The STE12 Homolog Is Required for Haploid Filamentation But Largely Dispensable for Mating and Virulence in Cryptococcus neoformans

Corresponding author: Joseph Heitman, 322 CARL Bldg., Box 3546, Research Dr., Duke University Medical Center, Durham, NC 27710., heitm001{at}duke.edu (E-mail)

Communicating editor: A. P. MITCHELL

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Cryptococcus neoformans is a fungal pathogen that causes meningitis in immunocompromised hosts. The organism has a known sexual cycle, and strains of the MAT mating type are more virulent than isogenic MATa strains in mice, and they are more common in the environment and infected hosts. A C. neoformans homolog of the STE12 transcription factor that regulates mating, filamentation, and virulence in Saccharomyces cerevisiae and Candida albicans was identified previously, found to be encoded by a novel region of the MAT mating type locus, and shown to enhance filamentous growth when overexpressed. We have disrupted the C. neoformans STE12 gene in a pathogenic serotype A isolate. ste12 mutant strains exhibit a severe defect in filamentation and sporulation (haploid fruiting) in response to nitrogen starvation. In contrast, ste12 mutant strains have only modest mating defects and are fully virulent in two animal models compared to the STE12 wild-type strain. In genetic epistasis experiments, STE12 functions in a MAP kinase cascade to regulate fruiting, but not mating. Thus, the C. neoformans STE12 transcription factor homolog plays a specialized function in haploid fruiting, but it is dispensable or redundant for mating and virulence. The association of the MAT locus with virulence may involve additional genes, and other transcription factors that regulate mating and virulence remain to be identified.

MANY yeast and fungi are dimorphic and grow as either budding yeast cells or filamentous hyphal forms. For example, in response to nitrogen starvation, the yeast Saccharomyces cerevisiae differentiates to a filamentous pseudohyphal form (GIMENO et al. 1992 ). Similarly, the human pathogen Candida albicans, which typically grows in vitro as a yeast, exhibits both hyphal and pseudohyphal growth in response to multiple environmental signals, and mutants that do not filament are avirulent (LO et al. 1997 ). The plant fungal pathogen Ustilago maydis also has a prominent hyphal growth form, and the filamentous dikaryon is the infectious form of this organism (BANUETT 1998 ). Recent studies reveal that filamentous growth in these diverse microorganisms is regulated by two conserved signal transduction pathways: a MAP kinase cascade and a pathway involving a conserved G protein, adenylyl cyclase, cAMP, and cAMP-dependent protein kinase (LIU et al. 1993 ; GOLD et al. 1994 ; LORENZ and HEITMAN 1997 ; ALSPAUGH et al. 1998 ; DURRENBERGER et al. 1998 ; KRONSTAD et al. 1998 ; KRUGER et al. 1998 ; ROBERTSON and FINK 1998 ; PAN and HEITMAN 1999 ).

Cryptococcus neoformans is an opportunistic fungal pathogen with worldwide distribution. It is the leading cause of fungal meningitis, which is uniformly fatal if untreated (CASADEVALL and PERFECT 1998 ). C. neoformans is a basidiomycete with a known sexual cycle (KWON-CHUNG 1975 , KWON-CHUNG 1976 ). The organism is heterothallic and has a bipolar mating system with both MAT and MATa mating types. Mating requires two signals: pheromone and nutrient deprivation. Mating results in cell fusion, filamentation, and nuclear migration to form heterokaryotic cells linked by fused clamp connections. The tips of the hyphae differentiate to form rounded structures, called basidia, in which nuclear fusion and meiosis occur and chains of spores are then produced by budding from the surface of the basidia (reviewed in ALSPAUGH et al. 1999 ).

Interestingly, mating type of C. neoformans is linked both to prevalence in the environment and to virulence. Isolates of the MAT mating type are 40-fold more abundant in the environment than MATa strains, and MAT isolates are 30-fold more prevalent than MATa strains in clinical specimens from infected individuals (KWON-CHUNG and BENNETT 1978 ). Moreover, in a direct comparison of congenic strains, a MAT strain was more virulent than a MATa strain in a murine model of systemic cryptococcal infection (KWON-CHUNG et al. 1992A ).

In response to severe nitrogen starvation and desiccation, MAT strains of C. neoformans differentiate to a filamentous form known as monokaryotic (or haploid) fruiting (WICKES et al. 1996 ). This filamentous growth form involves the production of filaments, with nuclear migration along the monokaryotic hyphal cells, which are joined by unfused clamp connections. Similar to mating, terminal basidia form at the ends of the hyphae. However, in contrast to sporulation following mating, only short chains of spores are produced by monokaryotic fruiting, and all spores are of the MAT mating type. MATa strains do not undergo haploid filamentation.

Infection by C. neoformans is likely to be initiated by inhalation of the infectious particle, thought to be either small, dessicated yeast cells or spores, both of which are small enough to efficiently deposit into the alveoli of the lung (NEILSON et al. 1977 ). It has been proposed that one role of monokaryotic fruiting may be to generate infectious spores, and that this might explain the increased prevalence of MAT strains in the environment and in clinical samples (WICKES et al. 1996 ). However, this hypothesis fails to explain the association of MAT and virulence because MAT cells are inherently more virulent than MATa cells in a murine model in which the organism is delivered intravenously, bypassing entry through the lung (KWON-CHUNG et al. 1992A ). Alternative roles for haploid fruiting might be to generate spores to survive environmental stress under conditions when a mating partner is not readily available, or to disperse MAT spores to locate more distant MATa mating partners.

The link between MAT mating type and both environmental prevalence and virulence has focused research interest on the structure and encoded products of the MAT locus. An ~40-kb region of the MAT locus was identified previously and found to contain a gene encoding a small peptide pheromone homolog (MOORE and EDMAN 1993 ). A second distinct region of the MAT locus was found to encode a homolog of the STE12 transcription factor (WICKES et al. 1997 ), a homeobox DNA-binding protein (YUAN and FIELDS 1991 ) known to regulate mating, invasive growth, and filamentation in S. cerevisiae and filamentation and virulence of C. albicans (LIU et al. 1993 , LIU et al. 1994 ; LO et al. 1997 ). Overexpression of the C. neoformans STE12 homolog stimulates monokaryotic fruiting and production of the virulence factor melanin in a serotype D strain (WICKES et al. 1997 ).

Here, we have isolated the STE12 gene homolog from a pathogenic serotype A isolate of C. neoformans and disrupted the gene by homologous recombination. ste12 mutant strains exhibited only modest defects in mating. On the other hand, ste12 mutant strains were completely defective for monokaryotic fruiting, and reintroduction of the wild-type STE12 gene complemented these mutant phenotypes. Overexpression of STE12 suppressed the haploid fruiting defect but not the mating defect of a mutant lacking the Gß protein GPB1. Lastly, ste12 mutant strains were fully virulent in the rabbit meningitis model and in the murine systemic infection model at several different inoculum sizes. Our findings reveal that the C. neoformans STE12 homolog is specialized to regulate filamentous growth during monokaryotic fruiting, but does not play a primary role in either mating or virulence during systemic or central nervous system infections. We propose that a conserved MAP kinase cascade regulates monokaryotic fruiting in C. neoformans, and that other transcription factors that regulate mating and virulence, as well as additional genes within the MAT locus linked to mating and virulence, remain to be identified.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains, media, and growth conditions:
The C. neoformans serotype A MAT strain H99 and its derivative, the ade2 mutant M049, were used in this study (PERFECT et al. 1993 ). All mating assays of these strains were performed with the serotype D MATa strain JEC20 or congenic auxotrophic derivatives (MOORE and EDMAN 1993 ). The serotype A strain 450 from Eric Jacobson (DUMC 133.96) was generously provided by Wiley Schell through the Duke University Mycology Research Unit Fungal Collection, and the congenic serotype D MAT strain JEC21 and MATa strain JEC20 were provided by Jeff Edman. ura5 auxotrophic derivatives of the wild-type strain H99, the ste12 mutant, and the ste12 + STE12 reconstituted strains were isolated by selection for 5-FOA resistance, as described previously (KWON-CHUNG et al. 1992B ).

Standard yeast media and genetic manipulations were as described (SHERMAN 1991 ). Limiting nitrogen medium contains 0.17% yeast nitrogen base without amino acids or ammonium sulfate, 2% dextrose, 2% Bacto-agar and 50 µM ammonium (SLAD; GIMENO et al. 1992 ). All other media were prepared as described (WICKES et al. 1996 ; ALSPAUGH et al. 1997 ).

Isolation of the C. neoformans serotype A STE12 gene:
Two primers, 5'-TCTGAGGAATCTCAAACCGG (1557) and 5'-TCGTCGTTGGTTCCAACCC (1558), from the C. neoformans serotype D STE12 conserved homeodomain region (WICKES et al. 1997 ) were used to PCR amplify a 420-bp fragment of the STE12 gene from genomic DNA. The resulting 420-bp fragment was used as a probe for Southern blot analysis of genomic DNA isolated from the serotype A strain H99. A 10-kb HindIII fragment hybridized to this 420-bp STE12 probe, and a HindIII 8- to 12-kb size-selected genomic library was constructed in the plasmid pBluescript. Clones containing the desired HindIII fragment were identified by colony hybridization. Nylon membranes (Zeta-probe blotting membranes; Bio-Rad, Richmond, CA) were used for colony lifts, and subsequent hybridization and washes were as described (SAMBROOK et al. 1989 ). Sequence analysis revealed that this HindIII fragment was missing the 3' end of the STE12 gene. A 6- to 8-kb size-selected KpnI library was subsequently constructed in pBluescript to retrieve the 3' end of the STE12 gene by colony hybridization. Overlapping regions of both the 10-kb HindIII fragment and the 7-kb KpnI fragments were sequenced across the STE12 gene at the Duke University DNA sequencing facility using synthetic primers.

Preparation of probes:
Probes used for library screening, Southern blots, and Northern blots were prepared from restriction fragments or PCR-amplified fragments, as described in the text. Probes were labeled using the DIG nonradioactive nucleic acid labeling system (Boehringer Mannheim, Indianapolis).

Southern hybridization:
Genomic DNA (10 µg) was digested with various restriction enzymes and electrophoresed in 0.8% agarose DNA gels. The DNA was then transferred to nylon membranes (Boehringer Mannheim) using standard protocols (SAMBROOK et al. 1989 ). Hybridization, washing, and detection of hybridized bands were performed according to the instructions from the manufacturer (Boehringer Mannheim).

3'-RACE analysis:
The C terminus of the STE12 protein was confirmed, and the polyadenylation site of the STE12 transcript established by RACE was performed using the 3'-RACE system kit from GIBCO-BRL (Gaithersburg, MD). First-strand cDNA synthesis was performed using 10 µg of total RNA from C. neoformans serotype A strain H99 grown in YPD liquid medium at 30°. The nested PCR was performed using an internal primer, 5'-GGAGAGATACGAGTGAAC (2520), the universal amplification primer (UAP) 5'-CUACUACUACUAGGCCACGCGTCGACTAGTAC, as well as a PCR program of 3 min at 94°, 30 cycles of 45 sec at 94°, 25 sec at 55°, 3 min at 72°, and a final 5-min extension step at 72°. The resulting 700-bp PCR RACE product was ligated into the TA vector (Invitrogen, Carlsbad, CA) and sequenced.

Disruption of the C. neoformans serotype A STE12 gene:
The full-length, 2.9-kb serotype D genomic ADE2 gene was blunt-ended and ligated into the blunted XbaI site of a 3.9-kb SacI-BamHI STE12 gene fragment cloned in pBluescript (see Figure 1). The 3.9-kb SacI-BamHI fragment of STE12 had been cloned by a three-way ligation of a 1.5-kb SacI-HindIII and a 2.4-kb HindIII-BamHI fragment into pBluescript vector cleaved with SacI and BamHI. Note that the ADE2 gene was inserted into the STE12 gene such that the two genes were transcribed in opposite directions. A total of 5 µg of the ste12::ADE2 disruption plasmid was precipitated onto 0.6-µg gold microcarrier beads (Bio-Rad) and biolistically transformed into the ade2 mutant strain M049 as described (TOFFALETTI et al. 1993 ). Transformants were selected on regeneration medium (synthetic medium with 1 M sorbitol and lacking adenine).

View larger version (22K): In this window In a new window Download PPT slide   Figure 1. Identification and disruption of the serotype A STE12 gene. (A) A restriction map of the the locus containing the STE12 gene and the targeting disruption construct is depicted. The 7700-bp region containing the STE12 gene was sequenced, and the GenBank accession number is AF121343. Restriction mapping was done in regions not sequenced. Sites for a few restriction enzymes are listed. K, KpnI; B, BamHI; H, HindIII; E, EcoRI; R, EcoRV, X, XbaI; S, SacI. The arrow above the locus indicates the direction of the STE12 gene. Below the locus, the line labeled "probe" indicates the EcoRI-BamHI fragment that was used in the Southern and Northern hybridizations to confirm the gene disruption and reintroduction shown in B and C. For the ste12::ADE2 disruption allele, a 2.9-kb restriction fragment containing the C. neoformans serotype D ADE2 genomic locus was ligated into the XbaI site of the SacI-BamHI fragment containing the STE12 gene. (B) Southern blot of genomic DNA from the STE12 wild-type strain, the ste12 mutant strain, and the ste12 + STE12 reconstituted strain (lanes 1–3). Genomic DNA was cleaved with BamHI and electrophoresed in a 0.8% agarose gel. The probe used in this Southern blot is depicted in A. The wild-type STE12 gene corresponds to a 4-kb BamHI fragment (lane 1), the ste12::ADE2 disruption allele corresponds to a 7-kb BamHI fragment (lane 2), and the ste12 + STE12 reconstituted strain contains both the wild-type 4-kb and the 7-kb ste12::ADE2 fragments (lane 3). (C) Total RNA was isolated from the wild-type STE12 strain (lane 1), the ste12 mutant strain (lane 2), and the ste12 + STE12 reconstituted strain (lane 3) grown in low-nitrogen medium (SLAD; shown here) or YPD medium (data not shown). RNA was electrophoresed and analyzed by Northern hybridization with the STE12 gene as probe. Before RNA transfer, the gel was stained with ethidium bromide to determine the relative level of ribosomal RNA in each sample as a loading control.

Genomic DNA from 24 ADE+ transformants was prepared using the smash-and-grab protocol (PITKIN et al. 1996 ). PCR was performed using the following STE12-specific primers flanking the site of insertion of the ADE2 gene: 5'-CTGAGGAATCTCAAACCGGG (1671) and 5'-GCTTCGGATCATACCTGACG (2189). A PCR program of 5 min at 94°, 30 cycles of 30 sec at 94°, 30 sec at 55°, 3.5 min at 72° and a final 7 min at 72° was used to determine whether the STE12 gene had been disrupted. PCR amplification of the ste12::ADE2 disruption allele yielded a diagnostic 3400-bp fragment, whereas amplification of the wild-type STE12 gene yielded a 550-bp fragment. Transformants in which the 550-bp wild-type product was missing and the 3400-bp fragment was present were identified and confirmed to contain ste12::ADE2 gene replacements by Southern and Northern blot analyses.

Reintroduction of the wild-type serotype A STE12 gene:
A 6.3-kb EcoRV-BamHI fragment containing the wild-type STE12 gene was blunt-end ligated (as an EcoRV fragment where the second EcoRV site was adjacent to the BamHI site in the vector polylinker) into a blunt-ended XbaI site in the plasmid pTelHyg (COX et al. 1996 ). The resulting plasmid was cleaved with NotI to release a linear fragment and remove the telomeric DNA sequences present in pTelHyg. The fragment containing the STE12 gene and the linked hygromycin-B-resistance gene was purified and biolistically transformed into the ste12 deletion mutant strain as described (COX et al. 1996 ). Stable transformants were selected on YPD medium containing 200 µg/ml hygromycin B (Sigma, St. Louis). Transformants that contained the wild-type STE12 gene were identified by restoration of wild-type mating on SLAD medium with the MATa mating type tester strain JEC20. Genomic DNA isolated from four hygromycin-B-resistant transformants was analyzed by PCR with primers 1671 and 2189 described above, revealing that both the ste12::ADE2 disruption allele and the wild-type STE12 locus were now present in three of the four transformants tested. The presence of the wild-type STE12 gene was confirmed by Southern and Northern blot analyses.

Northern hybridization:
Total RNA from the C. neoformans wild-type STE12 strain H99, the ste12 mutant strain, and the ste12 + STE12 reconstituted strain grown in liquid YPD or liquid SLAD media was isolated using the Qiagen RNeasy Midi kit. Total RNA (15 µg) was electrophoresed in a 1% agarose formaldehyde gel and then transferred overnight by capillary action to a nylon membrane (Boehringer Mannheim) using standard protocols (SAMBROOK et al. 1989 ). Northern hybridization, washing, and detection of hybridized bands were performed according to the instructions from Boehringer Mannheim, and the probe used was the same as that described for the Southern blot analysis.

Mating and haploid filamentation assays:
Strains of opposite mating types were grown on YPD medium for 48 hr at 30°, cocultured on V8 or SLAD mating medium, and incubated at room temperature. Matings were scored microscopically for filamentation. For haploid filamentation assays, the Ras1-Q67L mutant protein (J. A. ALSPAUGH, L. M. CAVALLO, J. R. PERFECT and J. HEITMAN, unpublished results) was expressed by introducing linear DNA fragments containing the mutant RAS1 gene linked to either the hygromycin-B-resistance gene or the URA5 gene as selectable markers by biolistic transformation and selection for Ura⁺ or hygromycin-resistant colonies, respectively. Isolates containing the Ras1-Q67L mutant allele were identified by PCR amplification of genomic DNA with primers flanking the RAS1 gene and subsequent cleavage with XbaI to identify the restriction site introduced at the site of the Q67L mutation. Haploid filamentation was assayed by incubating spotted suspensions of cells on filament agar at 24° for up to 4 wk.

Quantitative mating assay:
As a measure of relative mating efficiency, the rate of recombinant basidiospore appearance was quantified within identical mating reactions of various serotype A strains. The STE12 wild type, three independent ste12 mutant strains, and two independent ste12 + STE12 reconstituted strains were grown in liquid YPD medium for 18 hr, pelleted, and resuspended in water at 10⁸ cells/ml. A similar cell suspension (10⁸ cells/ml) was made for the serotype D MATa strain JEC53 (MATa ura5 lys1). Mating mixtures for each of the serotype A strains (MAT URA5 LYS1) were made by mixing 100 µl of each cell suspension with 100 µl of the JEC53 cell suspension. Subsequently, 5 µl of each mating mixture were plated as drops on 10 identical V8 medium plates supplemented with uracil and lysine. These mating plates were incubated at 25°.

Every 48 hr, the mating reactions on one plate were excised as agar plugs, placed in 2 ml sterile water, and vortexed for 2 min. A total of 300 µl of this cell/spore suspension was plated on 5-fluoroorotic acid medium (SHERMAN 1991 ) lacking lysine to select for ura5 LYS1 recombinant progeny. Colonies were counted at 72 hr. Control incubations of parental strains alone failed to yield any recombinant isolates on this medium, indicating that all ura5 LYS1 isolates were attributable to mating.

Photomicroscopy:
All single-colony photographs were taken directly from Petri plates using a Nikon Eclipse E400 microscope with a x10 or x50 primary objective and a x2.5 trinocular camera adaptor for a final magnification of x25 or x125, as indicated.

Capsule induction:
All strains assayed for capsule production were incubated on YPD medium at 30° for 48 hr and subsequently inoculated into 10 ml low-iron medium with 56 µM ethylenediamine-di(o-hydroxy-phenylacetic acid) (VARTIVARIAN et al. 1993 ). After 72 hr incubation with shaking (250 rpm) at 30°, the cells were pelleted and resuspended in water at 2.5 x 10⁸ cells/ml (confirmed by duplicate hemacytometer counts). The cell suspensions were treated with 10% formalin and added to heparinized microhematocrit capillary tubes (Fisher, #02-668-66), the ends of which were sealed with clay. Capillary tubes were spun for 10 min in a Microhematocrit centrifuge, model MB (International Equipment Co.). Packed cell volume was measured analogously to a hematocrit as the length of the packed cell phase divided by the length within the capillary tube of the total suspension (GRANGER et al. 1985 ).

Melanin production:
Whole-cell laccase activity was assayed as described previously, with minor modifications (WILLIAMSON 1994 ). Cells were inoculated from fresh cultures grown on YPD agar into liquid minimal asparagine medium (MM) with 0.1% dextrose and incubated at 30° for 16 hr in a shaking incubator (250 rpm). Cells were pelleted, washed once with MM without dextrose, and resuspended in the same medium. Cells were incubated for an additional 5 hr at 30°, pelleted, washed once with water, and resuspended in 5 ml PBS at 25°. The cell suspensions were normalized to 10⁸ cells/ml in PBS at 25°. One-milliliter aliquots were removed and placed on ice to serve as T₀ controls. To the original cell suspensions, 1 mM epinephrine was added as a substrate for melanin production. The samples were incubated at 25° for 30 min, and the OD₄₇₅ of the supernatant was determined on a Beckman DU640 spectrophotometer. One unit of laccase activity was defined as 0.001 OD₄₇₅. Data points represent the mean of triplicate samples ± standard error.

Urease and phospholipase B production:
Strains were tested for urease activity using Christensen's urea agar and broth (Becton Dickinson, Cockeysville, MD). The tubes containing the suspensions of the transformants in Christensen's medium were incubated with agitation overnight at 30°. Christensen's medium contains 300 mM urea and phenol red as a pH indicator. In Christensen's medium, urease activity converts urea to ammonia and increases the pH, which is reflected by a color change of the medium from yellow to bright pink. C. albicans has no urease activity and was used as a negative control.

Phospholipase B activity was quantitated on egg yolk agar, as described (CHEN et al. 1997 ). Phospholipase activity was quantitated by measuring the zones of precipitation surrounding colonies grown on egg yolk agar. The zones of precipitation were controlled for colony size by dividing the diameter of the precipitation zone by the diameter of the colony.

Rabbit virulence model:
Strains were incubated on YPD medium for 48 hr at 37° and resuspended in phosphate-buffered saline [PBS, pH 7.4 (Sigma)] at 3 x 10⁸ cells/ml. Twelve New Zealand white male rabbits weighing 2–3 kg (four for each strain) were administered 2.5 mg/kg cortisone acetate intramuscularly before intrathecal injection of a 0.3-ml suspension of cryptococcal cells. Daily intramuscular cortisone injections were continued throughout the course of the experiment. Cerebrospinal fluid (CSF) was obtained by cisternal punctures at days 4, 7, 11, and 14 after infection. The CSF was immediately serially diluted, and each dilution was plated on YPD medium and incubated at 30° for 48 hr for quantitative analysis. Rabbits were sedated with 10 mg xylazine and 100 mg ketamine given intramuscularly before all cisternal inoculations or withdrawals (PERFECT et al. 1993 ). Student's t-test was used in evaluating the colony counts from the rabbit experiments.

Murine virulence model:
C. neoformans wild-type STE12, ste12 mutant, and ste12 + STE12 reconstituted strains were grown in liquid YPD media at 30° for 48 hr and resuspended in PBS, pH 7.4 (Sigma) at 1 x 10⁵, 1 x 10⁶, 1 x 10⁷, or 1 x 10⁸ cells/ml. A total of 0.1 ml of cell suspension (for a total of 10⁴, 10⁵, 10⁶, or 10⁷ cells) was injected into the lateral tail vein of each mouse (10 or 15 mice were used for the wild-type STE12 strain, and 10 were used for each of the ste12 mutant strains and the ste12 + STE12 reconstituted strain). Mice were 4- to 6-wk-old female BALB/c mice (Charles River Labs, Raleigh, NC). The number of mice were recorded twice daily. Mice that appeared moribund or in pain were killed using CO₂ inhalation. Survival data from the mouse experiments were analyzed by a Wilcoxon test for censored data using True Epistat (Standard Version, Epistat Services).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Isolation of the serotype A STE12 gene:
A C. neoformans homolog of the STE12 transcription factor was recently identified and found to stimulate haploid filamentation and melanin production when overexpressed (WICKES et al. 1997 ). These findings, and the fact that the STE12 gene is only present in MAT cells, suggested that STE12 might be responsible for the known association of the MAT mating type with virulence. However, the STE12 gene was not disrupted in these studies, and it has not been established if STE12 is required for either mating or virulence of this organism.

To determine the functions of STE12, we have capitalized on the ability to disrupt genes by homologous recombination at high efficiency in serotype A strains of C. neoformans (LODGE et al. 1994 ; ALSPAUGH et al. 1997 ; ODOM et al. 1997 ; DEL POETA et al. 1999 ). Because the STE12 gene was originally isolated from a serotype D strain, and genes derived from serotype A and D are sufficiently sequence divergent (~5%) to prevent efficient homologous recombination, we isolated the STE12 gene from the serotype A strain H99. To this end, the serotype D STE12 gene was used to probe a Southern blot of serotype A genomic DNA. A 10-kb HindIII fragment spanning part of the serotype A STE12 gene was isolated by screening a size-selected genomic library by colony hybridization. Sequence analysis revealed that this clone was missing the 3' region of the STE12 locus, and a second overlapping 7-kb KpnI fragment containing the 3' end of the STE12 locus was identified by genomic Southern and cloned. A restriction map for the resulting ~15-kb region spanning the serotype A STE12 locus is depicted in Figure 1A.

This region was further analyzed by genomic Southerns, dot blot hybridizations, and sequencing to reveal the STE12 gene and mating-type-specific and nonspecific regions. The serotype A STE12 gene is located in the central portion of the 15-kb region (Figure 1A). Sequence analysis of the genomic locus, cDNA, and 3'-RACE products revealed that the serotype A STE12 gene and product share extensive homology with the serotype D STE12 gene from C. neoformans and the STE12 genes and proteins of other yeast and fungi. The predicted STE12 protein product consists of 841 amino acids and contains an N-terminal homeobox DNA-binding domain with marked identity to those of the STE12 homologs from S. cerevisiae, C. albicans, and Kluvyeromyces lactis (Figure 2). In addition, the C-terminal domain of the STE12 protein shares identity with a variety of different zinc finger proteins. The serotype A STE12 locus is MAT specific: probes from this region hybridize to genomic DNA isolated from the serotype A MAT strain H99 and from the serotype D MAT strain JEC21, but not from the serotype D MATa strain JEC20 (data not shown). The STE12 locus is flanked on the right by an ~1-kb BamHI fragment that is unique in the serotype A strain H99 and additional sequences that are not mating type specific (Figure 1A; data not shown). On the left, the STE12 locus is flanked by additional sequences that hybridize poorly to genomic DNA from MAT or MATa serotype D strains; thus, additional sequences and the left border of this portion of the MAT locus likely remain to be identified (Figure 1A; data not shown).

View larger version (63K): In this window In a new window Download PPT slide   Figure 2. The C. neoformans STE12 protein contains a homeodomain DNA-binding region with marked identity to yeast STE12 homologs. The N-terminal region of the C. neoformans (C.n.) serotype A STE12 protein encompassing the homeodomain is aligned with the corresponding sequences of the S. cerevisiae (S.c.) STE12 protein, the C. albicans (C.a.) STE12 homolog CPH1, and the K. lactis (K.l.) STE12 protein. Identical amino acids are indicated in bold, similar amino acids are shaded, and nonidentical amino acids are in white. Note that a second distinct region of identity between the yeast STE12 proteins (DFPLDY) present at the C-terminal end of the aligned region is not conserved in the C. neoformans STE12 homolog.

Disruption and reintroduction of the STE12 gene:
The STE12 gene was disrupted by homologous recombination. The ADE2 selectable marker was cloned into an XbaI site in the STE12 open reading frame (Figure 1A). The resulting ste12::ADE2 disruption allele was introduced by biolistic transformation into the ade2 strain M049 derived from the pathogenic serotype A strain H99. Ade⁺ transformants were selected on synthetic medium lacking adenine and containing 1 M sorbitol. Approximately 400 transformants were obtained and analyzed initially for a presumptive sterile phenotype on V8 medium in crosses to the MATa mating type tester strain JEC20. No isolates with a pronounced sterile phenotype were obtained. Next, genomic DNA was isolated from 24 random transformants and analyzed for the presence of the ste12::ADE2 allele by PCR analysis with STE12-specific primers flanking the site of the ADE2 gene insertion and by Southern analysis. Out of 24 isolates, 1 lacked the wild-type STE12 locus and contained only the ste12::ADE2 disruption allele (Figure 1B). We subsequently identified 6 additional ste12::ADE2 isolates from 64 transformants, for an overall frequency of 8% homologous recombination (7 out of 88 Ade⁺ transformants). Southern blot hybrization confirmed that the wild-type STE12 locus had been replaced by the ste12::ADE2 disruption allele by homologous recombination (Figure 1B).

The wild-type STE12 gene was reintroduced into the ste12::ADE2 mutant strain by transformation. The wild-type STE12 gene on a 6.3-kb EcoRV-BamHI fragment was cloned adjacent to the hygromycin-B-resistance gene. The ste12::ADE2 strain was biolistically transformed with the STE12 gene and hygromycin-B-resistant isolates selected. PCR and Southern analysis confirmed that the wild-type STE12 gene had been introduced ectopically into the ste12 mutant strain (Figure 1B), and we designated the resulting reconstituted strain ste12 + STE12.

To examine expression of the STE12 gene, Northern analysis was performed on total RNA isolated from the isogenic STE12 wild-type parental, the ste12 mutant, and the ste12 + STE12 reconstituted strains (Figure 1C). In contrast to the STE12 wild-type strain, in which a 2.5-kb message was readily detectable, no message was detected in the ste12 mutant strain. This observation supports the conclusion that STE12 has been functionally deleted in the ste12::ADE2 mutant strain. After reintroduction of the wild-type STE12 gene by transformation, STE12 expression was restored in the ste12 + STE12 strain. By Northern blot analysis, the STE12 gene was found to be expressed at similar levels in cells grown in either YPD or SLAD low-nitrogen liquid medium (data not shown).

ste12 mutant strains exhibit modest mating defects:
We next analyzed the possible role of the STE12 homolog in mating in C. neoformans. In this organism, coculture of MAT and MATa strains on several different nutrient-limited media (V8 mating medium, SLAD, filament agar) results in cell fusion, filamentation, nuclear migration, basidia formation, nuclear fusion, meiosis, and sporulation. Thus, filamentation and sporulation can serve as a morphological assay of mating in C. neoformans.

When either wild-type or ste12 mutant MAT cells were mixed with a MATa mating type tester strain (JEC20) on V8 mating medium, abundant filaments and basidiospores were produced within 7 days of incubation at 24° (Figure 3). By comparison of mating mixtures at early time points after coculture, it was apparent that ste12 mutant strains have a modest mating defect compared to the isogenic STE12 parental strain (data not shown). This mating defect was manifested by the production of fewer filaments and basidia at early time points after coculture. However, within a few days of incubation, the overall extent of mating was similar (Figure 3). This mating defect was complemented by reintroduction of the wild-type STE12 gene; in fact, mating was enhanced in the ste12 + STE12 reconstituted strain compared to the STE12 wild-type parental strain (Figure 3). In a quantitative mating assay that monitors the production of prototrophic recombinant strains following coincubation of auxotrophic parental strains of opposite mating type (see MATERIALS AND METHODS), the ste12 mutation reduced the level of mating by ~10-fold compared to the isogenic STE12 wild-type and the ste12 + STE12 reconstituted strains. These findings indicate that STE12 plays a modest but not absolute role in mating.

View larger version (87K): In this window In a new window Download PPT slide   Figure 3. The ste12 mutation confers modest defects in mating in C. neoformans. The isogenic STE12 wild-type, ste12 mutant, and ste12 + STE12 reconstituted MAT strains were cocultured in cell patches with the MATa strain JEC20 on V8 mating medium or nitrogen-limiting mating media (SLAD) for 7 days at 25°. The edges of the mating mixtures were photographed (x25).

Mating also occurs on defined, nutrient-limiting medium with reduced levels of ammonium sulfate, either SLAD medium (50 µM ammonium sulfate) or filament agar (no added nitrogen source). The ste12 mutant strain exhibited a more pronounced mating defect on these synthetic mating media than on V8 agar, but it still eventually produced filaments and basidiospores (Figure 3). We note that the mating defect conferred by the ste12 mutation was more subtle than the pronounced mating defects conferred by mutations in several other components of the mating pathway that we have analyzed, including GPA1, GPB1, STE20, SCH9, CPK1, and CNA1 (ALSPAUGH et al. 1997 ; P. WANG and J. HEITMAN, unpublished results; R. C. DAVIDSON and J. HEITMAN, unpublished results; M. C. CRUZ and J. HEITMAN, unpublished results).

STE12 is required for haploid fruiting:
Nitrogen starvation and desiccation induces some C. neoformans MAT strains to undergo haploid filamentation and sporulation (WICKES et al. 1996 ). Under these same conditions, MATa strains do not differentiate and, thus, genes within or linked to the MAT locus are required for haploid filamentation. We therefore tested whether the STE12 gene is required for haploid fruiting.

The MAT serotype A strain H99 has not been observed to undergo haploid filamentation when incubated on filament agar at room temperature for up to several months (our unpublished observations). In comparison, the MAT serotype A strain 450 (also known as NIH38 or DUMC 133.96, provided by Eric Jacobsen) undergoes robust monokaryotic fruiting. This atypical strain forms abundant filaments with unfused clamp connections on V8, SLAD, or filament agar within 48 hr, and these filaments produce abundant basidia and basidiospores. The MAT serotype D strain JEC21 undergoes haploid fruiting with filamentation and basidiospore production in 1–4 wk on filament agar. Hence, the serotype A MAT strain H99 appears to be lacking elements required for haploid filamentation, whereas haploid fruiting is derepressed in the serotype A MAT strain 450. Similar strain differences have been noted in the ability of certain laboratory strains of S. cerevisiae to undergo pseudohyphal differentiation (GIMENO et al. 1992 ).

Recently, we have found that expression of a dominant active allele of the C. neoformans homolog of the small G protein Ras1 stimulates haploid filamentation in strain H99 and in the Ade⁺ reconstituted versions of the ade2 mutant strain M049 (J. A. ALSPAUGH, L. M. CAVALLO, J. R. PERFECT and J. HEITMAN, unpublished results). Thus, isolates of strain H99 expressing the dominant activated Ras1-Q67L mutant allele undergo haploid fruiting to produce filaments and basidiospores in 3–14 days on filament agar (Figure 4). Strikingly, when the Ras1-Q67L mutant allele was introduced into several independent isogenic ste12 mutant strains, no haploid filamentation was ever observed in multiple independent transformants incubated for up to 1 mo (Figure 4). Haploid filamentation in response to the dominant active Ras1-Q67L mutant allele was restored in the ste12 + STE12 reconstituted strain (Figure 4), indicating that the defect in monokaryotic fruiting is attributable to the ste12 mutation.

View larger version (55K): In this window In a new window Download PPT slide   Figure 4. STE12 is required for haploid fruiting in C. neoformans. The isogenic STE12 wild-type (H99), the ste12 mutant, and the ste12 + STE12 reconstituted strains were transformed with the gene encoding the dominant active Ras1-Q67L mutant protein. Transformants were grown on dry, nitrogen-limiting filament agar medium to induce haploid filamentation for 14 days at 24°, and the edges of cell patches were photographed (x125).

Genetic epistasis experiments provide evidence that STE12 functions in a MAP kinase cascade and regulates haploid fruiting, but not mating. We have recently found that mutant strains lacking the G-protein ß subunit GPB1 are sterile and are defective in monokaryotic fruiting in response to nitrogen starvation (P. WANG, J. R. PERFECT and J. HEITMAN, unpublished results). Here, we performed epistasis analysis to test if STE12 functions downstream of GPB1 in a signaling cascade regulating mating and morphogenesis. When a GAL7-STE12 fusion gene was introduced into the gpb1::ADE2 mutant strain expressing the Ras1-Q67L mutant protein, haploid fruiting was restored when STE12 expression was induced on galactose filament agar, but no haploid fruiting was observed when STE12 expression was repressed on glucose filament agar (Figure 5). Thus, overexpression of STE12 suppresses the haploid fruiting defect of mutant cells lacking GPB1. In contrast, the GAL7-STE12 fusion gene failed to restore mating of gpb1 mutant cells coincubated with the MATa mating partner JEC20 (data not shown). Taken together, these observations provide additional evidence that STE12 is a component of a MAP kinase cascade and is specialized to regulate haploid fruiting, but is either not required for mating or is redundant with other factors that regulate mating.

View larger version (80K): In this window In a new window Download PPT slide   Figure 5. STE12 restores haploid fruiting in gpb1 mutant cells. The G-protein ß subunit GPB1 is required for mating and Ras1-Q67L-induced haploid filamentation in C. neoformans (P. WANG, J. R. PERFECT and J. HEITMAN, unpublished results). For haploid filamentation, the gpb1 RAS1Q67L strain and the gpb1 RAS1Q67L strain transformed with plasmid pCGS-1 expressing the GAL7-STE12 fusion gene were grown on glucose filament agar and galactose filament agar. Plates were photographed (x25) after 3 days of incubation at 24°.

Effects of the ste12 mutation on virulence factor production:
We next tested whether the ste12 mutation has any effect on known virulence factors of C. neoformans. The virulence traits tested included prototrophy, growth at elevated temperature, and the production of melanin, urease, phospholipase B, and polysaccharide capsule.

The isogenic STE12, ste12 mutant, and ste12 + STE12 reconstituted strains all grow on minimal medium (YNB) and, thus, have no novel auxotrophic phenotypes (data not shown). All three strains were equally viable at 37 or 39° (data not shown) and produced melanin to similar extents on either minimal DOPA agar or niger birdseed agar (Figure 6C). A quantitative spectrophotometric assay of whole-cell phenoloxidase activity confirmed that all three strains produce equivalent amounts of melanin (Figure 6D). The ste12 mutant strain also produced two potential virulence factors, urease and phospholipase B, at similar levels to the STE12 wild-type and the ste12 + STE12 strains (data not shown; see MATERIALS AND METHODS).

View larger version (50K): In this window In a new window Download PPT slide   Figure 6. The role of STE12 in capsule and melanin production. (A) Cerebrospinal fluid was isolated at 10 days after infection from rabbits inoculated with the STE12 wild-type, the ste12 mutant, or the ste12 + STE12 reconstituted strain and was analyzed by India ink counterstain to examine the size of the polysaccharide capsule surrounding the yeast cells. Representative cells were photographed (x200). (B) Capsule size was quantified by measuring the packed cell volume of cell suspensions (normalized to 108 cells/ml) for each of the three strains grown under capsule-inducing, low-iron conditions in liquid medium for 72 hr. Each data point represents the mean of triplicate samples, and error bars indicate standard error of the mean. (C) Melanin production by the STE12 wild-type, ste12 mutant, and ste12 + STE12 reconstituted strains was assessed by plating cell suspensions onto Niger seed agar medium and incubating at 30° for 5 days; the cell patches were photographed (x5). (D) To quantify melanin production by the isogenic STE12 wild-type, ste12 mutant, and ste12 + STE12 strains, whole-cell laccase enzyme activity was determined by incubating glucose-starved cells with 1 mM epinephrine at 25° for 25 min. The optical density of the resulting cell supernatants was measured at 475 nM. Each data point represents the mean of triplicate samples, and error bars indicate standard error of the mean.

We did observe that ste12 mutant cells consistently produced less capsule compared to the isogenic STE12 wild-type and the ste12 + STE12 reconstituted strains, as revealed by India ink stains of CSF fluid from infected rabbits (Figure 6A). This effect was quantified by photomicroscopy and by measuring the packed cell volume (cryptocrit) of cells grown under iron-limiting conditions in vitro. The average yeast capsule size was reduced by ~50% under capsule-inducing conditions in several independent ste12 mutant strains compared to the STE12 wild-type and the ste12 + STE12 reconstituted strains (Figure 6A). By the cryptocrit assay, which measures packed cell volume and, therefore, detects both cell and capsule sizes, the capsule size of the ste12 mutant strain was also found to be reduced compared to the STE12 wild-type and the ste12 + STE12 reconstituted strains (Figure 6B). This observation suggests that STE12 contributes to regulate capsule production.

STE12 is dispensable for virulence in animal models:
Because the MAT-mating-type locus is linked to virulence of C. neoformans, and the ste12 mutation caused a modest defect in capsule induction in vivo and in vitro, we tested the contribution of the STE12 gene to virulence in two animal models of cryptococcal infection. First, we employed the rabbit meningitis model (PERFECT et al. 1980 ) that had been previously used to study the virulence of several different mutant strains (LODGE et al. 1994 ; ALSPAUGH et al. 1997 ; ODOM et al. 1997 ). Rabbits were immunosuppressed with corticosteroids and inoculated intrathecally with the STE12 wild-type strain, the ste12 mutant strain, and the ste12 + STE12 reconstituted strain. Cerebrospinal fluid was recovered on days 4, 7, 11, and 14 following infection, and the number of surviving yeast colonies were determined by quantitative plating. Importantly, quantitative yeast counts are correlated with morbidity in this animal model system (PERFECT et al. 1980 ). We found no difference in quantitative yeast counts obtained from animals infected with the STE12 wild-type, the ste12 mutant, or the ste12 + STE12 reconstituted strains over the course of a 2-wk infection (Figure 7A). Yeasts recovered from animals infected with the ste12 mutant strain exhibited the modest mating defect of the ste12 mutant strain and, thus, the ste12 mutation had not reverted or been suppressed during the course of the infection. These observations suggested that the STE12 homolog is not required for virulence of this C. neoformans serotype A strain in the vertebrate central nervous system under these conditions.

View larger version (18K): In this window In a new window Download PPT slide   Figure 7. STE12 is dispensable for virulence of C. neoformans in animal models. (A) Rabbits were immunosuppressed with corticosteroids and inoculated intrathecally with 108 cells of the isogenic STE12 wild-type, the ste12 mutant, and ste12 + STE12 reconstituted strains. CSF was removed on days 4, 7, 11, and 14 after inoculation, and the number of surviving organisms was determined by serial dilution and plating on YPD medium. Each data point represents the mean of all cultures for each strain, and the strandard error of the mean is indicated. (B) A total of 107 cells each of the wild-type STE12, the ste12 mutant, and the ste12 + STE12 reconstituted strains were injected into the tail vein of mice. A total of 15 mice were infected with the wild-type strain, and groups of 10 mice were each infected with the mutant and reconstituted strains. The percentages of mice surviving from each group were plotted over the course of this 50-day experiment.

Next, we assessed the contribution of the STE12 homolog to virulence in the murine systemic infection model. Groups of 10–15 mice each were infected with 10⁷ cells of the STE12 wild-type, the ste12 mutant, and the ste12 + STE12 reconstituted strains by tail vein injection. Moribund mice were sacrificed in this model of systemic cryptococcal infection; in general, the majority of infected mice exhibited hydrocephalus as a consequence of cryptococcal central nervous system (CNS) infection. As shown in Figure 7B, we again observed no difference between the virulence of the wild-type STE12, the ste12 mutant, and the ste12 + STE12 reconstituted strains, as measured by cumulative survival.

To address whether a virulence defect might be apparent with a smaller initial inoculum, groups of 10 animals each were infected with the wild-type, ste12 mutant, and ste12 + STE12 reconstituted strains at inoculum sizes of 10⁶, 10⁵, or 10⁴ cells per infected animal (Figure 8). In these studies, the time course of lethal infections was somewhat prolonged, as expected, but the virulence of the three strains was comparable at equivalent inoculum sizes, with no apparent virulence defect of the ste12 mutant strain (Figure 8).

View larger version (25K): In this window In a new window Download PPT slide   Figure 8. The ste12 mutant is as virulent as the wild type at different inoculum sizes. The wild-type STE12, the ste12 mutant, and the ste12 + STE12 reconstituted strains were injected into the tail vein of mice at four different inocula: 107 cells (A), 106 cells (B), 105 cells (C), and 104 cells (D). A total of 10 mice were infected with each strain. A second independent ste12 mutant strain was included in the experiment at an inoculum of 106 cells per infected animal (B, inverted open triangles). The percentages of mice surviving from each group were plotted over the course of this 65-day experiment. The wild-type strain is depicted with solid circles, the ste12 mutant strain with open circles, and the ste12 + STE12 reconstituted strain with closed, inverted triangles.

Finally, to exclude that a suppressor mutation might be present in the ste12 mutant strain that had been tested thus far, a second independent ste12 mutant strain was assessed for virulence. In this case, the virulence of the isogenic wild-type and the two independent ste12 mutant strains was comparable at an inoculum size of 10⁶ cells per infected animal (Figure 8B). Taken together, these findings indicate that in a pathogenic serotype A strain of C. neoformans, the STE12 homolog is not required for pathogenicity in two different animal model systems involving both systemic and CNS infection, two vertebrate species, and a range of infectious doses.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We are interested in elucidating the signal transduction pathways that regulate mating, filamentation, and virulence in the pathogenic basidiomycete C. neoformans. Here, we describe our findings with respect to one gene encoded by the MAT locus, which encodes a homolog of the STE12 transcription factor identified previously by WICKES et al. 1997 .

In the studies reported here, the gene encoding the STE12 homolog was isolated from the pathogenic serotype A strain H99 and disrupted by transformation and homologous recombination. As previously reported for the serotype D STE12 gene (WICKES et al. 1997 ), the serotype A STE12 protein contains a homeodomain that shares marked sequence identity with the STE12 homologs from S. cerevisiae, C. albicans, and K. lactis (Figure 2). In contrast, other regions of STE12, including C-terminal regions that interact with the Dig1 and Dig2 repressors, or with the Mcm1 or Tec1 coactivator proteins in S. cerevisiae, do not appear to be conserved in the C. neoformans STE12 homolog (KIRKMAN-CORREIA et al. 1993 ; PI et al. 1997 ). Either the STE12 protein functions with a different set of partner proteins in C. neoformans, or these protein-protein interaction surfaces are more divergent than the corresponding STE12 homeodomain protein-DNA interaction surface. Homologs of the Ste12 regulators in S. cerevisiae, including Dig1, Dig2, Mcm1, and Tec1, have not yet been identified in C. neoformans, and further studies will be required to define STE12-interacting proteins in this pathogenic fungus.

Interestingly, the C. neoformans ste12 mutant strains exhibited only modest defects in mating on standard V8 mating medium, manifested as a reduction in filament production at early time points and an ~10-fold reduction in the production of prototrophic progeny in a quantitative mating assay. On a synthetic medium that supports mating to a lesser extent (SLAD), the ste12 mating defect was more pronounced, but was not as severe as the mating defects observed with a variety of other mutations, including those in the G protein GPA1 (ALSPAUGH et al. 1997 ) or the Gß protein GPB1 (P. WANG, J. R. PERFECT and J. HEITMAN, unpublished results). We note that these matings involve a divergent MAT serotype A and a MATa serotype D partner because no MATa serotype A strains have yet been identified. Thus, even under very stringent mating conditions (serotype D partner, poorer mating medium), the ste12 mutant strains are able to mate. These considerations likely explain why we observe a subtle mating defect conferred by the ste12 mutation in serotype A, whereas no mating defect was apparent when a serotype D ste12 mutant was crossed to a congenic wild-type serotype D strain on V8 mating medium (Y. CHANG, B. L. WICKES and K. J. KWON-CHUNG, personal communication). Our observations suggest that the STE12 homolog plays only a modest role in mating (Figure 9), or that its functions are redundant with other signaling components or pathways. Our findings are also analogous to previous studies of S. cerevisiae ste12 and C. albicans cph1 mutants, which revealed only partial defects in filamentous growth and virulence, indicating that other proteins may partially overlap with these STE12 functions (LIU et al. 1993 , LIU et al. 1994 ; LO et al. 1997 ).

View larger version (14K): In this window In a new window Download PPT slide   Figure 9. Role of STE12 in signaling pathways regulating C. neoformans mating, filamentation, and virulence. Signaling pathways that regulate C. neoformans responses to environmental signals are depicted. STE12 is required to regulate responses to nitrogen starvation and desiccation that trigger haploid filamentation and sporulation. Here, both the RAS1 protein and the MAP kinase cascade are depicted as signaling upstream of STE12. Haploid fruiting is also regulated by elements of the pheromone-activated MAP kinase cascade, including the G-protein ß subunit GPB1. Mating is regulated by both pheromone and nutrients via pathways involving the Gß protein GPB1 and the G protein GPA1. STE12 plays only a modest role in regulating mating, which is indicated by a dashed line. Finally, the G protein GPA1 and cAMP regulate melanin and capsule production and virulence. By comparison, the ste12 mutation has only modest effects on capsule synthesis and is not required for virulence in two different animal models.

The C. neoformans ste12 mutant strain exhibits a severe defect in monokaryotic fruiting. In response to nitrogen starvation, some MAT strains of C. neoformans filament and sporulate in a process called haploid or monokaryotic fruiting (WICKES et al. 1996 ). We have found that expression of a dominant-active form of the Ras1 protein, Q67L, triggers haploid filamentation in strain H99 in response to nitrogen limitation (J. A. ALSPAUGH, L. M. CAVALLO and J. HEITMAN, unpublished results). We have used this finding to demonstrate that the ste12 mutation blocks haploid fruiting in strain H99 (Figure 4). Complementary studies by Kwon-Chung and colleagues have revealed that the STE12 gene is similarly required for haploid fruiting in serotype D strains of C. neoformans (Y. CHANG, B. L. WICKES and K. J. KWON-CHUNG, personal communication). These complementary findings reveal that STE12 is required for haploid fruiting. In contrast, the STE12 protein is absolutely required for mating in S. cerevisiae (HARTWELL 1980 ), but only partially required for filamentous growth in either S. cerevisiae or C. albicans because of overlapping, redundant signaling pathways (LIU et al. 1993 , LIU et al. 1994 ; LO et al. 1997 ). This situation appears to be reversed in C. neoformans, in that the STE12 homolog is largely dispensable for mating, but required for filamentation. Furthermore, these observations suggest that monokaryotic fruiting is either a simpler linear signaling pathway, or that more than one signaling pathway is absolutely required for differentiation.

Our genetic epistasis experiments provide further support for the hypothesis that STE12 is specialized to regulate filamentation, but not mating. Overexpression of STE12 restored haploid fruiting in mutant strains lacking the Gß protein GPB1, but did not restore mating (Figure 5). Taken together, these findings suggest that STE12 regulates expression of genes necessary for haploid fruiting, but is either not required or is redundant with other factors that regulate mating-specific gene transcription.

Finally, we found no virulence defect of the C. neoformans ste12 mutant strains in two animal models. The first model involves intrathecal inoculation of immunosuppressed rabbits, with recovery of CSF and quantitative yeast counts of CSF over a 2-wk course of infection. There was no difference in survival in the central nervous system between the wild-type STE12, the ste12 mutant, and the ste12 + STE12 reconstituted strains. We also tested the ste12 mutant strain in a murine systemic infection model. In this case, with four different inoculum sizes over a range of four orders of magnitude and two different independent ste12 mutant strains, there was again no difference in the virulence of the ste12 mutant strains compared to either the STE12 wild-type or the ste12 + STE12 reconstituted strains.

These findings in the host are in accordance with our observations that the ste12 mutant strain has little or no defect with respect to most of the known or potential virulence factors of C. neoformans. The only exception was a modest reduction in the capsule size of ste12 mutant cells induced in vitro for capsule production or recovered from the CSF of infected rabbits. However, this subtle defect in capsule production did not impact on virulence in animal models. Similarly, differences in capsule size in nonisogenic strains was previously shown to not have an impact on virulence (DYKSTRA et al. 1977 ). We cannot exclude that in certain strains or serotypes in which capsule or other factors might be limiting for virulence, the ste12 mutation could have an impact on virulence, but even under these conditions, STE12 is unlikely to be a major determinant for virulence of C. neoformans. In C. albicans, mutations in the STE12 homolog Cph1 have only modest effects on virulence, whereas mutants lacking both Cph1 and the Phd1 homolog Efg1 are avirulent (LO et al. 1997 ). Similar redundant signaling pathways may operate in C. neoformans.

It has been suggested that filamentation and monokaryotic fruiting might be required for virulence. On the other hand, the pathogenic serotype A isolate H99 does not undergo haploid fruiting under a variety of different culture conditions and yet is fully virulent in rabbits inoculated intrathecally, or in mice inoculated intravenously or intratracheally (G. HUFFNAGLE, personal communication). Thus, haploid fruiting may not be required for virulence. In addition, C. neoformans in the host is found virtually exclusively as budding yeast, and filamentous forms are rare (ALSPAUGH et al. 1999 ).

In conclusion, our studies reveal that a conserved homolog of the STE12 transcription factor that regulates mating, filamentation, and virulence in S. cerevisiae and C. albicans plays a prominent role in regulating monokaryotic fruiting in response to nitrogen starvation in C. neoformans (Figure 9). Our epistasis analysis begins to delineate elements of a MAP kinase cascade that regulates STE12. Target genes regulated by STE12 remain to be identified. The finding that the STE12 homolog is specific to MAT cells and encoded by a novel region of the MAT locus was initially puzzling because it was not clear how MATa cells would mate without STE12. Our finding that the STE12 homolog is specialized to regulate the MAT-specific filamentation response and does not play a primary role in mating suggests that other transcription factors that regulate mating of both MAT and MATa cells remain to be discovered. For example, mating in C. neoformans may be regulated by a homolog of the Tec1 transcription factor (GAVRIAS et al. 1996 ; MADHANI and FINK 1997 ), which regulates filamentous growth in conjunction with STE12 in S. cerevisiae, or the high mobility group (HMG) class transcription factors Ste11 and Prf1, which regulate mating in Schizosaccharomyces pombe and U. maydis (SUGIMOTO et al. 1991 ; HARTMANN et al. 1996 ). Clearly, much remains to be learned from the signaling pathways that regulate differentiation, mating, and virulence of this important human pathogen, which cannot be directly inferred from studies of other yeast and fungal model systems.

	ACKNOWLEDGMENTS

We thank M. Cristina Cruz and Maria Elena Cardenas for advice and discussions, Cristl Arndt for technical assistance, June Kwon Chung for discussion of results before publication, Brian Wickes for the GAL7-STE12 plasmid pCGS-1, and Rey Sia and Maria Elena Cardenas for comments on the manuscript. These studies were supported in part by R01 grants AI39115, AI41937, AI28388, and AI42159 (J.P. and J.H.), P01 grant AI44975 to the Duke University Mycology Research Unit (G.C., J.P., and J.H.), and KO8 grant AI01556 (J.A.A.), all from the National Institutes of Allergy and Infectious Diseases. Gary Cox is a Burroughs Wellcome New Investigator in Molecular Pathogenic Mycology. Joseph Heitman is an associate investigator of the Howard Hughes Medical Institute and a Burroughs Wellcome Scholar in Molecular Pathogenic Mycology.

Manuscript received April 17, 1999; Accepted for publication August 17, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALSPAUGH, J. A., J. R. PERFECT, and J. HEITMAN, 1997  Cryptococcus neoformans mating and virulence are regulated by the G-protein subunit GPA1 and cAMP. Genes Dev. 11:3206-3217[Abstract/Free Full Text].

ALSPAUGH, J. A., J. R. PERFECT, and J. HEITMAN, 1998  Signal transduction pathways regulating differentiation and pathogenicity of Cryptococcus neoformans.. Fungal Genet. Biol. 25:1-14[Medline].

ALSPAUGH, J. A., R. C. DAVIDSON and J. HEITMAN, 1999 Morphogenesis of Cryptococcus neoformans, in Dimorphism in Human Pathogenic and Apathogenic Yeasts, edited by J. F. ERNST and A. SCHMIDT. Karger, Basel, Switzerland (in press).

BANUETT, F., 1998  Signalling in the yeasts: an informational cascade with links to the filamentous fungi. Microbiol. Mol. Biol. Rev. 62:249-274[Abstract/Free Full Text].

CASADEVALL, A., and J. R. PERFECT, 1998 Cryptococcus neoformans. American Society for Microbiology, Washington, DC.

CHEN, S. C. A., M. MULLER, J. Z. ZHOU, L. C. WRIGHT, and T. C. SORRELL, 1997  Phospholipase activity in Cryptococcus neoformans: a new virulence factor? J. Infect. Dis. 175:414-420[Medline].

COX, G. M., D. L. TOFFALETTI, and J. R. PERFECT, 1996  Dominant selection system for use in Cryptococcus neoformans.. J. Med. Vet. Mycol. 34:385-391[Medline].

DEL POETA, M., D. L. TOFFALETTI, T. H. RUDE, C. C. DYKSTRA, and J. HEITMAN et al., 1999  Topoisomerase I is essential in Cryptococcus neoformans: role in pathobiology and as an antifungal target. Genetics 152:167-178[Abstract/Free Full Text].

DÜRRENBERGER, F., K. WONG, and J. W. KRONSTAD, 1998  Identification of a cAMP-dependent protein kinase catalytic subunit required for virulence and morphogenesis in Ustilago maydis.. Proc. Natl. Acad. Sci. USA 95:5684-5689[Abstract/Free Full Text].

DYKSTRA, M. A., L. FRIEDMAN, and J. W. MURPHY, 1977  Capsule size of Cryptococcus neoformans: control and relationship to virulence. Infect. Immun. 16:129-135