Roles for Stress Response and Cell Wall Biosynthesis Pathways in Caspofungin Tolerance in Cryptococcus neoformans

Limited antifungal diversity and availability are growing problems for the treatment of fungal infections in the face of increasing drug resistance. The echinocandins, one of the newest classes of antifungal drugs, inhibit production of a crucial cell wall component. However, these compounds do not effectively inhibit the growth of the opportunistic fungal pathogen Cryptococcus neoformans, despite potent inhibition of the target enzyme in vitro. Therefore, we performed a forward genetic screen to identify cellular processes that mediate the relative tolerance of this organism to the echinocandin drug caspofungin. Through these studies, we identified 14 genetic mutants that enhance caspofungin antifungal activity. Rather than directly affecting caspofungin antifungal activity, these mutations seem to prevent the activation of various stress-induced compensatory cellular processes. For example, the pfa4Δ mutant has defects in the palmitoylation and localization of many of its target proteins, including the Ras1 GTPase and the Chs3 chitin synthase, which are both required for caspofungin tolerance. Similarly, we have confirmed the link between caspofungin treatment and calcineurin signaling in this organism, but we suggest a deeper mechanism in which caspofungin tolerance is mediated by multiple pathways downstream of calcineurin function. In summary, we describe here several pathways in C. neoformans that contribute to the complex caspofungin tolerance phenotype in this organism.

in clinical settings for the treatment of infections caused by diverse fungi. However, echinocandins do not have potent antifungal activity against the fungal pathogen Cryptococcus neoformans, whose growth is only inhibited at high levels of caspofungin that are not clinically achievable in patients ; Bartizal et al. 1997;Espinel-Ingroff 1998). The fact that this drug is so ineffective against this fungus is surprising for a number of reasons. First, the gene that encodes the b-1,3-glucan synthase catalytic subunit in C. neoformans, FKS1, is essential in this organism (Thompson et al. 1999). Additionally, the C. neoformans enzyme is highly sensitive to caspofungin in vitro, even potentially at lower concentrations than species that are clinically susceptible to these drugs, such as Aspergillus species (Maligie and Selitrennikoff 2005). Based on these data, caspofungin could be expected to be an effective inhibitor of C. neoformans growth.
Given these observations, several investigators have tried to explain the discrepancy between the high sensitivity of the target enzyme's activity and the high tolerance of the organism to caspofungin. Recent work using a fluorescently tagged form of caspofungin has suggested that intracellular concentrations of caspofungin are low in wild-type C. neoformans (Huang et al. 2016). Therefore, the cell wall or the polysaccharide capsule could prevent the accessibility of caspofungin, a high-molecular weight drug, to its target enzyme. Alternatively, caspofungin could be entering the cell but be rapidly eliminated by degradation or through the action of a multidrug resistance pump. We hypothesized that interrupting specific C. neoformans cell processes would render this fungus more susceptible to caspofungin, even at these relatively low effective concentrations. To this end, we screened through two targeted gene deletion collections to identify C. neoformans mutants that were hypersensitive to caspofungin relative to the wild-type strain. We also demonstrate possible mechanisms for how the cellular processes controlled by these genes might affect echinocandin resistance in C. neoformans.

Strains, media, and growth conditions
The collections used for the caspofungin sensitivity screen consist of both the 2008 Madhani and 2015 Madhani plate collections, which were purchased from the Fungal Genetics Stock Center (Liu et al. 2008;Chun and Madhani 2010). The wild-type (WT) strain used in this study is the clinical strain H99 (Perfect et al. 1980). Newly generated strains, as well as strains from alternate sources, are listed in Supplemental Material, Table S1. A similar screen using a subset of these mutant strains was also recently performed (Huang et al. 2016).

Strain creation
To generate the new and independent mutant strains used in this study, targeted gene deletion constructs were designed to replace the entire open reading frame with the neomycin (NEO) or nourseothricin (NAT) dominant selectable markers. Each knockout construct was generated using PCR overlapextension and a split selectable marker as described previously (Davidson et al. 2002;Kim et al. 2012). All constructs were transformed into the C. neoformans H99 strain by biolistic transformation, as previously described (Toffaletti et al. 1993). All deletion primers used in this study can be found in Table S2A.
Upon transformation, strains were selected on YPD medium containing either NAT or NEO. Deletion mutants were checked by a combination of positive and negative confirmation PCRs demonstrating replacement of the WT locus with the mutant allele, followed by Southern blotting to confirm single integration of the deletion constructs (data not shown).
Plasmids used in this study can be found in Table S3. Cloning primers can be found in Table S2B. The overexpressed GFP-RHO1 construct was generated by cloning the RHO1 gene plus its native terminator into the BamHI site of the pCN19 vector, which contains the histone H3 promoter and GFP with the NAT selection marker. The endogenous FKS1-GFP construct was engineered by cloning the following into the pUC19 vector: the FKS1 gene (without promoter), the GFP gene, the FKS1 terminator, and the NEO marker flanked by genomic sequences to target this construct to the FKS1 locus. These plasmids were transformed as described, and selected on NEO or NAT, and transformants were confirmed by a positive PCR to document the presence of the introduced allele. In the case of the eFKS1-GFP construct, PCRs to confirm integration of the construct into the endogenous FKS1 locus were performed.

Caspofungin sensitivity primary screen
A small pilot screen revealed that the calcineurin B subunit mutant was hypersensitive to caspofungin compared to the WT, agreeing with published data (Del Poeta et al. 2000). This strain became the standard against which to measure other potentially caspofungin-susceptible strains. To determine the optimal conditions under which to perform the caspofungin sensitivity screen, we assessed WT vs. cnb1D growth in YNB medium at caspofungin concentrations between 5 and 50 mg/ ml with shaking at 150 rpm at 30°(Cancidas; Merck). We identified 15 mg/ml caspofungin to be a concentration at which the cnb1D mutant strain was markedly impaired for growth and the WT grew robustly. We then screened the strain collections of 3880 isolates, first preincubating in YPD liquid medium for 16 hr with shaking (150 rpm) at 30°. Cultures were diluted 1:10 in 96-well plates containing either YNB or YNB + caspofungin (15 mg/ml). Strains were incubated with shaking (150 rpm) at 30°for 24 hr, and growth was assessed by measuring OD 600 on a FLUOStar Optima plate reader (BMG Labtech). Plates were also pinreplicated to YPD plates to assess strain viability after incubation with caspofungin.
After screening, strains were divided into four groups based on caspofungin susceptibility: (1) strains that were inviable postcaspofungin treatment; (2) strains that were viable after caspofungin treatment, but that had significantly decreased growth (OD 600 that was .2 SD less than the average WT OD 600 ); (3) strains that did not have significantly different growth from WT in caspofungin medium; and (4) strains that did not grow in YNB.

Disc-diffusion secondary screen
To confirm the caspofungin sensitivity of the mutants identified in the above screen, each mutant was incubated overnight in 150 ml of YPD in a 96-well plate with shaking, along with the WT and cna1D mutant strains as controls. Strains were diluted 1:200 in PBS, then 75 ml per strain was spread onto YNB agar in 6-well plates. Sterile filter discs were placed in the center of each well and 5 ml of 7.5 mg/ml caspofungin was added to each disc. Plates were incubated at 30°for 3 days. After incubation, plates were imaged and zones of inhibition were measured. Mutant phenotypes were classified as WT-like, cna1D-like, or intermediate.

Minimal inhibitory concentration assay
Minimal inhibitory concentration (MIC) assays were performed according to modified Clinical Laboratory Standards Institute standard methods for broth microdilution testing of antifungal susceptibility (Pfaller et al. 1990;Clinical and Laboratory Standards Institute 2008). In brief, cells were diluted in phosphate-buffered saline (PBS) to an OD 600 of 0.25, then diluted 1:100 in YNB medium. Caspofungin was diluted in PBS. Next, 23 working stocks of caspofungin were prepared in YNB medium, then caspofungin was serially diluted twofold in 100 ml YNB medium, following which 100 ml of 1:100 dilution of cells was added to diluted drug in 96-well plates. The final concentration range of caspofungin was 200-0.39 mg/ml. Plates were incubated for 48 hr without shaking at 30°or 35°. After 48 hr, OD 600 was measured on a FLUOStar Optima plate reader. MIC 50 values were calculated by calculating relative growth using (drug-treated OD 600 /untreated OD 600 ), with MIC 50 corresponding to a $ 50% decrease in relative growth.

Checkerboard assay
Checkerboard assays to assess antifungal drug synergy using the fractional inhibitory concentration index (FIC) for combinations of compounds were performed as described (National Committee for Clinical Laboratory Standards 1992; Franzot and Casadevall 1997). In brief, the WT strain H99 was inoculated from plated colonies into PBS at an OD 600 of 0.25, and subsequently diluted 1:100 into RPMI medium. Nikkomycin Z stock was diluted in PBS (nikkomycin Z; Sigma [Sigma Chemical], St. Louis, MO). Manumycin A, clorgyline, and 2-bromopalmitate (2BP) were diluted in DMSO (manumycin A, Bioviotica; clorgyline, Sigma; and 2BP, Sigma). Caspofungin was diluted for a final concentration range between 100 and 1.5625 mg/ml, and the test drugs were diluted to the following final concentration ranges: nikkomycin Z 400 to 0.78125 mg/ml, manumycin A 40 to 0.078 mM, 2BP 400 to 0.78125 mM, and clorgyline 100 to 1.5625 mM. Assays were incubated at 30 and 37°. FIC index values for a combination of compounds A and B were calculated as: where an FIC index value of , 1.0 is considered synergistic (with , 0.5 considered strongly synergistic), additive if the value was 1.0, autonomous if the value was between 1.0 and 2.0, and antagonistic if the FIC index was . 2.0 (Franzot and Casadevall 1997).  ) and Picard to realign reads before SNP calling using the UnifiedGenotyper component of GATK with the haploid ploidy setting. The resulting variant call formats were filtered using VCFtools (Danecek et al. 2011) and annotated for variant effect using SnpEff (Cingolani et al. 2012). Heterozygous calls were removed as presumed mismapped repetitive regions. Raw reads are available on the National Center for Biotechnology Information (NCBI) Sequence Read Archive under accession number PRJNA501913.

Chitin and chitosan assay
The chitin and chitosan contents of C. neoformans cell walls were assessed as described . Briefly, cells were incubated overnight in YPD. Cultured cells were then diluted to an OD of 0.8 in SC or SC + 15 mg/ml caspofungin, and incubated with shaking at 30°for 6 hr. Cells were divided and lyophilized, then either mock treated or treated with acetic anhydride to acetylate chitosan to form chitin. Cell walls were then digested with 5 mg/ml chitinase for 72 hr. Monomer levels of N-acetylglucosamine were assessed by a p-dimethylaminobenzaldehyde colorimetric assay and read on a FLUOStar Optima plate reader. Acetic anhydride samples represented levels of both chitin and chitosan in the cell wall, while untreated samples represented chitin alone. Chitosan levels were calculated as the difference between the acetic anhydride-treated and the untreated samples. Data were analyzed using a two-way ANOVA, followed by Student's t-tests to determine statistical significance.

Cell wall staining and microscopy
Cells were prepared for cell wall staining as described (Ost et al. 2017). WT cells were cultured overnight in YPD medium. Overnight cultures were diluted to an OD 600 of 1 in 15 ml SC or SC plus caspofungin (5, 10, or 20 mg/ml caspofungin), and incubated with shaking at 30°. At the indicated timepoints, 1 ml aliquots of each culture were collected and stained with calcofluor white (CFW). Cells were pelleted at 5000 rpm for 2 min, then resuspended in 100 ml PBS + 25 mg/ml CFW and incubated in the dark at room temperature for 10 min. Cells were then washed two times with PBS and resuspended in 50 ml PBS for imaging. Strains were imaged on a Zeiss ([Carl Zeiss], Thornwood, NY) Axio Imager A1 fluorescence microscope equipped with an Axio-Cam MRM digital camera to capture both DIC and fluorescent images. Cell wall-staining fluorescence intensity was analyzed using Fiji software, and the mean gray values were analyzed (Schindelin et al. 2012). Data presented represent the average fluorescence values. Data were analyzed using a two-way ANOVA, followed by Student's t-tests to determine statistical significance. For fluorescent fusion protein microscopy, strains were incubated in SC medium for 18 hr at 30°with shaking. These cultures were pelleted at 3000 rpm for 5 min, and resuspended in SC or SC + 15 mg/ml caspofungin. Strains were incubated with shaking at 30°for 90 min, with aliquots collected for imaging at 15-min intervals. Aliquots were incubated with NucBlue Live Ready Probes reagent for 5 min, pelleted at 5000 rpm for 2 min, then resuspended in 50 ml SC (Thermo Fisher Scientific). Strains were imaged on a Zeiss Axio Imager A1 fluorescence microscope equipped with an Axio-Cam MRM digital camera to capture both DIC and fluorescent images. For the Fks1-GFP and GFP-Rho1 localization experiments, cultures were incubated for 18 hr at 30°with shaking in SC medium. These cultures were normalized to an OD 600 of 2.0 in SC plus 0, 5, 10, or 15 mg/ml caspofungin, and imaged at 20-min intervals for 1.5 hr.

RNA preparation and quantitative real-time PCR
The WT strain was grown in YPD for 18 hr at 30°with shaking. Cells were then inoculated at an OD 600 of 1.5 into 5 ml SC, SC + 10 mg/ml caspofungin, or SC + 15 mg/ml caspofungin. Cultures were incubated at 30°with shaking for 90 min, then cells were harvested by centrifugation at 3000 rpm for 5 min and lyophilized. RNA was isolated using a RNeasy Plant Mini Kit (QIAGEN, Valencia, CA), with the addition of bead beating for 1 min prior to lysis and on-column DNase treatment (QIAGEN). cDNA was prepared using the AffinityScript QPCR cDNA synthesis kit using oligo-dT primers to bias for mRNA transcripts (Agilent Genomics). Quantitative real-time PCR was performed using PowerUp SYBR Green Master mix (Applied Biosystems, Foster City, CA) on a QuantStudio 6 Flex system. Real-time PCR primers are listed in Table S2C (Esher et al. 2018). Data were analyzed using a two-way ANOVA, followed by Student's t-tests to determine statistical significance.

Data availability
Strains and plasmids are available upon request. File S1 contains a list and descriptions of all supplemental files. Sequence data are available at the NCBI Sequence Read Archive under the accession number PRJNA501913. Supplemental material available at FigShare: https://doi.org/10.25386/ genetics.8085290

Results
Initial screen for processes contributing to caspofungin tolerance in C. neoformans We performed a forward genetic screen of targeted deletion mutants to identify cellular processes that contribute to C. neoformans tolerance to caspofungin treatment. Using two screening methods in sequence, we screened 3880 mutants for altered growth during caspofungin treatment at 30° (Liu et al. 2008;Chun and Madhani 2010). We specifically chose this permissive incubation temperature for our initial screens to potentially capture pathways that may be required for thermotolerance, and for which mutant strains would therefore not be viable at higher temperatures. Our screening methods yielded 54 mutants that appeared to have reduced caspofungin tolerance relative to the WT strain. We then performed broth microdilution assays to compare caspofungin susceptibility to the WT strain, as well as to a cna1D strain with a mutation in the calcineurin A subunit gene. Strains with altered calcineurin function are known to be more susceptible to caspofungin (Del Poeta et al. 2000;Kraus et al. 2003). Of the mutants identified in the primary screen, 14 were confirmed to have caspofungin susceptibility similar to or greater than the cna1D mutant strain (Table 1). This strict screening criterion enabled us to focus our studies exclusively on more highly drugsusceptible mutants (Table 1). The remaining strains displayed only minimal increases in sensitivity to caspofungin and they were not tested further. Of note, we identified the cnb1D calcineurin regulatory subunit mutant, which phenocopies the cna1D mutant, as a caspofungin-hypersensitive mutant in our screen, validating our screening methods (Odom et al. 1997;Del Poeta et al. 2000;Fox et al. 2001;Kraus et al. 2003).
Within the list of sensitive mutants, we identified multiple biological processes that seem to be important for caspofungin tolerance based on gene ontology (GO) analysis (see Table S4 for full GO analysis) (Stajich et al. 2012). Of note, the msh1D mutant strain displayed a very low caspofungin MIC. This strain has a defect in a predicted homolog of a MutS mismatch repair protein most likely functioning on mitochondrial DNA (Sia and Kirkpatrick 2005). Interestingly, independently created C. neoformans msh1D mutants did not display the caspofungin-susceptible phenotype of the original mutant, suggesting that secondary mutations in this strain were the cause of increased drug susceptibility. Indeed, whole-genome sequencing of the original msh1D mutant strain revealed several additional mutations in both the nuclear and mitochondrial genomes, as predicted for a strain with potential defects in DNA repair mechanisms (Table S5).
Calcineurin signaling plays a role in caspofungin tolerance in C. neoformans In previous in vitro studies, the calcineurin inhibitor FK506 has demonstrated synergistic interactions with caspofungin against C. neoformans (Del Poeta et al. 2000). In our caspofungin sensitivity screen, we identified a mutant of the calcineurin B regulatory subunit, cnb1D, to be highly sensitive to caspofungin, as seen in previous work (Kraus et al. 2003). Indeed, mutants of both the calcineurin A catalytic and B regulatory subunits-cna1D and cnb1D, respectively-exhibit an eightfold increase in caspofungin sensitivity when compared to the WT strain (Table 1 and  Table 2). Therefore, we strove to identify the mechanism by which this phosphatase is facilitating caspofungin tolerance in C. neoformans by examining the activation of a known target of calcineurin signaling in C. neoformans, the Crz1 transcription factor.
The Crz1 transcription factor is activated through dephosphorylation by calcineurin, and Crz1 mediates many of the known transcriptional effects of this calcineurin signaling (Lev et al. 2012). We assessed the effect of caspofungin on Crz1 nuclear localization by examining mCherry-tagged Crz1 fusion protein (Crz1-mCherry) colocalization with the GFP-Nop1 nucleolar marker, as well as with DAPI staining (Chow et al. 2017). In untreated cells, Crz1-mCherry remains localized in the cytosol. We determined that Crz1-mCherry localizes to the nucleus after 45 min of incubation in 15 mg/ml of caspofungin ( Figure 1, A and B). The nuclear localization of Crz1-mCherry suggests that Crz1 is likely being activated under these conditions. However, the crz1D mutant displayed a caspofungin MIC more similar to WT than to the cna1D or cnb1D mutant strains (Table  2). Together, these results suggest that, although Crz1 is being activated during caspofungin treatment, caspofungin tolerance is mediated in part through calcineurin-dependent but Crz1-independent targets. Therefore, we assessed the caspofungin susceptibility of several strains with loss-offunction mutations in downstream calcineurin targets (Table S6) (Park et al. 2016). Unlike the cna1D or cnb1D calcineurin subunit mutants, none of the tested strains deficient in individual downstream calcineurin effector genes displayed altered caspofungin susceptibility. These data are consistent with emerging reports that C. neoformans calcineurin acts through both Crz1-dependent and -independent processes (Lev et al. 2012;Chow et al. 2017). Alternatively, there may be functional redundancy among the calcineurin effectors with respect to caspofungin susceptibility. The Pfa4 palmitoyltransferase plays a role in caspofungin tolerance through regulation of its target proteins The Pfa4 palmitoyltransferase is required for the addition of palmitoyl groups to various C. neoformans proteins (Nichols et al. 2015;Santiago-Tirado et al. 2015). This post-translational modification is necessary for the proper localization and function of these target proteins. In our MIC assays, we documented a fourfold increase in caspofungin susceptibility of the pfa4D strain at 30°. Given the temperature sensitivity of this mutant strain (Nichols et al. 2015;Santiago-Tirado et al. 2015), we would not have isolated it in screens at higher incubation temperatures. The caspofungin sensitivity of the pfa4D mutant was more pronounced at 35°, a temperature at which the pfa4D mutant grows but with a modest growth delay compared to WT (Table 1 and Table 3) (Nichols et al. 2015;Santiago-Tirado et al. 2015). Additionally, since Pfa4-mediated palmitoylation also seems to play a role in caspofungin tolerance in C. neoformans, we assessed whether there might be synergy between caspofungin and inhibitors of palmitoyltransferases. 2BP is a competitive inhibitor of palmitoyltransferases with antifungal activity against Aspergillus fumigatus (Jennings et al. 2008;Fortwendel et al. 2012). In contrast, we found that 2BP displayed limited antifungal activity alone against C. neoformans, with an MIC of 25-50 mM. However, 2BP displayed synergy with caspofungin at both 30 and 37° (Table  4). This observed pharmacologic synergy between caspofungin and 2BP is consistent with the synthetic effect on growth inhibition between caspofungin and the pfa4D mutation.
Since Pfa4 is responsible for the regulation of various functions within the cell, it is likely that this caspofungin sensitivity is due to dysregulation of one or more Pfa4 palmitoylation targets. Moreover, the additional growth and drug-sensitivity phenotypes at 35°are in concordance with known targets of Pfa4 being required for full thermotolerance in C. neoformans (Nichols et al. 2015). Therefore, the caspofungin susceptibility caused by mutations in several Pfa4-regulated gene products was assessed.
The thermotolerance arm of the Ras signaling pathway is required for caspofungin tolerance Pfa4 palmitoylates the C. neoformans Ras1 GTPase, and this post-translational modification is required for the proper subcellular localization of Ras1, as well as its function (Nichols et al. 2015). To determine whether the caspofungin susceptibility of the pfa4D mutant is reflected in this downstream target pathway, we assessed caspofungin susceptibility of the ras1D mutant, as well as for mutants in the Ras1 morphogenesis (mediated by Rac proteins) and Ras1 thermotolerance subpathways (mediated by Cdc24/Cdc42 and the septin proteins) (Figure 2) (Waugh et al. 2002;Nichols et al. 2007;Ballou et al. 2009Ballou et al. , 2013a. The ras1D mutant is fourfold more sensitive to caspofungin than WT. Similar increases in caspofungin susceptibility were also noted for the cdc42D and cdc24D mutants, as well as the cdc3D and cdc12D septin mutants, which are further downstream effectors of the C. neoformans Ras thermotolerance pathway (Nichols et al. 2007;Ballou et al. 2009Ballou et al. , 2013a. These proteins mediate dynamic actin cytoskeletal changes associated with budding and cell division, and a robust septin complex is required for C. neoformans to grow in the presence of cell stresses such as elevated temperature. Therefore, each of these signaling proteins is required for cryptococcal viability at elevated temperatures and other states of cell stress. Accordingly, the changes in caspofungin susceptibility in the corresponding mutant strains were only observed at 35°, not 30°, indicating that the elements of this pathway are necessary for caspofungin tolerance under conditions at which the pathway is activated ( Table 2 and Table 3). In contrast, the Rac1 and Rac2 proteins are not required for caspofungin tolerance. These Ras1-mediated GTPases are involved in a distinct signaling pathway controlling morphological transitions, such as hyphal formation during mating (Ballou et al. 2013b).
Although specific inhibition of fungal Ras activity is limited by the highly conserved nature of this protein, recent investigations have explored inhibitors of Ras-modifying enzymes as antifungal agents (Hast et al. 2011;Selvig et al. 2013;Esher et al. 2016;Pianalto and Alspaugh 2016). In addition to palmitoylation, Ras-like GTPases also require the post-translational addition of lipophilic prenyl groups to C-terminal cysteine residues for their proper localization and function. These lipid additions are catalyzed by protein-farnesyltransferases and/or protein geranylgeranyltransferases (Palsuledesai and Distefano 2014). Additionally, protein-farnesyltransferase inhibitors (FTIs) could be affecting the function of not only Ras proteins, but other farnesylated proteins. Therefore, we assessed the pharmacological synergy between caspofungin and the FTI manumycin A (Hast et al. 2011). Manumycin A treatment has been demonstrated previously to disrupt the plasma membrane localization of GFP-Ras1 in C. neoformans, consistent with its biochemical activity as an FTI (Hast et al. 2011). In checkerboard MIC assays, manumycin A displayed synergy with caspofungin, with FIC indices ranging from 0.508 to 0.75 at 30°, and from 0.562 to 0.75 at 37° (Table 4) (Franzot and Casadevall 1997). Given the limited intrinsic antifungal activity of this first-generation FTI, these results suggest that Ras inhibitors with greater anticryptococcal activity might be promising coadministered agents to augment the effect of caspofungin against C. neoformans.
Chitin synthesis and the Chs3 chitin synthase are involved in the response to caspofungin treatment One of the most prominent targets of Pfa4 palmitoyltransferase activity is the Chs3 chitin synthase, which is responsible for the biosynthesis of chitin that is destined to become chitosan in the cell wall Baker et al. 2007). In addition to the pfa4D mutant strain, our screen also identified the chitin synthase regulator mutant strain csr2D as a highly caspofungin-sensitive strain (Table 1). Csr2 and Chs3 coregulate chitin synthesis in C. neoformans . Given the altered caspofungin tolerance of two mutants predicted to affect the function of the Chs3 chitin synthase, we assessed the caspofungin susceptibility of a chs3D strain to determine if misregulation of this protein, and a lack of proper chitin and chitosan deposition, might contribute to the caspofungin susceptibility of the pfa4D mutant strain. The chs3D mutant strain was eightfold more sensitive to caspofungin than the WT strain (Table 2). These results are consistent with observations in A. fumigatus and Candida species in which compensatory increases in cell wall chitin are induced upon exposure of these fungi to caspofungin (Walker et al. 2008(Walker et al. , 2012Fortwendel et al. 2010;Verwer et al. 2012). To further explore the interaction of caspofungin treatment and a response involving chitin biosynthesis in the cell wall, we used the chito-oligomer-binding dye CFW to assess changes in C. neoformans cell wall chitin content in response to caspofungin treatment. We observed a dosedependent increase in CFW staining for cells treated with caspofungin compared to untreated cells, with the most significant difference being at the highest concentration of caspofungin (Figure 3, A and B). Accordingly, when we assessed cell wall chitin and chitosan composition using an in vitro colorimetric assay , we found that C. neoformans cells treated with 15 mg/ml caspofungin have an 2.5-fold increase in both chitin and chitosan compared to untreated cells ( Figure 3C).
Since the chs3D chitin synthase mutant strain is so susceptible to the effects of caspofungin, we hypothesized that the coadministration of caspofungin with the chitin synthase inhibitor nikkomycin Z might result in a similar synergistic antifungal effect. We tested antifungal drug interactions using a checkerboard MIC assay, using varying concentrations of each drug in combination. In this assay, we did not observe synergy between caspofungin and nikkomycin Z, one of the few known inhibitors of chitin synthesis in some fungi (Gaughran et al. 1994;Li and Rinaldi 1999). The FIC index, a marker of drug interaction, was 1.0, suggesting an additive effect between the two compounds rather than drug synergy (Table 4). However, historically, nikkomycin Z has not been an effective inhibitor of C. neoformans growth, potentially due to the redundancy in chitin synthase genes in this organism (Li and Rinaldi 1999;Kraus et al. 2003;Banks et al. 2005). Eight chitin synthases have been identified in C. neoformans, with both distinct and overlapping functions . While the specific biochemical target of nikkomycin Z has yet to be identified in C. neoformans, treatment with nikkomycin Z activates the Pkc1 cell wall integrity pathway, suggesting that nikkomycin Z inhibits cell wall biosynthesis (Kraus et al. 2003). Given the very limited antifungal activity of nikkomycin Z against C. neoformans and our new genetic studies, we propose that there is potential for synergy between caspofungin and future, more effective chitin synthase inhibitors, or other cell wall biosynthesis inhibitors. Cell wall gene expression is altered during caspofungin treatment Given the caspofungin-induced changes in cell wall chitin content, we explored whether the expression of cell wall biosynthesis genes was altered in response to caspofungin treatment. We assessed the transcript abundance for genes involved in the synthesis of chitin (CHS1, CHS2, CHS3, CHS4, CHS5, CHS6, CHS7, and CHS8), chitosan (CDA1, CDA2, and CDA3), a-1,3-glucan (AGS1), b-1,3-glucan (FKS1), and b-1,6-glucan (KRE6 and SKN7) (Thompson et al. 1999;Banks et al. 2005;Reese et al. 2007;Gilbert et al. 2010;Esher et al. 2018). Many of these cell wall biosynthesis genes demonstrate altered regulation in response to caspofungin treatment ( Figure 4). CHS1, CHS2, CHS4, CHS7, SKN1, and CDA1 all exhibited significant increases in expression, especially at the highest concentration of caspofungin tested in this experiment. These results may reflect some of the compensatory processes seen in other fungi, such as Candida species and A. fumigatus, in which increased cell wall chitin has been proposed to be a mechanism for these fungi to overcome the effects of caspofungin treatment (Walker et al. 2008;Fortwendel et al. 2010). However, we also noted the potential involvement of more diverse cell wall components. Additionally, the upregulation of SKN1 suggests a potential role for b-1,6-glucan synthesis in the caspofungin tolerance mechanism for C. neoformans. In light of previous data showing that caspofungin exposure results in a decrease in b-1,6-glucan staining via immunoelectron microscopy, there is the potential that the upregulation of b-1,6-glucan biosynthetic genes could represent transcriptional compensation for secondary cell wall effects after drug treatment (Feldmesser et al. 2000). Interestingly, CHS5, CHS6, CDA2, and CDA3 all displayed decreased expression during caspofungin treatment. Therefore, though not all cell wall-associated genes are upregulated in response to caspofungin, there seems to be a decisive alteration in the expression of cell wall biosynthesis and modification genes, reflecting a coordinated compensatory response to the cell wall inhibitor caspofungin.

Fks1-GFP localization is not altered during caspofungin treatment
Since we did not observe statistically significant differences in FKS1 transcript levels after exposure to caspofungin, we assessed whether treatment with this drug has a direct effect on the cellular localization of the b-1,3-glucan synthase. We examined the localization of the two components of the complex: the Fks1 catalytic subunit and the Rho1 regulatory subunit, each as a fusion protein tagged with GFP. We treated C. neoformans strains expressing either endogenous Fks1-GFP or overexpressed GFP-Rho1 in a gradient of caspofungin concentrations, and assessed localization over 1 hr of treatment ( Figure S1). Both Fks1-GFP and GFP-Rho1 localized to endomembranes, including structures that resemble perinuclear endoplasmic reticulum staining . Similarly, both proteins also localized to the outer edge of the cell. Fks1-GFP, in particular, appeared to localize to the plasma membrane, but in distinct patches rather than along the entire cell surface. Therefore, there was no significant difference in Fks1-GFP or GFP-Rho1 localization in response to caspofungin treatment.

Inhibition of efflux pumps increases potency of caspofungin
Recent studies suggest that caspofungin is not efficiently maintained at high intracellular concentrations in C. neoformans (Huang et al. 2016). Therefore, we hypothesized that this organism might be preventing the intracellular accumulation of caspofungin through the activity of one or more efflux pumps. To address this hypothesis, we performed a drug interaction assay to determine whether caspofungin potency might be enhanced in the presence of clorgyline, a monoamine oxidase inhibitor that also acts as an inhibitor of ATP-binding cassette (ABC) transporters in other fungal species (Holmes et al. 2012). This pharmacological approach was pursued since there are . 50 predicted ABC transporters and . 150 predicted major facilitator superfamily proteins encoded in the C. neoformans genome (Loftus et al. 2005;Holmes et al. 2016). We found that the efflux inhibitor clorgyline and caspofungin demonstrated a strong synergistic interaction, with an FIC index ranging from 0.25 to 0.281 at 30°and 0.18-0.625 at 37° (Table 4). These data suggest that drug efflux might be a contributor to the innate tolerance of C. neoformans to caspofungin treatment.

Discussion
Despite the medical importance of fungal infections, few novel antifungal drugs have been introduced in recent years. Echinocandins, one of the most recently approved classes of antifungal drugs, were especially promising given their limited toxicity profile and novel mechanism of action. However, these agents display limited efficacy against several fungal classes, including the thermally dimorphic fungi and Cryptococcus species Bartizal et al. 1997;Espinel-Ingroff 1998). Our screen, using a collection of C. neoformans mutant strains, revealed that cryptococcal tolerance to the echinocandin drug caspofungin is likely a multifaceted phenomenon. We identified several pathways and processes important for full tolerance of caspofungin, including those associated with cell surface integrity, such as chitin biosynthesis and the phosphatase Ppg1, or general stress tolerance, such as calcineurin. Indeed, one of the most caspofungin-sensitive mutants that we identified was a ppg1D mutant. Ppg1 is a phosphatase that has been shown to be necessary for full cell wall integrity, even under nonstress conditions, but that is also required for survival in cell wall stress conditions . However, Ppg1 is not a part of the traditional cell wall integrity MAP kinase pathway in C. neoformans. Interestingly, we did not see hypersensitivity among mutants in the cell wall integrity MAPK pathway, which suggests that noncanonical cell wall integrity signaling could be playing a role in the response to caspofungin-induced cell wall stress . We also found multiple proteins that have potential antioxidant roles, such as an Rdl2 Rhodanese homolog and a putative Lys7 homolog, which typically partners with superoxide dismutase 1 (Culotta et al. 1997;Orozco et al. 2012). We chose to focus our initial evaluation on strains with mutations in cellular pathways previously implicated in C. neoformans pathogenesis.
Another recent study assaying C. neoformans mutants for caspofungin sensitivity identified several other processes that were not identified in this study. The differences in the caspofungin-susceptible C. neoformans mutants identified in the two different studies are likely explained by variations in experimental conditions. First, the two screens used overlapping but different mutant libraries for the screens, as well as different cutoffs for defining altered caspofungin susceptibility. Also, we specifically chose to perform our screens at 30°to identify mutants that might have thermotolerance defects; for example, mutants such as the calcineurin subunit cnb1D mutant would not have been discovered at a higher temperature given its intrinsic inability to grow at higher temperatures.
In their study, Huang et al. determined that mutation of the Cdc50 lipid flippase b-subunit increased the sensitivity of this organism to caspofungin and fluconazole (Huang et al. 2016). By using a fluorescently labeled caspofungin molecule, they also observed that the cdc50D mutant strain displayed increased drug uptake compared to WT strains, suggesting that the altered membrane integrity and lipid content in the mutant strain might allow better penetration of the drug into the cell, and, therefore, higher efficacy. This observation implies that overcoming limitations of entry into or accumulation of the drug in the cryptococcal cell might enhance fungal cell killing. To further explore the importance of drug efflux, we assessed the effect of the efflux pump inhibitor clorgyline on caspofungin activity, and we found strong synergy between these two compounds. We specifically employed clorgyline as a pharmacological inhibitor of ABC transporters to address the potential functional redundancy of Figure 2 Ras1 signaling partner mutants are differentially affected by caspofungin treatment. Model of Ras signaling in C. neoformans. Paired paralogs are represented as the major (larger oval) and minor (smaller oval) paralog (Ballou et al. 2013a). Caspofungin minimal inhibitory concentrations at 35°for each component are presented within each oval. The green ovals represent mutants that displayed increased caspofungin susceptibility, while the gray ovals display wild-type caspofungin susceptibility.
the large number of C. neoformans genes likely to encode efflux pumps (Loftus et al. 2005;Holmes et al. 2016).
In validation of our screening methods, we identified the calcineurin B subunit mutant as a hypersensitive strain. Calcineurin signaling is required for full virulence of C. neoformans and is involved in the response to elevated temperatures, as well as the response to caspofungin and other cellular stresses (Odom et al. 1997;Cruz et al. 2000;Del Poeta et al. 2000;Fox et al. 2001). Similarly, calcineurin signaling has been shown to be required for paradoxical growth in the presence of high Cells were incubated in SC medium at 30°with 0, 5, or 20 mg/ml CSF over 24 hr. At each timepoint, cells were harvested and stained with CFW to assess total cell wall chito-oligomer content, then imaged on a Zeiss AxioVision epifluorescence microscope. Images were masked for fluorescence and average fluorescent intensity for each cell was quantified using ImageJ. Each timepoint represents quantification of . 80 cells over at least three images. Error bars represent SEM. Statistics performed were two-way ANOVA, followed by pair-wise Student's t-tests. * P , 0.05 and *** P , 0.001. (B) Representative images of cells treated as above with 0, 5, or 20 mg/ml CSF, and then stained for chitin content with CFW at 4 hr post-CSF treatment. (C) Quantification of cell wall chitin and chitosan using a DMAB colorimetric assay. Cells were incubated in SC 6 15 mg/ml CSF at 30°for 6 hr, then harvested and the assay performed as described. Error bars represent SEM. Statistics performed were two-way ANOVA, followed by pair-wise Student's t-tests. * P , 0.05 and *** P , 0.001. CFW, calcofluor white; CSF, caspofungin; DMAB, p-dimethylaminobenzaldehyde; GlcNAc, N-acetylglucosamine; ns, not significant.
concentrations of caspofungin in Candida species and A. fumigatus (Walker et al. 2008(Walker et al. , 2012Fortwendel et al. 2010;Lee et al. 2011;Juvvadi et al. 2015). In C. neoformans, calcineurin is known to regulate one transcription factor, Crz1, and many calcineurin-dependent cellular processes are likely mediated by Crz1 (Lev et al. 2012). Importantly, recent work identified several genes involved in cell wall biosynthesis and remodeling that are transcriptionally upregulated by Crz1 and Cna1 in response to thermal stress (Chow et al. 2017). However, this work also revealed that a significant number of genes are regulated differentially between the cna1D and the crz1D transcriptomes, suggesting that other targets of Cna1 activity could have effects on the transcriptional response to stress. Additional work demonstrated that calcineurin likely has multiple downstream targets in addition to Crz1, including proteins that localize to stress granules and P-bodies, and are involved in post-transcriptional responses to thermal stress (Park et al. 2016). Our work suggests that the calcineurin-regulated response to caspofungin likely involves both the activity of the Crz1 transcription factor as well as Crz1-independent calcineurin targets. Indeed, it has also been demonstrated in C. albicans that Crz1 is only partially responsible for the calcineurin-mediated echinocandin tolerance seen in that organism (Singh et al. 2009). Future work could probe Crz1-independent calcineurin targets for roles in caspofungin tolerance in this organism.
We also tested multiple processes regulated by the Pfa4 palmitoyltransferase to explore whether the drug-susceptible phenotype of the palmitoyltransferase mutant could be attributed to dysfunction of one or more of its downstream targets. Interestingly, we determined that the C. neoformans Ras1-mediated thermotolerance pathway is required for tolerance to caspofungin treatment in this organism. This signaling pathway responds to extracellular cues to control the activation of the Cdc42 GTPase, a protein that directs actin cytoskeleton polarization as well as directional growth and budding ). The septin proteins Cdc3 and Cdc12 are required for cytokinesis through their role in septum formation. Additionally, Cdc3 and Cdc12 are necessary for normal cell morphogenesis at elevated temperatures, and the localization of these proteins is mediated by Cdc42. In the absence of any of these pathway proteins, the cryptococcal cell is growth-impaired under many stressful conditions, such as mammalian physiological temperatures (Nichols et al. 2007;Ballou et al. 2009Ballou et al. , 2013a. Although mutants of genes encoding two members of this pathway, RAS1 and CDC42, were present in the collections tested in this study, their growth phenotypes are only apparent at elevated temperatures. As a result, these mutants would not be expected to display caspofungin susceptibility at the lower temperatures tested, emphasizing the importance of screening for drug sensitivity in multiple incubation conditions. Since Ras and Ras-related proteins are required for pathogenesis in many fungal systems, there have been concerted efforts to identify small molecules that inhibit Ras protein function. One of the most promising directions includes targeting protein prenyltransferases, enzymes that add lipid moieties to the C-terminal regions of Ras-like GTPases, directing the localization of these proteins to cellular membranes, which are their sites of function (Vallim et al. 2004;Nichols et al. 2009;Fortwendel et al. 2012;Esher et al. 2016). Prenylation inhibitors are predicted to alter the localization and function of many cellular proteins in addition to Ras, potentially having broad antiproliferative effects in divergent cell types. Structural studies have identified fungal-specific features of farnesyltransferase enzymes, suggesting that Figure 4 Cell wall biosynthesis gene expression is altered in response to CSF. Wild-type C. neoformans cells were incubated in SC medium containing 0, 10, or 15 mg/ml CSF for 90 min, followed by RNA purification and quantitative real-time PCR for the indicated target genes, using the GPD1 gene as an internal control. The results represent average values for biological triplicate samples. Statistical significance was assessed using two-way ANOVA, followed by Student's t-test for pair-wise comparisons. * P , 0.05, ** P , 0.01, *** P , 0.001, and **** P , 0.0001. CSF, caspofungin.
antifungal specificity could be engineered into FTIs (Hast et al. 2011;Mabanglo et al. 2014). Antifungal susceptibility testing demonstrated in vitro synergy between the activities of caspofungin and the FTI manumycin A, the basis of which was identified in our screen. Importantly, the currently available FTIs tested have limited efficacy as antifungal compounds when used alone. However, as newer and more potent antifungal FTIs are identified, these agents might be used as adjunctive therapies to enhance the effect of caspofungin.
We also determined that chitin biosynthesis is upregulated during caspofungin treatment and that the Chs3 chitin synthase is required for caspofungin tolerance. As caspofungin inhibits the biosynthesis of the cell wall component b-1,3glucan, upregulation of other cell wall biosynthesis or cell wall-modifying enzymes may compensate for altered glucan content. Indeed, in A. fumigatus, cell wall chitin deposition, as well as the expression of chitin synthase genes, increases during caspofungin treatment (Fortwendel et al. 2009(Fortwendel et al. , 2010. Additionally, Candida clinical isolates with naturally increased levels of chitin, as well as strains grown in chitininducing conditions, demonstrate increased survival during caspofungin treatment (Walker et al. 2008(Walker et al. , 2012Lee et al. 2011). A similar phenomenon appears to be occurring in C. neoformans: during caspofungin treatment, we measured increased levels of chitin both by CFW staining and by an in vitro biochemical quantification of chitin monomers. There is a similar increase in chitosan levels. As Cryptococcus species typically display higher levels of chitosan in the cell wall than other pathogenic fungal species , these inducible cell wall changes in chitin/chitosan content likely represent a conserved mechanism by which C. neoformans adapts to caspofungin treatment. Additionally, the increased baseline chitosan levels in the C. neoformans cell wall may also contribute to its innate tolerance to this drug.
In conclusion, several varied cellular processes were represented among the strains with enhanced caspofungin susceptibility. It is possible that the loss of some of these processes enhances the intracellular accumulation of the drug, preventing the low concentrations of caspofungin in C. neoformans cells suggested in prior studies. Alternatively, compromising these cellular pathways may render the C. neoformans cell susceptible to lower effective concentrations of this drug. As new fungal-specific inhibitors are developed for processes such as Ras protein localization and chito-oligomer synthesis, these agents may provide promising new directions for combination antifungal therapy.