Rewiring of Signaling Networks Modulating Thermotolerance in the Human Pathogen Cryptococcus neoformans

Abstract

Thermotolerance is a crucial virulence attribute for human pathogens, including the fungus Cryptococcus neoformans that causes fatal meningitis in humans. Loss of the protein kinase Sch9 increases C. neoformans thermotolerance, but its regulatory mechanism has remained unknown. Here, we studied the Sch9-dependent and Sch9-independent signaling networks modulating C. neoformans thermotolerance by using genome-wide transcriptome analysis and reverse genetic approaches. During temperature upshift, genes encoding for molecular chaperones and heat shock proteins were upregulated, whereas those for translation, transcription, and sterol biosynthesis were highly suppressed. In this process, Sch9 regulated basal expression levels or induced/repressed expression levels of some temperature-responsive genes, including heat shock transcription factor (HSF1) and heat shock proteins (HSP104 and SSA1). Notably, we found that the HSF1 transcript abundance decreased but the Hsf1 protein became transiently phosphorylated during temperature upshift. Nevertheless, Hsf1 is essential for growth and its overexpression promoted C. neoformans thermotolerance. Transcriptome analysis using an HSF1 overexpressing strain revealed a dual role of Hsf1 in the oxidative stress response and thermotolerance. Chromatin immunoprecipitation demonstrated that Hsf1 binds to the step-type like heat shock element (HSE) of its target genes more efficiently than to the perfect- or gap-type HSE. This study provides insight into the thermotolerance of C. neoformans by elucidating the regulatory mechanisms of Sch9 and Hsf1 through the genome-scale identification of temperature-dependent genes.

Hsf1, Sch9, transcriptome analysis, high temperature, Cryptococcus neoformans

RESPONSE and adaptation to their host’s physiological temperature, 37°, are key virulence attributes for most human fungal pathogens that infect from natural environments at ambient temperature. Cryptococcus neoformans is an example of such a pathogen. This basidiomycetous fungus exists in diverse environment niches and generates infectious yeast cells or spores through bisexual and unisexual differentiation, which are inhaled through the respiratory tract in the host and initially colonize the lungs (Idnurm et al. 2005; Lin and Heitman 2006). C. neoformans is subsequently disseminated into the bloodstream, breaches the central nervous system, and eventually elicits fatal meningoencephalitis, responsible for an estimated hundreds of thousands of deaths annually (Park et al. 2009). During this infectious process, the ability to sense and adapt to temperature upshift are crucial factors for the pathogen to establish the initial colonization of the lungs. Therefore, most mutants that are severely defective in thermotolerance tend to be highly attenuated in virulence. Due to such reasons, signaling cascades governing thermotolerance in fungi have been intensively investigated toward their exploitation as potential antifungal drug targets (Bahn et al. 2007; Bahn and Jung 2013). In C. neoformans, temperature-sensing signaling cascades include the calmodulin/calcineurin (Fox et al. 2001; Kraus et al. 2005), Ras (Alspaugh et al. 2000; Nichols et al. 2007), HOG (Bahn et al. 2005, 2006), Msi1-like protein (Yang et al. 2012), protein kinase C 1 (Pkc1)/Mpk1 MAPK (Kraus et al. 2003; Gerik et al. 2005), and unfolded protein response (UPR) (Cheon et al. 2011) pathways.

Unlike other temperature-sensing signaling components described above, a protein kinase, Sch9, suppresses C. neoformans thermotolerance (Wang et al. 2004). SCH9 deletion increases C. neoformans thermotolerance. Hence, the sch9∆ mutant survives well up to 41° in contrast to the wild-type (WT) strain (Wang et al. 2004; Kim et al. 2009). Sch9 also negatively regulates the production of the polysaccharide capsule, which is also a key virulence factor for C. neoformans (Wang et al. 2004). Nevertheless, the virulence of the sch9∆ mutant is attenuated, possibly due to a greater susceptibility to oxidative stress (Wang et al. 2004; Kim et al. 2009). SCH9 expression is induced by oxidative stress (Kim et al. 2009). The role of Sch9 in thermotolerance and capsule synthesis appears to be dependent on protein kinase A (PKA), although the mechanistic connection of Sch9 to the cAMP signaling pathway is not clear (Wang et al. 2004). Therefore, how Sch9 governs C. neoformans thermotolerance remains unknown.

The functions of Sch9 have been well characterized in the model yeast Saccharomyces cerevisiae. Sch9 is a Ser/Thr protein kinase that plays important roles in nutrient sensing, stress response, and modulation of the chronological lifespan. In response to nutrients, the target of rapamycin complex 1 (TORC1), which is a key nutrient sensor in eukaryotes that governs ribosome biogenesis and translation, directly phosphorylates Sch9. Sch9 appears to be required to maintain the optimal activity of RNA polymerases, presumably by promoting the recruitment of the catalytic subunit Rpa190 (Huber et al. 2009). Furthermore, Sch9 is essential for the proper processing of the 35S ribosomal RNA (rRNA) into 25S, 18S, and 5.8S rRNA and at least one component of the processome such as Rps6 (Huber et al. 2009). Sch9 has a synthetic relationship with the cAMP/PKA pathway, which is another nutrient-sensing pathway (Roosen et al. 2005). Sch9 plays an important role in regulating cellular stress through the Rim15 kinase or interaction with the Sko1–Hog1 complex. TORC1-activated Sch9 phosphorylates and promotes cytoplasmic retention of Rim15. Upon inhibition of TORC1 and Sch9, Rim15 is translocated to the nucleus and activates the Msn2/4 transcription factors (TFs), which govern general stress responses. Furthermore, in response to osmotic stress, Sch9 appears to act as a chromatin-associated transcriptional activator complex that consists of Sch9, Sko1, and Hog1. Sch9 interacts with both Sko1 and Hog1, and Sch9 phosphorylates Sko1. Sch9 is also involved in lipid signaling: it is homologous to mammalian protein kinase B (PKB)/Akt protein kinase with a pleckstrin homology (PH) domain, which binds to the lipid second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3). Upon binding to PIP3, PKB is phosphorylated and activated by phosphoinositide-dependent kinase 1 (PDK1) and by TORC2. Yeast has two PDK1 homologs, Pkh1/2, which are required for the function of yeast Sch9. Sch9 is also required for regulation of genes involved in mitochondrial functions, the tricarboxylic acid cycle, and the induction of autophagy.

In this study, we aimed to identify Sch9-dependent and Sch9-independent signaling cascades and downstream target genes involved in governing C. neoformans thermotolerance and thereby address a potential molecular mechanism for Sch9-mediated thermotolerance. To this end, we performed comparative transcriptome analysis of the WT strain and sch9∆ mutant during temperature upshift from 25° to 37° or 40°, with DNA microarray analysis. Among Sch9-dependent temperature-regulated genes, HSF1, encoding an ortholog of the yeast heat shock transcription factor 1 (Hsf1), was particularly interesting because it appeared to function both as a transcriptional repressor and activator in thermotolerance and oxidative stress response in C. neoformans.

Materials and Methods

Strains and culture conditions

The strains and primers used in their generation or analysis in this study are listed in Supplemental Material, Table S1 and Table S2. Strains were cultured in yeast extract–peptone–dextrose (YPD) medium. Yeast nitrogen base (YNB) medium containing 25 µM CuSO₄ or 200 µM bathocuproinedisulfonic acid (BCS) was used to reduce or induce HSF1 expression in the P_CTR4:HSF1 strain. The diploid C. neoformans strain AI187 (ade2/ADE2 ura5/URA5 MATa/MATα) was used as the parental strain for constructing the heterozygous HSF1/hsf1 mutant.

Total RNA isolation and DNA microarray analysis

For the DNA microarray analysis under temperature-shift conditions, total RNA was isolated as follows: WT (H99) strain, sch9Δ, ire1Δ, and hxl1Δ mutants were grown in 50 ml YPD medium at 30° for 16 hr. Then, 5 ml of the overnight culture media was inoculated into 100 ml of fresh YPD medium and further incubated for 5–6 hr at 25° until the optical density at 600 nm (OD₆₀₀) of the culture medium reached ∼1.0. For the zero time point, 50 ml of the 100-ml culture was sampled; the remaining 50 ml of culture medium was pelleted by centrifugation and the supernatant medium was discarded. Prewarmed (37° or 40°) fresh YPD medium was added into the remaining pellet and further incubated for 30 min at 37° or 40°. The culture was then frozen in liquid nitrogen and lyophilized overnight. Three independent cultures for each strain were prepared for RNA isolation as biological replicates for DNA microarray analysis. Total RNA was isolated with the RiboEx reagent (GeneAll Biotechnology), as previously described (Ko et al. 2009). For control, total RNA prepared from WT and mutant cells was pooled (pooled reference RNA).

For the DNA microarray analysis using the constitutive HSF1 overexpression strain, total RNA was isolated as follows: WT (H99) and the constitutive HSF1 overexpressing strain (P_H3:HSF1, YSB2200) were grown in YPD medium at 30° for 16 hr. Five milliliters of this culture were reinoculated into 95 ml of fresh YPD liquid medium. The cultures were further incubated at 30° until OD₆₀₀ reached ∼1.0. The cultures were then immediately centrifuged, frozen in liquid nitrogen, and lyophilized overnight. Three independent cultures for each strain were prepared for total RNA preparation as biological replicates for DNA microarray. Total RNA was prepared by using the easy-BLUE reagent (Invitrogen, Carlsbad, CA), as previously described (Ko et al. 2009). For control, total RNA prepared from WT and HSF1 overexpressing strain was pooled (pooled reference RNA).

For the complementary DNA (cDNA) synthesis, the total RNA concentration was adjusted to 1 µg/µl with diethylpyrocarbonate (DEPC)-treated water and 15 µl of the total RNA was used. The cDNA was synthesized by using reverse transcriptase (Fermentas) and labeled with Cy5/Cy3 agents (Amersham, Piscataway, NJ), as previously described (Ko et al. 2009). For DNA microarray analysis, we used a C. neoformans serotype D 70-mer microarray slide containing 7946 probes (Duke University, Durham, NC). Among the total 6980 genes in the H99 strain, 6302 genes are matched to 6756 spots (including multiple spots for a single gene) in this gene chip with an e-value of 1e⁻⁶ (90% coverage) by BLASTn search. Using the serotype A gene sequence, each S. cerevisiae gene name or ID listed in the SI Tables was identified by BLASTp search (e-value cutoff: e⁻⁶). Three independent DNA microarrays with three independent biological replicates were performed. After hybridization and washing, the microarray slides were scanned with a GenePix 4000B scanner (Axon Instrument) using GenePix Pro software (version 4.0) and analyzed with Acuity software (version 4.0) and the GenePix array list (Gal) file. For hierarchical and statistical analyses, data transported from GenePix software were analyzed with Acuity software by employing LOWESS normalization, reliable gene filtering (>95% filtering), hierarchical clustering, zero transformation, ANOVA analysis (P < 0.05), and Microsoft Excel software.

Construction of the hsp104∆, ssa1∆, ssa2∆, hsp104∆+HSP104, and ssa1∆+SSA1 strains

The HSP104 gene (CNAG_07347) was disrupted as follows by the homologous replacement of its open reading frame with a piece of DNA containing a dominant drug resistance gene marker. During the first round of PCR, primer pairs used for the amplification of the 5′ and 3′ flanking regions of the HSP104 gene were B5125/B5126 and B5127/B5128, respectively. The NAT STM#125 dominant marker was amplified with the primer pair M13Fe/M13Re. For the second PCR to create a continuous fragment, the hsp104Δ::NAT construct was amplified using the B5125/B5128 primer pair, and the three gel-extracted DNA fragments from the first-round PCR as the template. C. neoformans strain H99 was biolistically transformed with the deletion allele (Toffaletti et al. 1993). To identify the desired hsp104Δ mutant, diagnostic PCR was performed with the primer pair B79/B5124 and Southern blot analysis conducted with a gene-specific probe amplified with the primer pair B5125/B5129.

The SSA1 gene (CNAG_01727) was disrupted as follows. For the first round of PCR, primer pairs used for the amplification of the 5′ and 3′ flanking regions of the SSA1 gene were B5135/B5136 and B5137/B5138, respectively. The NAT STM#169 dominant marker was amplified with primer pair M13Fe/M13Re. For the overlap PCR, the ssa1Δ construct was amplified using the B5135/B5138 primer pair on the three gel-extracted DNA fragments from the first-round PCR. C. neoformans strain H99 was biolistically transformed with the deletion allele. To identify the desired ssa1Δ mutant, diagnostic PCR with primer pair B79/B5139 and Southern blot analysis with a gene-specific probe PCR amplified with primer pair B5135/B5140 were performed.

The SSA2 gene (CNAG_01750) was disrupted as follows. For the first round of PCR, primer pairs used for the amplification of the 5′ and 3′ flanking regions of the SSA2 gene were B5219/B5220 and B5221/B5222, respectively. The NAT STM#177 dominant marker was amplified with primer pair M13Fe/M13Re. During the second round of overlap PCR, the ssa2Δ construct was amplified by using the B5219/B5222 primer pair and the three gel-extracted DNA fragments from the first-round PCR. C. neoformans strain H99 was biolistically transformed with the deletion allele. To identify the desired ssa2Δ mutant, diagnostic PCR with primer pair B79/B5223 and Southern blot analysis with a gene-specific probe PCR amplified with primer pair B5219/B5224 were performed.

To complement the hsp104Δ mutant, the WT HSP104 (CNAG_07437) gene was amplified by two fragments with B7785/B8001 and B8002/B8003 primer pairs. Each fragment was cloned into TOPcloner TA vector to construct pHSP104-F₁ and pHSP104-F₂ and confirmed by sequencing. StuI and XbaI digested HSP104-F₂ was cloned into pHSP104-F₁ to construct pHSP104. XhoI and XbaI digested HSP104 was cloned into pJAF12 to construct pJAF12–HSP104. The plasmid pJAF12–HSP104 was linearized by NsiI digestion and introduced into the hsp104Δ (YSB2261) strain by biolistic transformation. Ectopic insertion of HSP104 was confirmed by diagnostic PCR with a primer pair, B7789/B8001.

To complement the ssa1Δ mutant, the WT SSA1 (CNAG_01727) gene was amplified by two fragments with B7778/B8004 and B8005/B8006 primer pairs. Each fragment was cloned into TOPcloner TA vector to construct pSSA1-F₁ and pSSA1-F₂ and confirmed by sequencing. XhoI and XbaI digested SSA1-F₂ was cloned into pSSA1-F₁ to construct pSSA1. KpnI- and NotI-digested SSA1 was cloned into pJAF12 to construct pJAF12–SSA1. The plasmid pJAF12–SSA1 was linearized by NsiI digestion and introduced into the ssa1Δ (YSB2235) strain by biolistic transformation. Ectopic insertion of SSA1 was confirmed by diagnostic PCR with a primer pair, B7783/B8004.

Construction of the P_CTR4:HSF1 promoter replacement strain

The promoter replacement strain for HSF1 was constructed by using the promoter of CTR4 as follows. During the first round of PCR, the 5′ flanking region [HSF1 promoter region spanning the area from −756 to −1, relative to the ATG start codon (+1 to +3)] and the 3′ flanking region (HSF1 gene, +1 to +971) were amplified with primer pairs B1892/B5190 and B5193/B5145, respectively. The combined NAT–CTR4 promoter DNA was amplified by PCR with two primers B354 and B355. During the second round of overlap PCR, the P_CTR4:HSF1 construct was amplified with B1892 and B5145 primers on the three gel-extracted DNA fragments from the first round. C. neoformans strain H99 was then biolistically transformed with the replacement allele. Stable transformants were selected on YPD medium containing nourseothricin (100 µg/ml). The P_CTR4:HSF1 strains were confirmed by diagnostic screening PCR and Southern blot analysis using HSF1-specific probe amplified with primer B5144 and B5145.

Construction of the constitutive HSF1 overexpression strain

The constitutive HSF1 overexpression strain drives transcription of the gene using the histone H3 promoter of C. neoformans serotype D strain JEC21. During the first round of PCR, the 5′ flanking region [HSF1 promoter region spanning the area from −756 to −1, relative to the ATG start codon (+1 to +3)] and the 3′ flanking region (HSF1 gene, +1 to +971) were amplified with primer pairs B1892/B5190 and B5191/B5145, respectively. The NEO–H3 promoter fragment was amplified with primer pair B4017/B4018. During the second round of overlap PCR, the P_H3:HSF1 construct was amplified using primer pair B1892/B5145. C. neoformans strain H99 was biolistically transformed with the replacement allele. Stable transformants were selected on YPD medium containing G418 (50 µg/ml). The P_H3:HSF1 strain was confirmed by diagnostic PCR and Southern blot analysis using HSF1-specific probe amplified with primers B5144 and B5145. HSF1 expression levels in the P_H3:HSF1 strain were measured by Northern blot analysis using a HSF1-specific probe amplified by PCR using primers B5067 and B5068.

Construction of the HSF1:4xFLAG epitope-tagged strain

The HSF1:4xFLAG tagging strain was constructed by using a plasmid containing a 4xFLAG and the NEO dominant selection marker (So et al. 2016). During the first round of PCR, the C-terminal region of the HSF1 and the HSF1 terminator region were amplified with primer pairs B5460/B6141 and B6142/B5740, respectively. The 4xFLAG–NEO construct was amplified with primer pair B5431/B1966. In the second round of overlap PCR, the HSF1:4xFLAG:NEO construct was amplified by using the B5460/B5740 primer pair from the three gel-extracted DNA fragments from the first-round PCR. WT H99 and sch9Δ strains (YSB619 and YSB620) were biolistically transformed with the construct. Stable transformants were selected on YPD medium containing G418 (50 µg/ml). The HSF1:4xFLAG and sch9Δ HSF1:4xFLAG strains were confirmed by diagnostic PCR and Southern blot analysis. The production of Hsf1:4xFLAG-tagged protein was confirmed by Western blot analysis using monoclonal anti-FLAG M2 antibody produced in mouse (Sigma-Aldrich, St. Louis, MO) and the secondary detection with goat anti-mouse IgG-HRP (Santa Cruz Biotechnology, Santa Cruz, CA).

Hsf1 phosphorylation and λ-phosphatase assay

The HSF1:4xFLAG strain was cultured overnight at 30° in 50 ml YPD liquid media. The culture was 1/10 diluted in 50 ml fresh YPD liquid media until OD₆₀₀ reached ∼1.0 at 25°. Then 25-ml aliquots were further incubated at 25° or 37° for 30 min. The cells were sampled by centrifugation and the whole cell extracts (WCEs) were prepared with lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5 mM EDTA, 0.5% Triton X-100 supplemented with protease inhibitor cocktail]. The samples were incubated with PMP buffer [50 mM HEPES (pH 7.5), 100 mM NaCl, 2 mM DTT, 0.01% Brij 35, 1 mM MnCl₂] and 400 units of λ-protein phosphatase (New England BioLabs) (with or without phosphatase inhibitor cocktail) for 1 hr at 30°. The proteins were resolved on 8% SDS-PAGE gels and transferred to PVDF membranes (Bio-Rad, Hercules, CA). Membranes were assayed by Western blot employing mouse monoclonal anti-FLAG M2 mouse antibodies (Sigma-Aldrich), followed by anti-mouse antibody conjugated to horseradish peroxidase, and ECL Western blotting detection reagent.

Construction of the heterozygous HSF1/hsf1∆ mutant

A hsf1∆::NAT construct consisted of 1.5-kb flanking regions fused with the NAT cassette, made by in vivo recombination in S. cerevisiae in strain FY834 by combining three DNA fragments with plasmid pRS426 (Colot et al. 2006; Ianiri and Idnurm 2015). HSF1 5′ and 3′ flanking regions were amplified with primers ALID2267–ALID2268 and ALID2269–ALID2270 from C. neoformans genomic DNA, respectively, while the NAT cassette was amplified off plasmid pAI3 with primers ai006–ai290 (Idnurm et al. 2004). Primers were designed to replace a 2017-bp region of the HSF1 gene, including the entire functional HSF-type DNA-binding domain. The construct was amplified from S. cerevisiae transformants using primers ALID1229 and ALID1230. The replacement allele was introduced into strain AI187 by biolistic transformation. Transformants were selected on YPD + nourseothricin (100 µg/ml) and screened for homologous recombination events by PCR using primers specific for the NAT gene (ai37 and ai270) in combination with screening primers (ALID2271 and ALID2272) located outside the homologous recombination regions. Primers used are listed in Table S2. The heterozygous strain HSF1/hsf1Δ was grown on Murashige–Skoog (MS) medium to induce meiosis and sporulation. After 3 weeks of incubation in the dark, 46 haploid basidiospores from mixed populations were dissected onto YPD agar. Resultant colonies were grown for 3–4 days at 30°, transferred to a 96-well plate containing 100 µl of YPD, and tested for the four genetic markers that segregate during meiosis [nourseothricin resistance (NAT^R) or sensitivity (NAT^S), ura5/URA5, ade2/ADE2, and MATa/MATα] by spotting 3 µl of a cellular suspension onto YPD + nourseothricin (100 µg/ml), YNB + adenine (20 mg/liter), or YNB + uracil (40 mg/liter); the mating type marker was scored by crossing haploid progenies to strains KN99a and KN99α on MS media supplemented with adenine and uracil, and by evaluating the formation of sexual structures under a microscope.

Chromatin immunoprecipitation analysis

Strain HSF1:4xFLAG (YSB3160) was grown overnight and then diluted to an OD₆₀₀ of 0.2 in two 50-ml cultures. The cells were grown for 3 hr at 30°. One culture was then placed at 37° and the other was kept at 30° for 30 min. Formaldehyde was added at a final percentage of 1% and incubated at room temperature for 20 min for cross-linking. Cross-linking was quenched by adding glycine to a final concentration of 125 mM and incubated for 5 min at room temperature. Cells were pelleted and washed once in 10 ml of phosphate-buffered saline (PBS) containing 125 mM glycine. The washed pellets were resuspended in 1 ml chromatin immunoprecipitation (ChIP) lysis buffer (50 mM HEPES-KOH, pH 7.5; 140 mM NaCl; 1% Triton X-100; 1 mM EDTA) and placed into a bead beating tube with ∼200 µl of acid-washed glass beads. Cells were lysed by bead beating three times for 1 min with 1 min on ice in a Mini Bead Beater 15 (BioSpec). The lysate was then aspirated and added to a 15-ml conical tube and 1 ml more of ChIP lysis buffer was added. The lysates were then sonicated five times to shear the DNA, with 30 sec on and 1 min off with icing in between. Lysates were then cleared by centrifugation for 10 min at 20,000 × g at 4°. The supernatant was transferred into an Eppendorf tube and centrifuged once more to further clear the lysate. The resulting supernatant was the input for the ChIP experiment.

Thirty microliters of supernatant were transferred to a new tube as the WCE and these samples were kept at 4° during the duration of the experiment. To prepare the immunoprecipitation samples, 250 µl of the lysate was added to 250 µl of cold ChIP lysis buffer and 120 µl of 3.6 M NaCl. To these samples, 50 µl of protein A agarose (Protein A Plus Agarose, Pierce Chemical, Rockford, IL) beads was added, and the mixtures were incubated for 1 hr at 4°, while rotating them to clear proteins that nonspecifically bind to the beads. During this time, FLAG beads (EZview Red Anti-FLAG M2 Affinity Gel, Sigma-Aldrich) were blocked through incubation in PBS with 5% milk for 1 hr. Lysates with agarose beads were then centrifuged at 10,000 × g for 1 min at 4° and the supernatant was transferred to a new tube. Blocked FLAG beads were washed three times in ChIP lysis buffer and then resuspended into 100 µl. Fifty microliters of the blocked FLAG bead slurry was added to the lysates and incubated overnight at 4° while rotating. The following day, the beads were washed two times in 1 ml ChIP lysis buffer, once in 1 ml ChIP lysis buffer containing 0.5 M NaCl, once in 1 ml ChIP wash buffer (10 mM Tris-HCl, pH 8.0; 0.75 M LiCl; 0.5% NP-40; 1 mM EDTA), and then two times in 1 ml TE, all at 4°. The immunoprecipitated Hsf1-4xFLAG was eluted by adding 150 µl of ChIP elution buffer (50 mM Tris-HCl, pH 8.0; 10 mM EDTA; 1% SDS) and incubated for 10 min at 65°. Tubes were then centrifuged at 10,000 × g for 1 min and 125 µl of the supernatant was transferred to a new tube. Ninety-five microliters of ChIP elution buffer was added to the WCE. The IP and WCE tubes were then incubated at 65° overnight to disrupt the formaldehyde cross-links. The next day, 5 µl of proteinase K (BioLine) was added to the samples and incubated at 37° for 90 min. The samples were then purified using a QIAGEN PCR purification kit as per manufacturer’s instructions (QIAGEN, Hilden, Germany) and eluted in 50 µl of water.

Wax moth virulence assay

To prepare the inocula, C. neoformans strain H99 and P_H3:HSF1 strains (YSB2200 and YSB2201) were grown overnight at 30° in YPD medium, washed three times with PBS, and resuspended in PBS for cell counting by hemocytometer. Galleria mellonella caterpillars in the final instar larval stage were used within 7 days from the day of shipment (Vanderhorst Wholesale, St. Marys, OH). For each group, 15 G. mellonella larvae, ranging from 200 to 300 mg in body weight, were randomly chosen. Four microliters of 10⁶ cells/ml (4000 C. neoformans cells per larva) were inoculated through the distal prolegs of caterpillars by using a 100-µl Hamilton syringe equipped with a 10-µl size needle and repeating dispenser (PB600-1, Hamilton). PBS was injected as a noninfection control. After injection, caterpillars were incubated in Petri dishes in humidified plastic containers and monitored daily. Caterpillars were considered dead when they displayed no movement upon touch. Caterpillars that transformed into pupa during the experiment were censored for statistical analysis. Survival curves were drawn using GraphPad Prism 6 (GraphPad, San Diego, CA) and analyzed with the log-rank (Mantel–Cox) test.

Data and reagent availability

The microarray data generated in this study were submitted to the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/). The accession numbers are GSE66508 (DNA microarray analysis upon temperature-shift conditions) and GSE66509 (DNA microarray analysis using HSF1-overexpressing strain). The strains used in this study (listed in Table S1) can be obtained from Yong-Sun Bahn’s laboratory upon request.

Results

Transcriptome analysis to elucidate the Sch9-dependent thermotolerance regulon in C. neoformans

To elucidate the Sch9-mediated thermotolerance signaling mechanism in C. neoformans, we performed comparative transcriptome analysis of the serotype A WT strain (H99) and sch9∆ mutants during temperature upshift from 25° to 37° or 40° using DNA microarrays (Table S3). First, we analyzed microarray data from the WT strain to identify temperature-regulated genes in C. neoformans. A total of 1831 genes exhibited significantly different expression patterns (ANOVA test, P < 0.05) during temperature shift from 25° to 37° or 40° (Figure 1A). Among them, 1671 genes showed more than twofold induction or reduction at 37° (1691 at 40°), strongly suggesting that a significant portion of the C. neoformans genome is remodeled to adapt to different growth temperatures (Table S4).

Figure 1

Global expression profiles of Sch9-dependent and Sch9-independent genes during temperature upshift. (A) Gene clustering and functional categorization of temperature-regulated genes using the Eukaryotic Orthologous Groups of proteins (KOG) (http://www.ncbi.nlm.nih.gov/COG/) database. A total of 1831 genes, whose expression was differentially regulated upon temperature upshift, (ANOVA test, P < 0.05) are categorized by KOG functional description. (B) The inserted table summarizes the pattern of gene expression change in the WT (H99) and sch9Δ (YSB619) strains under temperature upshift from 25° to 37°. (C) Functional categories of genes differentially regulated by Sch9. A total of 313 genes, whose basal and temperature-mediated increased/decreased expression levels are significantly different in the sch9Δ mutant compared to the WT strain, are categorized by KOG functional description.

Next, we compared the transcript profiles of the sch9∆ mutant to those of the WT strain during temperature upshift to find Sch9-dependent and Sch9-independent genes. Verifying our array quality, SCH9 expression levels were very low in the sch9∆ mutant compared to those in the WT strain (−5.71 at 25°, −2.98 at 37°, and −6.61 at 40°; log₂ scale) (Table S3). There could be three classes of Sch9-dependent genes. The first class included genes whose basal expression levels are affected by Sch9. The second class included those whose temperature-induced or temperature-repressed expression levels were affected by Sch9. The third class included genes whose basal and temperature-induced/repressed expression levels were both affected. For this comparison, we first performed an ANOVA test using data from WT at 25°, WT at 37°, sch9∆ at 25°, and sch9∆ at 37° and identified gene sets using a P < 0.05 cutoff. A total of 2909 genes had significantly affected expression under these conditions. We performed further analyses with these genes.

First, we compared basal transcript profiles (at 25°) between the WT strain and sch9∆ mutant. A total of 1032 genes exhibited more than a twofold induction (583 genes) or reduction (449 genes) in the sch9∆ mutants compared to the WT strain (Table S5). This result indicates that Sch9 significantly affects the basal expression of a variety of genes in C. neoformans.

Second, we compared transcription profiles that showed an increase or decrease in expression during temperature upshift from 25° to 37° between the WT strain and sch9∆ mutant. The genes were divided into four groups according to expression levels: significantly induced more than twofold or less than twofold and significantly reduced more than twofold or less than twofold in WT and sch9Δ strains during temperature upshift. We found that ∼76% of genes (2218 of 2909 genes) showed similar expression patterns in WT and sch9Δ strains (Figure 1B). The expression of a total of 246 genes was reduced or induced in the sch9Δ mutant, but not in the WT strain. In contrast, the expression of 427 genes was reduced or induced in the WT strain, but not in the sch9Δ mutant. Interestingly, 12 genes were induced in the WT strain, but their expression was reduced in the sch9Δ mutant, while six genes were oppositely regulated (Figure 1B), suggesting that these 18 genes might be tightly regulated by the Sch9 pathway (Table S6). Last, we searched for genes whose basal and induced/reduced transcription levels upon temperature upshift were both regulated by Sch9. To identify these gene categories, we compared the results from basal transcriptome profiles at 25° (Table S5) and from induced or reduced transcriptome profiles (Table S6) during temperature upshift from 25° to 37°. A total of 313 genes were identified in this group, including those involved in post-translational modifications and chaperone functions, signal transduction, carbohydrate metabolism and transport, or intracellular trafficking and secretion (Figure 1C and Table S7). In conclusion, Sch9 serves as a master regulator of genes under basal and temperature-upshift conditions.

Identification of thermotolerance-regulating signaling pathways in C. neoformans

Our microarray data revealed that 1671 genes (from 25° to 37°) and 1691 genes (from 25° to 40°) were induced or reduced more than twofold in WT strain during temperature upshift. To identify thermotolerance-regulating signaling pathways in C. neoformans, we compared these temperature-regulated genes to our previous large-scale TF and kinase phenome-based studies (Jung et al. 2015; Lee et al. 2016). Nine TFs and 17 kinases were regulated in transcript levels by temperature upshift and required for thermotolerance of C. neoformans (Table 1). Most of these TF (8 of 9) and kinase (14 of 17) genes were also involved in the virulence or infectivity of C. neoformans (Jung et al. 2015; Lee et al. 2016). This suggested that the genes involved in adaptation to temperature upshift are crucial for virulence of C. neoformans.

Temperature-regulated TFs and kinases in C. neoformans

Table 1

Temperature-regulated TFs and kinases in C. neoformans

H99 ID (meaning identifier)	Designated name	Class								In vivo virulence dataa
			Gene expression (log2) normalized to WT 25°			Growth (°)a				Insect		Mouse
			WT 25°	WT 37°	WT 40°	25	30	37	39	Virulence	RMSb	Virulence	STMc score
CNAG_01438	MBS2	TF	0	−2.97	−2.221	0	0	0	−1	0	1	−1	−2.44
CNAG_06134	BZP1(HXL1)	TF	0	3.525	3.256	0	0	−3	−3	−3	>1.67	−3	−6.82
CNAG_07924	MCM1	TF	0	−1.07	−1.394	0	0	0	1	0	1	0	−0.78
CNAG_00156	SP1(CRZ1)	TF	0	2.169	2.628	0	0	0	−2	−1	1.37	0	−0.49
CNAG_00514	GAT6	TF	0	1.688	−1.416	0	0	0	−1	0	0.99	1	1.52
CNAG_01626	ADA2	TF	0	2.364	3.095	0	0	0	−2	−3	>1.43	0	1.45
CNAG_04804	SRE1	TF	0	2.252	1.816	0	0	0	−1	−3	2.04	−2	−4.64
CNAG_07724	CUF1	TF	0	1.734	0.963	−2	−2	−2	−2	0	1.2	2	5.66
CNAG_05222	NRG1	TF	0	1.151	1.017	−1	−1	−1	−2	−1	1.33	−3	−6.73
CNAG_00363	TCO6	Kinase	0	1.064	2.659	0	0	−1	−1	0	1.14	−1	−3.58
CNAG_00405	KIC1	Kinase	0	0.978	1.188	0	−2	−2	−2	−2	1.46	−3	−11.09
CNAG_00769	PBS2	Kinase	0	1.558	0.976	0	0	0	−1	0	1.08	0	−1.9
CNAG_02233	MEC1	Kinase	0	3.369	−3.31	1	−2	−3	−3	−3	1.92	−2	−7.58
CNAG_02357	MKK2	Kinase	0	2.78	2.367	0	0	−1	−3	−3	2	0	0.37
CNAG_02680	VPS15	Kinase	0	2.414	2.066	0	−1	−3	−3	−3	1.9	−3	−9
CNAG_02712	BUD32	Kinase	0	5.653	−3.412	−1	−3	−1	−1	−3	2.4	−1	−3.38
CNAG_02859	POS5	Kinase	0	1.695	1.39	−1	−2	0	−1	−3	1.9	−2	−6.71
CNAG_03367	URK1	Kinase	0	1.955	−1.776	0	0	0	−1	−2	1.5	−2	−7.52
CNAG_03567	CBK1	Kinase	0	0.931	−1.991	−1	−1	−3	−3	−2	1.71	0	−0.18
CNAG_03670	IRE1	Kinase	0	2.478	1.722	0	−1	−3	−3	−3	2.3	−1	−2.8
CNAG_04215	MET3	Kinase	0	−8.12	−7.356	−1	0	−1	−1	−3	1.95	−2	−5.27
CNAG_04316	UTR1	Kinase	0	1.214	1.887	−1	0	−2	−2	−2	1.7	−1	−3.53
CNAG_05386	PRO1	Kinase	0	2.642	0.912	0	0	−2	0	0	1	0	−1.56
CNAG_05753	ARG5,6	Kinase	0	1.921	−1.605	0	0	−2	−3	−2	1.75	−1	−3.05
CNAG_06301	SCH9	Kinase	0	2.917	3.049	0	0	0	1	−2	1.4	−1	−2.13
CNAG_06632	ABC1	Kinase	0	2.093	0.875	0	0	0	−1	0	1.1	0	−1.95

a

The growth data and in vivo virulence data are obtained from previous reports (Jung et al. 2015; Lee et al. 2016).

b

Relative median survival (RMS) means that mutant median survival day is divided by WT median survival day. The RMS scores for two-independent strains for each TF or kinase mutants were described as an average value.

c

Signature-tagged mutagensis (STM) score means the ratio of values comparing the threshold cycle (Ct) of output genomic DNA from each lung of the killed mouse with the threshold cycle of input genomic DNA. The STM score of each mutant was calculated using the variation of the 2^−ΔΔCt.

The probabilistic functional gene network of C. neoformans, CryptoNet (http://www.inetbio.org/cryptonet/) (Kim et al. 2015), revealed that the thermotolerance-regulating TFs and kinases are highly interconnected (Figure 2), suggesting that multiple signaling pathways cooperate to promote the thermotolerance of the pathogen. These include the following reported signaling pathways: Hog1 and Mpk1 MAPK, calcineurin, Rim101/Nrg1, UPR, and RAM pathways (Figure 2). Interestingly, we found a novel thermotolerance-regulating pathway, the NAD kinase pathway, in C. neoformans. In S. cerevisiae, there are three NAD kinases, Utr1, Yef1, and Pos5 (Zhang 2015). Utr1 and Yef1 are cytosolic and Pos5 is a mitochondrial NAD kinase (Kawai et al. 2001; Outten and Culotta 2003; Shi et al. 2005). UTR1 and POS5 were induced more than twofold during temperature upshift and required for growth at high temperature in C. neoformans (Table 1). YEF1 was also induced by temperature upshift, but not involved in thermotolerance (Table S4) (Lee et al. 2016).

Figure 2

Identification of novel temperature-regulated signaling pathways in C. neoformans. Prediction of temperature-regulating pathways in C. neoformans by CryptoNet. Nine TFs and 17 kinases were used to find temperature-regulating signaling pathways. The nodes that have an edge to another node are indicated.

Induction of diverse molecular chaperones and heat shock proteins for high-temperature adaptation of C. neoformans

The DNA microarray data showed that genes whose expression was most dramatically regulated during temperature upshift belonged to the molecular chaperone family (Table S8). Molecular chaperones play a role in mediating proper folding of newly synthesized polypeptides under physiological conditions and also prevent protein denaturation or assist refolding of denatured proteins under stressed conditions (Liberek et al. 2008). Our array revealed that cytoplasmic, ER-, and mitochondria-resident molecular chaperone genes were all highly upregulated upon temperature upshift (Table S8). As a cytoplasmic molecular chaperone complex, genes encoding a heat shock protein (HSP) Hsp104 (CNAG_07347), Ydj1p/Hsp40 (CNAG_03944), and Ssa1/Hsp70 (CNAG_01750) were upregulated during the temperature upshift. Genes encoding ER-resident molecular chaperones, LHS1 (CNAG_03899) and KAR2 (CNAG_06443), were also upregulated, further supporting our previous finding that the UPR pathway in the ER is activated in response to temperature upshift (Cheon et al. 2011). Interestingly, protein refolding within the ER is known to also require the cytoplasmic chaperone Hsp104 (Hänninen et al. 1999). Furthermore, a mitochondria-resident molecular chaperone gene, SSC1 (CNAG_05199), was also induced (Table S8). Sch9 appears to be involved in basal expression levels of some of these molecular chaperone genes, but does not significantly regulate their induction upon temperature upshift (Table S8).

Among these genes, HSP104 was chosen for verification of the array data, as it was the most highly expressed gene. We found that HSP104 expression was strongly induced within 10 min upon temperature upshift and that its induction was maintained for more than 2 hr (Figure 3A). HSP104 appeared to present alternatively spliced transcripts upon temperature upshift (Figure S1A). HSP104 expression was also induced in the sch9∆ mutant upon temperature upshift, but its rapid induction was modestly delayed, suggesting that Sch9 plays a minor role in HSP104 induction. To address whether Hsp104 plays a role in thermotolerance, we deleted HSP104 in C. neoformans (Figure S2) and analyzed the hsp104∆ mutant phenotype. The hsp104∆ mutant did not show any growth defects at temperatures ranging from 25° to 39° and was as resistant as the WT strain to most of the environmental stresses and antifungal drugs we tested (Figure S3). However, the hsp104∆ mutants exhibited a much greater susceptibility to extreme heat shock at 55° than the WT strain (Figure 3B). Complementation of the hsp104∆ mutants with the WT HSP104 gene restored the heat shock resistance, suggesting that Hsp104 is required for the response and adaptation of C. neoformans to extreme thermal shock, rather than mild thermal stresses.

Figure 3

HSP and ergosterol biosynthesis genes are regulated by temperature upshift. (A) HSP104 is induced under temperature upshift from 25° to 37° in both WT and sch9Δ cells. qRT-PCR was performed using total RNA isolated from the WT (H99) and sch9Δ (YSB619) strains upon temperature upshift from 25° to 37° with HSP104-specific primers. The HSP104 expression levels were normalized to the ACT1 expression levels. Error bars indicate SD. qRT-PCR was performed twice with three technical replicates and representative data are presented. (B) Hsp104 is required for growth in extreme heat shock conditions (55°). The WT, hsp104Δ (YSB2261), and hsp104Δ+HSP104 (YSB4689) strains were grown overnight at 30° in YPD liquid medium, serially diluted (1–10⁴ dilutions), and spotted on YPD plates. The YPD plates were placed in a 55° incubator for the indicated amount of time, then further incubated at 30° for 2 days, and photographed. (C) Expression of ERG11 and ERG2 is reduced during temperature upshift. qRT-PCR were performed using total RNA isolated from the WT and sch9Δ strains upon temperature upshift from 25° to 37°. For qRT-PCR, ERG11 and ERG2 expression levels were normalized by using ACT1 expression levels as controls. Data were collected from three technical replicates. Error bars represent SD. This experiment was repeated twice and one representative experiment is presented. (D) ERG11 expression is induced by sterol depletion in Sch9-independent manner. For the qRT-PCR, total RNA was isolated from fluconazole-treated (10 μg/ml for 90 min) or nontreated WT and sch9∆ strains. The relative ERG11 expression levels were quantitatively measured by qRT-PCR with ERG11-specific primers after normalization with ACT1 expression levels. Error bars represent SD. This experiment was repeated twice and one representative experiment is presented.

In addition to these cellular molecular chaperones, a number of genes encoding other HSPs and chaperone-like proteins were identified as temperature-regulated genes in C. neoformans (Table S9). These include HSP10, HSP12, HSP60, HSP78, and HSP90. Similar to the regulation of molecular chaperones, Sch9 was mainly involved in the regulation of basal expression levels of a number of HSP genes (Table S9). In conclusion, diverse cellular molecular chaperones and HSPs are all upregulated during the adaptation to high temperature in C. neoformans.

Repression of ribosome/translation and ergosterol biosynthesis genes of C. neoformans in response to temperature upshift

Our array data revealed that expression of genes involved in ribosomal biosynthesis and translation significantly decreased during the temperature upshift (Figure 1A). In addition, most ergosterol biosynthetic genes were significantly downregulated during the temperature upshift in both WT and sch9∆ strains (Table S10). We confirmed this finding by assessing the transcript levels of ERG11 and ERG2 by quantitative RT-PCR (qRT-PCR) and Northern blot analysis, showing that expression levels were significantly reduced in the WT and sch9∆ strains during temperature upshift (Figure 3C and Figure S1B). We previously reported that, compared to the WT strain, the sch9∆ mutant showed increased susceptibility to azole drugs such as fluconazole and ketoconazole (Kim et al. 2009). We next assessed whether ERG11 induction upon azole treatment depends on Sch9. Similar levels of fluconazole-induced ERG11 expression were observed for the WT and the sch9∆ mutant (Figure 3D). In conclusion, ergosterol biosynthesis is tightly regulated under temperature upshift and sterol-depleted conditions in a Sch9-independent manner.

Identification of Hsf1 and Hsp90 as temperature-regulated genes in C. neoformans

The most notable, but rather unexpected, finding of this array analysis was that a gene encoding HSF1 (CNAG_07460) was downregulated during temperature upshift in C. neoformans (Table S9). Hsf1 is a TF that orchestrates transcriptional activation and synthesis of HSPs in response to environmental stresses. Hsf1 is known to be constitutively expressed in most tissues and cell types of eukaryotic organisms and is mainly regulated by post-translational mechanisms (Anckar and Sistonen 2011). Therefore, HSF1 downregulation upon temperature upshift in C. neoformans was rather counterintuitive and surprising because a number of HSP and chaperone genes appeared to be upregulated during the temperature upshift as described above. In contrast, HSP90 (CNAG_06150.2), which is a putative molecular chaperone that may be controlled by Hsf1, was induced by the temperature upshift.

To confirm the gene expression findings from the microarray experiments, we performed Northern blot analysis. In agreement with the data, HSF1 was downregulated during the temperature upshift (either from 25° to 37° or 25° to 40°) in the WT strain (Figure S4A). In the sch9∆ mutant, HSF1 was less downregulated during the temperature upshift (Figure S4A), particularly at 40°. To further analyze the expression patterns of HSF1, we monitored HSF1 expression at different time points during the temperature upshift (Figure 4A). In the WT strain, HSF1 was gradually downregulated after temperature upshift from 25° to 37° (Figure 4A). In the sch9∆ mutant, HSF1 was also downregulated upon temperature upshift although its basal expression levels became lower in the mutant than WT (Figure 4A).

Figure 4

HSF1 is downregulated during temperature upshift partly in an Sch9-dependent manner. (A) HSF1 expression gradually decreased upon temperature upshift, partially in a Sch9-dependent manner. WT (H99) and sch9∆ (YSB619) strains grown at 30° to the logarithmic phase (OD₆₀₀ ∼ 1.0) were further incubated at 37° or 40° and then sampled at the indicated time point for total RNA isolation For qRT-PCR, HSF1 expression levels were normalized by using ACT1 expression levels as controls. Data were collected from three technical replicates. Error bars represent SD. This experiment was repeated twice and one representative experiment is presented. (B) Hsf1 protein levels also decreased upon temperature upshift, partially in a Sch9-dependent manner. For Western blot analysis, whole cell lysate was isolated from HSF1:4xFLAG (YSB3160) and sch9∆ HSF1:4xFLAG (YSB3338) strains cultured as described in A. The Western blot membrane was hybridized first with anti-FLAG antibody and developed. After deprobing, the same membrane was rehybridized with an anti-β-actin antibody as a loading control. Each Hsf1/β-actin is a value relative to that of the zero time WT strain set to 1.0. (C) Hsf1 exhibits electrophoretic mobility shift during temperature upshift. For the Western blot, whole cell extract from HSF1:4xFLAG and sch9Δ HSF1:4xFLAG strains were used. Black arrow indicates phosphorylated Hsf1-4xFLAG protein (P-Hsf1) and white arrow indicates dephosphorylated Hsf1-4xFLAG protein (Hsf1). (D) The reduced mobility of Hsf1-4xFLAG by temperature upshift is caused by phosphorylation. For the preparation of whole cell extract, HSF1:4xFLAG strains grown at 25° to the logarithmic phase (OD₆₀₀ ∼ 1.0) were further incubated at 25° or 37° for 30 min and then sampled. The whole cell extract was incubated at 30° for 1 hr with or without λ-phosphatase and phosphatase inhibitors. The Western blot membrane was hybridized with anti-FLAG antibody and developed. (E) HSP90 is induced by temperature upshift in both Sch9-independent and -dependent manners. The same RNA used in A was used for qRT-PCR with HSP90-specific primers. Data were collected from three technical replicates. Error bars represent SD. This experiment was repeated twice. Representative image from independent experiments is shown. (F) Overexpression of HSF1 affects the expression of HSP90 in C. neoformans. For qRT-PCR, HSP90 expression levels were normalized by using ACT1 expression levels as controls. Data were collected from three technical replicates. Error bars represent SD. Representative image from independent experiments is shown.

To test whether Hsf1 protein levels were similarly reduced as HSF1 transcript levels during temperature upshift, we generated strains expressing epitope-tagged Hsf1 (HSF1:4xFLAG) in the WT and sch9∆ backgrounds and performed Western blot analysis. The Hsf1 protein levels were also gradually decreased (up to 50%) by temperature upshift and then maintained in the HSF1:4xFLAG strain (Figure 4B). Interestingly, the Hsf1 protein levels in the sch9∆ HSF1:4xFLAG strain was slightly lower under basal conditions (∼70%), but were not significantly changed during temperature upshift in sch9∆ HSF1:4xFLAG. This suggests that Sch9 plays a role in HSF1 downregulation during temperature upshift, but may be required for HSF1 messenger RNA (mRNA) stability under basal conditions.

To address whether Hsf1 is phosphorylated during temperature upshift, we monitored the electrophoretic mobility shift of Hsf1:4xFLAG by Western blot analysis. We found that the Hsf1:4xFLAG protein exhibited reduced mobility after temperature upshift in both HSF1:4xFLAG and sch9Δ HSF1:4xFLAG strains (Figure 4C), suggesting that Hsf1 undergoes phosphorylation upon temperature upshift. In the sch9Δ mutant, Hsf1 phosphorylation appeared to be slightly delayed. To confirm that the reduced mobility of Hsf1:4xFLAG was caused by phosphorylation, we incubated the immune-isolated Hsf1:4xFLAG with λ-phosphatase in the presence or absence of phosphatase inhibitors and analyzed them by Western blot analysis. The reduced mobility of Hsf1:4xFLAG by temperature upshift was reversed with λ-phosphatase, but not with λ-phosphatase in the presence of phosphatase inhibitors (Figure 4D). In conclusion, Hsf1 undergoes phosphorylation upon temperature upshift in C. neoformans.

One of the well-known Hsf1 target genes in other organisms is HSP90. HSP90 transcription is activated by Hsf1, which is negatively regulated by interaction with Hsp90 as an autoregulatory loop (Leach et al. 2012). Interestingly, HSP90 expression was rapidly induced in both WT and sch9∆ strains by temperature upshift, which is in stark contrast to HSF1 expression patterns (Figure 4E and Figure S5). Overexpression of HSF1 did not affect the basal expression levels of HSP90, but increased induced expression levels of HSP90 upon temperature upshift (Figure 4F). These results implied that the increased amount of phosphorylated Hsf1 proteins in response to high temperature may enhance HSP90 induction. Taken together, our data showed that expression of HSF1 and HSP90 is differentially regulated during temperature upshift in C. neoformans.

Hsf1 is essential for C. neoformans growth

Whether Hsf1 promotes or represses C. neoformans thermotolerance remained unclear. Given the fact that HSF1 was downregulated during temperature upshift, Hsf1 may repress genes involved in C. neoformans thermotolerance. In this case, HSF1 disruption may increase C. neoformans thermotolerance, although Hsf1 is known to be essential for the viability of most eukaryotic organisms reported thus far. Therefore, we attempted to disrupt the HSF1 gene in C. neoformans. However, repeated attempts to delete HSF1 were not successful, implying that HSF1 may be indispensable for C. neoformans growth. To confirm the essentiality of Hsf1, two complementary strategies were performed.

First, a conditionally null HSF1 mutant was generated using the copper-regulated promoter of the CTR4 gene (P_CTR4:HSF1) (Figure 5A and Figure S6). Under inducing conditions with the specific copper chelator BCS, the P_CTR4:HSF1 strain grew as well as the WT strain (Figure 5B). However, under repressing conditions with CuSO₄, the growth of the P_CTR4:HSF1 strain was severely impaired, indicating that Hsf1 is critical for C. neoformans viability (Figure 5B).

Figure 5

HSF1 is required for C. neoformans viability. (A) Schematic representation of the P_CTR4:HSF1 promoter replacement strain. A construct containing the promoter of the CTR4 gene and NAT cassette was inserted in front of the ATG start codon of the HSF1 gene. The letter H denoted in the diagram indicates the HindIII restriction enzyme site used in Southern blot analysis. (B) Hsf1 is required for the growth of C. neoformans. WT strain H99 and P_CTR4:HSF1 (YSB2340, YSB2341, YSB2342, and YSB2343) strains were cultured in YPD liquid medium at 30° and spotted on YNB medium containing 200 µM BCS or 25 µM CuSO₄ for induction or reduction, respectively, of HSF1. The plates were further incubated at 30° for 2 days and photographed. (C) Schematic representation of the HSF1 gene before and after the targeted replacement in a diploid C. neoformans strain. Note the replacement of the functional domain HSF1-type DNA binding by the NAT cassette. (D) PCR analysis with screening primers in combination with primers specific for the NAT cassette, ALID2271-ai37 and ALID2272-ai270, producing two amplicons of expected size (2592 and 2452 bp, respectively) only for the heterozygote HSF1/hsf1Δ. (E) Seven colonies obtained from 46 basidiospores dissected from the heterozygote HSF/hsf1Δ were grown on four media types for segregation analysis. All progeny being nourseothricin sensitive indicates the inability of the hsf1::NAT deletion strains to grow, suggesting that HSF1 is essential for C. neoformans viability.

The second approach generated a heterozygous hsf1Δ/HSF1 strain in a diploid background, and then we analyzed the haploid progeny derived after meiosis for the presence or absence of the NAT-dominant marker. To generate a heterozygous HSF1/hsf1Δ, the hsf1Δ::NAT deletion allele was generated and the well-characterized diploid strain AI187 of C. neoformans was transformed by biolistic transformation according to a recently reported strategy (Idnurm 2010; Ianiri and Idnurm 2015). A representation of the target gene and the mutated copy of HSF1 are presented in Figure 5C. Nourseothricin resistant (NAT^R) transformants were screened by PCR for correct homologous recombination events and one transformant producing two amplicons of the expected size was selected for further experiments as the heterozygous hsf1Δ/HSF1 strain (Figure 5D). This heterozygous mutant was grown on medium to induce sporulation for classical genetic analysis of the progeny generated. Forty-six basidiospores were dissected and the 7 of them that germinated were tested for the four markers [NAT^R or nourseothricin sensitivity (NAT^S), ura5 or URA5, ade2 or ADE2, MATa or MATα] that segregate during meiosis. All 7 progeny were unable to grow on selective medium containing nourseothricin (Figure 5E), further confirming that the HSF1 gene is required for C. neoformans viability.

Hsf1 promotes C. neoformans thermotolerance

If Hsf1 serves as an activator of thermotolerance in C. neoformans, its overexpression may enhance the viability of the pathogen at high temperature. To address this hypothesis, we constructed an HSF1-overexpressing strain using the constitutively active histone H3 promoter (P_H3:HSF1 strain; Figure 6, A and B). HSF1 overexpression in the P_H3:HSF1 strain was confirmed by Northern blot analysis (Figure 6C). Surprisingly, HSF1 overexpression significantly enhanced the growth rate of C. neoformans at high temperature (Figure 6D), strongly suggesting that Hsf1 serves as an activator of thermotolerance in C. neoformans. Furthermore, HSF1 overexpression in the sch9∆ mutant (P_H3:HSF1 sch9∆) further increased the thermotolerance of the sch9∆ mutant (Figure 6E).

Figure 6

HSF1 overexpression promotes C. neoformans thermotolerance. (A) Schematic representation for the construction of P_H3:HSF1 overexpression strain. The promoter of the histone H3 gene and the NEO drug resistance cassette were inserted in front of the ATG start codon of the HSF1 gene. The letter K denoted in the diagram indicates the KpnI restriction enzyme site used in Southern blot analysis. (B) Southern blot analysis for confirmation of P_H3:HSF1 strains. The genomic DNA from the WT strain H99 (lane 1) and P_H3:HSF1 strains, including YSB2200 (lane 2), YSB2201 (lane 3), and YSB2202 (lane 4), was isolated and digested with KpnI for analysis. (C) Northern blot analysis confirms the constitutive overexpression of HSF1 in P_H3:HSF1 strains. Total RNA from WT and P_H3:HSF1 strains (YSB2200, YSB2201, and YSB2202 in lanes 1–3) was used for Northern blot analysis. The membrane was hybridized with a radioactively labeled HSF1-specific probe (B and C). (D) HSF1 overexpression increases C. neoformans thermotolerance. WT, hog1Δ (YSB64), sch9Δ (YSB619 and YSB620), and P_H3:HSF1 (YSB2200 and YSB2201) strains were grown on YPD agar medium. The cells were incubated at the indicated temperature for 3 days and photographed. (E) HSF1 overexpression slightly increases the thermotolerance of the sch9Δ strain. WT, hog1Δ (YSB64), sch9Δ (YSB619), P_H3:HSF1 (YSB2200, YSB2201, and YSB2202), and sch9Δ P_H3:HSF1 (YSB2269 and YSB2270) strains were cultured at 30° in YPD medium for 16 hr, serially diluted (1–10⁴ dilution), spotted on YPD agar medium, further incubated at the indicated temperature for 4 days, and photographed. (F) HSF1 overexpression promotes HSP104 expression. Total RNA from WT, P_H3:HSF1 (YSB2200), and hsp104Δ (YSB2260) strains was used for Northern blot analysis. The membrane was hybridized with radioactively labeled HSP104-specific probe. (G) HSF1 overexpression increases SSA1 and SSA2 expression levels. For qRT-PCR, SSA1 and SSA2 expression levels were normalized by using ACT1 expression levels as controls. Data were collected from three technical replicates. Error bars represent SD. This experiment was repeated twice and one representative experiment is presented. (H) Ssa1 is important for C. neoformans thermotolerance. WT, ssa1Δ (YSB2235), and ssa1Δ+SSA1 (YSB4702) strains were cultured at 30° in YPD medium for 16 hr, serially diluted (1–10⁴ dilution), spotted on YPD agar medium, further incubated at the indicated temperature for 2 days, and photographed. The two images split by a horizontal white line in each spot assay were obtained from the same plate.

To investigate further the positive role of Hsf1 in C. neoformans thermotolerance, we determined whether HSF1 overexpression could induce the expression of temperature-responsive genes. To this end, we chose the HSP104 and two SSA1-like genes, whose expression is among the most highly induced by temperature upshift (Table S8 and S9). Northern blot analysis showed that HSP104 expression was induced by HSF1 overexpression (Figure 6F), suggesting that Hsf1 promoted HSP104 expression, required for heat shock resistance.

C. neoformans contains two SSA1-like genes. One is CNAG_01727 and the other is CNAG_01750. Adler et al. (2011) first characterized CNAG_01727 (named Ssa1), which is orthologous to an Hsp70 member protein, Ssa1, in S. cerevisiae. Intriguingly, Ssa1 is a DNA-binding transcriptional coactivator of laccase in C. neoformans (Zhang et al. 2006). CNAG_01750 also encodes an Ssa1-like protein in C. neoformans, named SSA2. To discriminate SSA1 and SSA2 transcriptional regulation patterns, we performed qRT-PCR with primer sets specific to each gene (Figure 6G). Basal SSA1 expression levels were much higher than those of SSA2 (Figure 6G). To address the role of Ssa1 and Ssa2 in C. neoformans thermotolerance, we generated ssa1Δ and ssa2Δ mutants. Interestingly, SSA2 expression appeared to be enhanced in the absence of SSA1 (in the ssa1∆ mutant), implying that Ssa2 may complement Hsp70 function in C. neoformans. In the P_H3:HSF1 strain, both SSA1 and SSA2 expression were highly induced, strongly suggesting that Hsf1 promotes the expression of the Hsp70 genes (Figure 6G). The ssa1∆ mutant, but not the ssa2∆ mutant, showed greatly enhanced thermosensitivity (Figure 6H and Figure S7), whereas ssa1∆+SSA1-complemented strains restored thermotolerance to WT levels, further suggesting that Ssa1 is a major Hsp70 in C. neoformans. Taken together, Hsf1 may promote C. neoformans thermotolerance by inducing two major HSPs, Hsp104 and Ssa1.

Hsf1 governs a subset of temperature and oxidative stress responsive genes as both activator and repressor

To further elucidate the downstream network governed by Hsf1, we performed additional DNA microarray analyses using the HSF1-overexpressing strain. We found that a total of 722 genes were differentially modulated at statistically significant levels (P < 0.05, ANOVA) (Table S11). Among these, 56 genes showed more than twofold induction or reduction in the P_H3:HSF1 strain. Sixteen of the 56 Hsf1-responsive genes have orthologs in S. cerevisiae (Table S12). The array data further support that Hsf1 has dual functions as repressor and activator of genes required for C. neoformans thermotolerance. A number of genes, whose expression was induced by temperature upshift (25°–37°), were downregulated by HSF1 overexpression (Table S13). These include HSP12, which is a small HSP.

Notably, the expression of SRX1, which encodes a sulfiredoxin, was downregulated by HSF1 overexpression (Table S11). We recently showed that SRX1 is rapidly induced by peroxide stress and required for the recycling of peroxiredoxin (Tsa1) and oxidative stress response and adaptation (Upadhya et al. 2013a). The qRT-PCR analysis confirmed that basal SRX1 expression levels were reduced by HSF1 overexpression (Figure 7A). However, peroxide-dependent SRX1 induction was normally observed in the P_H3:HSF1 strain like in the WT strain (Figure 7B), indicating that Hsf1 represses only the basal expression of SRX1. The P_H3:HSF1 strain was as resistant to H₂O₂ and tBOOH as the WT strain, supporting these data (Figure 7C). Nevertheless, HSF1 overexpression further exacerbated the resistance of the sch9∆ mutant to H₂O₂ (Figure 7C). In addition, HSF1 overexpression increased the sensitivity to diamide, which is a thiol-specific oxidant (Figure 7C), indicating that Hsf1 plays a negative role in the oxidative stress response. However, we also found that HSF1 expression was indeed induced by oxidative stresses (Figure 7D). The fact that HSF1 overexpression decreased resistance to oxidative stresses prompted us to examine whether the virulence of the P_H3:HSF1 strain is attenuated. In the insect host model system using G. mellonella (set at 37°), the P_H3:HSF1 strain exhibited WT levels of virulence (Figure 7E), indicating that HSF1 overexpression did not affect C. neoformans virulence.

Figure 7

Hsf1 plays dual roles in oxidative stress response and adaptation. (A) HSF1 overexpression decreases the SRX1 basal expression levels. The qRT-PCR was performed with three technical replicates. Error bars represent SD. This experiment was repeated twice and one representative experiment is presented. (B) Hsf1 does not regulate H₂O₂-mediated induction of SRX1. WT and P_H3:HSF1 (YSB2200 and YSB2201) strains grown to logarithmic phase (OD₆₀₀ ∼ 1.0) were further incubated in YPD medium containing 2.5 mM H₂O₂ for 30 min and total RNA was isolated for Northern blot analysis. (C) HSF1 overexpression increases susceptibility to oxidative stresses. WT, sch9Δ (YSB619), P_H3:HSF1 (YSB2200 and YSB2201), and sch9Δ P_H3:HSF1 (YSB2269 and YSB2270) strains were cultured at 30° in YPD medium for 16 hr, serially diluted (1–10⁴ dilution), and spotted on YPD agar medium containing indicated concentrations of H₂O₂, tBOOH, or diamide. The plates were incubated at 30° for 2 days and photographed. The two images split by a horizontal white line in each spot assay were obtained from the same plate. (D) HSF1 expression is induced by peroxide stresses. WT and sch9Δ (YSB619) strains grown at 30° to the logarithmic phase (OD₆₀₀ ∼ 1.0) were further incubated in YPD medium containing 2.5 mM H₂O₂, 2.5 mM diamide, or 0.6 mM tBOOH for 30 min. Total RNA was isolated from treated and nontreated cells for Northern blot analysis. (E) HSF1 overexpression does not affect the virulence of C. neoformans. WT and P_H3:HSF1 strains (YSB2200 and YSB2201) were used for the wax moth virulence assay. PBS was used as a noninfection control for injection.

Hsf1 binds to conserved and divergent HSEs in the promoter of temperature-regulated genes in C. neoformans

In eukaryotic cells, Hsf1 directly binds to the conserved HSEs in the promoter region of target HSP genes. Three types of HSEs have been reported thus far: perfect (P), gap (G), and step (S) types. The P-type HSE consists of at least three contiguous inverted repeats of the nGAAn unit (e.g., nTTCnnGAAnnTTCn) (Amin et al. 1988; Xiao and Lis 1988). The G-type HSE has two inverted units of nGAAn, followed by another unit after a gap of 5 bp [e.g., nTTCnnGAAn(5 bp)nGAAn] (Santoro et al. 1998). The S-type HSE is composed of three direct repeats of nTTCn (or nGAAn) interrupted by 5 bp (Yamamoto et al. 2005). To address whether C. neoformans Hsf1 is able to bind to known HSEs, we scanned the promoter regions of Hsf1 target genes identified by both our DNA microarray analysis and HSE study in S. cerevisiae (Yamamoto et al. 2005). Among these, four genes (SSA1, SSA2, HSP78, and KAR2) had a step-type-like HSE and a single gene (STI1) had a perfect- or gap-type-like HSE (Figure 8A). Only SSA2 had a highly conserved step-type HSE (gTTCtcgagagTTCtggtggaTTCg).

Figure 8

ChIP PCR reveals direct target genes of Hsf1. (A) Schematic representation of the HSEs in the promoter region of heat shock protein (HSP78) and chaperone genes (STI1, SSA1, SSA2, and KAR2). Red characters on the sequence indicate the conserved Hsf1 binding motif (nGAAn or nTTCn) and yellow characters indicate the divergent Hsf1 binding motif. The HSE was predicted by alignment of perfect-type, gap-type, or step-type motif with the sequence of the promoter region of each gene by Clustal X2 software. (B) STI1, SSA1, SSA2, HSP78, and KAR2 are direct targets of Hsf1. qPCR was performed with three technical replicates. The error bars indicate SD. (C) Hsf1 does not regulate its own transcription. qPCR was performed with three technical replicates. The error bars indicate SD.

To test whether Hsf1 interacts with these candidate binding sites, we performed ChIP with the C. neoformans strain expressing Hsf1-4xFLAG, followed by qPCR with primers flanking the putative HSE. Notably, Hsf1 appeared to bind to the promoter regions of four genes having step-type HSEs (SSA1, SSA2, HSP78, and KAR2) even at 30° (Figure 8B). However, upon temperature upshift to 37°, its binding capacity dramatically increased, suggesting that Hsf1 directly binds to these target genes as a transcriptional activator (Figure 8B). Even the Hsf1-binding capacity to STI1, presenting a perfect/gap-type-like HSE, was also increased, albeit to a lesser extent, by temperature upshift. By contrast, Hsf1 did not bind to unrelated control genes (ACT1 and β-tubulin genes) at either 30° or 37°, indicating that Hsf1-binding is specific to its binding elements. In conclusion, Hsf1 appeared to bind to the step-type HSE more strongly than to the perfect- or gap-type HSE in C. neoformans.

Next, we addressed whether Hsf1 regulates its own transcription to explain why HSF1 expression is downregulated during temperature upshift in C. neoformans. In Candida albicans, Hsp90 interacts with Hsf1 and negatively regulates Hsf1 activity (Leach et al. 2012). Therefore, it is possible that increased expression of Hsp90 may inhibit Hsf1, which subsequently reduces autoinduction of HSF1. To address this hypothesis, we first checked whether the promoter region of HSF1 contains HSEs. By alignment using Clustal X2, we found several conserved HSEs on the 1-kb upstream region of HSF1 (Table S14). Perfect type 1 (from −394 to −382), the reverse sequence of perfect type 1 (from −795 to −783), gap type 3 (from −134 to −117), step type (from −197 to −175) and the reverse sequence of step type (from −410 to −388) were well-conserved HSEs in this region. We checked the location of adjacent genes on the genome, to exclude that these putative HSEs are located on the coding sequence of the adjacent gene. By comparing the location of the HSF1 promoter region to that of HSEs, we found that two HSEs, gap type 3 and step type HSEs, were located within the HSF1 promoter region. We performed ChIP followed by qPCR using primers for the promoter region of HSF1 (from −229 to −130), which can cover the gap-type HSE and step-type HSE, to identify whether Hsf1 can bind to these HSEs. Hsf1 did not appear to bind to these HSEs in its promoter region at both 30° and 37° conditions (Figure 8C). In conclusion, HSF1 expression did not appear to be autoregulated during temperature upshift in C. neoformans.

Discussion

In this study, we focused on identifying genes that regulate thermotolerance, dependent or independent of Sch9, by performing DNA microarray-based transcriptome analyses and characterizing how the Sch9 protein kinase and one of its downstream TFs, Hsf1, govern C. neoformans thermotolerance. Our findings are summarized in Figure 9. In response to high temperature (37°–40°), several molecular chaperone and HSP genes are upregulated in C. neoformans, while genes involved in ribosome biogenesis/translation and ergosterol biosynthesis are downregulated. During this process, Sch9 appears to play both positive and negative roles in controlling the expression of a plethora of genes involved in thermotolerance, including HSP104. We found that the evolutionarily conserved heat shock factor, Hsf1, is an essential protein and undergoes transcriptional downregulation during temperature upshift, which is partly regulated by Sch9. Nevertheless, Hsf1 is transiently phosphorylated upon temperature upshift and promotes C. neoformans thermotolerance by regulating a subset of thermotolerance genes. In contrast, HSF1 expression is induced by oxidative stress, but Hsf1 represses resistance to oxidative stresses. Therefore, Hsf1 likely functions both as a transcriptional activator and repressor for thermotolerance and oxidative stress responses in C. neoformans.

Figure 9

Proposed regulatory mechanism of Sch9 and Hsf1 in thermotolerance of C. neoformans. C. neoformans Sch9 regulates thermotolerance in an Hsf1-dependent and Hsf1-independent manner. In response to temperature upshift, HSF1 transcript levels decrease partly in an Sch9-dependent manner. However, under basal conditions, protein levels, not transcript levels, of Hsf1 appear to be positively regulated by Sch9, suggesting that Sch9 may be involved in Hsf1 stability. Hsf1 is required for growth and control genes involved in oxidative stress response and thermotolerance. Hsf1 may regulate thermotolerance both positively and negatively. Hsf1 directly binds to the step-type-like HSEs within the promoter region of HSPs even under unstressed conditions and its binding markedly increases upon temperature upshift. Hsf1 increases the expression of KAR2, a known downstream factor of the UPR pathway. The UPR pathway is also involved in C. neoformans thermotolerance. In conclusion, Sch9 is involved in thermotolerance by regulating Hsf1, the UPR pathway, and other unknown HSP factors in C. neoformans.

We compared our array data of the WT strain to those from other microarray analyses performed independently to identify temperature-regulated or heat-shock responsive genes in C. neoformans (Steen et al. 2002; Kraus et al. 2004; Chow et al. 2007). Steen et al. (2002) demonstrated that the tag representing cyclophilin A (CPA1 and CPA2) is expressed at higher levels at 37° than at 25°. Our array data also confirmed this finding (Table S3 and Table S4). Kraus et al. (2004) found that the expression of a number of genes, including SLG1, MGA2, CLC1, RDS1, SMG1, TPS1, and RIM15, are induced at 37° and that genes like ILV5, ILV2, and URA2 are repressed at 37°. We also found that SLG1, MGA2, RDS1, SMG1, TPS1, and RIM15 expression was induced during temperature upshift, while ILV2, ILV5, and URA2 expression was repressed (Table S3 and Table S4). Chow et al. (2007) demonstrated that a number of heat shock genes (KAR2, HSC8202, HSP104, HSP78, SSA4, STI101, SSE2, and LHS1) are induced during temperature shift. Our array analysis confirmed that expression of some of the heat shock genes, including KAR2, HSP104, and HSP78, was strongly induced by temperature shift and also revealed that a number of other heat shock genes, including HSP12, SSB2, SSC1, SSA1-like gene (CNAG_01727), SSE1, AHA1, CDC37, and HSP10, also showed this effect. Chow et al. (2007) reported that the expression of genes involved in ergosterol biosynthesis (ERG3, ERG4, ERG502, ERG25, and OSH6) is repressed by heat. In addition to the reported ergosterol biosynthesis genes (ERG3, ERG4, and ERG25), our study also revealed that the expression of a number of other ergosterol biosynthesis genes, including ERG2, ERG5, ERG6, ERG8, ERG10, ERG11, ERG13, ERG20, ERG24, NCP1, and MVD1, was repressed by high temperature.

One common finding made by this and previous studies was that genes involved in oxidative stress response are generally upregulated at high temperatures. Steen et al. (2002) reported that a peroxiredoxin (TSA1) and a superoxide dismutase (SOD) are highly expressed at 37°. Similarly, Chow et al. (2007) reported that the expression of the genes involved in the thioredoxin system (TRR1, thioredoxin reductase; TRX3, thioredoxin) is induced by the heat shock response. In addition, Kraus et al. (2004) reported that several catalase genes were induced by high temperature. Supporting these data, we also found that the expression of the thioredoxin system genes, including TSA1, TRR1, and TRX3, was highly induced by high temperatures. Furthermore, our analysis also revealed that SRX1, encoding a sulfiredoxin that is required for the recycling of Tsa1 (Upadhya et al. 2013b), was also induced by high temperature. In addition, catalase (CNAG_04981 and 00575) and the two superoxide dismutase (SOD1 and SOD2) genes were strongly induced by high temperatures. In conclusion, our microarray data for identifying temperature-regulated genes in C. neoformans are highly consistent with those of previous transcriptome analyses, supporting the quality of our data, and also they identified previously unnoticed, temperature-regulated genes in the pathogen.

The most notable finding of this study is that HSF1 expression is downregulated, partly through Sch9, during temperature upshift, while HSF1 expression is rapidly induced in response to oxidative stress. Therefore, HSF1 expression likely undergoes dynamic regulation through either transcriptional control or mRNA stability control. This is a rather unexpected finding as Hsf1 is largely controlled post-translationally in most eukaryotic systems. In S. cerevisiae there is a genetic interaction between Sch9 and Hsf1 (Morano and Thiele 1999). The loss of SCH9 suppresses the temperature sensitivity of an hsf1 mutant strain having a C-terminal truncation of Hsf1. However, Sch9 does not directly modulate Hsf1 activity, but instead represses Hsp90 chaperone activity. Therefore, deletion of SCH9 derepresses Hsp90 chaperone complex activity, which results in the restoration of the thermotolerance of the hsf1 mutation. Hsp90 is known to be directly activated by Hsf1 and also serves as a key negative feedback regulator of Hsf1 in most fungi (Leach et al. 2012). In our study, we also found that HSP90 expression was strongly induced upon temperature upshift. The fact that overexpression of HSF1 did not alter basal expression levels of HSP90 but enhanced induced levels of HSP90 in response to high temperature indicates that Hsf1 may also upregulate HSP90 induction as a negative feedback regulation in C. neoformans. Our study demonstrated that Sch9 partly controls HSF1 expression directly or indirectly, but whether Sch9 directly affects Hsf1 activity per se remains unknown. It is possible that Sch9 may affect HSF1 expression levels through interaction with Hsp90 in C. neoformans, although Sch9 does not belong to the known fungal Hsp90 interactome (Zhao et al. 2005; McClellan et al. 2007; Diezmann et al. 2012). This possibility needs to be further addressed in future studies.

In spite of reduced expression of HSF1 during temperature upshift, Hsf1 promotes C. neoformans thermotolerance, as demonstrated by the increased thermoresistance of the HSF1 overexpressing strain. Furthermore, Hsf1 undergoes phosphorylation transiently during adaptation to high temperature. These findings implicate that Hsf1 may also undergo post-translational regulation, including oligomerization and/or phosphorylation, in response to thermal stress. In animal cells, in the absence of external stresses, Hsf1 is maintained as a monomeric, inactive form without DNA-binding and trans-activating capacity and shuttles between the nucleoplasm and cytoplasm. In response to proteotoxic stresses (high temperature, oxidative stress, and heavy metals), Hsf1 undergoes monomer-to-trimer transition, accumulates in the nucleus, and then binds to HSEs, which contain arrays of 5-bp sequences (nGAAn). Subsequently, Hsf1 is post-translationally modified (including phosphorylation) to achieve trans-activating competence and ultimately activates diverse HSPs (Anckar and Sistonen 2011). In this model, Hsp90, a molecular chaperone, plays a key role in controlling Hsf1 activation by interacting with unfolded proteins. However, in S. cerevisiae, the trimeric Hsf1 constitutively binds to the HSE even in unstressed cells and undergoes cooperative Hsf1 interaction with the HSE, through hyperphosphorylation, to increase its transcriptional activity in response to heat shock (Sorger et al. 1987; Jakobsen and Pelham 1988; Erkine et al. 1999). The regulatory mechanism of C. neoformans Hsf1 seemed to be more similar to that of S. cerevisiae Hsf1 than that of mammalian Hsf1. Our ChIP–PCR analysis demonstrated that in C. neoformans Hsf1 appeared to bind constitutively to the HSE of some HSP genes (e.g., SSA1, SSA2, HSP78, and KAR2) under unstressed conditions (30°), but exhibited increased binding at high temperature (37°). How Hsf1 precisely undergoes post-translational control, such as oligomerization, in response to thermal and oxidative stresses, should be addressed in future studies.

The role of Hsf1 in oxidative stress response has also been suggested in mammals and S. cerevisiae. In mammals, a reduction of HSF1 perturbs redox homeostasis and increases mitochondrial oxidative damage (Yan et al. 2002). In S. cerevisiae, oxidative stress induces a more sustained phosphorylation of Hsf1 and expression of CUP1, which encodes a metallothionein protein and protects cells from toxic heavy metals and other cytotoxic agents, while heat shock only induces transient phosphorylation of Hsf1 and CUP1 expression, indicating that Hsf1 mediates distinct cellular responses to the two different stresses (Liu and Thiele 1996). In this process, Skn7, which has a similar domain structure as that of Hsf1, directly interacts with Hsf1 to induce HSPs in response to oxidative stresses (Raitt et al. 2000). Although interaction between Skn7 and Hsf1 is not known at this point in C. neoformans, the previous findings that Skn7 is involved in oxidative stress response in C. neoformans (Wormley et al. 2005; Bahn et al. 2006; Coenjaerts et al. 2006) indicates a potential cooperative interaction between the two TFs. Therefore, it is possible that Hsf1 may mediate distinct transcriptional and post-transcriptional responses to varying environmental stresses in C. neoformans.

Based on our ChIP–PCR-based analysis, Hsf1 appears to preferentially bind to the step-type HSE than perfect- or gap-type HSEs. However, C. neoformans HSEs present some variations compared to S. cerevisiae HSEs. For example, the Hsf1-binding capacity of some HSP genes presenting less-conserved step-type HSEs, including SSA1, HSP78, and KAR2, was as strong as that of SSA2 presenting well-conserved step-type HSE. Therefore, C. neoformans Hsf1 may have a more flexible DNA binding motif than S. cerevisiae Hsf1. Whether Hsf1 could bind a different DNA binding motif(s) during oxidative stress responses remains to be elucidated. More comprehensive genome-wide analyses of Hsf1 binding elements (e.g., using ChIP-sequencing methods) are needed in the future.

In conclusion, the data presented by this study provide new insights into the regulatory mechanism of thermotolerance of C. neoformans, a pathogen responsible for meningitis in humans worldwide.

Acknowledgments

This work was supported by National Research Foundation of Korea grants (2015R1A2A1A15055687) from the Ministry of Science, ICT and Future Planning and the Strategic Initiative for Microbiomes in Agriculture and Food funded by Ministry of Agriculture, Food and Rural Affairs (916006-2) (to Y.-S.B.). The work was also supported in part by National Institutes of Health grants R21-AI094364 (to A.I.), R37-AI39115-19 and R01-AI50113-12 (to J.H.), R01-GM041840 (to D.J.T.), and F32-GM100678 (to R.A.F.).

Footnotes

Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10.1534/genetics.116.190595/-/DC1.

Communicating editor: A. P. Mitchell

Author notes

1

Present address: Molecular Microbiology Section, Laboratory of Clinical Infectious Diseases, NIAID, NIH, Bethesda, MD 20892.

2

Present address: Department of Microbiology and Immunology, Virginia Commonwealth University, Richmond, VA 23298.