Characterization and Regulation of the Trehalose Synthesis Pathway and Its Importance in the Pathogenicity of Cryptococcus neoformans

Abstract

The disaccharide trehalose has been found to play diverse roles, from energy source to stress protectant, and this sugar is found in organisms as diverse as bacteria, fungi, plants, and invertebrates but not in mammals. Recent studies in the pathobiology of Cryptococcus neoformans identified the presence of a functioning trehalose pathway during infection and suggested its importance for C. neoformans survival in the host. Therefore, in C. neoformans we created null mutants of the trehalose-6-phosphate (T6P) synthase (TPS1), trehalose-6-phophate phosphatase (TPS2), and neutral trehalase (NTH1) genes. We found that both TPS1 and TPS2 are required for high-temperature (37°C) growth and glycolysis but that the block at TPS2 results in the apparent toxic accumulation of T6P, which makes this enzyme a fungicidal target. Sorbitol suppresses the growth defect in the tps1 and tps2 mutants at 37°C, which supports the hypothesis that these sugars (trehalose and sorbitol) act primarily as stress protectants for proteins and membranes during exposure to high temperatures in C. neoformans. The essential nature of this pathway for disease was confirmed when a tps1 mutant strain was found to be avirulent in both rabbits and mice. Furthermore, in the system of the invertebrate C. elegans, in which high in vivo temperature is no longer an environmental factor, attenuation in virulence was still noted with the tps1 mutant, and this supports the hypothesis that the trehalose pathway in C. neoformans is involved in more host survival mechanisms than simply high-temperature stresses and glycolysis. These studies in C. neoformans and previous studies in other pathogenic fungi support the view of the trehalose pathway as a selective fungicidal target for use in antifungal development.

Cryptococcus neoformans is an opportunistic heterobasidiomycetous fungus, with a propensity for producing meningoencephalitis in immunocompromised hosts. For such an opportunistic pathogen to successfully colonize a host, the fungus must survive different environmental conditions which bring the yeast cells under great stresses that might not be part of their natural environmental niche. Cryptococcus neoformans primarily produces infection through entry into the lungs, followed by dissemination into the bloodstream and ultimately invasion of the subarachnoid space as the yeast cells cross the blood-brain barrier. In order for colonization and/or disease in the host to be successful, this yeast must be able to detect changes in its extracellular environment and respond to these signals appropriately (44). During infection the yeast needs to utilize reserve carbohydrates for energy and requires protection of its proteins against damaging insults, such as temperature elevations, nutrition and osmolarity changes, and hypoxic or oxidative stresses from the host. Interestingly, protective mechanisms against many of these stresses that are essential for cellular survival in yeasts have been proposed to be mediated by the disaccharide trehalose (15, 55).

Trehalose (α-glucopyranosyl-α-d-glucopyranoside) is produced by numerous organisms such as bacteria, fungi, plants, and invertebrates (20), but this sugar has never been detected in mammals. Trehalose is composed of two glucose molecules linked at their 1-carbons and is rapidly produced under various cellular stresses. Two recent studies have identified both trehalose and the expression of genes in its pathway at the sites of cryptococcal infections, and these results reinforce the potential importance of the trehalose pathway in C. neoformans survival within the host (28, 50).

Trehalose is hypothesized to protect the yeast cell by increasing its resistance to both internal and external stresses. Several previous studies have shown that as yeast cells come under stress, there is a corresponding increase in intracellular trehalose. The disaccharide appears to increase resistance to heat (18, 35), dehydration and desiccation (25, 30), and other stresses (3), primarily by preventing the denaturation of certain important proteins (30). In addition, trehalose suppresses the aggregation of proteins already denatured, favoring their reactivation by molecular chaperones such as heat shock proteins (14). Trehalose can also act as a membrane protectant during stressful conditions (14). On the other hand, some investigators have suggested that trehalose might also act as a reserve carbon source (23, 31, 36, 51) and that trehalose-6-phosphate (T6P) can regulate hexokinase II and thus glycolytic flux in some fungi (5). Although its use as an energy source remains somewhat controversial, it has been definitively linked to sporulation in fungi (37).

From biochemical and genetic studies in the yeast, Saccharomyces cerevisiae, the trehalose pathway has been shown to involve a complex of seven genes. Trehalose is synthesized by the enzymes T6P synthase, encoded by the TPS1 (trehalose-6-phosphate synthase) gene, and T6P phosphatase, encoded by the TPS2 gene, which converts glucose-6-phosphate via T6P to trehalose (Fig. 1). The reaction is aided by two regulatory subunits encoded by the genes TSL1 and TPS3, which are highly homologous to each other (4, 18). Trehalose is then shuttled to wherever it is needed within the cell. Once utilized, spent trehalose is transported back into the cytosol, where it is hydrolyzed by a neutral trehalase, encoded by the gene NTH1 (32), which degrades the disaccharide into two molecules of glucose. A second, seemingly redundant, neutral trehalase gene, NTH2, is found in some fungal species (57) although little is known about its function. The neutral trehalases have a pH optimum of 7.0 and demonstrate activation by cyclic AMP dephosphorylation (2) and deactivation by phosphatases (10). A third trehalase (ATH1), which is an acid vacuolar, extracellular glycoprotein (17) with a pH optimum of 4.5, has also been identified (39). Very little is known about the regulation of ATH1, but it is believed to be involved in the utilization of exogenous trehalose (41) rather than its direct hydrolysis. Studies have shown that all the genes are essential for optimal enzymatic activity of the trehalose pathway, since a mutation in any gene will either reduce or completely abolish trehalose synthesis in S. cerevisiae (4). From studies in various yeasts, the trehalose pathway seems to act as a carefully regulated recycling system in which both trehalose and T6P levels are carefully controlled by a fine balance of the synthesizing and hydrolyzing enzymes (21).

FIG. 1.

Diagram of trehalose pathway in Saccharomyces cerevisiae.

Because of its potential influence on the yeast stress response in the host environment and the uniqueness of its features in comparison to those of mammalian systems, this study examines the impact of the trehalose pathway on survival mechanisms for C. neoformans, through the isolation and mutation of the synthesizing genes TPS1 and TPS2 and the degrading gene NTH1. The mutation of each of these genes enabled us to study the importance of each enzyme on C. neoformans pathobiology. The synthesizing enzymes of trehalose are clearly important to the ability of C. neoformans to grow at mammalian body temperatures and thus are an important component of the virulence composite; on the other hand, the trehalose-degrading gene has no apparent phenotype linked to the virulence composite.

MATERIALS AND METHODS

Strains, media, and reagents.

C. neoformans strains were wild-type H99, a serotype A clinical isolate (46), and H99R, a spontaneous uracil auxotroph derived from H99 (11). The tps1 and tps2 mutant strains were created from H99R, while the nth1 mutant originated from H99. H99, H99R, and the nth1 strain were maintained on nonselective YEPD medium (1% yeast extract, 2% peptone, 2% glucose, 2% agar), while the tps1 and tps2 strains were both maintained on uracil-depleted medium (yeast nitrogen base [YNB] without amino acids [Difco, Franklin Lakes, NJ]; 2% glucose; 2% agar; the l-amino acids adenine, tryptophan, histidine, arginine, and methionine at 20 mg liter⁻¹ [final concentration]; leucine at 60 mg liter⁻¹; threonine at 200 mg liter⁻¹; phenylalanine at 50 mg liter⁻¹; tyrosine and lysine at 30 mg liter⁻¹ [Sigma, St Louis, MO]). Prior to the studies, however, all strains were grown on nonselective YEPD medium plates.

Molecular biology techniques.

Genomic DNA was isolated using the techniques described by Pitkin et al. (48). RNA was isolated using the Trizol reagent and protocol (Gibco BRL, Rockville, MD). DNA probes were labeled with [³²P]dCTP (New England Nuclear, Pittsburgh, PA) using a random primer labeling kit (Gibco BRL, Rockville, MD).

For Southern hybridizations, a 2.3-kb TPS1 probe was generated by digestion of a TOPO (Invitrogen, Carlsbad, CA)-cloned TPS1 gene with EcoRI. For Northern analysis, a 600-bp probe was generated from an exonic region using the primers TPS5 (5′-CCA CAC ATA TGA TTA TGC-3′) and TPS3 (5′-GGT GGG GCA CTT ACC AAG-3′). For both Southern and Northern analyses of TPS2, a 751-bp region of an exon was isolated by PCR using the primers TPS2 Probe 5 (5′-GCC ATC AAG CTC TTG AAA AAG ACG-3′) and TPS2 Probe 3 (5′-CTC CGG TAT AAT ATT TGA AAT TTC C-3′). For Southern analysis of the NTH1 gene, a 446-bp probe was amplified using the primers NTH4F (5′-GGC ACA TAT ATG TTA TCG AAT CTT CTT CAG G-3′) and NTH6R (5′-CCA TTG CAA GAG CGA GGA TGC CTG G-3′). A 604-bp fragment was also amplified using the primers NTHATG#1 (5′-ATG AGC CCA CCC AAT GGT GTG CCA AGG-3′) and NTH4R (5′-CCT GAA GAA GAT TCG ATA ACA TAT ATG TGC C-3′). For Northern analysis of NTH1, a 569-bp exonic region was amplified using the primers NTH1 Exon 5′ (5′-CTT TAT GGC CCT GGG CCT CTT GG-3′) and NTH1 Exon 3′ (5′-CTT GTA TAA GAG CGA GTT CAA ATC G-3′). For Southern analysis of the URA5 gene, a 1,538-bp probe was amplified using the primers URA5 5′ (5′-GTG TCC GAT CGA CAT GAT CCA CGG-3′) and URA5 3′ (5′-GCT TGC CTC CAG GAG GTG GGA GGG-3′).

All restriction enzymes were either from Gibco BRL, Rockville, MD, or New England Biolabs, Beverly, MA. Primers were made by Integrated DNA Technologies, Coralville, IA.

Identification and characterization of the C. neoformans genes encoding T6P synthase, T6P phosphatase, and neutral trehalase.

An 800-bp fragment of the C. neoformans TPS1 gene was generated by degenerative PCR using H99 genomic DNA and the primers Tresyn#1 (5′-TGG GTN CAY GAY TAY CAC-3′) and Tresyn#2 (5′-CTG RTA YTA YTC YTC NAC RTC- 3′) (N = A or T or C or G; Y = C or T; R = A or G). This Taq-amplified PCR fragment was cloned into a pCR2.1-TOPO vector for rapid cloning of PCR products (Invitrogen, Carlsbad, CA) and sequenced. For the isolation of the entire TPS1 gene, a genomic library of H99 in EMBL3 was screened with the amplified 800-bp probe fragment described above. Positive plaques were purified through three rounds of repeated screening, and three positive λ plaques were purified using the plate lysate method. The clones were digested with SalI, and a 6.5-kb fragment, a 5.0-kb fragment, and a 4.0-kb fragment were isolated, cloned into SalI sites in pBluescript SK, and sequenced.

For the isolation of TPS2 and NTH1 genes, a BLAST search of the Cryptococcus neoformans H99 genome sequencing databases (Duke Center for Genome Technology [http://cneo.genetics.duke.edu/] and Cryptococcus neoformans Sequencing Project, Center for Genome Research [http://www.broad.mit.edu]) using homologous genes from other fungal species was carried out to identify sequences that were used to design primers for amplification of full-length sequences. Identification of potential motifs was completed using MOTIF software (Genome Net, Japan [http://motif.genome.jp/]).

Generation of site-directed mutants of tps1, tps2, and nth1 by biolistic DNA delivery.

A plasmid containing a disrupted tps1 gene was generated. The entire 2.3-kb TPS1 gene was isolated by digestion of the cloned 4.0-kb lambda fragment with EcoRI. After cloning into a TOPO vector (Invitrogen, Carlsbad, CA), the plasmid was linearized by digestion with SmaI and then treated with shrimp alkaline phosphatase (Gibco BRL, Rockville, MD). A plasmid containing the URA5 cassette, pRCD69, was also digested with SmaI, and the 2.5-kb URA5 fragment was then ligated into the digested TPS1 gene. The entire TPS1::URA5 fragment was then isolated from the plasmid and treated with a dideoxy treatment to reduce recircularization (13). This disruption construct was then used to generate the tps1 mutant by biolistic DNA delivery into H99R and selection on YNB-uracil dropout plates containing 1 M sorbitol, as described by Toffaletti et al. (52). Transformants were then transferred to YNB-uracil dropout-glucose and YNB-uracil dropout-galactose plates in order to identify those unable to grow at 37°C on glucose. The site-directed mutants were confirmed by colony PCR and Southern blotting.

The tps2 insertional mutant was constructed by blunt-end ligation of a TOPO plasmid containing an 800-bp fragment of the TPS2 gene and a 2.5-kb URA5 cassette. The 800-bp TPS2 gene fragment was isolated using the primers TPS2-2 (5′-CGG TAT GCT GGG AGC CAA CC-3′) and TPS2R (5′-CGT GAG GGT TGA TCT GTA GAG C-3′). The BbsI-digested TOPO-TPS2 vector was treated with shrimp alkaline phosphatase prior to the ligation. The final construct was isolated and then introduced by biolistic DNA delivery into H99R. Transformants with the tps2 gene were identified similarly to tps1 mutants, by their subsequent temperature sensitivity (TS) to 37°C, and confirmed by colony PCR and Southern blotting.

Both the tps1 and tps2 strains were then reconstituted back to wild type by replacing the URA5 fragment with the original TPS1 and TPS2 gene fragments, respectively, using biolistic DNA delivery onto YPD plates containing 1 M sorbitol (Sigma, St Louis, MO). The entire 2.3-kb TPS1 gene was again isolated from its vector and used for the reconstitution of the tps1 mutant. Primers used to amplify the 822-bp TPS2 fragment used for the reconstitution were TPS2 5′ Recon (5′-GCG TTC GAC ATG TTT GTA GGG CTG-3′) and TPS2 3′ Recon (5′-CCC TTA TTG ATG GCT TGG GCA ACG CCC-3′). After 2 to 3 h on YPD-sorbitol plates, both the TPS1 and TPS2 reconstituted transformants were selected by growth on 5-fluoroorotic acid plates. Colonies that demonstrated uracil auxotrophy and resumed normal growth at 37°C were selected. Southern analysis demonstrated homologous recombination at the TPS1 locus with no additional ectopic integrations of TPS1. In the tps2 reconstitution event, analysis showed similar recombination at the TPS2 native locus with no additional ectopic integrations. These transformations produced TPS1::ura5 or TPS2::ura5 auxotrophic strains and thus could not be assessed for virulence in vivo. In order to correct the auxotrophy, the URA5 gene cassette was isolated and introduced into each strain via biolistic DNA delivery. Cells that were able to grow on uracil-deficient medium, and thus were prototrophic, were selected for further study.

The neutral trehalase mutant, the nth1 strain, was generated using a neomycin gene cassette, which encodes resistance to the antibiotic G418. Use of this marker allows the utilization of wild-type H99 rather than an auxotrophic strain. The nth1 mutant was produced by deletion of the gene from the ATG start site to its stop codon and replacement of the gene with the neomycin cassette. The deletion mutant was constructed using a PCR overlap technique (16). Two fragments were synthesized by PCR; 1 kb of 5′-untranslated region (UTR) sequence upstream of the ATG start codon, and 1 kb of 3′-UTR sequence downstream of the stop codon. The primers used to generate the upstream fragment were NTH1-1A (5′-GCG TCC GAA GCC GAA TGG AGG TTC G-3′) and NTH1-2 (5′-CCG TGT TAA TAC AGA TAA ACC AAG GGC TTG ACT CTC ATG GAG AGC-3′), while the primers used to generate the downstream fragment were NTH1-5 (5′-GCT CAC ATC CTC GCA GCA AGG CCG GAC TTC ACA-3′) and NTH1-6A (5′-GCT GGA CCA GTC CTC CAA TGT CGG-3′). The two fragments were held together with a third fragment: the neomycin gene cassette which was amplified using the primers NTH1-3 (5′-GCT CTC CAT GAG AGT CAA GCC CTT GGT TTA TCT GTA TTA ACA CGG-3′) and NTH1-4 (5′-CTG GCA GAT GTG AAG TCC GGC CTT GCT GCG AGG-3′). The entire construct was introduced by biolistic DNA delivery into H99 cells on YEPD-sorbitol plates and then incubated at 30°C for 16 to 20 h. Liquid YEPD medium was then used to resuspend the cells, which were then replated onto medium containing G418 and further incubated for 3 to 4 days at 30°C. Yeast colonies with replacement of the NTH1 gene by the neomycin cassette were identified by PCR and confirmed by colony Southern blotting.

Northern blot analysis.

Northern blots were performed on H99 and tps1 strains after cells were grown overnight and then placed in the following conditions and grown for 1 h or 5 h. Conditions were 30°C (RPMI 1640 with NaM0PS and 0.75 M NaCl, pH 7.3), 37°C (RPMI 1640 containing NaMOPS, pH 7.3, air), and 5% CO₂ at 30°C (RPMI 1640 containing NaHC0₃ and NaMOPS, pH 7.3). Northern blots were performed on H99 and mutant (cna1, cpa2, cpa1/2, and mga2) strains, after cells were grown overnight at 30°C and suspended into RPMI (pH 7.3) for 2 h at 37°C. Northern blots were performed on H99 grown at 30°C and 37°C with yeast extract peptone and various sugars, under heat shock (42°C for 30 min) and cold shock (4°C for 30 min) conditions.

All yeast cells were lyophilized, and RNA was extracted with Trizol. RNA was electrophoresed across a 1.5% formaldehyde agarose gel, blotted onto Nytran membranes, and probed with ³²P-labeled fragments of TPS1 (517 bp), TPS2 (751 bp), NTH1 (567 bp), and a control glyceraldehyde-3-phosphate dehydrogenase gene.

Quantification of trehalose and T6P using NMR spectra.

Lyophilized yeasts were suspended in 0.3 ml phosphate-buffered saline (PBS) (pH 7.2, at room temperature) made up in 50% deuterated water (D₂O/PBS; Australian Nuclear Science and Technology Organization, Lucas Heights, Australia) containing 5 mM p-aminobenzoic acid for quantification. The suspension was immediately transferred to a 5-mm-diameter nuclear magnetic resonance (NMR) tube. For ³¹P-NMR studies, cells were incubated on Sabouraud dextrose agar (Difco Labs, Detroit, MI) plates at temperatures of 30°C and 37°C for 24 and 48 h in triplicate. Cells were then suspended in yeast nitrogen broth (Difco) made up with 50% D₂O containing 1% glucose, buffered at pH 7.0 with 0.345% (wt/vol) 3-(N-morpholino)propanesulphonic acid (Sigma Chemica, St Louis, MO) and transferred to a 10-mm-diameter NMR tube. ¹H-NMR spectra were obtained at 37°C on a Bruker Avance 360-MHz NMR spectrometer using a 5-mm-diameter (¹H, ¹³C) inverse-detection dual-frequency probe. Two-dimensional homo- and heteronuclear correlation spectra were acquired for assignment of ¹H-NMR resonances to specific compounds.

Trehalose determinations were estimated by integration of ¹H, ¹H correlation spectroscopy cross peaks (34). Acquisition parameters were as follows: spectral width in t₂, 3600 Hz; t₂ time domain, 2K; 256 increments of eight acquisitions each; relaxation delay, 3 s. Sine-bell window functions were applied in the t₁ dimension, and Gaussian-Lorentzian window functions were applied in the t₂ dimension. Zero filling was used to expand the data matrix to 1K in the t₁ dimension. The cross peak integrals for glucose and trehalose (H-1/H-2 as well as H-2/H-3) were compared to those of p-aminobenzoic acid. To account for the differences in relaxation times and ³J (¹H, ¹H) coupling constants between the studied cross peaks, calibration factors were determined from a mixture of pure chemicals of known concentration. The experimental error of the method was found to be in the order of 10 to 20% and confirms previous reports (9) Despite the recognized limitations associated with the measurement of peak volumes in two-dimensional NMR spectra, the concentration estimates are considered to be more meaningful than qualitative descriptions of peak intensities.

The ³¹P-NMR studies were performed for determination of the relative amount of T6P. ³¹P-NMR spectra were obtained at 37°C on a Bruker Avance 400-MHz NMR spectrometer using a 10-mm-diameter broadband probe. Acquisition parameters were as follows: operating frequency, 161.98 MHz; 30^o excitation pulse; 5-s relaxation delay; 512 transients to acquire a 64K flame ionization detection; spectral width, 16,000 Hz. A solution of 5 mM dimethyl methylphosphonate (Fluka, Buchs, Germany) in a 1-mm-diameter insert was used as a chemical shift (δ = 40.5 ppm) and concentration standard.

In vitro characterization of the tps1, tps2, and nth1 mutants.

Three null mutant (tps1, tps2, and nth1) strains were examined for basic growth rate on full-nutrient medium, temperature sensitivity, capsule production, urease production, and melanin synthesis. Growth at various environmental temperatures was assessed for fungistatic or fungicidal activity. Growth rate was measured by inoculating the strain to grow in YEPD (2% glucose) or YEPG (2% galactose) broth with aeration for 48 h at 30°C and 37°C with periodic subculture for numbers of viable yeast cells. Samples were diluted in PBS and plated on YEPD plates for calculating numbers of CFU. Cell growth was plotted against time, and the gradient of the log phase of the cells was designated as the generation time. Temperature sensitivity was assessed according to whether the cells grew at 37°C or above. If cell counts declined over 48 h, the strain was designated as fungicidal; if yeast cell counts remained constant over this time, the strain was designated as fungistatic. For detecting suppression of the TS phenotype, the mutants were grown on YEPD agar plates with 1 M sorbitol added at 37°C.

All strains were placed in an anaerobic chamber for 5 days in YEPD broth at 25°C and examined for turbidity indicating quantitative growth and then plated for growth in air at the end of 5 days. To create oxidative stress, strains were incubated in 0.125 mM, 0.25 mM, 0.5 mM, and 1 mM t-butylhydroperoxide (t-B00H) in YEPD for 48 h at 30°C with shaking, and quantitative cultures were performed. To induce osmotic stress, strains were placed in 1 M sorbitol in PBS for 48 h at 30°C, and quantitative cultures were performed and compared before and after incubation. Urease production was determined by inoculating Christiansen's liquid broth with each strain and examining for the appearance of pink medium. Melanin synthesis was determined by diluting colonies onto Niger seed agar plates, incubating at 30°C, and examining for the appearance of brown colonies. Impact on capsule production was analyzed by growing yeast cells on minimal iron (53a) medium and using India ink examinations to measure capsule sizes of 10 random yeast cells by microscopy at a magnification of ×400 daily for 2 weeks.

The tps1 and tps2 mutant strains, along with H99, were analyzed for any synergistic antifungal activity with the stress of antifungal compounds, such as amphotericin B, caspofungin, fluconazole, and flucytosine. Tests were carried out in accordance with NCCLS protocol M-27A (40) with some modifications to determine MIC₈₀, MIC₁₀₀, and minimum fungicidal concentration (MFC) values.

Macrophage-Cryptococcus intracellular growth assay.

The macrophage cell line J7748.1 was used. Cells (5 × 10⁵/ml) were activated with 100 units/ml of murine gamma interferon and 0.15 μg/ml of lipopolysaccharide. These cells were added to equal numbers of yeast cells and incubated with 10 μg/ml of 18B7, an anti-GXM monoclonal antibody (generously supplied by Arturo Casadevall) for opsonization. In 96-well plates, the yeast and macrophage monolayers were incubated for 2 h and then washed with tissue culture medium to remove extracellular yeast cells. The monolayers were incubated for 24 h at 37°C and lysed with 1% sodium dodecyl sulfate to release yeast cells, which were plated on YEPD agar plates for quantification. Yeast cells were also seeded into wells without macrophages and incubated for 24 h, and these were quantitated as extracellular growth controls.

In vivo characterization mutants.

New Zealand White male rabbits (2 to 3 kg) were housed in separate cages and provided with water and Purina rabbit chow ad libitum. H99, the tps1 strain, and the reconstituted TPS1 strain were prepared by growth at 30°C for 48 h in YEPD. The cells were pelleted at 13,000 × g at 4°C for 10 min, washed once, and resuspended in 0.015 M PBS to a concentration of 3.3 × 10⁸ cells ml⁻¹. The rabbits were treated intramuscularly with 1.2 mg betamethasone sodium phosphate-betamethasone acetate (Schering, Kenilworth, NJ) for 1 day prior to inoculation and then daily for 14 days. After sedation with ketamine (Fort Dodge, Fort Dodge, IA) and xylazine (Vedco, St Joseph, MO), 10⁸ cells in a volume of 0.3 ml of each strain were inoculated intracisternally into three different groups of rabbits (four rabbits per group). Rabbits were sedated on days 3, 7, 10, and 14 after inoculation, and cerebrospinal fluid (CSF) was withdrawn. Quantitative yeast cultures were performed by dilution of CSF with PBS and plating on YPD medium with incubation at 30°C for 72 h in order to count CFU (46), and yeast counts between groups were directly compared statistically using a t test for unpaired means.

In the murine survival model (12), BALB/c mice were inoculated intranasally with 10⁶ cells. The mice were housed in cages and provided with food and water ad libitum. Groups of 10 mice received wild-type H99, the tps1 strain, or the reconstituted TPS1 strain, and the animals were observed daily. At any sign of difficulties with daily grooming activities or significant weight loss, animals were sacrificed. Historical studies have shown that animals do not survive more than a few days after these symptoms appear. Results were analyzed statistically using the Kaplan-Meier method and log rank test. Animals surviving at day 60 were sacrificed, and lung/brain tissue samples were cultured quantitatively.

The in vivo assays using the Caenorhabditis elegans/C. neoformans model system was performed as previously described, with minor modifications. Briefly, C. neoformans (H99, tps1, tps1/TPS1, tps2, nth1) strains were inoculated into 2 ml of YEPD and grown at 30°C for 48 h. Ten microliters of the culture was spread on 35-mm-diameter tissue culture plates (Falcon, Franklin Lakes, NJ) containing brain heart infusion agar (Difco, Franklin Lakes, NJ). The plates were incubated at 30°C for 48 h. Ampicillin (100 μg/ml) was also included in the medium to selectively prevent growth of Escherichia coli OP50 carried over to the yeast-containing plates with transfer of worms. The standard C. elegans strain N2 Bristol was used for all studies. Nematodes were maintained at 15°C and propagated on E. coli OP50 using established procedures (8). Approximately 40 to 50 young adult worms were transferred from a lawn of E. coli OP50 on standard nematode growth medium to a lawn of the yeast to be tested, incubated at 25°C, and examined for viability at 24-h intervals with a dissecting microscope. Worms were considered dead when they did not respond to touch with a platinum wire pick. Each experiment was performed in triplicate. Plotting of killing curves and estimation of differences in survival (log rank and Wilcoxon tests) with the Kaplan-Meier method were performed using STATA (College Station, TX) 6 statistical software.

RESULTS

Isolation and characterization of C. neoformans TPS1, TPS2, and NTH1 genes.

A full-length TPS1 gene (GenBank accession number AY619946) was isolated, revealing a putative 2,489-bp open reading frame that encodes a predicted 671-amino-acid protein with seven introns ranging in size from 52 to 87 bp. Motif analysis of the promoter region of this gene indicates the presence of eight putative stress response elements (STREs). These elements have the core consensus sequence AGGGG or CCCCT, and they have been shown to mediate stress induction of genes through their promoter by many different stresses (32, 51, 59). Comparative genomic sequence analysis showed that the cryptococcal TPS1 gene had amino acid identities of between 59 and 62% with homologues in other fungal species, including Ustilago maydis (62%), Schizosaccharomyces pombe (62%), Aspergillus niger (62%), Emericella nidulans (61%), and Magnaportha grisea (59%).

The identified C. neoformans TPS2 gene is a putative 3,291-bp open reading frame that encodes a predicted 988-amino-acid protein with six introns ranging in size from 50 to 61 bp (GenBank accession number BK005414). Comparative genetic sequence analysis revealed amino acid identities of between 39 and 50% with TPS2 homologues in other fungal species, including Emericella nidulans (50%), Candida albicans (44%), Saccharomyces cerevisiae (40%), and Schizosaccharomyces pombe (39%). Motif analysis of the promoter region predicted the presence of four putative STREs within the 5′ UTR.

The neutral trehalase gene, NTH1, is a putative 3,127-bp open reading frame that encodes a predicted 826-amino-acid protein with 13 introns ranging in size from 37 to 55 bp (GenBank accession number BK005427). Comparative genetic sequence analysis showed it to have amino acid identities of between 45 and 62% with other NTH1 fungal species, including Ustilago maydis (62%), Aspergillus nidulans (54%), Schizosaccharomyces pombe (54%), Magnaportha grisea (53%), Saccharomyces cerevisiae (50%), and Candida albicans (45%). Motif analysis (Genome Net, Japan) of 120 bp of the putative promoter region suggests the presence of two STREs.

Neither a second neutral trehalase gene, NTH2, nor an acid trehalase gene, ATH1, could be identified by conducting sequence comparisons in BLAST of other fungal genes of the NTH2 or ATH1 strain with either H99 or JEC21 cryptococcal genome sequencing databases (H99 sequencing databases: Whitehead Institute [http://www.broad.mit.edu], Stanford [http://sequence-www.stanford.edu/group/C.neoformans/], Duke [http://cneo.genetics.duke.edu]; JEC21 sequencing database: TIGR [http://www.tigr.org/tdb/edb2/crypt/htmls/]).

Generation and reconstitution of site-directed gene mutants.

In order to study the pathophysiological impact of the genes TPS1 and TPS2 in C. neoformans, site-directed disruption mutants were generated. Southern blot analysis showed the insertion of a URA5 fragment into each gene, with no ectopic integrations, in the tps1 and tps2 mutants selected. Several transformants from each gene mutation were created for further characterization, and one was used for the reconstitution of each gene. Both the tps1 and tps2 strains were reconstituted back to wild type by the methods previously described (56). Southern blot analysis illustrated homologous recombination at the TPS1 locus with no additional ectopic integrations of the TPS1 gene fragment. In the tps2 reconstitution event, analysis showed similar recombination at the TPS2 native locus with no additional ectopic integrations. Both these transformations produced TPS1::ura5 or TPS2::ura5 auxotrophic strains, which were made prototrophic by introducing the URA5 gene into reconstituted strains. Southern blot analysis of a reconstituted tps1/TPS1 strain showed no additional ectopic integrations, and this strain was chosen for virulence studies. Similarly, analysis of the tps2/TPS2 reconstituted strain indicated homologous recombination at the TPS2 locus with no additional ectopic integrations.

The neutral trehalase mutant was constructed as a deletion mutation by insertion of the neomycin marker. Of 32 potential mutants, 5 showed the correct band upon PCR analysis. Southern blot analysis of the five mutants illustrated homologous recombination at the NTH1 locus with no additional ectopic integrations. One mutant was chosen for further analysis.

In vitro characterization of the tps1, tps2, and nth1 mutants.

In the characterization for a virulence phenotype, tps1, tps2, and nth1 mutants all show normal production of melanin, urease and capsule when placed under specific inducing conditions, and these phenotypes for all mutants were similar to those for H99.

The tps1 mutation in C. neoformans confers the TS phenotype such that growth of the tps1 strain is inhibited when cells are exposed to glucose-containing medium at 37°C (Fig. 2), but growth at 37°C was improved in galactose-containing medium (generation times, 23 and 13 h, respectively). Generation times for the tps2 and nth1 mutants at 30°C in fully nutritious medium with glucose or galactose were the same as those for H99. The tps1 mutant was slightly slower than H99 or the other mutants, with a generation time of 2.8 h in glucose or galactose compared to 2.0 h for H99.

FIG. 2.

Growth curves of strain H99 and the tps1, tps1/TPS1, and tps2 strains over 48 h at 37°C in YEP with either 2% glucose or 2% galactose.

The tps2 mutants also showed temperature sensitivity in growth at 37°C with glucose, but in contrast to the tps1 strain, which shows little growth over 48 h, the tps2 strain starts to die after 6 h at 37°C, and demonstrates >90% cell death by 48 h with a 10⁵-yeast-cell inoculum (Fig. 2). On the other hand, in galactose-containing medium at 37°C the tps2 mutant did grow with a generation time of 6.1 h, compared to 3.7 h for H99. For comparison, a calcineurin A mutant (the cna1 strain) which is known to be TS (42) shows some death over 48 h at 37°C but does not die as rapidly as the tps2 mutant (data not shown). In contrast, the neutral trehalase mutant, the nth1 strain, did not show any temperature sensitivity at elevated temperatures with 48 h growth at 37°C similar to H99 (data not shown). The TS phenotype of the tps1 and tps2 strains at 37°C in glucose-containing medium was completely suppressed in the presence of 1 M sorbitol in the glucose-containing medium (Fig. 3).

FIG. 3.

Serial dilutions of H99 and the nth1, tps1, and tps2 strains on YEPD (left) or YEPD with 1 M sorbitol (right) at 37°C for 72 h, showing the ability of sorbitol to rescue the tps1 and tps2 strains from growth defects.

In the macrophage-Cryptococcus growth assay, there were no differences in intracellular survival of tps1, tps2, and nth1 mutants compared to H99. There were direct effects of temperature on yeast cell growth in this host cell assay at 37°C, but these effects were observed on both intracellular and extracellular cells, and the ratios between these cell populations were the same as that for H99. None of the four strains (H99 and the tps1, tps2, and nth1) grew in an anaerobic environment, but after 5 days in reduced oxygen, all strains were able to recover viability.

Under the oxidative stress of t-B00H exposure only the tps1 strain showed some increased susceptibility compared to H99, the tps2 strain, and the nth1 strain, in that growth of the tps1 strain was inhibited at 0.5 mM t-B00H, compared to 1.0 mM t-B00H for the other strains. Similarly, in the presence of the osmotic stress of 1 M sorbitol solution at 30°C, only the tps1 mutant had a decrease in viability of approximately 10-fold over 48 h, while the tps2 and nth1 mutants were similar in survival to H99.

For a further appreciation of the impact of specific stresses and the trehalose pathway, the mutants were exposed to four antifungal drug classes. In vitro susceptibility testing of the tps1 and tps2 strains with amphotericin B, caspofungin, fluconazole, and flucytosine were determined at 30°C. The results are described in Table 1. MICs for H99 were 0.06 μg/ml for amphotericin B, >16 μg/ml for caspofungin, 8 μg/ml for fluconazole, and 8 μg/ml for flucytosine. MICs and MFCs for the tps1 and tps2 strains were similar to each other, and neither mutant showed more than a fourfold difference from MICs and MFCs measured for H99 for all antifungal agents, except that the tps1 strain was less susceptible to fluconazole than the other two strains by as indicated by MICs, but killing was similar as indicated by MFC.

TABLE 1.

In vitro susceptibility of three strains to four antifungal agents

Antifungal compound	MIC80			MIC100			MFC
	H99	tps1	tps2	H99	tps1	tps2	H99	tps1	tps2
Amphotericin B	0.0625	0.0312	0.0625	0.125	0.0625	0.125	0.25	0.25	0.25
Caspofungin	>16	16	8	>16	16	>16	>16	>16	>16
Fluconazole	8	32	2	8	64	4	64	>64	>64
Flucytosine	8	8	4	8	32	8	>64	>64	>64

The impact of these gene mutations on the trehalose pathway was examined by measuring two major products of the pathway: trehalose and T6P. The effects of the tps1 and tps2 mutations on the levels of trehalose and T6P within the cell under various conditions of temperature and time were determined. ¹H ¹H correlation spectroscopy NMR spectra were acquired to determine the total trehalose content (trehalose plus T6P) in the H99, tps1, and tps2 strains. ³¹P-NMR spectra were used to estimate the concentration of T6P relative to that of total trehalose. Data are summarized in Table 2. Trehalose levels peaked in H99 cells after 1 to 2 h of exposure at both 30°C and 37°C and were consistently higher at 37°C. After 24 h of exposure to each temperature, trehalose levels returned to those observed after the initial 20-minute exposure. Trehalose could not be detected in the tps1 or tps2 strains. As predicted, large amounts of T6P accumulated in the tps2 cells and were larger at 37°C than at 30°C. T6P was too low to detect by ³¹P NMR in the H99 and tps1 strains. We hypothesize that the drop in T6P levels in tps2 after 24 h of exposure was related to the death of the yeast cells at 37°C.

TABLE 2.

Measurements of trehalose and T6P in C. neoformans strains^a

Time	Amt of trehalose (μmol/108 cells) after exposure at:						Amt of T6P (μmol/108 cells)
	30°C			37°C			tps2 at 30°C	tps2 at 37°C
	H99	tps1	tps2	H99	tps1	tps2
20 min	52	ND	ND	50	ND	ND	87	-
40 min	66	ND	ND	83	ND	ND	110	175
1 h	81	ND	ND	92	ND	ND	135	200
2 h	77	ND	ND	90	ND	ND	151	260
3 h	60	ND	ND	83	ND	ND	170	280
4 h	58	ND	ND	73	ND	ND	-	211
24 h	45	ND	ND	68	ND	ND	147	75

^aND, not detected; -, not performed.

Gene expression studies of the trehalose pathway.

In order to appreciate the regulation of the trehalose pathway, we examined the transcription of TPS1, TPS2, and NTH1 in H99 in response to several environmental stimuli with 1- and 5-h exposures. As shown in Fig. 4, TPS1 is up-regulated approximately fourfold at 37°C compared to 30°C. TPS2 was induced at 37°C but not to same extent as TPS1, and NTH1 expression was not altered by temperature changes (data not shown). TPS1 is highly up-regulated when the yeast is exposed to either a heat or cold shock condition. TPS1 is turned on in the presence of glucose, but its transcription is turned off in the presence of other sugars, such as maltose (Fig. 4). Other environmental stresses, such as the presence of 5% CO₂ and an osmotic challenge with 0.75 M NaCl appeared to slightly (twofold) increase expression of the trehalose pathway genes, TPS1 and TPS2, but had no impact on NTH1 as shown by Northern blot analysis (data not shown).

FIG. 4.

Northern blot analysis of TPS1 expression in H99 under temperature, sugar, or heat/cold shock stress.

Because temperature sensitivity of the trehalose pathway is so important in both its regulation and survival at high temperatures for C. neoformans, an attempt was made to determine whether there were any links between this pathway and several signaling pathways associated with high-temperature growth. Therefore, we examined several signaling mutants (cna1 [42], mga1 [33], and cpa1/cpa2 [54]) for their involvement in expression of the trehalose pathway genes (TPS1, TPS2, and NTH1). We found that there was no apparent difference as shown by Northern blot analysis in the expression of the three trehalose pathway genes in these mutants compared to trehalose gene expression patterns in H99 at 2 h of exposure to 37°C (data not shown).

TPS1 activity is essential for virulence in mammals.

The impact of the initial enzyme in the trehalose synthesis pathway, TPS1, was studied in the immunosuppressed rabbit cryptococcal meningitis and murine inhalational models. The immunosuppressed rabbit routinely shows a sustained infection in the subarachnoid space over 14 days, and death occurs in the mouse model within 35 days with the wild-type H99 infection (13, 46).

H99, the tps1 strain, and the reconstituted tps1/TPS1 strain were inoculated intracisternally into immunosuppressed rabbits (Fig. 5). H99 and the reconstituted tps1/TPS1 strains showed an equal and sustained high level of infection with approximately 6.1 × 10⁵ CFU/ml of CSF cells, consistently found in the CSF of all animals, during the 14-day infection period. In striking contrast, the tps1 mutant showed only 1.8 × 10² CFU/ml CSF at day 3 and a complete clearance of viable yeast cells in the CSF by day 7 of infection (P was <0.01 compared to H99 and reconstituted tps1/TPS1 strains). In the murine survival model, H99, the tps1 strain, and the reconstituted tps1/TPS1 strains were inoculated intranasally. H99 cells produced death in 100% of mice by day 16. In contrast, all mice inoculated with the tps1 strain survived over 60 days of infection, after which they were all sacrificed and the lungs and brain were removed and cultured to determine the number of yeast cells present. No viable yeast cells were found in the lung or brain tissue of any of the animals, indicating that viable yeast cells had been completely cleared from the host. The survival of tps1 strain-infected mice was significantly longer than that of those infected with H99 (P < 0.01). The mice had cleared infection with the tps1 strain both symptomatically and microbiologically. The reconstituted tps1/TPS1 strain was slightly less virulent than the wild-type strain in the mouse model but remained significantly more virulent than the tps1 mutant strain (P < 0.01) (Fig. 6).

FIG. 5.

Comparative quantitative CSF yeast counts (means ± standard errors of the mean) over 2 weeks in immunosuppressed rabbits with cryptococcal meningitis. H99 and the reconstituted tps1/TPS1 strain produced a chronic infection, but the tps1 mutant strain was rapidly eliminated.

FIG. 6.

Survival graph of murine inhalational cryptococcosis with H99, tps1, and reconstituted tps1/TPS1 strains over 40 days. All of the mice with H99 and reconstituted tps1/TPS1 infection died but all of the mice with tps1 infection survived.

Impact of trehalose on the C. elegans model at room temperature.

In order to obtain insight into the possible evolutionary development of trehalose as a virulence factor and to determine whether, pathologically, this sugar had an impact on the virulence composite through other stress mechanisms besides its importance for exposure to high temperatures, we examined the killing of the nematode Caenorhabditis elegans by the tps1, reconstituted tps1/TPS1, tps2, and nth1 mutant strains and by H99.

In the C. elegans/C. neoformans system, the tps1 mutant strain was less virulent than H99 and the reconstituted tps1/TPS1 strains as determined by survival analysis (P < 0.02) and the results were similar to those of attenuation from the mammalian models with these strains. Furthermore, the tps2 and nth1 mutant strains had a virulence profile similar to that of strain H99 in the worm, and all were significantly more virulent than the tps1 strain (P < 0.01) (Fig. 7).

FIG. 7.

Survival of C. elegans N2 animals feeding on the C. neoformans tps1, tps2, and nth1 mutant strains compared to wild-type strain H99. Only the tps1 mutant has an attenuated phenotype.

DISCUSSION

The potential pathobiological importance of trehalose in cryptococcosis was identified by two in vivo screens. First, in a global transcriptional gene analysis, CSF was removed from rabbits infected with cryptococcal meningitis. With the use of serial analysis of gene expression for identification and quantitation of transcripts from yeast cells within the CSF, one of the genes most highly expressed was TPS1 (50). Second, the potential importance of the trehalose pathway for the yeast within the central nervous system was further supported when tissues surrounding cryptococcomas from infected rat brains and lungs were analyzed by NMR studies (28, 29). One of the most abundant metabolites identified in these cryptococcomas was trehalose. The functional consequences of the trehalose pathway on the pathogenicity of Cryptococcus neoformans were hypothesized to be related to either its energy capabilities or its stress protection.

Studies in other yeasts have shown that genes containing STREs are negatively regulated by signaling genes such as the one encoding protein kinase A and, conversely, mediate stress induction by increased binding of the transcriptional factors Msn2 and Msn4 (38, 49). Since the trehalose pathway is highly regulated by a fine balance of synthesizing and hydrolyzing enzymes, the identification and understanding of the regulatory components of this stress response pathway are likely to yield important control features in the molecular pathogenesis of C. neoformans. However, the protein kinase A gene is not likely to be controlling the trehalose pathway stress response since pka mutants in C. neoformans are not TS (19), and similarly, we have disrupted a putative Msn2/4 homologue in C. neoformans, and it was found not to be TS (unpublished data). We also examined whether several of the other known signaling pathways or transcriptional activators for high-temperature growth in C. neoformans (CNA1, CPA1/2, or MGA1) might participate in the regulation of the trehalose pathway, but by expression analysis of the trehalose genes in these mutants there did not appear to be any apparent transcriptional links with these genes. It remains unclear how the trehalose pathway is regulated in C. neoformans and which TS mutants might use trehalose production as their mechanisms for high-temperature growth.

From previous studies on the interaction of the trehalose pathway and glycolysis in other fungi, we hypothesized that the tps1 mutant would grow poorly at 37°C on glucose due to the hyperaccumulation of sugar phosphates and poor control of glycolysis through loss of hexokinase regulation and reduced production of ATP. For instance, in the fermentative yeast Saccharomyces cerevisiae the tps1 mutant cannot grow on glucose at all environmental temperatures (27), but in contrast, the human pathogen, Candida albicans, develops a growth defect at high temperatures (42°C) (58), and lack of trehalose to stabilize proteins at high temperatures may influence this phenotype. A tpsl mutant strain in C. neoformans was TS as predicted for a respiratory human-pathogenic yeast. Its improved vegetative growth at high temperatures on galactose rather than glucose support the hypothesis that T6P does down-regulate hexokinase and its control of glycolysis and ATP formation. When T6P is absent the energy production of this yeast is likely compromised. In contrast, the Schizosaccharomyces pombe tps1 mutant strain is able to grow at elevated temperatures, but it cannot germinate spores (6). Interestingly, it has recently been demonstrated that TPS1 is also critical for sporulation in C. neoformans (37). However, the impact on glycolysis cannot completely explain attenuated virulence phenotype of the tps1 mutant. First, galactose does not completely restore high-temperature growth to wild-type rates. Second, Idnurum et al. have shown that a C. neoformans pex1 mutant without peroxisomes cannot utilize glucose for growth in vitro but that the mutant grows normally on galactose, suggesting an impact on hexokinase activity. However, this mutant is not TS and is fully virulent in mice (unpublished data). Third, besides the TS phenotype, the tps1 mutant at lower environmental temperatures did not display any apparent deficiencies in classic virulence factors such as melanin, capsule, or urease production, but the tps1 mutant was slightly more susceptible than the tps2 mutant to osmotic and oxidative challenges. However, the reduced protection against these conditions did not translate into a detectable intracellular growth defect, which has been linked with a lower-virulence phenotype in C. neoformans (1, 11, 12, 24).

Since from prior studies the tps1 mutant phenotype has been partially attributed to the inability of the cell to produce sufficient trehalose to enable the yeast cell to survive under the stress of high temperatures, we attempted to suppress the tps1 mutant growth defect by the addition of exogenous trehalose to the culture medium. We were not able to reconstitute the high-temperature growth phenotype. Several possibilities which would explain these results exist. These include the following: (i) C. neoformans does not possess an acid trehalase responsible for the mobilization of exogenous trehalose, (ii) we were simply unable to introduce sufficient trehalose into the cell to reconstitute the growth defect through a poor transport system, or (iii) the exogenous trehalose is either unable to reach the intracellular target proteins or is being degraded en route. In C. neoformans databases we have been unable to identify any open reading frames that might be predicted to be an acid trehalase gene (ATH1). However, H99 does show positive, albeit slow, growth on minimal medium using trehalose as the sole carbohydrate (generation time, 14.3 h on minimal medium versus 5.4 h on a glucose-containing minimal medium). This finding suggests that C. neoformans possesses an acid trehalase and/or specific trehalose transport system, but the yeast cell may be inefficient in transporting trehalose into the cell, and indeed trehalose can be toxic to the cell in sufficient quantities and inefficiency of transport may offer some protection. On the other hand another sugar, sorbitol, was able to completely suppress the TS phenotype of the tps1 mutant on glucose. With sorbitol's known ability to stabilize proteins and cell walls and membranes, these findings support the hypothesis that trehalose at high temperatures protects proteins and/or outer membrane structures in C. neoformans.

A second important feature of the tps1 mutant is that this mutant no longer produces T6P. T6P is an intermediate in trehalose biosynthesis that has been shown to possess regulatory potential (5, 26). First, in Saccharomyces cerevisiae it has been shown that T6P inhibits a hexokinase critical for control of glycolysis (5), and this enzyme has an impact on growth. Our results indicate that this interaction also occurs in C. neoformans, but T6P may have other regulatory capacities. For instance, in C. neoformans var. grubii (H99) in this study there was no impact on capsule or melanin production. However, we have recently found that in Cryptococcus gattii the tps1 mutant has both melanin and capsule defects that are not observed in a tps2 mutant. These findings suggest that recent evolutionary divergence in regulatory use of T6P in this species allows it to control all major virulence phenotypes (P. Ngamskulrungroj, M. Chayakulkeeree, D. L. Toffaletti, W. Meyer, and J. R. Perfect, Abstr. 16th Int. Soc. Hum. Anim. Mycoses, abstr. P-0674, 2006).

Although research efforts have been devoted to studying tps1 mutants in fungi, S. cerevisiae, S. pombe, C. albicans, and A. nidulans tps2 mutants are TS and accumulate T6P after heat shock (7, 21, 22, 47, 53, 59) and give further insights into T6P. In C. albicans, the tps2 mutant flocculates during stationary growth, while in A. nidulans, the mutant produces nonviable germination of spores through a deficiency in chitin production. Furthermore, both of these phenotypes can be suppressed by the addition of osmoprotectants to the culture medium (7, 53, 59). We did not observe a flocculation phenotype in a C. neoformans-like tps2 strain of C. albicans, but we did find different phenotypes between tps1 and tps2 mutants at low and high temperatures. The tps2 mutant was both more resistant to osmotic and oxidative challenges but conversely more sensitive to high environmental temperatures than the tps1 mutant. For instance, the tps2 mutant may demonstrate a TS phenotype since the cell is unable to synthesize trehalose similar to the tps1 strain; however, a block at this particular site in the trehalose pathway also results in an apparent toxic buildup of T6P within the cell that appears to produce a fungicidal phenotype during high-temperature exposure. There is more accumulation of T6P in the tps2 mutant at 37°C than at 30°C. As in the tps1 mutant, this severe phenotype can be suppressed with sorbitol, and in C. albicans the suppression of the tps2 mutant with sorbitol was suggested to be secondary to an osmoprotectant effect on the cell wall (59). In C. neoformans, it is not clear how sorbitol protects the tps2 mutant at high temperatures, but strict osmotic protection is less likely since the tps2 mutant is not particularly susceptible to osmotic challenges at lower temperatures. Finally, we were unable to reproduce any yeast toxicity when we added to medium concentrations of T6P that were up to 100-fold higher than those measured within yeast cells of both H99 at 37°C and the tps2 strain at 30°C. These findings suggest that a specific inhibitory compound targeted to this enzyme would elicit a fungicidal rather than fungistatic response in C. neoformans but that it must reach its intracellular target. It was previously reported that both trehalose and T6P are unable to enter some cells in sufficient quantity to produce toxicity (51).

Biologically, several prominent differences between the tps1 and tps2 mutants have emerged. First, the tps1 strain continues to show a virulence defect, even when high temperature is removed from the infection model. This difference is illustrated when the tps1 strain remains attenuated for virulence in the C. elegans system, while the tps2 strain does not. Second, the buildup of T6P in the tps2 mutant in vitro appears to produce a more severe viability defect than simply the lack of trehalose production in the yeast. These differences in phenotype survival illustrate vividly that despite the linkage of metabolic proteins with a common final product, the intermediary products may also have profound effects on the pathobiology of the microorganism. T6P has regulatory properties that also contribute to the virulence composite of C. neoformans along with its direct effect on trehalose production and specific impact on glycolysis. Third, it is clear that combination antifungal drug therapy remains a cornerstone of treatment for cryptococcal meningitis, and in fact, the combination of amphotericin B and flucytosine is the treatment of choice for this infection (45), but our findings suggest that both TPS1 and TPS2 are probably not linked to standard antifungal drug pathways, and it seems unlikely that inhibitors of either target will be synergistic with present antifungal agents.

In contrast to the trehalose-synthesizing mutants, the hydrolyzing neutral trehalase mutant, the nth1 strain, does not possess any known virulence phenotypes defects in C. neoformans. Despite extensive searches, we were unable to identify a putative second neutral trehalase gene, NTH2, by sequence comparison, which might complement NTH1 function, but there may be alternative methods for trehalose degradation in C. neoformans. Despite studies in other species that suggest the careful elimination of this sugar for the health of the cells, we hypothesize that the pathophysiological impact of the synthesis of trehalose in C. neoformans will be more transparent and important in its controls than its degradation.

A series of genes that are required for TS growth at 37°C have been identified, and these include genes that are involved in signaling pathways, specific cellular functions, and basic metabolism (43). Despite substantial knowledge of the genetics of high-temperature growth in C. neoformans, the specific mechanisms of how these genes support high-temperature growth remain uncertain. In this study, where both tps1 and tps2 mutants were incapable of producing trehalose and growth at 37°C, we have identified a potential mechanism necessary for high-temperature growth. It will be important to determine whether other genes or pathways that produce a TS phenotype rely on trehalose production as a basic stress protectant mechanism for high-temperature survival. Recently, Wang et al. found in a two-hybrid assay with cyclophilin A as a promoter trap that the TPS2 gene was interacting with the gene CPA1 (P. Wang, Abstr. 104th Gen. Meet. Amer. Soc. Microbiol., abstr. F054, 2004), which is known for its importance in high-temperature growth of C. neoformans. However, our transcriptional studies with mutants in several signaling pathways associated with high-temperature growth, including the cpa1 mutant, did not appear to be linked with the three trehalose genes. Furthermore, we examined the possibility that a calcineurin mutant (cna1) which has a TS phenotype and also dies slowly at high temperature might not be able to produce trehalose in the cell. In fact, we found that a cna1 mutant produced amounts of trehalose similar to that produced by H99 at 30°C.

It is apparent from our animal experiments and in vitro phenotypic studies that the trehalose synthesis genes, TPS1 and TPS2, show great promise as drug targets for C. neoformans infections. TPS2 shows value as an antifungal target since it is fungicidal in vitro and it is important in virulence in another medically important fungus, C. albicans (58). Similarly, the C. neoformans tps1 mutant was completely and rapidly eliminated from both immunosuppressed and immunocompetent hosts in this study. As noted for the severely immunosuppressed rabbit meningitis model, in which the poor inflammatory response approximates that of an AIDS patient, the tps1 mutant was rapidly and completely eliminated at an immunologically privileged body site. Furthermore, the trehalose pathway is simple and relatively well understood and is not found in mammals. Since the pathway is essential for C. neoformans survival in the mammalian host, it could be used either as a target for antifungal drugs or for creation of an attenuated TS strain for use as a live immunizing vaccine.