The Cryptococcus neoformans Rho-GDP Dissociation Inhibitor Mediates Intracellular Survival and Virulence

Abstract

Rho-GDP dissociation inhibitors (Rho-GDI) are repressors of Rho-type monomeric GTPases that control fundamental cellular processes, such as cytoskeletal arrangement, vesicle trafficking, and polarized growth. We identified and altered the expression of the gene encoding a Rho-GDI homolog in the human fungal pathogen Cryptococcus neoformans and investigated its impact on pathogenicity in animal models of cryptococcosis. Consistent with its predicted function to inhibit and sequester Rho-type GTPases, overexpression of RDI1 results in cytosolic localization of Cdc42. Likely as a result of this finding, RDI1-overexpressing strains exhibited altered morphology compared to that of the wild type, with apparent defects in maintaining proper cell polarity and cytokinesis. RDI1 deletion resulted in increased vacuole size in tissue culture medium and aberrant cell morphology at neutral pH. Maintenance of normal cell morphology is vital for C. neoformans pathogenicity. Accordingly, the rdi1Δ mutant strain also showed reduced intracellular survival in macrophages and severe attenuation of virulence in two murine models of cryptococcosis. This reduction in virulence of the rdi1Δ mutant occurs in the absence of major growth defects in rich medium and with classical virulence-associated phenotypes.

Cryptococcus neoformans is an important fungal pathogen among immunocompromised patients, causing disease primarily in patients with AIDS, organ transplants, or immunosuppressive therapy (6). It is also emerging as a primary pathogen in immunocompetent individuals (33, 52). For example, in Vancouver, Canada, a newly identified genotype of Cryptococcus gattii (a species that is closely related to C. neoformans) is responsible for a significant outbreak of cryptococcal disease in human and animal populations (22, 23, 33).

Interestingly, this fungus proliferates both intracellularly and extracellularly in infected hosts (19). Several factors are required for intracellular survival. The polysaccharide capsule is an established virulence-associated phenotype of Cryptococcus, and recent studies have demonstrated the importance of polysaccharide production in intracellular survival (18). Likewise, extracellular phospholipase B activity is required for full virulence (11). The abilities of the fungus to regulate oxidative stress responses (7) and produce melanin (58) are also important for intracellular survival. Lastly, maintenance of C. neoformans within alveolar macrophages has been shown to be crucial for disease progression (32). Indeed, C. neoformans can remain dormant within macrophages for decades after the initial infection (26).

Many of the basic cellular processes that allow the survival of microbial pathogens in their hosts are controlled by highly conserved signal transduction pathways, which regulate fundamental growth signals in various microorganisms. These conserved signaling molecules include Rho-type GTPases, such as Cdc42 and Rac proteins. Homologs of Cdc42 and Rac in C. neoformans mediate the downstream activities of the Ras signal transduction cascade, which has been shown to regulate high temperature growth, mating, and cell morphology (3, 42, 54, 59). Since Cdc42 and Rac are likely central regulators of cell growth and survival in the host, we elected to study those processes that control the activities of these proteins.

Rho-GDP dissociation inhibitors (Rho-GDI) are important regulators of monomeric GTPases, including Rho1 and Cdc42 (34, 41). Rho-GDIs inhibit the activities of these monomeric GTPases by binding their GDP-bound forms and sequestering them in the cytosol. In this state, these GTPases remain inactive until released from Rho-GDI inhibition. Although Rho-GDI homologs regulate key signaling events in the cell, deletion of the genes encoding these proteins in several fungal systems has thus far resulted in no apparent phenotypes among the mutant strains (38, 41). Interestingly, deletion of RDI1 from the human pathogenic yeast Candida albicans results in reduced filamentation only when coupled with deletions of the Cdc42 GTPase-activating proteins Rga2 and Bem3 (9). Therefore, even though Rho-GDI homologs regulate central signaling proteins involved in basic growth processes, deletion of the genes encoding Rho-GDI proteins results in few phenotypes in the fungal systems studied to date. In contrast, overexpression of Rho-GDI homologs results in dramatic phenotypes, including severe morphological defects and death (38, 41).

Interestingly, Rho family signaling has been shown to be important for disease processes in other plant and animal pathogens. Rho signaling pathways are active in trophozoites of the pathogenic amoeba Entamoeba histolytica, inducing the remodeling of actin required for host tissue invasion (20). Likewise, deletion of CDC42 or RAC1 in the basidiomycete plant pathogen Ustilago maydis results in strains which are nonpathogenic. Deletion of RAC1 impedes hyphal extension, which is a requirement for pathogenicity in maize. Curiously, CDC42 deletion also renders the fungus nonpathogenic on maize, albeit with no discernible morphological abnormalities (37).

We therefore hypothesize that perturbation of Rho family signaling in C. neoformans will result in virulence defects. We report that deletion of the Rho family regulatory protein Rdi1 results in severely attenuated virulence in both inhalational and intravenous mouse models of cryptococcal disease. This attenuation occurs in the absence of defects in the classical pathogenicity factors of capsule, melanin production, high-temperature growth, or sensitivity to oxidative and nitrosative stress. In contrast, C. neoformans Rdi1 helps to control Cdc42 protein localization. Altering Rdi1 expression affects cell morphology, vacuole morphology, intracellular survival in macrophages, and survival in mice.

MATERIALS AND METHODS

Strains and media.

Three independent Cryptococcus neoformans rdi1Δ mutant strains, MPC2, MPC3, and MPC9 (MATα rdi1Δ::nat), were created in the serotype A wild-type (WT) strain H99 (46). Strains were grown using standard yeast media as described previously (51) or tissue culture medium (Dulbecco's modified Eagle's medium [DMEM] with or without 22 mM NaHCO₃ buffer, 10% NCTC 109, 10% fetal calf serum, 1% minimal essential medium nonessential amino acid solution, and 1% penicillin-streptomycin).

Cloning and overexpression of the RDI1 gene.

The C. neoformans RDI1 homolog (NCBI GeneID CNG02620) was identified with a BLASTP search of the C. neoformans H99 genome (Broad Institute, Cambridge, MA), using the sequence of the Saccharomyces cerevisiae Rdi1p protein as a query. The genomic sequence of C. neoformans RDI1, along with annotation, was procured from the Cryptococcus neoformans Serotype A Genome Database at the Broad Institute (http://www.broad.mit.edu/annotation/genome/cryptococcus_neoformans.2/Home.html). An alignment of the C. neoformans Rdi1 protein sequence to those of other Rho-GDI proteins was performed using Clustal X version 2.0.6 (36) with default settings.

The C. neoformans RDI1 gene was amplified from genomic DNA using primers AA1246 and AA1247 for cloning into an expression vector harboring the C. neoformans serotype A GAL7 promoter sequence and the neomycin resistance marker (Table 1). The In-Fusion cloning kit was used according to the manufacturer's instructions (Clontech Laboratories, Inc., Mountain View, CA). The resulting plasmid, pMP12, was then introduced into C. neoformans H99 using biolistic transformation as previously described (53). Stable transformants were selected by plating cells on yeast-peptone-dextrose (YPD) medium containing 200 μg/ml Geneticin (Invitrogen Corp., Carlsbad, CA). Overexpression was verified for three transformants by quantitative real-time PCR using primers AA952 and AA953 (Table 1) as previously described (13). One of these transformants, strain MPC16, was selected for further study.

TABLE 1.

Primers used in this study

Primer	Sequence	Description
AA0756	GGTAAGAAAGAGGTCGGAGT	RDI1 5′-end-flanking forward primer
AA0757	GTCATAGCTGTTTCCTGGCATCTTTGGGAGATGAAT	RDI1 5′-end-flanking reverse primer
AA0758	CTGGCCGTCGTTTTACACGTTTTGTAGCCTTTGTGT	RDI1 3′-end-flanking forward primer
AA0759	TTGCACTTCAAGTTTGTGAG	RDI1 3′-end-flanking reverse primer
M13 forward (−20)	GTAAAACGACGGCCAG	natR marker forward primera
M13 reverse	CAGGAAACAGCTATGAC	natR marker reverse primer
AA1047	GTCATAGCTGTTTCCTGATGGACGTACCGAGACTG	RDI1 locus reverse primer
AA1049	CAGTCTCGGTACGTCCATCAGGAAACAGCTATGAC	neoR marker forward primerb
AA1050	ACACAAAGGCTACAAAACGTGTAAAACGACGGCCAG	neoR marker reverse primer
AA0952	GAGTACTCTGTCGGCATCAC	RDI1 quantitative PCR forward primer
AA0953	TTTTGTGTAAGGTTCCTGCT	RDI1 quantitative PCR reverse primer
AA1246	CTGGCGGCCGCTCGAATGTCCAACCAACAGGTAAG	RDI1 overexpression forward primer
AA1247	TAGATGCATGCTCGAATGGACGTACCGAGACTG	RDI1 overexpression reverse primer

^aThe nat marker is cloned into pCR2.1-TOPO (Invitrogen Corp., Carlsbad, CA); the M13 forward (−20) and M13 reverse primers are used to amplify the marker.

^bThe neo marker is amplified from pJAF1 (23).

Cdc42 localization assay.

We created a green fluorescent protein (GFP)-Cdc42 fusion to evaluate the interaction of Cdc42 and Rdi1 (47). A histone H3 promoter-GFP fusion (30) was cloned into the nourseothricin resistance gene-containing plasmid pCH233 (a gift from Christina Hull), utilizing EcoICRI and BamHI restriction sites to create the plasmid pCN19. The CDC42 gene from C. neoformans H99 was then amplified by PCR using primers modified with BamHI restriction sites on their 5′ and 3′ ends and subsequently cloned into pCN19 at the BamHI site to create the plasmid pCN20. pCN20 was then biolistically transformed into C. neoformans strain MPC16 as previously described (53).

Disruption of the RDI1 gene and creation of an rdi1Δ mutant.

The linear gene deletion construct for disruption of RDI1 was produced using PCR overlap extension as described previously (14), using the nourseothricin resistance gene (nat^R) for positive selection of transformants (39). Primers AA0756 and AA0757 were used to amplify the 5′-end-flanking region of the RDI1 gene, primers AA0758 and AA0759 were used to amplify the 3′-end-flanking region, and M13 forward (−20) and M13 reverse primers were used to amplify the nat^R resistance marker (Table 1). Identification of the 5′- and 3′-end-flanking regions of RDI1 was accomplished by searching the C. neoformans H99 genome sequence with the sequence of the S. cerevisiae Rdi1 protein and selecting the regions that were 1 kb upstream and downstream of the region of highest homology.

The rdi1Δ::nat deletion construct was biolistically transformed into strain H99 as described previously (53). Stable transformants were selected by plating them on YPD medium containing 100 μg/ml nourseothricin (Werner BioAgents, Jena, Germany). Confirmation of rdi1Δ mutants was accomplished by Southern hybridization, using DNA regions inside and outside the replaced RDI1 locus as probes (data not shown) using standard methods (49). One of these mutants, MPC3, was selected for further investigation.

The reconstituted strain MPC11 (MATα rdi1Δ::nat RDI1-neo) was made by biolistically transforming the rdi1Δ mutant MPC3 with a construct containing the native RDI1 gene and the neomycin resistance marker (neo^R) (5, 23). Briefly, the native RDI1 locus was amplified with 1 kb of flanking DNA at the 5′ end and 381 bp at the 3′ end using primers AA0756 and AA1047 (Table 1). This PCR product was fused to the neo^R marker (Table 1, primers AA1049 and AA1050) via overlap PCR as previously described (14). A 1-kb flanking DNA starting 58 bp from the RDI1 stop codon was amplified using primers AA0758 and AA0759 (Table 1), and this PCR product was then fused by overlap PCR to the RDI1-neo^R PCR product to yield the reconstitution construct. The RDI1 reconstitution construct was biolistically transformed into MPC3. Stable transformants were selected by plating cells on YPD medium containing 200 μg/ml Geneticin (Invitrogen Corp., Carlsbad, CA). The reconstituted strains were tested for reversion of mutant phenotypes in tissue culture medium and by quantitative PCR to assess RDI1 expression.

Urease activity assay.

WT and rdi1Δ mutant strains were assessed for urease activity, a known virulence determinant (12). One colony of each strain was added to sterile, deionized H₂O in a microcentrifuge tube and vortexed vigorously. One BBL Taxo urease differentiation disk (Becton, Dickinson, & Co., Sparks, MD) was added to each tube by a sterile technique. The tubes were then incubated at 30°C and checked at 10 min and 30 min for urease activity according to the manufacturer's instructions.

Capsule and melanin visualizations.

C. neoformans WT, rdi1Δ, and rdi1Δ::RDI1 strains were incubated at 37°C in 5% CO₂ for 3 days in tissue culture flasks (Corning, Inc., Corning, NY) containing 20 ml DMEM for capsule induction. Following capsule induction, the cultures were formalin fixed, washed once with 20 ml sterile phosphate-buffered saline (PBS), and resuspended in 1 ml sterile PBS for visualization of capsule. India ink preparations were made for each strain on glass slides for visualization. Images were collected at a ×100 magnification as described below. Total (cell and capsule) and cell-only diameter measurements were obtained along the same axis for 50 cells for each strain (27). Melanin production was assessed by growth on niger seed agar as described previously (1).

Visualizations of actin and vacuoles.

Strains were grown under tissue culture conditions as described for capsule measurement for 2 to 3 days. Actin and DNA were visualized as previously described (56), with the following modifications. Cells were fixed in formaldehyde (10% final concentration) for 30 min. Cells were washed three times in PBS, permeabilized in 1% Triton X-100 in PBS for 10 min, washed three times in PBS, and stained with 8 μg/ml 4,6-diamidino-2-phenylindole (DAPI) (D-21490; Molecular Probes) to visualize DNA or 10 μg/ml rhodamine-conjugated phalloidin (P-195; Sigma-Aldrich, St. Louis, MO) to visualize filamentous actin.

To visualize vacuoles, cells were stained with the lipophilic dye N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino) phenyl) hexatrienyl) pyridinium dibromide (FM 4-64) (T-3166; Invitrogen Corporation, Carlsbad, CA) as described elsewhere (55), with the following modifications. WT and rdi1Δ strains were incubated in DMEM without NaHCO₃ at 37°C for 3 days without shaking. FM 4-64 was added to growing cells at a final concentration of 10 μg/ml and incubated for 30 min at 37°C. Cells were washed three times in PBS to remove dye and incubated for 1 h prior to visualization.

Impact of neutral pH on cell morphology.

Due to the effect of tissue culture medium on vacuole morphology, we sought to determine if this phenotype was pH related. C. neoformans WT, rdi1Δ, and rdi1Δ::RDI1 strains were incubated in yeast nitrogen base (YNB) broth at either pH 5 or pH 7, buffered with 50 mM succinic acid or 50 mM Na-MOPS (morpholinepropanesulfonic acid), respectively. The broth cultures were then grown in a 24-well plate at 37°C in 5% CO₂ for 2 days. Capsule induction was assessed by India ink counterstaining.

Effect of the rdi1Δ mutation on mating.

To determine the effect of the rdi1Δ mutation on mating, the rdi1Δ mutant strain MPC3 (rdi1Δ MATα) was crossed with WT strain KN99a (MATa) (44). Progeny of this cross were evaluated for movement of the rdi1Δ mutation into the MATa background, producing strain MPC5 (rdi1Δ MATa). A bilateral cross was then performed by pairing MPC3 and MPC5, along with rdi1Δ × WT crosses in both MAT backgrounds as controls. All matings were performed on V8 agar as previously described (3).

Survival in macrophages.

We assessed the effect of RDI1 deletion on the survival of C. neoformans in macrophages as previously described (10). WT, rdi1Δ, and rdi1Δ::RDI1 strains were incubated overnight at 30°C at 150 rpm to serve as inocula for the macrophage killing assay. All inocula were quantitatively cultured to document the normalization of C. neoformans cells in the macrophage coculture measurements. The phagocytosis index was determined for each strain in J774A.1 macrophages as described previously (10, 18).

Briefly, 1 × 10⁵ activated macrophages (strain J774A.1) were coincubated in 96-well microtiter plates with 1 × 10⁵ cells of each C. neoformans strain for 1 h at 37°C in 5% CO₂ and DMEM to allow phagocytosis to occur. Survival in macrophages was assayed in triplicate wells for each strain; the experiment was independently performed five times with identical results. The cells of each C. neoformans strain were opsonized with a monoclonal antibody against capsule (18B7 [a gift from Arturo Casadevall]) prior to coincubation with macrophages (60). Following the 1-h coincubation, the macrophages were washed twice with sterile PBS to remove extracellular yeasts. A 150-μl sample of DMEM was added to each coculture, and the cultures were incubated for a further 18 to 24 h. After incubation, the macrophages were lysed with 0.5% sodium dodecyl sulfate. A 50-μl sample of a 1:10 dilution of each lysate was plated on YPD agar plates and incubated at 30°C for colony formation. CFU were counted on each plate, and the numbers were normalized to the numbers of cells inoculated for each strain to determine yeast survival (10). Statistical significance of differences in intracellular survival in macrophages was determined using Student's t test.

In vivo virulence assessment.

The virulence of the rdi1Δ mutant strain was determined using both inhalational and intravenous mouse models. For the inhalational model, A/JCr mice (10 mice per C. neoformans strain) were infected with 10⁵ CFU of either WT (H99), mutant (rdi1Δ), or reconstituted (RDI1::rdi1Δ) strains of C. neoformans in a volume of 50 μl sterile PBS by following established protocols (12). The mice were followed closely for signs of worsening infection and euthanized at predetermined endpoints correlating with an eventually lethal infection (weight loss of ≥15%, neurological symptoms, and an inability to access food/water). The intravenous model was performed similarly, with inoculation of mice with 10⁵ CFU by tail vein injection of either the WT or rdi1Δ mutant strain. Statistical significance between the survival curves was determined using the log rank test.

Microscopy.

Brightfield, differential interference microscopy, and fluorescent images were captured with a Zeiss Axioskop 2 Plus fluorescence microscope equipped with an AxioCam MRM digital camera, using the appropriate filter set. Additional brightfield and capsule images were captured with a Nikon Eclipse E400 microscope equipped with a Nikon CoolPix 990 digital camera.

RESULTS

Identification of a Rho-GDI gene in C. neoformans.

BLASTP analysis of the C. neoformans genome identified a single protein with 39% identity to Rdi1p in S. cerevisiae. The corresponding C. neoformans gene encodes a putative protein of 205 amino acids and contains the RHO-GDI six-element fingerprint signature common to proteins of this type (EMBL-EBI accession no. IPR000406 [http://www.ebi.ac.uk/interpro/IEntry?ac=IPR000406]). An alignment of C. neoformans RDI1 with other Rho-GDI proteins revealed conservation of Rho protein interaction residues (Fig. 1). Interestingly, C. neoformans RDI1 lies adjacent to the RHO1 gene, encoding a Rho-GTPase. The two genes are divergently transcribed, with only 328 nucleotides separating their start codons, suggesting shared transcriptional regulatory elements.

FIG. 1.

Multiple-sequence alignment of predicted protein sequences for various Rho-GDI. The predicted amino acid sequence for C. neoformans Rdi1 (boldface) was aligned with other Rho-GDI proteins from various species using the Clustal X software package (36). Default settings were used for the alignment. Amino acid residues shown to be important in the interaction of Rho GTPase and the bovine and human Rdi2 proteins (48) are identified with filled circles. H. sapiens, Homo sapiens; D. melanogaster, Drosophila melanogaster; C. elegans, Caenorhabditis elegans; A. fumigatus, Aspergillus fumigatus; S. pombe, Schizosaccharomyces pombe; D. discoideum, Dictyostelium discoideum; A. thaliana, Arabidopsis thaliana.

Overexpression of RDI1 affects cell morphology and Cdc42 localization.

C. neoformans RDI1 was overexpressed in WT strain H99 using the galactose-inducible GAL7 promoter from C. neoformans H99. Overexpression of RDI1 resulted in altered cell polarization in approximately 5% of rdi1Δ cells (Fig. 2), similar to that observed for ROM2 mutants (24). The aberrantly polarized cells displayed short hypha-like projections. These projections were observed in <1% of WT cells under identical conditions.

FIG. 2.

Overexpression of RDI1 results in altered polarization. WT and Gal7(p)::RDI1 strains were grown in YPD broth with either glucose or galactose as a carbon source. No differences were observed between strains on glucose. However, the Gal7(p)::RDI1 strain showed a higher percentage of hyperpolarized cells (arrow) on galactose (Rdi1-inducing conditions) than on glucose. Bar = 50 μm.

Additionally, overexpression of RDI1 modified the localization of GFP-Cdc42. GFP-Cdc42 localizes primarily to the cell membrane as well as to various internal membranes in the presence of WT levels of Rdi1 (Fig. 3, left). However, cell membrane localization of GFP-Cdc42 was lost upon overexpression of RDI1 (Fig. 3, right), with a corresponding shift to cytosolic localization. This activity is predicted for a Rho-GDI protein and suggests that C. neoformans Rdi1 plays a major role in inhibiting the membrane localization and function of Cdc42 and other Rho GTPases.

FIG. 3.

RDI1 overexpression alters GFP-Cdc42 localization. The subcellular localization of C. neoformans Cdc42 was monitored using a constitutively expressed GFP-Cdc42 construct (47). A galactose-inducible RDI1 allele was coexpressed in this strain. (Left) Under non-Rdi1-inducing conditions (with glucose), GFP-Cdc42 displays prominent cell surface localization. (Right) In contrast, Rdi1-inducing conditions (with galactose) result in a loss of the cell surface localization of GFP-Cdc42 and a corresponding cytoplasmic pattern of fluorescence. Bar = 20 μm.

Deletion of the Rho-GDI RDI1 in C. neoformans H99. (i) General growth phenotypes.

To study its biological function in relation to pathogenicity, the C. neoformans RDI1 gene was deleted from the WT strain H99. Three independent rdi1Δ mutants were created, and all phenotypes were identical in the strains. Therefore, one of these, MPC3, was chosen as the rdi1Δ mutant for these studies. Additionally, we created a reconstituted rdi1Δ::RDI1 strain (MPC11) by reintroducing a WT copy of the RDI1 gene at the native locus. All rdi1Δ mutant phenotypes were entirely complemented in this reconstituted strain.

No differences in growth and colony morphology were observed between the rdi1Δ mutant and the WT when incubated on either YPD or YNB agar medium at 25°C, 30°C, 37°C, or 39°C. The rdi1Δ mutant strain also grew equally to the WT when challenged with various ionic or cell wall stressors (1.5 M NaCl, 0.5 mM CuSO₄, 5 mM CuSO₄, 0.025% sodium dodecyl sulfate, 100 μg/ml calcofluor white, and 20 mM caffeine). Additionally, no differences in urease activity were observed between the rdi1Δ mutant and the WT (data not shown).

(ii) Actin skeleton phenotype.

Rho-GDI proteins in other microorganisms control the activities of proteins involved in cell polarity and morphogenesis (41, 47, 48). Interestingly, no morphological defects were observed in the rdi1Δ mutant strain compared to the WT in the yeast cell phase during incubation in YPD medium. There were no differences in actin organization or cell polarity of budding yeast cells observed between the WT and rdi1Δ mutant strains when they were stained with rhodamine-conjugated phalloidin: normal actin localization was observed in emerging buds and regions of cytokinesis in all strains (data not shown).

(iii) Capsule and melanin phenotypes.

Several inducible phenotypes are required for C. neoformans pathogenesis, including melanin and capsule production. When incubated on melanin-inducing niger seed agar, no differences in levels of melanin production were observed between the WT and rdi1Δ strains. When incubated under capsule-inducing conditions (DMEM containing 22 mM NaHCO₃; 37°C, 5% CO₂), no difference in capsule size was observed between the rdi1Δ mutant and WT strains as assessed by India ink counterstaining and direct measurement of capsule diameter (data not shown). Additionally, no differences in growth rate were observed between the rdi1Δ mutant and the WT under these conditions (data not shown).

(iv) pH impact on cell morphology.

In contrast, vacuole morphology of the rdi1Δ mutant strain grown in DMEM containing 22 mM NaHCO₃ and incubated at 37°C with 5% CO₂ was noticeably altered. When viewed microscopically, the rdi1Δ mutant displayed prominent vacuoles, as visualized by the staining of cellular membranes (Fig. 4). In contrast, WT cells displayed a normal vacuole appearance when incubated under these conditions, as did the reconstituted rdi1Δ::RDI1 strain. To determine which component of DMEM contributed to this altered vacuole phenotype, we grew WT and rdi1Δ mutant cells in YNB broth with and without CO₂ or at various pHs.

FIG. 4.

RDI1 deletion elicits altered vacuolation in C. neoformans. Staining with the lipophilic dye FM 4-64 (T-3166; Molecular Probes) defines the vacuoles in WT and rdi1Δ cells. Cryptococcus strains were incubated in the tissue culture medium DMEM at 37°C without NaHCO₃ for 3 days. Staining and microscopy were performed as previously described (55), with the following modifications: FM 4-64 was added to growing cells at a final concentration of 10 μg/ml and incubated for 30 min at 37°C. Cells were washed three times in PBS to remove unincorporated dye and incubated for 1 h prior to visualization. As shown, rdi1Δ cells exhibit altered vacuolation (arrows) compared to the WT. Bar = 50 μm.

Further analysis of this phenomenon revealed a severe defect in maintaining cell polarity in the rdi1Δ mutant grown at neutral pH (Fig. 5). Strains were grown in YNB broth at either pH 5 or pH 7. The pH was buffered using succinic acid and Na-MOPS, respectively. As shown in Fig. 5, rdi1Δ mutant cells exhibit aberrant cell morphology when grown in YNB broth at pH 7 (compared to both the WT and reconstituted strain) yet displayed normal growth and morphology in the same medium at pH 5 (data not shown).

FIG. 5.

RDI1 deletion elicits altered morphology at neutral pH. WT and rdi1Δ mutant strains of C. neoformans were grown in YNB at either pH 5 or pH 7 at 37°C for 2 days. No differences in growth were observed between the strains at pH 5 (data not shown). At pH 7, the rdi1Δ mutant exhibited an aberrant cell morphology. Quantitative plate counts and vital staining with trypan blue showed that these aberrantly shaped mutant cells were still viable (data not shown). Bar = 50 μm.

(v) Basidiospore production.

Mating is a highly regulated process in C. neoformans and is governed in part by a Ras1 signal transduction pathway leading to filamentation and response to pheromone production (59). Ras1-mediated mating has been shown to involve Rac1 GTPase (54), which in other systems is regulated in part by Rdi1 (15). Therefore, we hypothesized that the loss of Rdi1 would result in changes in mating filament morphology. As shown in Fig. 6, the rdi1Δ mutant was capable of fusion and producing viable basidiospores in a unilateral cross with the appropriate WT mating partner. Similarly, a bilateral mating of rdi1Δ mutant strains resulted in normal fusion and mating filament morphology with normally fused clamp connections. However, the basidia produced in the bilateral cross were bare, with no basidiospores produced even after 21 days of incubation (Fig. 6, inset).

FIG. 6.

An rdi1Δ mutant bilateral cross reveals a sporulation defect. The rdi1Δ mutation was moved into the KN99a MATa background via conventional mating. The resulting strain, MPC5 (rdi1Δ MATa), was then mated with the parental rdi1Δ mutant strain MPC3. Sporulation occurred normally in crosses between the WT and rdi1Δ mutant strains (see left inset). However, the bilateral cross of two rdi1Δ mutants yielded naked basidia (right inset). Bar = 20 μm.

Effect of RDI1 deletion on survival in macrophages.

The rdi1Δ mutant strain grew normally under many in vitro conditions and displayed normal induction of classical virulence factors. However, given the abnormal cellular morphology in tissue culture medium and in YNB medium at neutral pH, we assessed the rdi1Δ mutant for the ability to survive when cultured with macrophages. We challenged J477A.1 macrophages with WT, rdi1Δ, and rdi1Δ::RDI1 strains to evaluate various aspects of the macrophage-yeast interaction. The rdi1Δ mutant strain exhibited an 87% reduction in intracellular survival compared to that of the WT after 24 h of coincubation with the macrophages (Fig. 7) (P < 0.01). The reconstituted rdi1Δ::RDI1 strain demonstrated full restoration of survival within macrophages. Since extracellular yeasts were removed after the first hour of coincubation, these results measured only the intracellular survival of the C. neoformans cells. This difference in survival is not due to medium effects, as there was no difference in growth rate among the strains in this medium (data not shown). Similarly, the intracellular growth differences are not due to altered uptake by the macrophages; there was no difference between the strains in the phagocytic index, a measure of macrophage phagocytosis efficiency (Materials and Methods; data not shown). Additionally, no differences were observed in vitro between the WT and rdi1Δ strains in susceptibility to either nitric oxide (NO) or hydrogen peroxide (H₂O₂), two species likely encountered in the context of host immune cells (MIC = 250 μM H₂O₂; MIC = 125 μM NaNO₂ [pH 4]).

FIG. 7.

Rdi1 mediates survival and proliferation in macrophages. Activated J774A.1 macrophages were challenged with either the WT, rdi1Δ, or rdi1Δ::RDI1 strain of C. neoformans (multiplicity of infection, 1:1). After 60 min of coincubation at 37°C in 5% CO₂, extracellular yeasts were removed and the cocultures were incubated at 37°C in 5% CO₂ overnight. Viable C. neoformans organisms were plated on YPD medium and incubated at 30°C for 2 days for CFU quantification. Data from one representative experiment of five independent and identical experiments are shown. Strains were assayed in triplicate. Error bars represent standard deviations.

Therefore, the rdi1Δ mutation does not result in striking growth phenotypes in many standard media, nor are classical C. neoformans virulence-associated phenotypes affected by this mutation. Although the rdi1Δ mutant is engulfed by macrophages in a manner identical to that of the WT, the mutant demonstrates a striking defect in intracellular survival.

Effect of the rdi1Δ mutation on virulence in vivo.

We next sought to determine if the intracellular survival defect in the rdi1Δ mutant in vitro would translate into altered virulence in vivo. Therefore, the virulence of the rdi1Δ mutant strain was tested in a murine inhalation model of cryptococcosis. A/JCr mice were challenged by intranasal inoculation with 10⁵ CFU of either the WT, rdi1Δ, or rdi1Δ::RDI1 strain. In contrast to the WT, which caused a lethal infection in all mice by 22 days postinoculation (mean survival = 18.5 days), the rdi1Δ mutant demonstrated a significant virulence defect, allowing infected animals to survive a mean of 43.2 days (P < 0.0001) (Fig. 8A). Reconstitution of the rdi1Δ mutant with the native RDI1 gene completely restored its virulence.

FIG. 8.

An rdi1Δ mutant displays severe attenuation of virulence in inhalational and intravenous murine models of disseminated cryptococcosis. In separate virulence experiments, inbred A/JCr mice were infected either intranasally (10 mice per strain) (A) or intravenously (8 mice per strain) (B) with the WT, rdi1Δ, or rdi1Δ::RDI1 strain of C. neoformans. The mice were observed over the course of the experiment for clinical signs correlating with eventual mortality (P < 0.0001; log rank test).

To ensure that the virulence defect in the rdi1Δ strain was not due merely to a defect in alveolar delivery or yeast migration from the lungs, we pursued two additional experiments. First, we harvested lungs from mice infected with the WT and rdi1Δ mutant at 17 days after inoculation. Samples from all infected mice demonstrated abundant yeast cells within the alveolar spaces by histopathological analysis using standard hematoxylin/eosin and methenamine silver staining (data not shown). Second, to bypass the lungs as a site of initial infection, virulence was also assessed by intravenous inoculation. A/JCr mice were infected by tail vein injection with 10⁵ CFU of either the WT or rdi1Δ strain. The rdi1Δ mutant exhibited considerable attenuation compared to the WT, with mortality occurring, on average, 17.9 days postinoculation, compared to 6.6 days for WT-infected mice (P < 0.0001) (Fig. 8B). Therefore, two separate models of systemic cryptococcosis with distinct sites of inoculation have confirmed a virulence defect in the rdi1Δ mutant.

DISCUSSION

Rho-GDI are important negative regulators of various monomeric GTPases in eukaryotic cells (15). The GTPases regulated by Rho-GDI participate in a number of crucial cellular processes, including cytoskeletal rearrangement, vesicle trafficking, and bud site selection (as reviewed in references 4, 25, 31, and 45). Because Rho-GDI-regulated proteins are crucial to cell growth and survival, we investigated the effect of alterations to Rho-GDI activity on survival in vivo using the human pathogenic yeast Cryptococcus neoformans. Studying this process in a microbial pathogen allowed us to determine the effect of host stress on Rdi1 activity.

As observed with other fungi, our studies demonstrate that Rho-GDI deletion yields no major changes in growth morphology under a variety of laboratory conditions (9, 38, 41). The growth of the rdi1Δ mutant is similar to that of the WT when either incubated at various temperatures or challenged with various cell wall or ionic stressors. These results suggest that the yeast is able to utilize functional redundancy in Rho-type GTPase regulation to compensate for the lack of Rdi1-mediated inhibition under laboratory growth conditions. Indeed, the regulation of Rho-type GTPases is complex, involving guanine exchange factors (activation) and GTPase-activating proteins (inhibition), in addition to GDP dissociation inhibitors, such as Rdi1. The paucity of aberrant phenotypes in the rdi1Δ mutant underscores the multifaceted regulation of Rho-type GTPase activity, of which Rho-GDIs are a component.

Rdi1 and the C. neoformans mating response.

We were able to detect alterations in the phenotypes in the rdi1Δ mutant that correlate with aspects of the pathogenic and developmental lifestyle of C. neoformans. For example, the ability to reproduce sexually has been associated with pathogenicity (1, 35). Indeed, the main infectious propagule of the fungus may be the sexual spore (8, 17). We found that deletion of RDI1 in both mating partners resulted in abolition of spore production (Fig. 6). A functional copy of RDI1 in either parental strain allowed normal mating and spore production.

Sexual reproduction in C. neoformans is a highly regulated process involving both heterotrimeric G protein and monomeric GTPase signal transduction cascades (1-3, 43, 57). One of these cascades, the Ras1 signal cascade, is responsible for high temperature growth and sexual reproduction in C. neoformans (3). This cascade includes the Rho-type GTPase Rac1, which is part of the sexual reproduction fork of the Ras1 signal cascade (54). Rac1 is involved in the mating response. Our model predicts that deletion of RDI1 approximates overexpression of Rho-type GTPases such as Rac1 and Cdc42. Therefore, it is likely that the dysregulated activity of one or more of these Rho-type GTPases is responsible for the failure of spore formation in the rdi1Δ mutant strain.

Rdi1 regulation of growth under host physiological conditions.

In addition to its effect on mating, C. neoformans Rdi1 is also required for growth and survival under host physiological conditions and for pathogenicity. The rdi1Δ mutant was unable to establish normal cell morphology under in vitro conditions which approximate the in vivo environment (pH 7, 37°C, 5% CO₂). This altered cell morphology appears to be pH dependent, as the rdi1Δ mutant grown at pH 5 was identical to the WT. The link between Rdi1 and pH may be via Cdc42. We demonstrated a shift in Cdc42 localization by Rdi1 overexpression in C. neoformans (Fig. 3). This Rdi1-dependent localization of GFP-Cdc42 is predicted in current models of Rho-GDI signaling (47). Therefore, alterations in Rdi1 activity are predicted to affect processes regulated by Cdc42.

Cdc42 proteins participate in pH response processes. Cdc42 has been shown to affect cell polarity in a pH-dependent fashion in fibroblasts via a proton-dependent positive-feedback loop (21). In the fibroblast, Cdc42 induces H⁺ efflux via the Na-H⁺ antiporter Nhe1 (29). This Nhe1-mediated H⁺ efflux induces guanine nucleotide exchange by the Cdc42 guanine exchange factor and, therefore, activation of Cdc42 (21). The buffering capacity of the medium may serve to tip the balance of H⁺ efflux in C. neoformans, thus upregulating the activity of the Cdc42 guanine exchange factor. Left unchecked by the rdi1Δ mutation, this proposed increased Cdc42 activity may explain the aberrant morphology seen in the rdi1Δ mutant at pH 7. Interestingly, there is one protein homolog of human Nhe1 in the C. neoformans genome (GeneID CND04380; 27% identity to human Nhe1). It is currently unknown if this Nhe1 homolog is influenced by the activity of Rdi1 in C. neoformans.

Cdc42 is also an important regulator of vesicle docking in budding yeast (16, 40). In C. neoformans, growth of the rdi1Δ mutant under host physiological conditions (37°C with 5% CO₂ in tissue culture medium) revealed an apparent defect in vesicle trafficking (Fig. 4). The mutant strain forms large vacuoles as the cell grows, in contrast to the WT (Fig. 4). In budding yeast, Rdi1 blocks vacuole fusion via removal of Cdc42 and Rho1 from vacuole membranes (16). Our data complement the study by Eitzen et al. by demonstrating a loss of inhibition of vesicle fusion in the rdi1Δ mutant when grown under host physiological conditions (16).

Rdi1 is required for C. neoformans pathogenesis.

The rdi1Δ mutant displays severely attenuated virulence in both inhalational and intravenous mouse models of disease. Although the major inducible virulence factors (melanin and capsule) are unaffected in the rdi1Δ mutant, this strain does exhibit a striking defect in survival in macrophages and growth at pH 7, two specific aspects of the host environment. The intracellular defect is unlikely to be linked to the pH growth defect, as the phagolysosome of macrophages has been determined to be pH 5.6 (28), a pH at which the rdi1Δ mutant is phenotypically normal. The intracellular growth defect may be due to an alteration in stress response signaling, as seen in an SKN7 deletion strain (7). The skn7Δ mutant is phenotypically normal for the known virulence factors important to C. neoformans disease, yet this mutant is reduced in intracellular survival in endothelial cells.

Predicted downstream effectors of Rho-GDI proteins also affect virulence of plant fungal pathogens. For example, in the rye pathogen Claviceps purpurea, a Δcdc42 strain was nonpathogenic in the absence of morphological and host invasion defects (50). Since Rdi1 interacts with Cdc42 in C. neoformans, it may be that the loss of Rdi1 repression in C. neoformans destabilizes Cdc42 function sufficiently to disrupt growth and polarization signals mediated by this protein when encountering the host environment.

In summary, we have demonstrated that the Rho-GDI of C. neoformans, Rdi1, is a negative regulator of Rho-type GTPases and that this protein is crucial in regulating cell morphology and vesicle trafficking. These important processes are regulated by cascades with redundant control mechanisms, including numerous activating and inhibiting proteins. Given Rdi1's role in regulating these important cascades, it is interesting that the deletion of RDI1 genes in various fungi does not result in more-profound phenotypes for growth and differentiation. However, our demonstration of the major defect of the C. neoformans rdi1Δ mutant in host survival underscores the central role that Rdi1 proteins and Rho-type GTPases play in important biological processes in microbial cells. Future investigations into the role of Rdi1 in maintaining pathogenicity in C. neoformans will provide valuable insights into the mechanisms of pathogen survival within infected hosts.