Cryptococcus neoformans requires the TVF1 gene for thermotolerance and virulence

Abstract

Cryptococcus neoformans is the primary causative agent of cryptococcosis. Since C. neoformans thrives in environments and its optimal growth temperature is 25–30°C, it needs to adapt to heat stress in order to cause infection in mammalian hosts. In this study, we aimed to investigate the role of an uncharacterized gene, CNAG_03308. Although the CNAG_03308 deletion strain grew as well as the parent strain KN99, it produced yeast cells with abnormal morphology at 37°C and failed to propagate at 39°C. Furthermore, the deletion strain exhibited slower growth at 37°C in the presence of congo red, which is a cell wall stressor. When cultured at 39°C, the deletion strain showed strong staining with fluorescent probes for cell wall chitin and chitosan, including FITC-labeled wheat germ agglutinin, Eosin Y, and calcofluor white. The transmission electron microscopy of the deletion strain revealed a thickened inner layer of the cell wall containing chitin and chitosan under heat stress. This cell-surface altered deletion strain induced dendritic cells to secrete more interleukin (IL)-6 and IL-23 than the control strains under heat stress. In a murine infection study, C57BL/6 mice infected with the deletion strain exhibited lower mortality and lower fungal burden in the lungs and brain compared to those infected with the control strains. Based on these findings, we concluded that CNAG_03308 gene is necessary for C. neoformans to adapt to heat stress both in vitro and in the host environment. Therefore, we designated the CNAG_03308 gene as TVF1, which stands for thermotolerance and virulence-related factor 1.

Introduction

Cryptococcus neoformans is a fungal pathogen ubiquitous in our environment. The fungus causes cryptococcosis primarily in patients with readily identifiable immunodeficiencies such as AIDS. It is estimated that more than 220 000 people globally are affected by the disease annually, with approximately 180 000 death.¹

The environmental source of C. neoformans includes soil, tree, and bird droppings, and its optimal growth temperature is 25°C–30°C. In the host environment, the temperature can be higher than 37°C, which is a heat-stress condition for cryptococcal cells. The genes involved in heat tolerance are necessary for adaptation to the host environment, and deletion of these genes resulted in reduced virulence in an experimental infection study.^2,3 In a recent study, Stempinski and coworkers identified 46 heat-sensitive strains derived from C. neoformans standard strain H99 by screening a deletion library of 4031 genes, and revealed the gene network required for heat tolerance.⁴ For example, strains lacking G proteins such as septins are included in the 46 heat-sensitive strains. In the H99 strain, four CDC genes (CDC3, CDC10, CDC11, and CDC12) code for a septin, and the septin-deficient strains grow as well as the wild-type at optimal temperatures, while they show incomplete cell division, pseudo-hyphal morphology, or chains of multiple cells at 37°C.^4,5 In Saccharomyces cerevisiae and Candida albicans, septins are required for normal division regardless of growth temperature, and in particular, defects in CDC3 and CDC12 are lethal.⁶ In C. neoformans, however, the deficiency of CDC3 or CDC12 results in the loss of septin complexes at the bud neck under the optimum temperature, but grows normally.⁵ It is poorly understood why C. neoformans specifically requires septins for mitotic growth only at high temperature. Rho-like G proteins Cdc42 (CNAG_05348) and Cdc420 (CNAG_05968) and aniline-like protein Bud4 (CNAG_06902) are required for the formation of the septin complex, and their deficiency results in mitotic abnormalities at 37°C, like septin deficiency.^7,8 The double-deficient strains cdc42∆/cdc420∆ and single-deficient strains bud4∆ proliferate as well as the wild-type strain at optimal temperatures of 25°C–30°C, although they fail to grow at 37°C.^7,8Cryptococcus neoformans also possesses Rho-like G proteins, Rac1 (CNAG_02883) and Rac2 (CNAG_05998), but their single deficiency does not affect proliferation at 37°C.⁹ The Ras-like G proteins Ras1 (CNAG_00293) and Ras2 (CNAG_04761) and the guanine nucleotide exchange factor Cdc24 (CNAG_04243) are known to regulate the activation of Rho-like G proteins. Cryptococcus neoformans lacking Ras1, Ras2, or Cdc24 does not place actin properly in daughter cells during cell division and fails to proliferate at 37°C.^10,11 Thus, these findings suggest that the Ras–Cdc42–Septin cascade is required for proper budding and proliferation in heat-stressed environments.

Rho-like G proteins Rho1 (CNAG_03315)/Rho10 (CNAG_03130)/Rho11 (CNAG_06606) and the associated PKC pathway, MAPK pathway, and calcineurin pathway seem to be involved in thermotolerance by different mechanisms from the Ras–Cdc42–Septin cascade.¹² Since those signaling pathways also play a role in sensing cell wall stress and regulating synthesis of cell wall components, strains deficient in the above pathways are sensitive to cell wall stress reagents such as calcofluor white (CFW) and congo red (CR), or cell membrane stress reagents such as fluconazole and sodium dodecyl sulfate (SDS).^2,12 In addition, the endoplasmic reticulum stress response (unfolded protein response [UPR]) is also required for thermotolerance since misfolded proteins accumulate in the endoplasmic reticulum under high-temperature conditions. In particular, (1) proteasomal degradation of aberrant proteins (Ubp5),¹³ (2) protein refolding by chaperones (Hsp104, Dnj1, and Ire1),^14–16 and (3) repression of protein synthesis by mRNA degradation (Rpb4 and Ccr4)^17,18 are responsible for UPR. Since UPR is also required for cell wall integrity, UPR-deficient mutants are sensitive to cell wall stress reagents such as CFW and CR.^16,19 Some UPR mutants increase the synthesis of chitin, chitosan, or glucan, and rearrange cell walls.^15,18 These findings indicate that the heat-stress response is closely related to cell wall integrity.

In our research series, we fortuitously found that the CNAG_03308-deficient strain is susceptible to heat and cell wall stress conditions. The putative protein coded by CNAG_03308 was found to have low homology to Axl2p of S. cerevisiae.²⁰ In the present study, we characterized the physiological role of CNAG_03308 in C. neoformans and demonstrated that the gene is necessary for thermotolerance and virulence, thus naming the gene thermotolerance and virulence-related factor 1 (TVF1).

Materials and methods

Ethics

All animal experiments followed the guidelines and policies of the Principles of Morality for Animal Experiments of the National Institute of Infectious Disease, Japan (approval number 121176).

Sequence comparison

Nucleotide and amino acid sequences were downloaded from FungiDB (fungidb.org/fungidb/app). Amino acid sequences were aligned in ClustalW (www.genome.jp/tools-bin/clustalw), and homology was evaluated in Sequence Manipulation Suite (www.bioinformatics.org/sms2/ident_sim.html). Amino acid motifs, domains, and their amino acid positions are depicted as presented in FungiDB.

Fungal strains

Cryptococcus neoformans strain lacking CNAG_04514 (MPK1) was obtained from the gene disruption library maintained at the Fungal Genetics Stock Center (FGSC; www.fgsc.net) and used as a positive control in various experiments.

To disrupt CNAG_03308 (CKF44_03308) nucleotide sequence in C. neoformans strain KN99 genome, the open reading frame (ORF) of CNAG_03308 was replaced by a gene disruption cassette containing the geneticin (G418) resistance gene. Both ends of the disruption cassette also possessed a homologous sequence (approximately 1 kb) to the upper and downstream of ORF for inserting it into the target locus via homologous recombination. The upstream homologous fragment was amplified with primers hAxl2-A (5′-GTGTCTGGTGATTGATACGTCG-3′) and hAxl2-B (5′-GCTAGTTTCTACATCTCTTCCGTGCAGAGGATTTGTC-GGAGTTGAG-3′) using genomic DNA of KN99 strain as a template. Similarly, for the amplification of the downstream fragment, primers hAxl2-C (5′-CGCCGCTCTCCAGCTCACATCCTCTGGTAAAGTAGGA-GGACGACG-3′) and hAxl2-D (5′-AATAGAGAGCTGCGAAGACC-3′) were used. The DNA fragment (approximately 1.8 kb) containing the G418-resistant gene was amplified with primers A1 (5′-CACGGAAGAGATGTAGAAACTAGC-3′) and B2 (5′-GAGGATGTGAGCTGGAGAGC-3′) using plasmid pJAF1²¹ as a template. Using those fragments, the two homologous regions, and the drug resistance gene fragment as templates, overlap PCR was done to generate a gene disruption cassette using primers hAxl2-A and hAxl2-D. This gene disruption cassette was introduced into KN99 using a helium-driven biolistic system (PDS-1000/He™, Bio-Rad), and transformants were screened on yeast extract-peptone-dextrose agar [YPD; 1% (w/v) yeast extract, 2% (w/v) bacto peptone, 2% (w/v) dextrose, 2% (w/v) agar; BD Difco] containing 120 µg/ml G418 (Thermo Fisher Scientific). To verify whether the disruption cassette was inserted into the targeted locus on the genome, PCR was performed with primers A1 and hAxl2-R6079 (5′-ACGCCGAGGCTCAAATTCAAG-3′) using the genomic DNA of transformants as a template. The presence or absence of target genes in transformants was confirmed by PCR using the primers hAxl2-seqF4 (5′-CGCCGCCGTCGAAATAATAAGAAC-3′) and hAxl2-R3149 (5′-GGGATCTGCTCTTCTCGCATC-3′) with genomic DNA from the transformants as templates.

To construct complemented strains, we constructed a restoring plasmid carrying the promoter, terminator, and coding region of CNAG_03308. For making the restoring plasmid, the gene locus containing CNAG_03308 in KN99 strain was amplified with primers hAxl2-G (5′-TTGGTACCGAGCTCGGATCCGCGGCCGCGTAGCGCA-GATCGTACATCTTG-3′) and hAxl2-H (5′-CACT-GGCGGCCGTTACTAGTTGACCGCTTTGACCACTTTC-3′), and then the amplified fragment was cloned into the SpeI/BamHI site of plasmid pJAF15²¹ carrying the hygromycin resistance gene. The cloned sequence of CNAG_03308 was analyzed using the following eight primers and confirmed to be consistent with the sequence presented in FungiDB: Seq-Vecotor-Fw1 (5′-CTACAGACAACAATACCATCCTTCC-3′), hAxl2-seqF1 (5′-GCATTGAGCAATAAGCTCTCC-3′), hAxl2-seqF2 (5′-TATGGTACACCTGCCTCAAG-3′), hAxl2-seqF3 (5′-CGATTGGCTATCCTTTAACAGGTG-3′), hAxl2-seqF4 (5′-CGCCGCCGTCGAAATAATAAGAAC-3′), hAxl2-seqF5 (5′-ACGGCATTCGAGACTCTTTCCAG-3′), hAxl2-seqF6 (5′-GAGTTTACGCCGCATCCAACG-3′), and hAxl2-seqF7 (5′-CGGGTGTGGTCATTAGTCTTG-3′). The restored plasmid was digested with NotI to obtain a linear fragment (the complementary cassette, approximately 7.1 kb) containing CNAG_03308 and the hygromycin resistance gene. This complementary fragment was introduced into the CNAG_03308 deficient strain using a helium-driven biolistic system, and then transformants were isolated on YPD agar containing 250 µg/ml hygromycin (Thermo Fisher Scientific). We also confirmed that the complementary strain did not grow on YPD agar containing geneticin because the G418 resistance gene in the deletant was replaced by the complementation cassette and disappeared from the genome. To check that the complementation cassette was inserted at the correct position and CNAG_03308 was reintroduced, PCR was performed using the genomic DNA of the transformant as a template, using primers A1 and hAxl2-R6079, or primers hAxl2-seqF4 and hAxl2-R3149.

Thermotolerance test

Cryptococcus neoformans strains were inoculated into YPD broth and incubated at 30°C overnight with shaking. After washing the cryptococcal cells with phosphate-buffered saline (PBS, Nacalai Tesque, Inc.), 1 × 10⁷ cells/ml fungal suspension and its serial dilutions were prepared. Five microliters of each suspension were spotted onto YPD agar plates and incubated for 2–6 days at 25°C, 30°C, 37°C, or 39°C.

For the other thermal tolerance test using a liquid medium, overnight culture was washed with PBS and inoculated into fresh YPD broth at a concentration of 1 × 10⁵ cells/ml and cultivated at 39°C for 7 days with shaking. The optical density of the culture was measured using a spectro photometer model AE-450 (ERMA Inc). This culture was transferred into a microtube, and the image of the microtube was captured with a scanner (GT-X970, Epson) before and after centrifugation.

Morphology

Cryptococcus neoformans strains were inoculated into YPD broth at 1 × 10⁶ cells/ml for 24 h with shaking. For the cultivation under 39°C, the culture scale was set to 10 ml because of their low yield at 39°C. After rinsing cryptococcal cells with PBS, 3 µl of the concentrated suspension was spotted onto a glass slide and sealed with a cover glass. Cell morphology was observed by the differential interference contrast with microscopy (IX81, Olympus). Five independent fields were photographed in each condition, and the number of cells was counted using ImageJ (https://imagej.nih.gov/ij/). The percentage of cells with four or more connected cells out of the total number of cells was calculated. Cryptococcal cells were also analyzed using FACSymphony™ A3 cell analyzer (BD Biosciences), and a dot plot was depicted using FlowJo software (Tree Star, Inc.).

Cell wall, plasma membrane, and osmotic stress test

Indicated amounts of CR (Wako Pure Chemical Industries, Ltd. #039-23192), CFW (MP Biomedicals #158067), SDS (Wako Pure Chemical Industries, Ltd.) or sodium chloride (NaCl; Wako Pure Chemical Industries, Ltd.) were mixed with YPD agar powder and autoclaved (121°C, >120 kPa, 15–20 min). Cryptococcus neoformans strains were cultivated in YPD broth for 1 day with shaking and washed with PBS, and then fungal suspension (1 × 10⁷ cells/ml) was prepared. Then, 10-fold serial dilutions were made, and 5 µl of each suspension were spotted onto the agar plate. The spotted plates were incubated at 30°C or 37°C for 2 days, and colony formation was photographed (GT-X970, Epson).

Cell wall staining with fluorescent probes

Cryptococcal cells were cultivated in the same manner as for the study of morphology. For staining with FITC-labeled wheat germ agglutinin (WGA-FITC; Vector Laboratories, Inc. #FL-1021), CFW (sigma #18909-100ML-F), and Dectin-1 Fc,²² cells were washed with PBS. McIlvain’s buffer pH 6.0 and Hanks’ Balanced Salt Solution containing calcium and magnesium (HBSS+: Gibco) were used for staining with Eosin Y (Wako Pure Chemical Industries, Ltd.) and Dectin-2 Fc (Enzo Life Science, Inc.), respectively. The concentrations of fluorescent probes and staining time were: 100 µg/ml WGA-FITC, 300 µg/ml Eosin Y, and 100 µg/ml CFW (1:10 dilution) for 30 min at room temperature. The staining with Dectin-1 Fc and Dectin-2 Fc was according to our previous study.²² For microscopy, an Olympus IX81 was used with filters WU (BP330-385, DM400, and BA420) and GFP (BP460–480, DM485, and BA495–540). For flow analysis, FACSymphony™ A3 cell analyzer or BD FACSCanto™ II was used, and flow data were analyzed by FlowJo. The amount of fluorescent probe binding to each strain was compared with the median fluorescence intensity (MedFI).

Transmission electron microscopy

Cryptococcal cells cultivated as described for morphological study were washed 3 times with distilled water and then suspended in fixation solution [0.1 m sodium cacodylate buffer pH 7.4 (Nacalai Tesque, Inc.), containing 2.5% (v/v) glutaraldehyde (Electron Microscopy Sciences, Inc.) and 2% (w/v) paraformaldehyde (Taab Laboratory Equipment Ltd)] and fixed for 3 days at 4°C. The samples were then post-fixed in the osmium fixation solution [1% (w/v) osmium tetroxide (Nisshin EM Co., Ltd.), 1.25% (w/v) potassium ferrocyanide (Nacalai Tesque, Inc.), and 5 m m calcium chloride (Wako Pure Chemical Industries, Ltd.)] and embedded in 2% (w/v) Low Melting Point Agarose (Nacalai Tesque, Inc.). After staining the block with 0.5% aqueous uranyl acetate (Merck Co., Ltd.), the specimens were dehydrated and embedded in Spurr resin (Polysciences, Inc.). Thin sections were mounted on copper grids and post-stained with saturated uranyl acetate and lead citrate. Specimens were observed using an HT7700 transmission electron microscope (TEM; Hitachi High-Technologies Corporation).

Immune stimulation to dendritic cells

The preparation of heat-killed (HK) cryptococcal cells and bone marrow-derived dendritic cells (BMDCs) and measurement of cytokines were carried out according to our previous study.²³ Briefly, fungal cells were cultured in YPD overnight and inactivated by heat treatment (94°C–98°C) for 1 h. BMDCs are nonadherent cells differentiated from bone marrow cells with 10 ng/ml granulocyte-monocyte colony-stimulating factor (GM-CSF) for 6–7 days. BMDCs (2 × 10⁵ cells/well) and HK-cryptococci were added to flat-bottomed 96-well plates (Corning #3595) and incubated for 24 h. Interleukin-6 (IL-6) OptEIA ELISA Kit (BD Biosciences) and IL-23 DuoSet ELISA Kit (R&D Systems) were used to measure the cytokines.

Animal study

Seven-week-old female C57BL/6 J mice were purchased from Japan SLC, Inc. The mice were reared in a specific pathogen-free environment at the animal experiment facility of the National Institute of Infectious Diseases of Japan. The organ fungal burden and mortality were measured according to our previous studies.^23,24 Briefly, yeast cells were cultured in YPD for 24 h with shaking and washed with PBS. The fungal suspension (3 × 10³ cells/30 µl) was intranasally administered. For anesthesia, 0.3 mg/kg medetomidine (Nippon Zenyaku Kogyo Co., Ltd.), 5 mg/kg midazolam (Maruishi Pharmaceutical Co., Ltd.), and 5 mg/kg butorphanol (Meiji Animal Health Co., Ltd.) were administered intraperitoneally, and 0.3 mg/kg atipamezole (Nippon Zenyaku Kogyo Co., Ltd.) was intraperitoneally administered to induce early recovery from anesthesia. Two and three weeks after infection, the mice were euthanized by carbon dioxide gas, and organs were removed and homogenized using the Omni Tissue Homogenizer Model TH115 (Omni International, Inc.). Fungal loads (colony-forming unit [CFU]) were determined by dilution plating onto YPD agar.

Histology

Lung samples were fixed in 10% neutral buffered formalin (Wako Pure Chemical Industries, Ltd. #060-01667). Paraffin embedding, sectioning (3-µm), and staining (hematoxylin–eosin [HE] and Grocott–Gomori’s methenamine silver stain [GMS]) were performed by Biopathology Institute Co., Japan. The specimens were imaged using a slide scanner SLIDEVIEW VS200 (Olympus), and the images were analyzed using the image analysis software OlyVia (Olympus).

Analysis of capsule formation, melanin production, and urease activity

The evaluation of capsule formation, melanin, and urease production followed previous studies.^15,23 Briefly, fungal cells growing in YPD medium were washed with PBS, and a fungal suspension (2 × 10⁸ cells/ml) was prepared. In a 25 cm² cell culture flask (Corning #430639), 5 ml of RPMI 1640 medium (serum-free, Nacalai Tesque, Inc. #06261-65), and 50 µl of the fungal suspension were added and incubated for 2 days at 37°C with 5% CO₂. The culture (1 ml) was centrifuged and suspended in 20 µl of India ink, and 3 µl of the suspension was spotted onto a glass slide and sealed with a cover glass. The capsules were observed under a microscope IX81 (Olympus).

For observation of melanin synthesis, fungal suspension (1 × 10⁷ cells/ml) was prepared after overnight cultivation in YPD, and 5 µl of the suspension was spotted onto cornmeal agar medium (BD Biosciences, #21132) in which 0.03% (w/v) caffeic acid and 0.5% (w/v) agar were added. The plates were incubated at 30°C for 6 days, and brown colonies were photographed using a scanner (Epson GT-X970).

In the urease assay, fungal suspension (1 × 10⁶ cells/ml) was prepared in a urease test medium (10 g/l glucose, 1 g/l asparagine monohydrate, 3 g/l KH₂PO₄, 1 g/l MgSO₄·7H₂O, 1 mg/l thiamine-HCl, 20 g/l urea, and 0.012 g/l phenol red). This suspension was added in 200 µl each to a 96-well plate (Corning #3595) and incubated at 37°C for 48 h. After incubation, the color change of liquid cultures was detected by a scanner (Epson GT-X970).

Statistical analysis

GraphPad Prism 7 (GraphPad Software, Inc.) was used for all statistical analyses. P-values of less than .05 were considered statistically significant.

Results

Cryptococcus neoformans lacking TVF1 grew as well as wild type at 25°C–37°C but barely at 39°C

The library of gene disruption strains obtained from FGSC contained tvf1∆. This mutant has been reported as unable to pass through the cell layer, mimicking the blood-brain barrier.²⁵ However, there is no other information available for CNAG_03308, and its function is poorly understood. The putative protein encoded by CNAG_03308 had 27.5% similarity with Axl2p of S. cerevisiae (Supplementary Figure 1A). Axl2p has a signal sequence and a transmembrane domain, and localizes at the budding site in S. cerevisiae.^20,26 The function of the Axl2p extracellular region is unknown, and the homology to the putative protein encoded in CNAG_03308 was 38.8%. The intracellular region of Axl2p interacts with Cdc42p and Bud4p and regulates septin localization and cell division.²⁶ The homology of the intracellular region was 17.0% (Supplementary Figure 1A). Cryptococcus neoformans requires septins for thermotolerance.^4,5 We initially found that the CNAG_03308-deficient strain from FGSC was susceptible to heat and cell wall stress. To analyze the physiological role of CNAG_03308 in detail using the reverse genetics strategy, we constructed a deficient as well as a complemented strain from the C. neoformans KN99 strain as a parental strain (Supplementary Figure 1B). A strain lacking MPK1 (CNAG_04514), which is sensitive to heat and to the reagents, including stress to the plasma membrane and cell wall, was used as a positive control in several experiments.^27,28 Colony formations of these strains on YPD agar were evaluated over time at 25°C–39°C (Fig. 1A). Since C. neoformans is an environmental fungus with an optimal growth temperature of 25°C–30°C, a host temperature of ≥37°C is heat-stressed environment. In fact, the growth rate of the KN99 strain slowed down as temperatures were increased to 37°C–39°C compared to that from 25°C to 30°C (Fig. 1A). The tvf1∆ strain formed colonies at 25°C–37°C as well as KN99 and its complemented strain, while barely growing at 39°C (Fig. 1A). The results were similar in YPD liquid medium, and the tvf1∆ strain grew poorly at 39°C (Fig. 1B). These results suggested that the TVF1 gene was required for C. neoformans to grow robustly at temperature higher than 37°C.

Figure 1.

Cryptococcus neoformans lacking CNAG_03308 (TVF1) barely grew at 39°C. The thermotolerance was evaluated in C. neoformans wild-type strain KN99, the TVF1-deficient strain (tvf1∆), and the complemented strain (tvf1∆+TVF1) cultivated in YPD agar (A) and broth (B). MPK1-deficient C. neoformans (mpk1∆) was placed as a positive control. Each experiment was conducted 3–4 times independently, and the representative results were shown. The bar graph shows the mean ± standard deviation of three experiments. *P < .05 vs. the wild-type and restored strains in analysis of variance (anova) with Tukey’s post hoc test.

TVF1 deficient strain failed to maintain normal cell morphology at 37°C–39°C and was defective in cell division

Heat-sensitive mutants of C. neoformans may exhibit incomplete cell division, unique connected cell form, and/or cell elongation at over 37°C.⁴ Next, we observed and compared the cell morphology of the tvf1∆ strain and control strains after growing at 30°C, 37°C, or 39°C (Fig. 2A). Although the growth of the tvf1∆ strain at 37°C was comparable to that of the wild-type and complemented strains (Fig. 1A), about 15% of tvf1∆ cells failed to divide and formed chains of four or more cells (Fig. 2A). The number of cells in chain increased significantly, reaching up to about 40% at 39°C (Fig. 2A). In the wild-type and complemented strains, no more than four connected cells were detected, but two or three connected cells—elongated or enlarged cells—were observed as the incubation temperature was raised (Fig. 2A). Since enlarged cells in chain should show high forward scatter (FSC) and side scatter (SSC) values in the flow analysis, we gated the population showing high FSC/SSC values and compared the percentage of such population (Fig. 2B). The increase in FSC and SSC values of wild-type and complemented strains as growth temperature was raised may represent an increase in cell enlargement and connected cells (Fig. 2B vs. 2A). At higher growth temperatures, the percentage of the subset showing high FSC/SSC increased, with the highest percentage in the tvf1∆ strain cultivated at 39°C (Fig. 2B). These results suggested that TVF1 is required for normal cell division and cell morphology under heat-stress conditions.

Figure 2.

TVF1-deficient strain failed healthy cell division at temperatures higher than 37°C. Cryptococcus neoformans strains were cultivated in YPD broth overnight, and their morphology was analyzed via microscopy with differential interference contrast (A) and flow cytometry (B). The experiment was repeated 3 times independently, and the representative photograph was shown. The percentage of cells with four or more connected cells out of the total number of cells was calculated, and the data from all three experiments were pooled to plot a bar graph of mean ± standard error (A). In the flow analysis, the population with high-intensity values of FSC and SSC was gated, and the percentage was shown as a bar graph of mean ± standard deviation. The graph and dot plot in flow analysis were representative of three independent experiments. *P < .05 in anova with Tukey’s post hoc test. nd, not detectable.

TVF1 deficient strain was more susceptible to cell wall stressor, CR, than control strains at 37°C

Heat-sensitive cryptococcal strains may be susceptible to the reagents that cause stress to cell membranes and cell walls at 25°C–30°C, and the integrity of the cell surface may be compromised.⁴ To verify whether TVF1 is involved in those stress responses, the tvf1∆ and its related strains were spotted on agar medium containing various stressors such as CR, CFW, SDS, and NaCl (Fig. 3). The loss of TVF1 did not affect the sensitivity toward these stressors at 30°C (Fig. 3, upper panels). The tvf1∆ cells were able to proliferate at 37°C, although some cells failed to divide. At 39°C, however, tvf1∆ cells barely proliferated (Figs. 1–2). Because the impact of each stressor at 39°C could not be evaluated, we examined the effects of the same stressors on tvf1∆ strain and its related controls at 37°C (Fig. 3, lower panels). The tvf1∆ strain was more susceptible to CR at 37°C than the wild-type and complemented strains, while the susceptibility to other stresses was comparable to the wild-type and complemented strains (Fig. 3, lower panels). Cryptococcus neoformans lacking MPK1 was more susceptible to CR, CFW, and SDS than the wild-type strain, which was in agreement with the previous report (Fig. 3).²⁸ These results suggest that C. neoformans requires TVF1 to maintain cell wall integrity under moderate heat-stress conditions.

Figure 3.

The TVF1-deficient cells were more susceptible to CR at 37°C than cells of the control strain. The susceptibility to the cell wall, cell membrane, and osmotic stressors was evaluated in YPD agar plates containing CR, CFW, SDS, or NaCl. The spotting plates were incubated for 2 days. The representative photographs were shown from three independent experiments.

TVF1-deficient strain has altered the properties of cell wall chitin–chitosan in a heat-stressed environment

Heat-sensitive cryptococcal strains, the UPR mutants in particular, increased the synthesis of chitin, chitosan, or glucan, and rearranged cell walls.^15,18 Next, we examined whether the tvf1∆ strain has altered the cell wall under heat stress by using various fluorescent probes that bind specifically to each component of the cell wall. The binding of the fluorescent probes to cryptococcal cells was then evaluated by microscopy and flow cytometry (Fig. 4). Wheat germ agglutinin (WGA) is a known lectin that can recognize chitin and chitosan exposed to fungal cell surfaces. In a steady state, the budding sites of cryptococcal cells are recognized by WGA.²⁹ In fact, we found FITC-labeled WGA primarily bound to the budding site of yeast cells at 30°C in TVF1-positive as well as deficient cells. At 39°C, however, only tvf1∆ strain showed more entire cells strongly stained with FITC-WGA (Fig. 4A). Flow analysis also revealed that the TVF1-deficient strain cultured at 39°C presented significantly higher fluorescence intensity by FITC-WGA (Fig. 4B), consistent with the results observed by the fluorescence microscopy. CFW and Eosin Y are small-molecule fluorescent probes that bind to chitin and chitosan, respectively, and uniformly stain the contour of cryptococcal cells in the steady state.^30–33 Both probes strongly labeled the TVF1-deficient strain cultured at 39°C, as in the case with FITC-WGA; not only the cell outline but also the entire cell was strongly stained (Fig. 4C–F, Supplementary Figure 2A). These results suggest that the TVF1-deficient strain alters the properties of chitin and chitosan in the cell wall in a heat-stressed environment.

Figure 4.

Thetvf1Δ cells subjected to heat stress were strongly labeled with the fluorescent probes for cell wall components, chitin and chitosan. The fungal cells were cultivated in YPD overnight, and labeled with WGA-FITC (A and B), Eosin Y (C and D), and CFW (E and F). The cells were observed with a fluorescent microscope (A, C, and E), and the fluorescence intensity was measured via flow cytometry (B, D, and F). The experiments were performed independently 3 times, and five photographs were taken per each condition. The histograms show the fluorescence intensity, and the MedFI was depicted as a bar graph (mean ± standard deviation). The representative data were shown. *P < .05 vs. wild-type and complementary strains at each temperature via anova Tukey’s post hoc test. ns, not significant.

The exposure of cell wall glucan on cryptococcal cells can be detected by specific antibodies or soluble recombinant immunoreceptor Dectin-1 Fc.^18,22 In response to heat stress, C. neoformans exposes glucans, although the percentage of exposed cells is about 10% even in the case of CCR4 deletion.¹⁸ This is modest compared to the fact that cell wall glucan is completely exposed when cryptococcal cells are cultured in the synthetic glucose medium referred to as YNB glucose medium.²² It has been shown that the loss of MPK1 accelerated the heat-stress-inducing glucan exposure in C. neoformans,²⁸ and we demonstrated similar findings using Detin-1 Fc (Supplementary Figure 2B). The TVF1-deficient strain also responded to heat stress and slightly increased glucan exposure, although the level of exposure was comparable to that of the wild-type and complemented strains (Supplementary Figure 2B). Another C-type lectin Dectin-2 is an immunoreceptor-recognizing cell wall mannan that can bind to cryptococcal cells cultivated in the YNB-glucose medium but not to those cultivated in YPD medium.²² We examined whether heat stress induced the exposure of cell wall mannan or Dectin-2 ligands on cryptococcal cells; however, Dectin-2 Fc did not bind to C. neoformans cultured in YPD medium regardless of the culture temperature or the status of TVF1 (Supplementary Figure 2C). These results suggested that TVF1 was not involved in the exposure of Dectin-1 and Dectin-2 ligands to the cell surface in a heat-stressed environment.

The following three layers are observed in the cell wall of C. neoformans in TEM; (1) the inner layer with high electron density containing chitin and chitosan, (2) the middle layer with low electron density, and (3) outer layer containing capsular polysaccharides.³⁴ The TVF1-deletant cultivated at 39°C presented significantly thicker inner and outer layers of the cell wall (50–100 nm thicker than the control strains) (Fig. 5). These findings appear to correlate with the observation that they were strongly stained by WGA-FITC, Eosin Y, and CFW (Fig. 4), and suggest that the TVF1-deficient strain produces a thickened cell wall inner layer that contains chitin and chitosan in a heat-stressed environment.

Figure 5.

The heat-stressed tvf1Δ cells thickened the inner layer of the cell wall. It has been known that the cryptococcal cell wall consists of the following three layers: (1) the inner layer with a high electron density containing chitin and chitosan, (2) the middle layer with a low electron density, and (3) the outer layer containing capsule polysaccharides. Each strain was cultivated in YPD overnight, and the ultrastructure of the cell wall was observed in TEM. Experiments were repeated independently 2 times, and total more than 15 cells were photographed. In each photograph, each layer was measured at 10 points, and the average value was taken as the representative value for the cell wall layer. The pooled data were depicted as bar graphs (mean ± standard deviation). *P < .05 vs. the wild-type and complemented strains in each temperature using anova with Tukey’s post hoc test. ns, not significant.

Heat-stressed TVF1-deficient cells stimulated dendritic cells to secrete IL-6 and IL-23

Cryptococcus neoformans cultured in YPD is not recognized by Dectin-1 and Dectin-2, even in acapsular strains, since neither glucan nor mannan, the pathogen-associated molecular patterns (PAMPs), are exposed.²² This is in contrast to other pathogenic fungi, such as C. albicans, which are usually recognized by Dectin-1 and Dectin-2.²² In other words, C. neoformans may be able to evade immune recognition by sophisticatedly altering its cell wall composition. It was conceivable that the TVF1-deficient cells alter immunogenic potential against innate immune cells such as dendritic cells due to an altered cell wall structure (Figs. 4 and 5). Thus, we examined whether heat-stressed TVF1-deficient cells more efficiently elicited dendritic cells to induce inflammatory cytokines IL-6 and IL-23 than other strains (Fig. 6). Considering the side effects of changing fungal properties during cultivation, heat-inactivated cryptococcal cells were used as a whole-cell antigen to stimulate dendritic cells. In addition, we compared the immune potential of fungal cells cultured at 30°C with that of those cultured at 37°C, the physiological temperature of the host at the time of infection. Since part of mpk1-deficient cryptococcal cells bound to Dectin-1 even in the steady state and the percentage of the cells that bound to Dectin-1 increased in response to heat stress (Supplementary Figure 2B), they were used as positive controls that might enhance the immunostimulatory activity. Fungal cells cultured at 30°C hardly stimulated dendritic cells regardless of their TVF1 status, and the levels of IL-6 and IL-23 secretion were comparable to those in the unstimulated group (Fig. 6). The TVF1-deficient strain cultured at 37°C induced significantly higher secretion of IL-6 and IL-23 by dendritic cells compared to the wild-type and complemented strains (Fig. 6). These results suggested that, without significant exposure of glucan or mannan, the TVF1-deficient strain subjected to heat stress enhanced immunogenic potential and induced cytokine secretion by dendritic cells.

Figure 6.

The heat-stressed tvf1∆ strain elicited IL-6 and IL-23 secretion by dendritic cells. BMDCs were cultivated in the presence of heat-inactivated cryptococcal cells for 24 h. The cytokine level in the culture supernatant was measured by enzyme-linked immunosorbent assay (ELISA). The ratio of BMDCs to fungi (multiplicity of infection [MOI]) is BMDCs:fungi = 1:1 or 1:10. Four-well copy experiments were conducted independently 2–3 times, and the representative data (mean ± standard deviation) are shown. *P < .05 vs. the wild-type and complemented strains in each control counterpart using anova with Tukey’s post hoc test. ns, not significant.

TVF1-deficient strain was less virulent in experimental animal

In general, the heat-stress response and the regulation of cell wall synthesis are associated with the pathogenicity of C. neoformans.^14,27 Highly immunogenic C. neoformans may be rapidly recognized and suppressed by innate immunity, resulting in less pathogenicity.³⁵ It has been shown that the TVF1-deficient cells did not pass through the cellular barrier, which mimics the blood–brain barrier, and it was speculated that they might be less pathogenic, but this hypothesis has not been verified in animal models.²⁵ In order to assess the role of TVF1 in pathogenicity, we compared the virulence of the TVF1-deficient strain with the wild-type and complemented strains in the experimental model of C57BL/6 J mice infected intranasally (Fig. 7). The fungal burden of TVF1-deficient cells was about 1000 times less abundant at 2 weeks post-infection in the lungs and about 100 times less at 3 weeks post-infection in the lungs and brain than the wild-type and complemented strains. These differences were statistically significant (Fig. 7A). From 2–3 weeks post-infection, the lung fungal burden of the TVF1-deficient cells increased, suggesting that the deficient strain proliferated slower in vivo than the wild-type and complementary strains. The median survival time of mice infected with the TVF1-deficient strain was 50 days, while that of the mice infected with the wild-type and complemented strains was 26 and 29 days, respectively (Fig. 7B). Log-rank test of the survival curves showed that the group infected with the TVF1-deficient strain survived a significantly longer period than those infected with the wild-type and complemented strains (Fig. 7B).

Figure 7.

Cryptococcus neoformans lacking TVF1 was less virulent than the control strains in experimental animals. The fungal cells were cultivated in YPD overnight, and the suspension (3000 cells/30 µl) was administered intranasally into C57BL/6 J mice. The fungal burden at 2 and 3 weeks post-infection (A; n = 3–4), survival rate (B; n = 8), and lung sections at 2 weeks post-infection stained with HE or Grocott–Gomori methenamine silver (GMS) (C; n = 3) were analyzed. The data from two independent experiments were pooled to depict bar graphs (mean ± standard error). The representative survival curve was shown from two independent experiments. *P < .05 vs. the wild-type and complemented strains using anova with Dunn’s or Dunnett’s post hoc test, #P < .05 vs. the wild-type and complemented strains using log-rank test.

Histological sections from 2 weeks post-infection revealed numerous yeast cells stained with Grocott–Gomori’s methenamine silver (GMS) in the lungs of mice infected with the wild-type and complemented strains, while considerably fewer yeast cells in the lungs of mice infected with the TVF1-deficient strain (Fig. 7C low-power field). Comparisons of the CFU presented in Figure 7A correspond with the histology shown in Figure 7C. In the lung sections stained with GMS, elongated cells and the cells in chains were seen with the TVF1-deficient strain, while only typical yeast cells were presented in the lungs of mice infected with the wild-type and complemented strains. These observations demonstrate that incomplete cell division of TVF1-deficient strain occurs in vivo as well as in vitro (Figs. 2 and 7C high-power field).

In the lung sections stained with HE, the immune cells, including granulocytes, macrophages, and lymphocytes infiltrated, around the fungal cells in each infected group (Fig. 7C). The multinucleated giant cells were also found, and some of them contained fungal cells (Fig. 7C, high-power field). In the lung infected with the deficient strain, overall inflammation was weaker than the other counterparts, and was correlated with the decreased number of fungal cells (Fig. 7C, low-power field). Consistent with these findings, the amount of interferon gamma (IFNγ) on 2 weeks post-infection was significantly lower in the lungs infected with the deficient strain than the other counterparts and only slightly higher than in the uninfected lungs (Supplementary Figure 3). These results suggested that the deficient strain grew slowly and elicited less inflammation in the lungs.

Taken together, these findings indicate that C. neoformans lacking TVF1 is less virulent. The general virulence-related factors such as capsule formation, melanin, and urease synthesis were comparable regardless of the TVF1 status and were unrelated to the differences in their virulence (Supplementary Figure 4).

Discussion

In the present study, we demonstrated that CNAG_03308 was required for heat-stress response and virulence in C. neoformans. Based on the findings, we advocate referring to CNAG_03308 as TVF1.

The tvf1Δ strain grew at 25°C–37°C, as well as the wild-type and complemented strains, whereas the proliferation was severely inhibited at 39°C, and the tvf1Δ strain cultured at 37°C–39°C showed unseparated cells and incomplete cell division. Why was this deficient strain susceptible to heat stress? In C. neoformans, mutants deficient in the Ras–Cdc42–Septin cascade show no abnormalities in proliferation at 25°C–30°C but suppressed proliferation at 37°C and mitotic failure, as seen in the tvf1-deficient strain.^5,8,11,36 Since Tvf1p is weakly homologous to Axl2p of S. cerevisiae (Supplementary Figure 1), which regulates the localization of septin,^20,26 it is possible that Tvf1p may also regulate the localization of septin, and that the tvf1Δ strain might fail to correctly regulate septin localization at the budding site. In future studies, it would be necessary to evaluate the localization of septin at the budding site in tvf1Δ cells. It should be noted that the tvf1Δ strains show no growth defects at 37°C, although they show mitotic failure at 37°C. The critical growth suppression of tvf1Δ strain at 39°C may be due to mitotic failure or other causes associated with the budding process. UPR is necessary to resolve the accumulation of denatured proteins in heat-stressed environments,^15,16,18 and whether Tvf1p is necessary for UPR also needs to be verified in future studies.

The TVF1-deficient strain cultured at 37°C induced cytokine secretion by dendritic cells, indicating that they increased the immunogenic potential in response to heat stress. It is assumed that the TVF1-deficient strain increased the contents of some PAMPs that are recognized by the pattern recognition receptors (PRRs) of innate immunity; however, the detailed molecular mechanism needs to be elucidated in the future study. It should be noted that TVF1-deficient strains presented normal capsule formation and no specific changes in binding with Dectin-1 and Dectin-2. We speculate that the general mechanism known so far cannot explain the heat stress-induced immunogenic enhancement in the TVF1-deficient strain. The binding of WGA-FITC, Eosin-Y, and CFW to the tvf1∆ strain was increased, and the inner layer of the cell wall was also thickened in response to the heat stress, indicating that the properties of the cell wall containing chitin and chitosan appeared to be altered. Are these structural changes in cell wall chitin/chitosan involved in the heat stress-induced immunogenic enhancement in the tvf1∆ strain? Several candidate immunoreceptors that recognize chitin have been identified, including TLR2 and LYSMD3.^37,38 Chitosan is known to activate the cGAS-STING pathway in dendritic cells, and promote Type-I IFN production and maturation of dendritic cells.³⁹ It would be a future subject whether these PRRs and immune cascades recognize the enhanced immunogenicity of tvf1∆ strain.

Cryptococcus neoformans required TVF1 for their pathogenicity. Why was the virulence reduced in the tvf1Δ strain? Since thermotolerance is required for the virulence of C. neoformans,¹⁴ the first possibility is that the reduced thermotolerance of the tvf1-deficient strain affected its virulence. Since the heat-stressed tvf1Δ cells were more immunogenic, the second possibility is that tvf1Δ cells were readily recognized by the innate immunity, which effectively suppressed their growth. Similar findings have been reported for mar1Δ and chs3Δ in C. neoformans, and may support this possibility.^35,40 The enhanced antigenicity of tvf1Δ strain was demonstrated by in vitro assay, whereas the overall inflammation was weaker in the lungs infected with the tvf1Δ strain than the other counterparts. This discrepancy may be due to the inability of tvf1Δ strain to proliferate in the lungs. Another possibility is that TVF1 may also regulate other virulence-related factors besides the heat-stress response and controlling antigenicity since tvf1Δ cells did not pass through the layer of human brain microvascular endothelial cells (HBMECs), mimicking the blood–brain barrier.²⁵ Li and coworkers also showed comparable proliferation between tvf1Δ and KN99 at 37°C, and pointed out that the inability of tvf1Δ passing through the HBMECs layer cannot be explained by its heat sensitivity.²⁵ Since the infection model using insect larvae allows us to evaluate the pathogenicity of tvf1Δ strain at the non-stress temperature of 25°C–30°C,^41,42 it may be possible to test via those infection models whether the weak thermal tolerance is responsible for the low virulence of tvf1Δ strain.

In this study, we investigated the physiological role of C. neoformans TVF1 through reverse genetics. To elucidate the pleiotropic function of Tvf1p in C. neoformans, it will be necessary to analyze the localization, physical interaction, and regulation of Tvf1p expression via biochemical and molecular genetic approaches in the future.

Author contributions

Keigo Ueno (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing), Akiko Nagamori (Investigation), Nahoko Oniyama Honkyu (Investigation), Michiyo Kataoka (Investigation, Methodology, Software, Writing – original draft, Writing – review & editing), Kiminori Shimizu (Resources, Writing – review & editing), Yun C. Chang (Funding acquisition, Methodology, Resources, Supervision, Writing – review & editing), Kyung J. Kwon-Chung (Funding acquisition, Resources, Supervision, Writing – review & editing), and Yoshitsugu Miyazaki (Funding acquisition, Supervision, Writing – review & editing).

Data availability

The datasets and materials in this study are available from the corresponding author upon reasonable request.

Conflict of interest

The authors report no conflict of interest. The authors alone are responsible for the content and the writing of the paper.