 strains were collected and resuspended, respectively, to a concentration of concentration of 108 cells ml−1, in 15 ml of YNB (without amino acids and ammonium sulphate) in a 50 ml tube and then cultivated at 37°C with shaking of 250 r.p.m. min−1. At the indicated time point, 100 μl of each strain suspension was taken out and used for determining cell viability based on cfu counting on YPD agar using 10-fold serial dilutions.
Capsule was assayed by India Ink microscopy after incubation on malt extract agar for 5 days at 30°C, laccase by incubation on noradrenaline asparagines agar and urease activity measured on Christensen's media as described previously (Panepinto et al., 2005). The virulence of the vps41Δ mutant, reconstituted vps41Δ and wild-type strains were evaluated in a previously described mouse meningoencephalitis model (Salas et al., 1996). In brief, each yeast strain was allowed to grow for 2 days on YPD agar at 30°C, and cells were collected, washed once with sterile water and suspended in sterile PBS attaining a concentration of 1 × 107 cells ml−1. For each yeast strain, 100 μl of a cell suspension (containing 106 cells) was tail vein-injected into 10 NIH-Swiss Albino female mice of 6–10 weeks old. Inoculated mice were monitored twice daily for survival for 30 days. Mice showing an inability to reach food or water were sacrificed.
Antibody staining and Western blotting
The method of Tucker and Casadevall (2002) and Eksi et al. (2004) was adapted as follows. Cells were grown to mid-log phase and either harvested or washed twice in sterile water and incubated for the indicated time in YNB without amino acids and ammonium sulphate at 37°C. Cells were fixed in 3% formaldehyde for 1 h at 4°C, washed extensively, and then subjected to spheroplasting using 40 mg ml−1 of lysing enzyme from Trichoderma harzianum (Amersham) in 1 M sorbitol, 10 mM sodium citrate, pH 5.8, as described (Varma and Kwon-Chung, 1991) for 4 h at 30°C. Cells were then washed in 1 M sorbitol, 10 mM sodium citrate buffer, and diluted 1:8 in PBS and dried on microscope slides, fixed in 100% anhydrous methanol at −60°C and dried. To quench autofluorescence, cells were incubated in 0.1 M glycine for 10 min. Cells were then incubated with a solution of 1:500 dilution of rabbit anti-yeast Atg8 (Abcam, Cambridge) in PBS containing 1 mg ml−1 bovine serum albumin (Sigma) at 4°C for 1 h, followed by extensive washing in PBS, then incubation for 1 h at 4°C with 1:1000 Alexafluor 594 chicken anti-rabbit antibody followed by three washes with PBS and immunofluorescence examined using an Olympus IX-70 microscope. Western blots using cryptococcal cell extract, anti-Atg8 antibody and a mouse anti-rabbit horseradish peroxidase antibody (Sigma) was performed to demonstrate immunoreactivity of cryptococcal Atg8 as described (Zhu et al., 2001).
Construction of pORA-XK cryptococcal shuttle vector
Because of limitations regarding size and the presence of unique restriction sites in available vectors, a C. neoformans shuttle vector was constructed for high-copy episomal retention and expression, based on the pPM8 shuttle vector derived previously (Mondon et al., 2000). The origin of replication from pBluescript SK (Stratagene) was PCR-amplified using primers: ORI-NotI-1000S, 5′-GCCGCCGCGGCCGCGGGGAGAGGCGGTTTG-3′ and ORI-XbaI-1942 A, 5′-GCCGCCTCTAGATTGATTTAAAACTTCATTTTTA-3′, and digested with NotI, XbaI. Plasmid pPM8 (a generous gift of K.J. Kwon-Chung) was then digested with NotI, XbaI, and two bands, the first corresponding to a Telomere/Kanamycin sequence and a second corresponding to the STAB sequence was gel-purified and the three fragments ligated and transformed into Escherichia coli. After confirmation by restriction mapping, the resultant plasmid was digested with XbaI and a double-stranded oligonuclotide was produced by temperature annealing and phosphorylation with T4 kinase using two oligonucleotides: RNAi-MCS-S, 5′-TATGGATCCGAATTCGAGCTCTAGACTCGA-3′ and RNAi-MCS-A, 5′-GATCTCGAGTCTAGAGCTCGAATTCGGATCCA-3′, and were ligated and transformed into E. coli cells, and the resultant 3 kb plasmid, pORA-XK, recovered, and verified by sequencing. For transformation into C. neoformans, a 1.3 kb PCR-amplified fragment of the URA5 gene described previously (Erickson et al., 2001) was ligated into the KpnI site of pORA-SK to produce pORAS-KU, then a fragment of the EF1-alpha terminator was PCR-amplified from H99 genomic DNA using primers: EF1-T-Age-A, 5′-GCCGCCACCGGTTGAGACACCTTCACCTTGAT-3′ and EF1-T-RI-S, 5′-GCCGCCGAATTCTCGCTTCTGTAGAGCCCAT-3′, digested with AgeI and EcoRI and ligated into compatible sites of pORA-SKU to produce pORA-SKUT. Next, a fragment of the Actin promoter was PCR-amplified from H99 genomic DNA using primers: ACT1-ATG-R1, 5′-CGCCGAATTCATAGACATGTTGGGCGAGTTTAC-3′ and ACT1-Bgl-S, 5′-GCCGCCAGATCTCCATTGCGCATGTACTCGC-3′, digested with BglII and EcoRI and ligated into compatible sites to produce the pORA-KUTAP C. neoformans expression vector and was sequence verified. Transformation of C. neoformans H99FOA19 was performed by electroporation after linearization with I-SceI and C. neoformans transformants were found to contain pORAS-KUTAP with a copy number of approximately 15–20 measured by the ratio of the intensity of the genomic to episomal URA5 band on uncut Southern blots, and was comparable to equivalent ratios obtained in transformants of the same cells using pPM8.
Localization of Vps41, vacuole and proteolytic vesicles of C. neoformans
To facilitate cellular localization studies in C. neoformans, a synthetic copy of GFP was constructed that utilized codon bias from C. neoformans. To produce this, a codon use table was determined for two highly expressed proteins in C. neoformans, SSA1 and EF1α, using annotated sequence from the H99 cryptococcal sequencing project http://www.broad.mit.edu and MacVector software. Synthetic overlapping oligonucleotides were then designed and synthesized and PCR amplification was used to produce the synthetic gene according to the method of Gao et al. (2003) which was introduced into a puc19 vector and verified by sequencing (GenBank DQ474233) and was inserted into an EcoRI site upstream of a 500 bp EF1-alpha terminator within pORAS-KUTAP using MunI sites at the 3′-end of the GFP to maintain the unique upstream EcoRI site for further use. A genomic copy of VPS41 spanning 1.5 kb upstream of the ATG start codon and terminating at amino acid 1038 in frame with was then introduced upstream of the Cneo-GFP construct previously inserted into pORA-S-KUTAP using BglII, EcoRI sites (which removes the ACT1 promoter) and the construct verified by sequencing. The cryptococcal shuttle vector was then linearized with I-SceI and transformed into a vps41Δ ura5– strain produced by standard methods (Varma et al., 1992). Expression of VPS41 was verified by Northern blot and complementation was verified by restoration of wild-type vacuolar morphology by FM4-64 staining described below. Expression of the VPS41 gene from S. cerevisiae was performed by insertion of a PCR-amplified copy of the gene from S. cerevisiae strain BJ-3505 (Kodak) using primers: 1, 5′-CAAGAATTCAATGACTACAGATAATCATCAG and 1, CAAGAATTCATGACTACAGATAATCATCAG into pORAS-KUTAP after digestion with EcoRI and ligation into compatible sites, then verified by sequencing. The plasmid was then linearized with I-SceI and transformed into a vps41Δ ura5 strain produced as described (Varma et al., 1992) and expression verified by Northern blot.
The vacuolar fluorescent stains, CMAC, CMAC-Arg, CMAC-Ala-Pro and MDY-64 (all from Invitrogen), were used to localize and determine the function of vacuolar proteases. Fungal cells were incubated under the described conditions, then incubated according to the manufacturer's directions, then embedded in 2% low-molecular-weight agarose under a coverslip, allowed 10 min to settle at 42°C, then cooled to 4°C on a Perkin Elmer PCR machine. Cells were then examined by epifluorescence and Nomarski or bright field optics using an Olympus IX-70 inverted microscope at optical wavelengths recommended for each dye according to the manufacturer.
Budding index and determination of cellular DNA content
Budding index was determined similar to the method of Zettel et al. (2003). Briefly, cells were harvested, vortexed for 30 s and the proportion of adherent buds measured. Methods for the determination of cellular DNA content was adapted from that by Ohkusu et al. (2001). Briefly, yeast cells were grown to mid-log phase and either harvested or washed twice in sterile distilled water and incubated for 24 h in YNB without amino acids and ammonium sulphate at 37°C in an orbital shaker. Harvested cells were washed once with sterile distilled water and resuspended in ice-cold 70% ethanol for fixation and stored at 4°C. Aliquots (1 ml) of the fixed cells were centrifuged in a microcentrifuge at 13 000 g for 30 s and resuspended in 1 ml of PBS and allowed to rehydrate for 5 min. Cells were centrifuged again and resuspended in 1 ml of propidium iodide (PI) staining solution (0.1% Triton X-100, 0.2 mg ml−1 DNAse-free RNAse, 5 μg ml−1 PI in PBS) and incubated at room temperature for 30 min. Stained cells were kept on ice until flow cytometry was performed. DNA content was examined in each cell using a laser-scanning-Beckman Coulter 500 cytometer with peak versus interval used to gate our aggregates with a 590 dichoic and a 610 fixed band path filter. DNA of cells were visualized by epifluorescence using the soft agar embedding technique described above after incubation with 1 μg ml−1 DAPI (Sigma-Aldrich) in PBS, pH 7.0. Assessment for metabolic activity was performed by incubating log-phase cells in YNB without amino acids and ammonium sulphate for 1.5 h at 30°C, followed by a 30 min incubation in the presence of 20 μM FUN 1 (Invitrogen) and observed for red staining vesicles according to the manufacturer's directions.
Statistical significance of mouse survival times was assessed by Kruskall–Wallis analysis (anova on Ranks). Pairwise analyses were performed post hoc by using Dunn's procedure. Budding index comparisons were performed by a chi-square two-sided test using a 2 × 2 contingency table.
This work was supported, in part, by United States Public Health Service Grants NIH AI49371 and AI45995 (to P.R.W.). We also thank TIGR (supported by NIH Grant 1 UO1 AI48594) and the Fungal Genomic Initiative for sequence information for the C. neoformans genome.