The role of Aspartyl aminopeptidase (Ape4) in Cryptococcus neoformans virulence and authophagy

In order to survive and cause disease, microbial pathogens must be able to proliferate at the temperature of their infected host. We identified novel microbial features associated with thermotolerance in the opportunistic fungal pathogen Cryptococcus neoformans using a random insertional mutagenesis strategy, screening for mutants with defective growth at 37°C. Among several thermosensitive mutants, we identified one bearing a disruption in a gene predicted to encode the Ape4 aspartyl aminopeptidase protein. Ape4 metalloproteases in other fungi, including Saccharomyces cerevisiae, are activated by nitrogen starvation, and they are required for autophagy and the cytoplasm-to-vacuole targeting (Cvt) pathway. However, none have been previously associated with altered growth at elevated temperatures. We demonstrated that the C. neoformans ape4 mutant does not grow at 37°C, and it also has defects in the expression of important virulence factors such as phospholipase production and capsule formation. C. neoformans Ape4 activity was required for this facultative intracellular pathogen to survive within macrophages, as well as for virulence in an animal model of cryptococcal infection. Similar to S. cerevisiae Ape4, the C. neoformans GFP-Ape4 fusion protein co-localized with intracytoplasmic vesicles during nitrogen depletion. APE4 expression was also induced by the combination of nutrient and thermal stress. Together these results suggest that autophagy is an important cellular process for this microbial pathogen to survive within the environment of the infected host.


Introduction
C. neoformans is a fungal pathogen with worldwide distribution; it is found in the environment on rotting wood and is often associated with bird excreta. This yeast can cause fatal respiratory and neurological infections, especially in immunocompromised populations. Recent surveys estimate more than 500,000 deaths from C. neoformans every year, especially in patients with AIDS and other diseases that compromise the immune system [1]. The immunocompromised

Strains, media and reagents
The ape4 mutant and ape4+APE4 reconstituted strains were generated in the C. neoformans var. grubii background (KN99α). The standard medium for growth was YPD (1% yeast extract, 2% peptone, 2% glucose and 2% bacteriological agar). To evaluate the influence of nitrogen source, synthetic dextrose (SD) medium was used, either with or without amino acids and ammonium sulfate (SIGMA cat. Y1250 and cat. Y1251 respectively).
In order to evaluate non-preferred nitrogen sources, the SD medium without amino acids and ammonium sulfate was supplemented with 10mM/mL of L-proline or uric acid. To verify the effect of non-preferred carbon source, the SD medium with amino acid and ammonium sulfate was supplemented with galactose (2%). For selection of mutant strains, the growth media were supplemented with Geneticin (Invitrogen-Cat. #11811-023) at 200 μg/ml final concentration.
CnAPE4 gene deletion and reconstitution APE4 (CNAG_01169) gene deletion was performed by substituting part of the coding region with an antibiotic resistance marker. The mutant allele construction was designed using PCR overlap extension as previously described [28] and introduced in KN99α by biolistic transformation [29]. The transformants were selected on YPD containing G418 (200 μg/mL), and homologous integration of the construct was detected by diagnostic PCR and Southern blot. Mutant reconstitution was performed by PCR amplification of the gene, which was introduced in the mutant strain by co-transformation with pZPHyg, (a plasmid bearing the hygromycin resistance cassette). The reconstituted strains, ape4+APE4, were confirmed by lack of growth on YPD supplemented with 200 μg/ml of Geneticin G418 and for the restoration of the ability to grow at 37˚C. Reconstituted strains were confirmed by diagnostic PCR and Southern blot. All primers used in the gene deletion procedure are described in supplementary S1 Table. GFP-APE4 gene fusion The APE4 coding region was amplified by PCR using oligonucleotides described in the S1 Table and fused in frame to the C-terminus of GFP (green fluorescent protein) using the plasmid pCN19 [30]. The GFP-APE4 construct was transformed into C. neoformans (KN99α) by biolistic transformation. The mutants were selected on YPD plate supplemented with 100 μg/ mL of nourseothricin (Jena Bioscience).

Confirmation of genetic manipulation
The C. neoformans (wild type, mutant and reconstituted strains) genomic DNA extractions were performed as described previously [31]. The confirmation of deletion and restoration of the CNAG_01169 gene (APE4) at its original locus was performed by Southern blot, using the techniques described by Sambrook et al. (1989). The detection of the band patterns of the wild individuals (KN99α), mutant (ape4) and reconstituted (ape4+APE4) was performed by chemiluminescence using a dig-based labeling kit, (PCR DIG Probe Synthesis Kit, Roche, cat. # 11636 090 910) and detection system (Roche, catalog number 11175041910) according to the manufacturer's instructions. We used as a probe the PCR fragment amplified with primers MAV143 and MAV144 (supplementary S1 Table).
Evaluation of high temperatures growth, melanin, capsule, phospholipase and urease Cultures of C. neoformans (wild-type, mutant and reconstituted) grown in YPD for 12 hours were centrifuged and the pelleted cells were washed with saline (NaCl 0.9%). The cell number was determined by counting in a Neubauer chamber, the cell concentration was adjusted to 2x10 6 cells/ml, followed by 10-fold serial dilution (four dilutions), from where 5μL aliquots were spotted on Petri dish plate containing appropriate media follow by incubation at the required temperature (30˚C or 37˚C).
The thermotolerance was evaluated in rich medium (YPD) at temperatures of 30˚C and 37˚C. The cell size assay was performed as follows: each Cryptococcus strain (KN99α, ape4 and ape4+APE4) was incubated at 30˚C for 16h to 18h in YPD medium, pelleted, and the inoculum washed three times with PBS. Cells were inoculated to a final OD 600nm of 0.5 into either YPD or SD-N-AA liquid media. Cultures were grown under agitation for 24 hours at 30˚C and 37˚C. This experiment was performed in duplicate. One hundred cells per experiment were measured using a 40x objective. Employing the Anatomic MIPro Standard software, the cell size (μm) was determined. As the Cryptococcus cell is not a perfect circle, we used the software ellipse area tool, where the cell edges were defined and the radius marked, to estimate the cell size (μm). The data generated by the software was treated statistically using ANOVA (GraphPad Prism 5 program). Melanin production was evaluated at 30˚C on Niger seed agar, as described [32]. The ability to produce phospholipase was evaluated at 30˚C as described by Price et al. [33]. Induction of capsule production was carried out at temperatures of 30˚C and 37˚C in CO 2 -Independent Medium (GIBCO, Cat. 18C45-088) at 150rpm [34]. Cell samples were collected at 24 hours, 48 hours and 72 hours and stained with India Ink (REMEL, cat R21518) and analyzed using light microscopy (Opton Model TIM-2005-T). Image recording was performed using a USB camera (USB Digital Camera, DV Mod 3000).
The urease activity was assayed at 30˚C in urea agar base as described before [35].The analysis of phospholipase production was performed according to the literature [36], by calculating Pz, which is the ratio between the diameter of the colony (Dc) and the diameter of the precipitation halo with colony (Dcp), (Pz = Dc / Dcp), where Pz = 1 indicate negative production, values 0.63>Pz <1 indicates positive production and Pz 0.63 indicates high production of phospholipase, All measurements were performed with the assistance of program MIPRO Standard v1.1. The data were treated statistically using ANOVA (GraphPad Prism 5 program).

Multi stress sensitivity assay
Multi-stress sensitivity was evaluated with YPD medium supplemented with 0.75M and 1.5M of NaCl or KCl. The cell wall sensitivity was evaluated on YPD plus 0.5% Congo Red plates. All plates were incubated at 30˚C.

Minimum inhibitory concentration
Sensitivity to antifungal drugs was determined for wild type, mutant and reconstituted strains by standard minimal inhibitory concentration (MIC) assays according to Clinical and Laboratory Standards Institute (CLSI M27-A2) protocols with slight modifications. Briefly, fresh cultures were diluted in saline (0.9% NaCl) and standardized to the turbidity level 1 on the McFarland scale, then inoculated on agar plates containing RPMI-1640 medium (Sigma) buffered with MOPS (Sigma) pH 7. Plates were air dried and E-test strips with fluconazole (Biomerieux, cat. 510818) or amphotericin B (Biomerieux, Cat. 526 318) were carefully laid on the top of the agar.

Cellular localization of Ape4
To localize the Gfp-Ape4 fusion protein, the transformants were incubated in liquid YPD for 18 hours at 30˚C with 150rpm. The cells were collect by centrifugation and washed twice with pre-heated (30˚C) saline (0.9% NaCl). Ten milliliter of suitable liquid medium (SD without nitrogen source or YPD) supplemented with 10μg/mL of FM4-64 dye (Invitrogen) as described before [37], was inoculated with a final cell concentration of 0.5 (OD 600 ).
The cultures were incubated in YPD at 30˚C and 37˚C and SD (without nitrogen source) at 30˚C and 37˚C for 2 hours. The samples from each treatment were analyzed using a fluorescent microscopy (Zeiss Axio Imager.A1 fluorescent microscope and AxioCam MR digital camera). Fifty cells per sample distributed in eight different visual fields were analyzed. Captured images were processed with the program ZEN, 2012 and subjected to statistical analysis (ANOVA, GraphPad Prism 5 program).

Growth curve
Cultures of KN99α, ape4 and ape4+APE4 grown in YPD broth overnight were washed twice in 0.9% NaCl at room temperature. After washing, the cells were suspended in rich medium (YPD) or Synthetic Dextrose (SD) supplemented with ammonium sulfate and amino acids (Y1250 Sigma), or 10 mM of either uric acid or L-proline. The effect of a non-preferred carbon source was evaluated in SD medium supplemented with galactose 2%. Cultures were standardized to OD 600 0.3 and incubated at 30˚C with constant shaking (150 rpm-TECNAL mod. TE 420). The optical density (OD 600 ) was determined at 0, 6, 8, 24, 48 and 72 hours (T0, T6, T8, T24, T48, T72,). The biological assay was performed in triplicate. The data registered in a Microsoft1 Excel worksheet were subjected to statistical analysis (ANOVA, GraphPad Prism 5 program).

APE4 transcriptional analysis by real time PCR
Total RNA was extracted from C. neoformans (KN99α) according to previously described protocols [38]. Inductions were carried out at temperatures of 30˚C and 37˚C in rich medium (YPD) and in synthetic medium (SD) supplemented with nitrogen source, and synthetic medium (SD) without nitrogen source. The cDNA was generated using the kit RevertAid H Minus First Strand cDNA Synthesis RNA (Thermo Scientific Cat. # K1632), using oligo dT primers and 5μg of total RNA.
The Real-Time PCR reactions were performed using cDNA (diluted 1:10), with 800 μM of target primers (MAV 205 and MAV 206) and 300 μM primers (AA 301 and AA 302) for the internal control GPDH1(glyceraldehyde-3-phosphate dehydrogenase), Varma and Know-Chung, 1999) and 1X Power SYBR Green master mix (Life Technologies). The quantification of transcripts was done on a comparative manner, using the ΔΔCt method [39] and standardized with GPDH1 as previously described by [40].

Intracellular viability assay
J774A.1 macrophages (ATCC TIB-67) were co-cultivated with yeast cells as described by [41] with modifications. Briefly, yeast cells were grown in YPD liquid cultures, opsonized by the 18B7 anti-GXM monoclonal antibody [42] at a concentration of 10 μg/μl in DMEM supplemented with 10% fetal bovine serum, and added to macrophage monolayers (50% confluent) in 96-well plates at a multiplicity of infection (MOI) of five cells per macrophage. Phagocytosis was allowed to proceed for two hours, and unincorporated yeast cells were then washed off once with PBS before fresh medium was added to each well.
Each yeast strain was added to three wells. After 24 hours of incubation (37˚C, 5% CO 2 ), the medium from each well was collected into a fresh tube, macrophages were quickly lysed with 0.5% SDS in PBS, the medium plus lysate were diluted 200-fold in PBS and 100 μl of the dilution were platted on YPD agar for CFU counting. This experiment was performed four times. Statistical validation: for each macrophage co-cultivation experiment, CFU counts were compared among strains by one-way ANOVA followed by Tukey's post-test for comparison of pairs. Differences were considered valid at a 95% confidence interval (p<0.05).

In vivo virulence assay
The murine inhalation model of cryptococcal infection [12,43] was used to access the virulence of the C. neoformans strains. All animal experimentation was performed under an established protocol prospectively approved by the Duke University Institutional Animal Care and Use Committee (Protocol A178-14-07). Prior to infection, animals were briefly anesthetized by isoflurane inhalation. The mice were then infected via nasal inhalation with the wild-type strain KN99α, mutant (ape4) or reconstituted (ape4+APE4) strains. The inoculum (10 5 cells in a 25μL per strain) was prepared from cultures which were incubated overnight in YPD at 30˚C. For each strain ten mice were infected.
Mice were examined twice daily and sacrificed according to clinical measures predicting mortality (weight loss >15%, inability to access food and water). Survival data were analyzed by the Kruskal-Wallis test. The ape4-infected mice were sacrificed at 40 days post-infection; lungs from two of the surviving mice were harvested, homogenized, serially diluted, and plated on YPD to access the number of viable cells in the lungs. These experiments followed the Duke University institutional guidelines for animal experimentation.

In silico analysis
The CNAG_01169 sequence was retrieved from the Cryptococcus neoformans genome database formerly located at the Broad Institute.
The supplementary S2 Table summarizes all accession entries for all microorganisms used in this in silico analysis.
All C. neoformans gene sequenced retrieved from the data base described above were employed as query in a reciprocal BLASTx to confirm that the best hit would be what we present at the supplementary S2 Table.

Deletion and reconstitution of the APE4 gene in C. neoformans
We created a collection of insertional mutants in C. neoformans using Agrobacterium tumefaciens, as previously reported [12]. In one of the mutants displaying impaired growth at 37˚C, we identified a mutation in the APE4 gene, encoding an aspartyl aminopeptidase (CNAG_01169).
BLAST search analysis at NCBI and SGD (Saccharomyces Genome Database) revealed that CNAG_01169 has an average amino acid identity of 44.6% compared to the S. cerevisiae aspartyl aminopeptidase 4 (YHR113w/APE4). When we used the CNAG_01169 nucleotide sequence in a BLASTx search against the S. cerevisiae at the SGD, the first hit found was YHR113 (E-value of 2.5e-68). The t-DNA bearing the drug resistance cassette (Neomycin) was inserted after the nucleotide sequences that encode the amino acid glutamine (Q209) in C. neoformans Ape4.
To confirm that the APE4 gene mutation was the reason for the observed phenotypes, we created an independent ape4 mutant by replacing the region encoding amino acid residues 181-334 with the G418-resistance cassette (neomycin phosphotransferase II). The disruption of the APE4 locus was confirmed by a Southern blot assay (S1 Fig) for two independent ape4 mutants. To further ensure any phenotypes observed in the ape4 strain were due to lack of a functional APE4 protein, we created an independent reconstituted strain by replacing the mutant ape4 allele with the wild type APE4 gene in a co-transformation with the hygromycin resistance cassette (pZHyg). S1 Fig shows a Southern blot assay for two reconstituted strains that no longer grew in presence of neomycin but which were resistant to Hygromycin and able to grow at 37˚C.

In silico analysis of Ape4 protein
The C. neoformans putative Ape4 protein (523 amino acids) is encoded by the CNAG_01169 gene located on chromosome 5. Protein domain analysis at Uniprot suggests that this protein has domains characteristics of the M18 family of metalloproteases, which have protease activity against aspartate and glutamate amino acids located at the amino-terminus of various proteins [16,44]. Consistent with this observation, the BLAST tool to locate Conserved Domains at NCBI indicated a zinc binding domains typical of the peptidases of this family (S2 Fig).
The C. neoformans Ape4 amino acid sequence was compared to its putative orthologues in Coprinopsis cinerea, Ustilago maydis, Neurospora crassa, Saccharomyces cerevisiae (S288c), Zea mays, Oryza sativa, Arabidopsis thaliana, Chlamydomonas reinhardtii, Homo sapiens, Mus musculus, Xenopus laevis, Danio rerio, Caenorhabiditis. elegans, Toxoplasma gondii and Plasmodium falciparum. The alignment showed that the Ape4 protein from C. neoformans has a high value of similarity to C. cinerea (57%), whereas, the lowest percentage of similarity (31.3%) was observed against the aspartyl aminopeptidase from P. falciparum (S3 Fig). This analysis suggests that the APE4 gene encodes a metalloprotease conserved in the basidio-and ascomycetes that we included in this analysis.
The ape4 mutants are sensitive to high temperature growth To confirm that the sensitivity to high temperatures by the ape4 mutant was associated with the absence of APE4 gene, serial dilutions of wild type (KN99α), mutant (ape4) and reconstituted (ape4+APE4) strains were spotted onto YPD agar medium and incubated at 30˚C and 37˚C for 72 hours. Fig 1A shows a thermos-sensitive growth defect specifically associated with the ape4 mutation. Another C. neoformans mutant, ras1, has its growth impaired at 37˚C, and this mutant also has increased cell size consistent with defects in cytoskeleton polarization and cell division [11]. Therefore, we decided to assess the impact of the ape4 mutation on cell size. We compared the morphology of the three strains after incubation at 30˚C and 37˚C in two separate media (YPD and SD-N-AA). We observed that at 30˚C in YPD medium the ape4 mutant strain was statistical indistinguishable from wild-type and reconstituted strains where the average cell size was about 4 μm. In the same medium at 37˚C, the average cell diameter was increased about 25% for all strains (about 5 μm) with no statistical difference observed among the strains (Fig 1B). When cells were subjected to nutritional deprivation (SD-N-AA), the cell diameter found for both WT and reconstituted strains was similar to that in rich medium (YPD) at 30˚C (4 μm) and 37˚C (5 μm). However, the ape4 mutant showed a significant increase (50%) in cell diameter when nutrition deprivation was combined with high temperature incubation (6 μm). Therefore, we observed that the lack of a functional APE4 gene led to a cell size increase at high temperature, which is more pronounced with concomitant nutritional deprivation. (Fig 1C) The ape4 mutation affects multiple virulence factors in C. neoformans The polysaccharide capsule of C. neoformans is necessary for protection against cell stress and dehydration, and it confers immunomodulatory properties during interaction with host cells during infection. C. neoformans mutants lacking capsule are often avirulent in animal model [45][46][47]. Previously, the autophagy process has been thought to be a process related to degradation of cell components as well as to promote programmed cell death. However, it has been related to other important cell developmental processes such as yeast sporulation, development of fruiting body in Dictyostelium discoideum, pupa development in D. melanogaster, and extension of cell life span in caloric restriction [48]. Therefore, we assessed if lack of the functional putative Ape4 protein would have any impact upon capsule production. We induced capsule in CO 2 -Independent medium at 30˚C, and we observed no significant difference in capsular volume for the first 48 hours for the wild type (KN99α), mutant (ape4) and reconstituted (ape4+APE4) strains. However, with longer incubations, the ape4 mutant displayed a marked decrease in surface capsule compared to control strains. At 37˚C, the difference in capsule volume was observed earlier, after 24 hours of incubation. Fig 2 illustrates the capsule production after 72 hours of incubation at 30˚C and 37˚C.
The production of phospholipase, a secreted virulence factor, was evaluated on egg yolk agar. The precipitation zones (Pz) as a measure of phospholipase activity were identical for wild type (Pz = 0.27 ± 0.004) and reconstituted (Pz = 0.26 ± 0.01) strains. However, the phospholipase activity for the ape4 mutant (Pz = 0.39 ± 0.01) was significantly reduced compared to wild type (p<0.05), similar to previously reported ura4 mutant with decreased phospholipase activity [12] (Fig 3). Deletion of APE4 gene did not have any impact on production of urease, melanin or mating. Therefore, we conclude that Ape4 is important for capsule and phospholipase production in C. neoformans.

APE4 gene is involved in multi-stress resistance
In response to stress, several genes are activated to enhance chitin deposition in the cell wall [49]. To explore the effects of Ape4 on processes that mediate cell wall formation under cell stress, serial dilution of the wild type, ape4 and ape4+APE4 strains were incubated on YPD medium containing Congo Red (0.5%w/v) at 30˚C for 72 hours. The ape4 mutant failed to growth in the presence of this dye (Fig 4A) suggesting that the lack of putative Ape4 protein affects normal cell wall formation. The ape4 mutant showed no sensitivity to alkaline pH (8.0) when compared to the wild and reconstituted strains. Similarly, the growth of the ape4 mutant was not impaired in the presence of 0.75M NaCl (Fig 4B), and its growth was only modestly reduced in presence of 0.75M KCl (Fig 4C). On the other hand, ape4 growth was abolished compared to the wild type and the reconstituted strains when incubated in the presence of either 1.5M NaCl or 1.5M KCl. This result indicates that the C. neoformans putative aspartyl aminopeptidase encoded by APE4 is important during response to osmotic/salt stress.

Lack of a functional putative Ape4 protein increases drug susceptibility
To assess the role of the C. neoformans Ape4 protein in antifungal drug tolerance, we performed an E-test (bioMérieux) to investigate the synergistic impact that the ape4 mutation would have in association to fluconazole and amphotericin B on yeast cell growth. The wild type (KN99α), mutant (ape4) and reconstituted (ape4+APE4) strains were equally sensitive to amphotericin B (0.125 μg/ml, Fig 5B), while the ape4 mutant was more sensitive to fluconazole (1 μg/ml) compared to wild type and the reconstituted strains (12 μg/ml) (ANOVA, t-Student's and Scott Knout, p<0,05) (Fig 5A). These results indicate that any disturbance on putative Ape4 protein should be considered in further studies regarding new drugs development against C. neoformans.

The putative APE4 coding gene is essential for fungal survival within macrophages for virulence
To assess the role of autophagy for survival of C. neoformans in macrophages, we co-incubated the three strains with J774A.1 cells and quantified fungal viability by quantitative culture (Hu et al., 2008). The ape4 mutant displayed a significant decrease in survival within these cells (Fig 6), similar to the vps34 mutant that is known to have a defect in autophagy signaling [26]. Also, Oliveira et al. (2016) showed that autophagy mechanisms were defective in the atg7 mutant, including its survival within the macrophage [27].
Given our data indicating the role of C. neoformans putative autophagy protein Ape4 on virulence trait expression and in vitro survival in macrophages, we tested whether the ape4 mutant would be virulent in an animal model of cryptococcal infection. Using a murine inhalational model of cryptococcosis [50], we compared the virulence of the wild type (KN99α), mutant (ape4), and the reconstituted (ape4+APE4) strains. As shown in Fig 7, infections with the wild type and reconstituted strains resulted in an average survival of 17.5 (±1.87) and 20 (±1.58) days post-infection, respectively, whereas the mutant (ape4) strain-infected mice survived 40 days post-infection. We harvested and homogenized lungs from two of these surviving mice, followed by serial dilution and quantitative cultures. Both animals appeared clinically healthy, however they had viable yeast fungal cells in their lungs (1.15×10 3 ±1.2×10 3 CFU/mg). This finding of minimally symptomatic infection was also observed by other authors studying other C. neoformans mutants with altered virulence [8,12,51]. The C. neoformans strain with a mutation in the putative autophagy regulator Atg7 also had its virulence attenuated in a murine model.

The APE4 expression pattern is influenced by nitrogen starvation and high temperature
In the yeast S. cerevisiae, nitrogen starvation modulates the expression of genes involved in autophagy [52,53]. Therefore, we investigated the C. neoformans putative autophagy gene APE4 transcription pattern under nutritional and temperature stress. The expression pattern of APE4 does not change at different temperatures when C. neoformans is incubated in either rich medium (YPD) or synthetic dextrose medium supplemented with an abundant nitrogen source (Fig 8A and 8B). However, lack of available nitrogen sources in association with high temperature leads to a nine-fold increase in APE4 transcription (Fig 8C). Therefore, we conclude that nutrient depletion in combination with high temperature has a more dramatic effect on the expression pattern of this gene, suggesting that APE4 is required for C. neoformans survival in this harsh environment.
The putative Aspartyl aminopeptidase (Ape4) of C. neoformans influences growth in non-preferred nitrogen source In S. cerevisiae, Ape4 is associated with degradation of proteins in nitrogen deprived conditions (Yuga, et al., 2011). The analysis of C. neoformans growth patterns on rich medium (YPD) revealed that at 30˚C there is no growth difference among wild type (KN99α), mutant (ape4) and reconstituted (ape4+APE4) strains (Fig 9A). While growth carried out at 30˚C in Synthetic Dextrose medium (SD) supplemented with ammonium sulfate and amino acids, or 10 mM of either uric acid or L-proline, showed that the ape4 mutant had grown significantly less when compared to the wild type and reconstituted strains (Fig 9B, 9C and 9D).    The nitrogen sources used (uric acid and L-proline), had a different effect on the growth of C. neoformans, and a higher optical density (OD) was observed for growth in uric acid ( Fig  9D). Even in these conditions, the growth of ape4 remained significantly lower when compared to the wild type and reconstituted strains. Likewise, the use of a non-preferred carbon source (SD + galactose) significantly decreased the growth of the ape4 compared to wild type and reconstituted strains in the same medium condition (Fig 10). In the presence of galactose, the lag phase for ape4 lasted for 48 hours, whereas on SD supplemented with dextrose ( Fig 9B) this mutant left the lag phase after 8 hours of incubation.
Ape4 is localized within the vesicles formed in response to nitrogen starvation and high temperatures In S. cerevisiae the expression of the genes encoding vacuolar hydrolases is influenced by nutritional status since, during nitrogen starvation, there is a significant increase in the amount of vacuolar proteins [54].
In rich medium, a small portion of S. cerevisiae Ape4 localizes in the vesicles; however, its vacuolar transport is accelerated by nutrient starvation, when cytosolic Ape4 redistributes to the vacuole to attend the yeast cells need for more active vacuolar degradation [16,55]. In order to check if C. neoformans Ape4 would behave like S. cerevisiae Ape4, we expressed a GFP-Ape4 fusion protein in KN99α (wild type). Fluorescent microscopic analysis showed that GFP-Ape4 was dispersed in the cytoplasm in 100% of cells analyzed at the optimal temperature and nutrition conditions (YPD medium at 30˚C, Fig 11). When the temperature was raised to 37˚C in YPD medium, Ape4 co-localizes within vesicles-like structures in 62% of cells analyzed. In nitrogen starvation (SD without nitrogen source), this percentage increases to 72% and 74% at 30˚C and 37˚C, respectively (ANOVA, Tukey's post test p<0.05) (Fig 11). These results indicate that elevated temperature and nitrogen deprivation likely work together to facilitate Ape4 localization to the vesicles. This stress-induced acceleration of Ape4 movement to the vacuole is similar to that described in S. cerevisiae, and it is proposed to be a response to increased cellular requirements for vacuolar degradation [16].

Transcription pattern of autophagy related genes in C. neoformans
In S. cerevisiae up to 34 autophagy-related (ATG) genes were reported [16,20,56,57]. We retrieved all the sequences for these reported genes from SGD (http://www.yeastgenome.org/) and used BlastX to identify likely homologs using the C. neoformans genome data base at Broad Institute (http://www.broadinstitute.org/). Out of these 34 autophagy-related (ATG) genes reported for S. cerevisiae, we found 21 homologs in C. neoformans (Supplementary S2  Table). This table shows that besides C. neoformans, other fungi also have fewer genes encoding autophagy-related proteins than S. cerevisiae. Among the Basidiomycetes, this number ranges from 21 to 23, whereas for the Ascomycetes Candida albicans and A. fumigatus these numbers are 27 and 29, respectively. However, the relevance of this reduced number of autophagy-related genes for the more distantly related basidiomycete fungi remains to be further explored. Above we reported how nutritional status and high temperature modulate the transcription pattern of APE4. Therefore, we asked how the transcriptional profile of the other C. neoformans autophagy-related genes is modulated by these conditions. In rich medium (YPD) the temperature change from 30˚C to 37˚C resulted in a significant increase in transcription (p<0.05) (two-fold and above) for 13 genes out of 21 evaluated, (Fig 12) suggesting that even under favorable nutrient conditions, high temperature stress, per se, is sufficient to induce the transcription of these genes. Next, we evaluated the impact of nitrogen starvation on the same set of genes. The results showed that synthetic dextrose (SD) without nitrogen at 30˚C leads to at least two-fold transcriptional induction for 15 genes out of 21, whereas among the fifteen, eight genes had more than five-fold induction and three had over ten-fold induction (ATG1, ATG4 and ATG22) (Fig 13). When this growth condition is repeated, but at 37˚C, we found that 18 genes were induced more than two-fold and among them, four had above five-fold induction while seven had over a ten-fold transcription induction (ATG1, ATG4, ATG7, ATG9, VPS15, APE4 and AMS1) (Fig 14). These results suggest that, to some extent, the autophagy related genes found in C. neoformans not only respond to lack of nitrogen source, but also to high temperature, suggesting a possible link between autophagy and thermal stress, which are conditions commonly found upon host infection. However, further studies must be done to confirm their role, if any, in the C. neoformans autophagy process.

Discussion
Previously, we reported the use of an insertional mutagenesis library to uncover traits involved in high temperature growth in C. neoformans [12]. We further investigated this library and found a gene encoding a putative aspartyl aminopeptidase, similar to S. cerevisiae APE4, which belongs to a M18 metalloprotease family [16,44]. In S. cerevisiae the Ape4 protein is part of the selective autophagy Cvt pathway, which is unique to this yeast [18,58]. Also Ape4 is described as one of the cargo proteins that is associated with Atg19 and is transported to the PAS site along with Ape1 and Ams1 when it enters the Cvt vesicle that later fuses to the autophagosome [16,44]. Cultures in stationary phase, as well as those experiencing nitrogen starvation, usually, have higher Cvt pathway activity, leading to increased vacuolar transport of Ape1 [59].
We confirmed the role of C. neoformans Ape4 in high temperature growth by independently generated ape4 mutants. These finding were consistent with Hu et al's (2008) description of autophagy-related genes impairing C. neoformans growth at 37˚C. Exploring further the effects of the ape4 mutation in this yeast we found that C. neoformans GFP-Ape4 fusion protein is also located within a vesicle, which is formed upon nitrogen starvation. Yuga et al Fig 13. C. neoformans autophagy-related genes respond to lack of nitrogen at the permissive temperature (30˚C). Fifteen of twenty-one genes tested had an increase in their expression pattern above two-fold when the yeast cells were subjected to nitrogen starvation. All fold values were calculated from triplicate assays and were statistically validated by ANOVA (p< 0.05).
https://doi.org/10.1371/journal.pone.0177461.g013 (2011) reported that S. cerevisiae Ape4 transport from the cytoplasm to the vacuole is accelerated by nutrient starvation to meet the requirements for more active vacuolar degradation [16]. Our data does not support a direct implication of Ape4 with autophagy in C. neoformans, but due to its nature as a possible protease we believe that C. neoformans Ape4 is somewhat involved with nutrient recycling during starvation in this yeast. It seems likely that its function takes place within vesicles during nitrogen deprivation, and this effect seems more dramatic when the cells are at higher temperature (37˚C), a situation where more protein is recruited to the vesicle. Also, the combination of elevated temperature and nitrogen deprivation leads to a 9-fold increase in APE4 transcript abundance, reiterating its role in response to nutritional and temperature stress.
The C. neoformans cells encounter a hostile environment upon host infection, confronting challenges including nutrient depletion and high temperature [7]. The autophagy processes may provide the means for the yeast cell to degrade cytoplasmic components, such as proteins and organelles, to recycle cellular nutrients during starvation until the cells are more adapted to the new environment [60]. In 2005 investigators demonstrated that two C. neoformans genes, ATG3 and ATG9, are up-regulated 2.6-(2 hours) and 3.2-fold (24 hours), during murine macrophage infection [2]. These genes are thought to be related to autophagy in C. neoformans, first implicating autophagy in cell adaptation in an unfavorable environment. Our results confirmed that the transcription of these two genes is indeed induced 8-and 22-fold, respectively, when C. neoformans is incubated under nitrogen shortage and high temperature. Recently, Oliveira et al. (2016) showed that the C. neoformans ATG7a putative autophagy regulator is important for virulence in this yeast [27]. Our transcription studies showed that this gene is induced about eightfold during nitrogen starvation (Fig 13).
Studies in other fungi species have demonstrated a complex interplay between autophagy related proteins and microbial pathogenesis/antifungal tolerance. In Aspergillus fumigatus, Richie et al. (2007) showed that lack of a functional autophagy process led to impaired radial growth and conidiophore development at 37˚C [61]. In C. albicans the atg9 mutant was deficient for autophagosome formation during nitrogen starvation, and it also showed defects in yeast-hyphae differentiation. However, this mutant was able to survive within the macrophage and was fully virulent in a mouse model of hematogeneously disseminated candidiasis, suggesting that this autophagy-related protein is dispensable for pathogenesis [62]. A recent study in C. albicans revealed that autophagy is important for ER stress response and tolerance to antifungal drugs [27]. Similarly, our studies demonstrated that the C. neoformans ape4 mutant is more sensitive to fluconazole, implicating autophagy in antifungal tolerance. Basse's group studying autophagy in Ustilago maydis uncovered that atg11 and atg8 are important for mitochondrial recycling [63,64]. Also, mutation of U. maydis atg8 rendered the mutant strain less virulent, with reduced teliospore production. A similar but less pronounced phenotype effect was observed for atg1. However the combination of both atg mutations exacerbated these phenotypes [65].
We performed an in silico analysis for C. neoformans genes predicted to play a role in autophagy, using the S. cerevisiae genes in a genome BLAST search (see supplementary S2 Table). Of 34 autophagy related genes described in S. cerevisiae [16] only 21 are clearly found in this pathogenic yeast genome. In depth analysis showed that either this process is functionally different in C. neoformans and S. cerevisiae, or there are functional orthologs in C. neoformans with limited sequence similarity to their yeast counterparts. These questions are even more compelling to explore given the importance of Ape4 in C. neoformans virulence. Therefore, further experiments must be carried out to uncover if this putative autophagy gene, APE4, has any role in C. neoformans autophagy.