|Home | About | Journals | Submit | Contact Us | Français|
The trehalose pathway is essential for stress tolerance and virulence in fungi. We investigated the importance of this pathway for virulence of the pathogenic yeast Cryptococcus gattii using the highly virulent Vancouver Island, Canada, outbreak strain R265. Three genes putatively involved in trehalose biosynthesis, TPS1 (trehalose-6-phosphate [T6P] synthase) and TPS2 (T6P phosphatase), and degradation, NTH1 (neutral trehalose), were deleted in this strain, creating the R265tps1Δ, R265tps2Δ, and R265nth1Δ mutants. As in Cryptococcus neoformans, cellular trehalose was reduced in the R265tps1Δ and R265tps2Δ mutants, which could not grow and died, respectively, at 37°C on yeast extract-peptone-dextrose agar, suggesting that T6P accumulation in R265tps2Δ is directly toxic. Characterizations of the cryptococcal hexokinases and trehalose mutants support their linkage to the control of glycolysis in this species. However, unlike C. neoformans, the C. gattii R265tps1Δ mutant demonstrated, in addition, defects in melanin and capsule production, supporting an influence of T6P on these virulence pathways. Attenuated virulence of the R265tps1Δ mutant was not due solely to its 37°C growth defect, as shown in worm studies and confirmed by suppressor mutants. Furthermore, an intact trehalose pathway controls protein secretion, mating, and cell wall integrity in C. gattii. Thus, the trehalose synthesis pathway plays a central role in the virulence composites of C. gattii through multiple mechanisms. Deletion of NTH1 had no effect on virulence, but inactivation of the synthesis genes, TPS1 and TPS2, has profound effects on survival of C. gattii in the invertebrate and mammalian hosts. These results highlight the central importance of this pathway in the virulence composites of both pathogenic cryptococcal species.
Cryptococcus gattii, a member of the Cryptococcus species complex, is a pathogenic basidiomycetous yeast known to cause diseases mainly in immunocompetent humans and animals. It is environmentally associated with a variety of trees in tropical and subtropical climates (5, 32, 36). Recently, an outbreak of cryptococcosis occurred among apparently nonimmunocompromised humans and a variety of animal species on Vancouver Island, western Canada, due to the C. gattii VGII molecular type that has raised the importance of studying the virulence traits of this species. The outbreak strains belonged to two submolecular types, VGIIa and VGIIb (28). One strain of the major population of the outbreak, R265 (type VGIIa), was found to be the most virulent of several tested strains (17). In contrast, another member of the Cryptococcus species complex, Cryptococcus neoformans, is known to be an opportunistic pathogen with a worldwide distribution and a major cause of fungal meningoencephalitis in immunosuppressed patients, especially human immunodeficiency virus-positive patients (43).
After a dramatic increase in cases of cryptococcosis as a consequence of the AIDS epidemic (6), molecular biological studies of Cryptococcus species began to prosper, and they became model yeasts for studies of fungal pathogenesis. A number of virulence factors of the Cryptococcus species complex have been identified, which include (i) melanin synthesis, (ii) production of a polysaccharide capsule, (iii) urease, (iv) phospholipase production, and (v) the ability to grow at 37°C (6). Through molecular studies, dozens of genes have been linked to the virulence composites of both pathogenic cryptococcal species (44). In contrast to those characterizing the virulence of C. neoformans, molecular studies characterizing the virulence of C. gattii are meager. Only a few genes, such as those for phospholipase B, superoxide dismutase, a transcription factor, and protein kinases, have been studied in C. gattii in connection with pathogenesis (19, 23, 35, 38, 48, 54). Despite the close evolutionary relationships of these species, several studies have shown differences in gene expressions between the two species (25), supporting potential differences in gene regulations and the use of signaling pathways for virulence gene expression.
The simple ability of pathogenic yeasts to withstand severe environmental stresses is mandatory for their survival in humans. For the successful establishment of infection in the mammalian host, efficient protective high-temperature survival mechanisms are indispensable. The nonreducing disaccharide trehalose has been reported to be a vital protector of proteins and a biological membrane stabilizer under a variety of stresses, including heat, cold, starvation, desiccation, osmotic or oxidative stress, exposure to toxicants, and hypoxia in yeasts (10). The disaccharide has been found in bacteria and certain eukaryotic microorganisms, such as fungi, plants, insects, and invertebrates, but not in vertebrates (20). This pathway, unique in yeasts compared to mammals, suggests that trehalose and its pathway might be an attractive potential drug target if it is essential to a microbe's survival in the host (15, 52).
In fungi, trehalose has been shown to be rapidly induced to increase an organism's resistance to both external and internal stresses (20). Despite consisting of only a few metabolites and simple enzymatic steps, its regulatory organization and processes are surprisingly complex (52). Based on studies of Saccharomyces cerevisiae, at least seven genes are involved in the trehalose pathway (20). Trehalose is synthesized by two steps, starting from uridine-diphosphoglucose and glucose-6-P (29). The two catalyzing enzymes, trehalose-6-phosphate (T6P) synthase and T6P phosphatase, form a complex with two other proteins, Tsl1 and Tps3, in order to synthesize trehalose. The last proteins are associated with the stability of the complex and are highly homologous to each other (1). Trehalose is transported to wherever it is needed within the cell. The neutral trehalase, encoded by the NTH1 gene, will then hydrolyze the utilized trehalose to two molecules of glucose after it is transported back into the cytosol (39).
Studies to determine the connections between trehalose and virulence have previously been conducted on certain pathogenic fungi, including Candida albicans (51, 55, 57), Magnaporthe grisea (16), and more recently, C. neoformans (37, 46). These studies showed that not only the phenotype of high-temperature growth, but also cell wall integrity and hyphal formation, are controlled by this pathway. Ultimately, this network had an impact on fungal pathogenicity (37, 46).
In the current study, we examined the function of the trehalose synthesis pathway in the C. gattii strain R265, a highly virulent strain from the Vancouver Island outbreak (17), through mutations of its synthesizing genes, TPS1 and TPS2, and a degrading enzyme gene, NTH1. Deletions of the synthesizing genes in C. gattii revealed a profound defect on high-temperature growth in the generated mutants and, thus, attenuated virulence in the mammalian host. No apparent phenotype was found in the deletion of the hydrolyzing gene. In contrast to that in C. neoformans, melanin synthesis, capsule production, mating, cell wall integrity, and protein secretion defects were uniquely observed in the tpslΔ mutant of C. gattii, demonstrating that this pathway has evolved differently in the two closely related cryptococcal species. The knowledge obtained from the current study allows for the mapping of a proposed trehalose synthesis pathway in C. gattii (Fig. (Fig.11).
Strains used in this study are listed in Table Table1.1. Mutants and complemented strains of the wild-type strain R265, a C. gattii, VGIIa, serotype B, mating type α clinical isolate from the Vancouver Island outbreak (28), and CBS1930, a C. gattii, VGII, serotype B veterinary isolate from Aruba (3), were created by biolistic transformation (50). A wild-type strain of C. neoformans var. grubii (strain H99) (45) and its mutants (46) were retrieved from the culture collection of the Duke University Mycological Research Unit (Table (Table1).1). All strains were maintained on nonselective yeast extract-peptone-dextrose (YPD) agar (1% yeast extract, 2% peptone, 2% glucose, 2% agar).
Genomic DNA was extracted according to a previously described method (47). RNA was extracted by using the TRIzol reagent according to the manufacturer's protocol (Invitrogen, CA).
A BLASTN search of the C. gattii R265 genome database using homologous genes from C. neoformans H99 (26, 46) was performed for the identification of the genes TPS1, TPS2, NTH1, HXK1, and HXK2 (Cryptococcus gattii, serotype B; Sequencing Project, Center for Genome Research [http://www.broad.mit.edu]). Identification of potential motifs was completed using the Web-based MOTIF software (Genome Net, Japan [http://motif.genome.jp/]).
Nourseothricin (NAT1) and G418 (NEO) resistance genes were used as dominant selectable markers in a biolistic transformation protocol with a Bio-Rad model PDS-1000/He biolistic particle delivery system as previously described (50). The PCR constructs for the deletion mutants were created by overlapping PCR (18). Null mutants of the five genes of either strain R265 or CBS1390 were created by using the same set of primers to create 5′ and 3′ flanking sequences of each gene (Table (Table2).2). The flanking sequences of each gene were used to generate overlapping PCR products with the NAT1 marker cassette for strain R265 or the NEO marker cassette for CBS1930 according to a previously described protocol (9, 18). For the gene disruption, an overlap PCR construct was introduced into strain R265 or CBS1930 by biolistic transformation. Homologous recombination events were identified by PCR with external primers of each gene and confirmed by Southern blot analysis. An overlap PCR product with the NEO marker linked to each of the wild-type genes was generated for complementation of the R265 mutants. The PCR product of the entire gene was introduced directly into each mutant. Identification of the transcript for each gene was confirmed by Northern blot analysis. The double mutants were created by introducing HXK1 and HXK2 constructs into the R265tps1Δ and R265tps2Δ mutants by using the neomycin resistant cassette as a selectable marker.
Due to the occurrence of occasional colonies with spontaneous recovery from a temperature-sensitive (ts) phenotype for the trehalose pathway mutants, suppressor mutants of the R265tps2Δ (R265tps2ΔS) and H99tps2Δ mutants (H99tps2ΔS) were isolated by recovering a yeast colony from the R265tps2Δ mutant and one from the H99tps2Δ mutant, respectively, which were able to grow well at 37°C on YPD agar. Both strains were then subjected to subsequent characterizations along with other mutants of the pathway.
Metabolite production by the yeast strains was followed by 1H, 13C, and 31P nuclear magnetic resonance (NMR) spectroscopy. Levels of trehalose, T6P, and mannitol of H99, R265, and their tps1Δ, tps2Δ, and tps2ΔS mutants grown at temperatures of 30°C and 37°C were measured by NMR. Each yeast strain was incubated on Sabouraud dextrose agar at the two temperatures and harvested at 12 and 28 h for the NMR measurement as described previously in detail (46). 31P NMR spectra were used for the quantification of T6P, and 1H, 1H COSY spectra were used for the estimation of total trehalose. A more detailed analysis of the utilization of glucose was performed using 13C NMR spectroscopy. Yeasts were initially grown on Sabouraud dextrose agar, harvested, and then incubated for 12 to 24 h at temperatures of 30°C and 37°C in yeast nitrogen base (YNB) broth (Difco) supplemented with 25% deuterated water (D2O) containing 1% [1-13C]glucose buffered at pH 7.0 with 0.345% (wt/vol) 3-(N-morpholino)propanesulfonic acid (Sigma Chemical, St. Louis, MO).
NMR spectra were acquired at 30°C or 37°C using a Bruker Avance 500-MHz NMR spectrometer (Bruker Biospin, Rheinstetten, Germany) and a 5-mm (1H, 13C) inverse-detection dual-frequency probe as well as a 5-mm broadband probe according to a previous study (24). Relaxation delays of 5× T1 were typically used for the acquisition of all 13C NMR spectra for quantification. Proton decoupling was achieved by inverse gated decoupling (GARP). Signals were assigned by using 2-D correlation spectra. All NMR spectra were processed and quantified using the Bruker software TopSpin.
The growth rates were determined by using equal amounts of cells from the R265, R265tps1Δ, R265tps2Δ, and R265nth1Δ strains, grown with aeration at 30°C and 37°C in YPD broth. The cells were counted at periodic intervals, every 6 h, by plating serially diluted yeast cells with phosphate-buffered saline to calculate the number of CFU. Cell growth was plotted against time to determine the log phase. In addition, the strains R265, R265tps1Δ, R265tps2Δ, R265tps2ΔS, R265nth1Δ, R265tps1Δ::TPS1, R265tps2Δ::TPS2, H99, and H99tps1Δ were grown in 2 ml of YPD overnight at 30°C with constant shaking. The cells were then counted with a hema cytometer and diluted to 108 CFU/ml. Due to the fact that suppression of the ts phenotype by sugar substitution or sorbitol supplementation was evident from the study of strain H99 (46), 10-fold serial dilutions were made in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 2 mM KH2PO4, pH 7.4), and 5 μl of this was plated onto Sabouraud agar (2% peptone, 2% agar) containing 2% glucose, 2% galactose, and YPD agar with or without 1 M sorbitol or 0.5 M mannitol and grown at 30°C and 37°C. The cell integrity of the strains was tested by growing all strains on YPD agar containing 1 mg/ml caffeine, 7 mg/ml calcofluor white, or 7 mg/ml Congo red (21, 42).
Niger seed, dopamine, and caffeic acid agars were used to test for melanin production (31). The strains R265, R265tps1Δ, R265tps2Δ, R265tps2ΔS, R265tps1Δ::TPS1, R265tps2Δ::TPS2, H99, and H99tps1Δ were diluted and plated on the selective media and grown at 30°C. The abilities to produce melanin were compared at 72 h by examining for brown colonies. Capsule production was determined by growing each strain in different capsule-inducing conditions as follows: Dulbecco's modified Eagle's medium either with 10% fetal bovine serum or in the presence of 5% CO2 at 37°C. Yeast cell preparations were performed according to a previously described established protocol (22, 56), and capsule production was determined by visually comparing the capsule sizes at 72 h using light microscopy and India ink staining.
Mating assays were carried out on V8 medium (5% [vol/vol] V8 juice, 2 mM KH2PO4, 4% [wt/vol] agar) (33). The ability to mate was tested against the C. gattii, VGII, mating type a strain CBS1930. Null mutants of all three genes were created in strain CBS1930 in order to test for bilateral mating ability. Strains were pregrown on YPD agar for 2 days, and yeast cells were removed and patched onto solid mating medium either alone or mixed with the mating type a tester strain (CBS1930 or its mutants). Plates were incubated at room temperature in the dark for 3 weeks and examined by light microscopy for filamentation and basidiospore formation.
To determine the extracellular protein secretion ability, the amount of protein secreted into broth by the strains R265, R265tps1Δ, R265tps2Δ, R265tps2ΔS, R265tps1Δ::TPS1, R265tps2Δ::TPS2, H99, and H99tps1Δ was measured according to a previously described method (12). Briefly, 104 yeast cells/ml of each strain were incubated in 2% glucose YNB broth (49) and grown to saturation at 30°C under constant agitation. Yeast cells were pelleted, and supernatants were used for protein quantification. Supernatants were dialyzed and stained with Coomassie blue. The protein concentration was determined in triplicate by comparing the optical density of 550 nm to an albumin standard curve.
Each strain was initially grown in YPD broth at 30°C for 24 h and then pelleted. Supernatant was discarded. The pellet was resuspended with fresh YPD broth and grown at 30 or 37°C for 1 h. The yeast cells were again pelleted and then lyophilized. RNA was extracted using 0.5-μm glass beads and vortexed in TRIzol reagent (Invitrogen, CA) according to the manufacturer's instructions. Transcription of TPS1, TPS2, HXK1, and HXK2 was determined at 37°C and that of LAC1, LAC2, and MPK1 was determined at 30°C by real-time PCR as described previously (11), using the primers listed in Table Table2.2. DNA and mRNA sequences of the TPS1 loci of the strains R265, R265tps2Δ, and R265tps2ΔS were determined by using the primers listed in Table Table22.
Growth of R265hxk1Δ, R265tps1Δhxk1Δ, R265tps2Δhxk1Δ, R265hxk2Δ, R265tps1Δhxk2Δ, R265tps2Δhxk2Δ, and R265hxk1Δ::HXK1 using glucose or galactose as a sole carbon source was tested on 2% glucose and 2% galactose YNB agar by using the dilution method as described above. Growth rates were determined in 2% glucose and 2% galactose YNB broth at 37°C. Yeast cell concentrations for each isolate were measured as the optical density at 600 nm at 0, 3, 9, 22, 32, 48, and 72 h as described previously (26).
In vivo characterization of the Δtps1, Δtps2, and Δnth1 mutants using a Caenorhabditis elegans/Cryptococcus gattii model was performed as previously described with C. neoformans (46). In brief, approximately 40 to 50 young adults of the standard C. elegans strain N2 Bristol, pregrown on a lawn of Escherichia coli OP50, were transferred to a lawn of each of the following C. gattii strains studied: R265, R265tps1Δ, R265tps1Δ::TPS1, R265tps2Δ, R265tps2Δ::TPS2, R265nth1Δ, and R265nth1Δ::NTH1. The cocultured plates were incubated at 25°C and examined for viability at 24-h intervals with a dissecting microscope. Plotting of the killing curves and estimation of differences in survival (log-rank and Wilcoxon tests) with the Kaplan-Meier method were performed using Stata 6 statistical software (Stata, College Station, TX).
A/JCr mice were inoculated intranasally with 105 yeast cells. The mice were housed and supplied with food and water ad libitum. Groups of 15 mice were inoculated either with strain R265 or the R265tps1Δ, R265tps2Δ, and R265nth1Δ mutants. Quantitative cultures of the lungs and brains were used to compare tissue burden between groups of mice (five animals each) that were sacrificed on days 4, 7, and 14 following inoculation. The lungs and brains were homogenized, diluted, and plated onto YPD with chloramphenicol (100 μg/ml) in order to determine the number of CFU/g of tissue and then plotted against time to determine the progression of the infection. The animal studies were repeated with BALB/c mice. The mice were inoculated with the strain R265, its mutants (R265tps1Δ, R265tps2Δ, and R265tps2ΔS), and reconstituted strains (R265tps1Δ::TPS1 and R265tps2Δ::TPS2). Mice were sacrificed on days 7 and 14 for quantitative yeast counts. Mean colony count statistical comparisons were made between groups with a Student t test, using SPSS15 software (LEAD Technologies, Inc.). All animal experiments were conducted in accordance with the approvals of the respective animal ethics committees.
The sequences for genes TPS1, TPS2, and NTH1 have been deposited in GenBank under accession numbers EU399551, EU399552, and EU399553, respectively. The sequences for the genes HXK1 and HXK2 have been deposited in GenBank under the accession numbers FJ716629 and FJ716630, respectively.
Genes TPS1, TPS2, and NTH1 were identified as putative 2,519-bp, 3,287-bp, and 3,313-bp open reading frames, respectively, with approximately 89% sequence similarity to the C. neoformans strain H99 for each gene, encoding predicted 673-, 990-, and 877-amino acid proteins, respectively. These genes appear to possess multiple intronic structures typical of cryptococcal genes. Although formal functional studies of the promoters were not performed, stress response (AGGGG motifs) and heat shock elements (repetitive NGAAN sequences) were noted in the sequences of the upstream area from the start codon for all three genes. Similar elements had been previously observed in the C. neoformans gene sequences (46).
To determine the effects of gene disruption on metabolites in the pathway of the targeted genes, metabolite production from [1-13C]glucose utilization of each mutant and the wild type was quantified and is shown in Fig. Fig.22 and Table Table3.3. As expected, trehalose was not detected in both the R265tps1Δ and R265tps2Δ mutants. An intermediate product of the pathway, T6P, was detected in the wild-type strain R265, and the concentration was determined to be <2 μmol/104 cells due to a low level of accuracy at these low concentrations. However, it was not detected in the R265tps1Δ mutant due to the absence of its synthase gene. On the other hand, the R265tps2Δ mutant showed a substantial accumulation of T6P due to the absence of its phosphatase gene (TPS2). These findings replicated what had been found in C. neoformans (strain H99) (46). Interestingly, neither trehalose nor T6P could be detected in either the R265tps2ΔS or H99tps2ΔS suppressor at high-temperature growth. This demonstrates that a functioning trehalose pathway is not always essential for the acquisition of a high-temperature growth phenotype. Furthermore, differences between the strains H99 and R265 and their mutants were also detected in their metabolites. While comparable levels of mannitol production were evident in strain H99 and its mutants, mannitol production was substantially increased in the mutants of strain R265 compared to the wild-type strain (Fig. (Fig.2;2; Table Table3).3). In addition, the production of trehalose is completely suppressed in the R265tps1Δ mutant, but some remaining traces of trehalose were detectable in the 13C NMR spectrum of the H99tps1Δ mutant (Fig. (Fig.2;2; Table Table33).
As the trehalose synthesis pathway is known to control glycolytic flux in other fungi, metabolites of other pathways using glucose as a precursor were measured. 13C labeling was incorporated into [1,6-13C]mannitol, [1,3-13C]glycerol, and [6-13C]-labeled hexoses (most likely [6-13C]glucose or [6-13C]trehalose). These compounds indicate metabolism and catabolism of [1-13C]glucose after conversion into glyceraldehyde-3-P. Interestingly, no changes in [1,3-13C]glycerol levels were observed in the mutants compared to the wild type. 13C labeling was also incorporated into acetate, ethanol, and fatty acids (Fig. (Fig.2).2). A marginally increased incorporation of 13C labeling in fatty acids was observed for the R265tps1Δ and H99tps1Δ mutants. For H99tps2ΔS and R265tps2ΔS, most 13C labeling was incorporated in [1,6-13C]mannitol, but in contrast to the other mutants, they did not show incorporation of the 13C label in glycerol, [6-13C]-labeled hexoses, fatty acids, ethanol, or acetate (Fig. (Fig.22).
As the trehalose synthesis pathway is essential for the high-temperature growth of several yeast species (20, 46), a test for their ability to grow at a human physiological temperature (37°C) was conducted. Growth of the wild type, R265, and that of its mutants were compared on YPD medium (2% glucose, 2% peptone, and 1% yeast extract). Similar to the H99 mutants (46), the R265tps1Δ and R265tps2Δ mutants exposed to 37°C either could not grow (fungistasis) or died, respectively (see Fig. S1 in the supplemental material). On the other hand, a 37°C growth defect was not found in R265nth1Δ. Furthermore, while the R265tps1Δ mutant showed a slightly slower growth rate than the wild type at 30°C (data not shown), no defective growth rate was observed in the R265tps2Δ and R265nth1Δ mutants and in all reconstituted strains of the R265 mutants at 30°C (Fig. (Fig.33).
In S. cerevisiae, the regulatory molecule T6P controls hexokinase II (HXKII) expression, which in turn affects glycolytic flux and produces the ts phenotype (20). For example, bypassing HXKII by replacing glucose with other sugars (i.e., galactose) in the medium can rescue the ts phenotype of trehalose mutants in S. cerevisiae (20) and C. neoformans (46). Thus, a growth comparison was performed on Sabouraud agar containing different sugars (2% peptone and 2% agar with 2% glucose or 2% galactose). In C. gattii, replacement of glucose in the medium by galactose did suppress the high-temperature growth defect of the R265tps2Δ mutant, but it had no impact on the R265tps1Δ mutant (Fig. (Fig.33).
Moreover, since a correlation between the presence of T6P and cell wall integrity at high temperatures had previously been reported (4), an osmotic stabilizer, either 1.0 M sorbitol or 0.5 M mannitol, both of which were able to suppress the ts phenotype of C. neoformans cell wall mutants (21), was supplemented in the medium. Similar to results with the galactose substitution, only the R265tps2Δ mutant was rescued (Fig. (Fig.33).
Since TPS1 and TPS2 genes were essential for heat tolerance, their transcriptional regulations at a human physiological temperature were examined. TPS1 and TPS2 genes of the wild-type strain R265 were induced approximately twofold (P values of <0.039 and 0.009, respectively) when exposed to 37°C compared to 30°C at a 1-h incubation in YPD (data not shown).
The ts suppressor of the R265Δtps2 mutant (R265tps2ΔS) occurred at a rate of 1 to 10 per million CFU when grown on glucose-rich medium at 37°C. To elucidate the mechanism of ts phenotypic suppression in the R265tps2ΔS strain, sequence and transcriptional comparison studies with the R265tps2Δ strain were undertaken. A Northern blot analysis confirmed the loss of the TPS2 transcript in the R265Δtps2S mutant. In addition, the TPS1 gene was sequenced in the suppressor strain and was found to have no mutations in its open reading frame. Reverse transcription-PCR for the TPS1 transcript detection in the suppressor strain showed that it was similarly expressed as in the wild-type R265. Moreover, mRNA levels of TPS1 and TPS2 for several R265tps2ΔS clones were identical to those for R265tps2Δ. These results suggested that the high-temperature growth adaptation of the R265tps2ΔS strain occurred mechanistically outside the trehalose pathway. This hypothesis was further supported by our hexokinase transcriptional studies of the suppression strain, which revealed a significant suppression of HXK2 (approximately 15 times less than R265 and R265tps2Δ, with no significant differences in HXK1 expression) (Fig. (Fig.4)4) and emphasizes the critical importance of glycolytic control during high-temperature growth.
Regulation of hexokinase function by the trehalose synthesis pathway has been shown in other fungi (20), and two hexokinase genes, HXK1 and HXK2, were previously characterized in C. neoformans (26). We identified the C. gattii genes HXK1 and HXK2 as putative 2,658-bp and 2,132-bp open reading frames, respectively, with approximately 85% sequence similarity to the C. neoformans strain H99 for each gene, encoding predicted 591-amino acid and 489-amino acid proteins, respectively. The following mutants were created: R265hxk1Δ, R265tps1Δhxk1Δ, R265tps2Δhxk1Δ, R265hxk2Δ, R265tps1Δhxk2Δ, R265tps2Δhxk2Δ, and R265hxk1Δ::HXK1.
Unlike HXK2 of C. neoformans, which is required for galactose utilization (26), HXK2 of C. gattii is not essential for either glucose or galactose utilization, as R265hxk2Δ had no growth defect when grown on 2% glucose or 2% galactose YNB (Fig. 5A and B). However, HXK1 of C. gattii was required for utilizing galactose as a sole carbon source. Reconstitution of strain R265hxk1Δ::HXK1 recovered growth on galactose to the wild-type level (data not shown). As predicted from the significant transcriptional suppression of HXK2 in the R265tps2ΔS strain, deletion of HXK2 suppressed the ts phenotype of both R265tps1Δ and R265tps2Δ strains (Fig. (Fig.5A).5A). R265tps2Δhxk2Δ growth at 37°C was similar to that of R265tps2ΔS (Fig. (Fig.5B).5B). As seen by a slight difference in colony size on the solid medium (data not shown), both R265tps2ΔS and R265tps2Δhxk2Δ also showed a slower growth rate than the wild-type strain in liquid medium. In addition, deletion of HXK1 also suppressed the ts phenotype of R265tps1Δ. However, the ts suppression by the hexokinase deletion of R265tps1Δ was lower compared to its effect on R265tps2Δ (Fig. 5A and B).
An effect on other unknown virulence factors of TPS1 beside temperature sensitivity was previously hypothesized in the C. neoformans strain H99 because of the hypovirulence of the H99tps1Δ mutant in the room temperature C. elegans model (46). Therefore, we tested expression of several known virulence factors in our R265tps1Δ mutants. The R265tps2ΔS mutant was also tested due to its loss of T6P, similar to R265tps1Δ. Unlike the H99tps1Δ mutant, the R265tps1Δ and R265tps2ΔS mutants showed a severe decrease in melanin production and capsule size in all tested induction media (Fig. (Fig.66 and and7).7). Impaired cell wall integrity was observed when the R265tps1Δ mutant was grown on known cell wall-perturbing agents (caffeine, calcofluor white, and Congo red), while the R265tps2ΔS mutant showed sensitivity only to calcofluor white and slight sensitivity to Congo red (Fig. (Fig.8).8). Defects in filamentation and sporulation were observed during bilateral crossings (mutant × mutant) of the R265tps1Δ mutant with the CBS1930tps1Δ mutant but not in unilateral crossings (mutant × wild type). None of these phenotypic virulence or morphological defects was observed in the R265tps2Δ and R265nth1Δ mutants or the reconstituted strains for all mutants (data not shown).
To investigate a possible mechanism of phenotypic differences, and since the extracellular protein secretion in C. gattii contains several proteins potentially indispensable for tissue invasion and growth, such as proteases and phospholipases (6), the amount of total protein secreted by each strain at 30°C growth to saturation was measured. The secreted protein concentrations of the R265tps1Δ and R265tps2ΔS mutants were significantly less than that of the wild-type strain R265 (P < 0.001). In contrast, the secreted protein concentrations of the R265tps2Δ and H99tps1Δ mutants were higher than that of the strain R265 or H99, respectively (P < 0.001) (Fig. (Fig.99).
In order to genetically link the trehalose pathway with the virulence phenotypes of melanin and cell wall integrity, transcriptions of the major cryptococcal laccase genes (LAC1 and LAC2, required for melanin production) and the MAP kinase gene (MPK1, required in the PKC pathway for cell wall integrity) were examined while strains were grown in YPD medium at 30°C. In the R265tps1Δ and R265tps2ΔS strains with no detectable T6P formation, the gene transcripts were reduced significantly (P < 0.05) compared to the wild-type R265 when grown in YPD at 30°C. This significant reduction was not seen in the R265tps2Δ mutant (P > 0.05) (Fig. (Fig.10).10). These transcription observations along with the phenotypic analysis of the mutants support T6P as a potential regulator of genes involved in both melanin production and cell wall integrity in C. gattii, which appears to be unique for C. gattii compared to C. neoformans.
In the C. elegans pathogenesis model, in which high-temperature growth is not a required survival factor for a pathogen, an attenuation of virulence was observed for both the R265tps1Δ and R265tps2Δ mutants. The R265tps1Δ mutant was substantially less virulent than the wild-type strain R265 and its reconstituted strains (P < 0.001). The R265tps2Δ mutant showed only a slight decrease in virulence in the C. elegans model compared to strain R265 (P < 0.001), and the hypovirulence status of the R265tps2Δ mutant was not found in a similar mutant in C. neoformans (46) (Fig. (Fig.11).11). The R265nth1Δ mutant was not attenuated in the C. elegans model. Restoration of wild-type virulence was observed in all of the reconstituted strains (data not shown).
To test for the overall impact of these genetic loci on the total virulence composite, murine inhalation models for yeast tissue survival were used. The R265tps1Δ, R265tps2Δ, and R265tps2ΔS mutants were found to be severely attenuated in the mice. These strains revealed a profound loss of virulence in that there was a rapid loss of viable yeasts in the host tissue. All yeasts were eliminated from the lungs and brains of mice by day 7 of the infection in both BALB/c and A/JCr murine models (P < 0.001). The R265nth1Δ mutant and both reconstituted strains, R265tps1Δ::TPS1 and R265tps2Δ::TPS2, remained fully virulent, with tissue yeast counts comparable to those of the wild-type strain R265 (Fig. (Fig.1212).
The ability to grow at human physiological temperatures has long been known as an essential virulence factor for any pathogenic microorganism for mammals. The trehalose synthesis pathway has been extensively reported in several fungal species as a major pathway controlling heat tolerance in yeasts (20). Therefore, the pathway has been validated to be critically necessary for the virulence potential of several pathogenic fungi (16, 46, 57). In the present study, the trehalose pathway in C. gattii had the same organization as in C. neoformans with the primary genes of T6P synthase (TPS1), T6P phosphatase (TPS2), and a neutral trehalase (NTH1), with similar sequence structures and impact on cellular T6P and trehalose levels. Despite a similar impact on high-temperature growth and response to mammalian host stresses for TPS1 and TPS2, several differences in their regulations, connections, and functions of genes in this pathway between these closely related species were found.
First, there was a change in mannitol production (two- to threefold) only in the R265tps1Δ and R265tps2Δ mutants of C. gattii compared to that in the wild-type strain. Production of mannitol might be a cellular protective mechanism as an oxidative scavenger (8) and/or osmotic stabilizer (7) for the yeast when trehalose production is compromised. For example, in Saccharomyces cerevisiae, stress protectants such as trehalose and glycerol are induced to protect against osmotic stresses (27), and thus mannitol might also be used in C. gattii as a factor for protection against environmental stresses. This hypothesis for mannitol use in stress protection was further emphasized by the suppression of the R265tps2Δ ts phenotype after further exogenous exposure to mannitol or sorbitol supplementation in the media.
Second, only the R265tps2Δ and not the R265tps1Δ ts phenotype could be suppressed by galactose substitution for glucose, while in contrast, the ts phenotype of both H99tps1Δ and H99tps2Δ could be suppressed by galactose substitution. This finding implies that TPS1 and TPS2 may control glycolytic fluxes in slightly different manners within the two species.
Third, although the trehalose pathway appeared to control virulence factors other than growth at 37°C in C. neoformans based on the outcomes in the worm model, specific identification of virulence factors was not found in the in vitro studies (46). On the other hand, several known virulence factors were clearly reduced in the C. gattii R265tps1Δ mutant and similarly in the R265tps2ΔS mutant.
These profound physiological and genetic network differences provide additional support for the rise of C. gattii to a separate species from C. neoformans (34). Our results emphasize that although these two cryptococcal species produce similar disease processes in humans and animals, some of their genetic connections are different, and further genetic differences may support their known diversity in ecology and their prevalence of disease in certain host risk groups (6).
Interestingly, we observed the relatively frequent formation of C. gattii R265Δtps2 suppressors with persistent defects in several virulence factor phenotypes. The discovery of these suppressors, which retained their ability to grow well at 37°C without producing any trehalose or T6P, suggests that the effects of intracellular trehalose on the support of high-temperature growth are not absolute. Moreover, the R265tps2Δ suppressors, instead of possessing high T6P levels, had both undetectable T6P and trehalose levels and a phenotype similar to that of the R265tps1Δ mutant despite an apparently normal TPS1 gene. This observation suggests an activation of an alternative metabolic pathway or alteration of a sugar transporter(s), resulting in either glucose-6-phosphate being no longer available for the Tps1 enzyme or TPS1 being posttranscriptionally modified. These possible mechanisms are supported by the loss of incorporated 13C in several end products of glucose-6-phosphate, including glycerol, 6-13C-labeled hexoses, fatty acids, ethanol, or acetate, in the R265tps2ΔS mutants, which indicates a general downregulation of the glycolytic pathway, which is needed to enable yeast growth at high temperatures in the presence of glucose. In support of this hypothesis, a transcriptional analysis showed a marked reduction of transcription of HXK2, which encodes one of the first enzymes in the glycolytic pathway, of the tps2ΔS mutant.
An interaction of the trehalose pathway and glycolysis has been demonstrated in several fungi (20). Since in C. gattii, the ts suppressor of the R265tps2Δ mutant in the presence of glucose was associated with HXK2 downregulation, we further characterized the connection of the two hexokinases, HXK1 and HXK2, with the trehalose pathway. For instance, the loss of the ts phenotype of R265tps2Δhxk2Δ directly supports a glycolytic link with the trehalose pathway and its impact on high-temperature growth through HXK2, possibly by reducing glucose-6-phosphate availability to make the toxic buildup of T6P in the R265tps2Δ mutant. On the other hand, TPS1 and its ability to produce T6P appear to control high-temperature growth and glycolysis through either HXK1 or HXK2. However, since ts phenotypic suppression of either tps1Δ or tps2Δ through the interruption of the hexokinase genes was not complete, additional trehalose pathway effects on other pathways evolved in high-temperature growth are suggested. One possibility is the PKC pathway, which is known to control both cell wall integrity and growth at 37°C in C. neoformans (30). We found that the gene for last kinase of this pathway (MPK1), in fact, was downregulated in both of the R265tps1Δ and R265tps2ΔS mutants. This connection could explain the possibility that another pathway is linked with trehalose for both the ts and cell wall integrity phenotypes. This linkage is also supported by the fact that sorbitol, a membrane stabilizer (21, 30), could suppress the ts growth of R265tps2Δ. However, a similar suppression was not found for the C. gattii R265tps1Δ mutant, in which growth was not recovered by the addition of sorbitol to the medium. This finding suggests that T6P, which is absent in the R265tps1Δ mutant, may regulate or interact with other stress pathways besides the PKC pathway.
It has been known that T6P is a major regulatory molecule for other metabolic pathways in yeasts (2, 53). Like that in C. neoformans (46), toxicity of intracellular T6P accumulation to the cryptococcal cell was vividly shown at 37°C in the R265Δtps2 mutant and prevented further studies of TCP's importance as a regulatory molecule. Therefore, we searched for comparative metabolic alterations in the R265tps1Δ, R265tps2Δ, and R265tps2ΔS mutants. Apart from trehalose and mannitol, in the R265tps1Δ and R265tps2Δ mutants, T6P was the only metabolite showing a major difference from the wild-type strain R265. Moreover, several in vitro virulence phenotypes, such as melanin production and cell wall integrity, together with the transcription of genes related to melanin production, LAC1 and LAC2, and MAP kinase gene MPK1, were significantly reduced in the R265tps1Δ and R265tps2ΔS mutants without T6P. In contrast to C. neoformans, C. gattii in the absence of TPS1 and/or T6P has profound defects in several virulence phenotypes, including melanin synthesis, capsule production, cell wall integrity, and fertility, and it possesses a general protein secretion defect. This impact of the pathway on other virulence factors is similar to the findings of C. neoformans, in which the cyclic AMP pathway has been shown to control both melanin and capsule production (13). Although we have not yet found a direct link of trehalose and the cyclic AMP pathways in C. gattii, it is clear that both pathways in C. gattii converge on the control of several virulence phenotypes. These observations suggest that T6P is a major regulatory molecule in C. gattii. Similar findings have been made in the plant pathogenic fungus Magnaporthe grisea, where TPS1 and T6P are able to regulate NADPH levels, the oxidative pentose phosphate pathway, expression of nitrogen source utilization, and direct virulence genes (53). It is clear that in C. gattii, TPS1 provides the gateway for all the major virulence phenotypes and that this pathway (TPS1 and TPS2) also impacts carbon utilization.
The total impact of the deletion of either TPS1 and TPS2 on the pathobiology of C. gattii for the mammalian host was hypothesized to be severe from our mechanistic studies. While the TPS1 block was found to be fungistatic for in vitro growth at 37°C, its impact on multiple virulence factors and pathways for fitness of C. gattii predicted a rapid loss of survival in the mammalian host. However, in the R265tps2Δ mutant we observed in vitro death at 37°C, with a propensity to select for colonies that were resistant to high-temperature lethality in glucose-containing environments. Therefore, we first tested the trehalose mutants in the C. elegans model in which high-temperature growth of yeasts is not a factor in their pathobiology. Similar to the H99tps1Δ mutant in C. neoformans, the C. gattii R265tps1Δ mutant was profoundly attenuated, reflecting multiple areas of impact for this locus on the pathobiological characteristics of C. gattii. In fact, the C. elegans model does not even detect survival differences in capsule and melanin mutants (E. Mylonakis, personal observations), thus confirming that TPS1 impacts multiple virulence characteristics other than just the three major virulence factors. Unlike in the C. neoformans studies (46), there was a slight attenuation of the R265tps2Δ mutant in the worm, but it is likely that the ts phenotype still remains the major feature of the R265tps2Δ mutant's ability to produce disease in the mammalian host.
In the mammalian model, we used two strains of mice to present different immune responses to the mutants: A/JCr mice with a complement deficiency and immunocompetent BALB/c mice. In both sets of mice, we found that the C. gattii R265tps1Δ, R265tps2Δ, and R265tps2ΔS mutants were severely attenuated with a rapid loss of yeast viability in the lungs and a lack of consistent brain involvement, which is in concordance with the C. neoformans H99tps1Δ mutant (46) in both host genetic backgrounds. These results compare favorably with the most severely attenuated cryptococcal mutants that we have studied in mice (44). The animals were able to completely eliminate the yeasts from both clinically important tissue sites. These findings support the fact that either TPS1 or TPS2 when used as an antifungal target would be considered fungicidal even with the frequent development of ts suppressors during the block of TPS2 (resistant to high temperatures). It is very unlikely that this process would become a clinically important drug resistance mechanism for the yeast during infection since the virulence of these suppressors was also severely attenuated.
In contrast to the profound effects of blocking trehalose synthesis, the neutral trehalase (NTH1) had no apparent impact on the C. neoformans or C. gattii virulence phenotypes or their virulence composite. In contrast to diminished thermotolerance of both neutral trehalase mutants of S. cerevisiae (39, 40), the R265nth1Δ mutant of C. gattii had no ts growth phenotype. The neutral trehalase mutants of C. neoformans (46) and C. albicans (14) also did not show ts growth patterns. Furthermore, trehalose levels were not increased in the trehalase mutants of C. albicans, and an arbuscular mycorrhizal fungus showed no changes in neutral trehalase RNA accumulation in response to stress treatments (41). Since the C. gattii R265nth1Δ mutant can grow on trehalose-only media (data not shown), it is possible that there are other NTH gene homologues in C. gattii that have not yet been identified. However, at present the synthesis part of the trehalose pathway appears to profoundly impact the virulence composite of both pathogenic cryptococcal species and not its degradation component.
Our results show that the trehalose synthesis pathway in C. gattii represents a complex network of cross talking with other pathways and is a major site for control of the virulence composite. From this study, it is clear that C. gattii and C. neoformans have a conserved trehalose pathway, but both pathways have evolutionarily diverged in certain connections that are manifested in some phenotypic differences between the two sibling species. However, the uniqueness of this pathway to fungi compared to mammals and its profound impact on the ability of pathogenic Cryptococcus species to produce disease support the potential of this pathway at several steps to become an ideal antifungal drug target(s).
We thank the following members of the Duke University Mycological Research Unit: Joseph Heitman, James Fraser, Gary Cox, Tom Rude, and Alexander Idnurm, and thank the members of Andrew Alspaugh's laboratory for their scientific advice.
This work was supported by Public Service Grants AI28388 and AI73896.
Editor: A. Casadevall
Published ahead of print on 3 August 2009.
§Supplemental material for this article may be found at http://iai.asm.org/.