Calcineurin-dependent contributions to fitness in the opportunistic pathogen Candida glabrata

ABSTRACT The protein phosphatase calcineurin is vital for the virulence of the opportunistic fungal pathogen Candida glabrata. The host-induced stresses that activate calcineurin signaling are unknown, as are the targets of calcineurin relevant to virulence. To potentially shed light on these processes, millions of transposon insertion mutants throughout the genome of C. glabrata were profiled en masse for fitness defects in the presence of FK506, a specific inhibitor of calcineurin. Eighty-seven specific gene deficiencies depended on calcineurin signaling for full viability in vitro both in wild-type and pdr1∆ null strains lacking pleiotropic drug resistance. Three genes involved in cell wall biosynthesis (FKS1, DCW1, FLC1) possess co-essential paralogs whose expression depended on calcineurin and Crz1 in response to micafungin, a clinical antifungal that interferes with cell wall biogenesis. Interestingly, 80% of the FK506-sensitive mutants were deficient in different aspects of vesicular trafficking, such as endocytosis, exocytosis, sorting, and biogenesis of secretory proteins in the endoplasmic reticulum (ER). In response to the experimental antifungal manogepix that blocks GPI-anchor biosynthesis in the ER, calcineurin signaling increased and strongly prevented cell death independent of Crz1, one of its major targets. Comparisons between manogepix, micafungin, and the ER-stressing tunicamycin reveal a correlation between the degree of calcineurin signaling and the degree of cell survival. These findings suggest that calcineurin plays major roles in mitigating stresses of vesicular trafficking. Such stresses may arise during host infection and in response to antifungal therapies. IMPORTANCE Calcineurin plays critical roles in the virulence of most pathogenic fungi. This study sheds light on those roles in the opportunistic pathogen Candida glabrata using a genome-wide analysis in vitro. The findings could lead to antifungal developments that also avoid immunosuppression.

Calcineurin also promotes resistance and tolerance in pathogenic fungi to several different classes of antifungals (17,18).For example, calcineurin activation and signaling promote tolerance (also called cell viability) during long-term exposure of yeasts to azole-class antifungals, which target ergosterol biosynthesis in the endoplasmic reticulum (ER) (19)(20)(21).Similar effects have been observed in response to non-clin ical antifungals such as tunicamycin and dithiothreitol, which perturb glycoprotein biosynthesis in the ER (20,22).The pro-survival effects of calcineurin during ER stress can operate independent of Crz1, a fungi-specific transcription factor unrelated to NFAT that is directly dephosphorylated by calcineurin (20,22).Several other known substrates of calcineurin in the model yeast S. cerevisiae were also not required for calcineurin-depend ent cell survival in response to ER stress (23) and the molecular mechanisms by which calcineurin promotes tolerance remain unknown.Calcineurin and Crz1 activation also promote echinocandin resistance by increasing the expression of FKS2 encoding a target of these drugs (17,18).Echinocandin resistance often arises through mutations within the coding sequences of FKS genes (24,25) although evidence suggests other pathways can contribute (26).Based on these findings, selective inhibitors of fungal calcineurin would enhance the potency or cidality of existing antifungals and diminish intrinsic virulence of fungal pathogens without causing immunosuppression in the host (27).
A better understanding of the stresses in fungi that are sensed by calcineurin and how the sensor promotes resistance, tolerance, and virulence could provide new approaches for antifungal interventions.Toward that goal, this study aims to identify cellular stresses in C. glabrata where calcineurin signaling becomes essential for fitness in vitro.These in vitro stresses may be representative of stresses encountered in vivo during host infections, thereby facilitating deeper studies of the inputs and outputs of the calcineurin signaling pathway.By profiling large pools of transposon insertion mutants in C. glabrata for hypersensitivity to FK506, a broad range of genotoxic stresses can be surveyed.Using this approach, we identify dozens of genes and several cellular processes whose disruptions cause dependence on calcineurin signaling for fitness.One such process is cell wall biosynthesis, in which calcineurin appears to induce paralogs that function in compensatory pathways and provide resistance to an echinocandin.GPI-anchor deficiency, which is also conferred by the experimental antifungal man ogepix, also produces a cellular stress for which calcineurin function compensates.Overall, deficiencies in vesicular trafficking were the most common cellular stresses that depended on calcineurin signaling.

Mutants with elevated dependence on calcineurin signaling
We sought to identify stresses in C. glabrata during which calcineurin becomes important for cell survival or proliferation.To do this, pools of Hermes transposon insertion mutants in the strain BG14 (28), a ura3∆ derivative of vaginal isolate BG2 (29), were grown to stationary phase, diluted 100-fold into fresh culture medium containing or lacking FK506, and then incubated for another 24 h.Each insertion site of more than 400,000 total in each pool was PCR amplified, sequenced, mapped, and tabulated gene-wise as described previously (30), and then each gene was charted under the two conditions (Fig. 1A).Most genes were equally represented by transposon reads in the two conditions, suggesting that their disruptions do not generate stresses that are compensated by calcineurin signaling (black dots in Fig. 1A).No disrupted genes were significantly enriched in the FK506 condition.However, dozens of disrupted genes were significantly underrepresented following FK506 exposure.Among those genes is the known FK506-sensitive gene FKS1 (red square) that was shown previously to depend on calcineurin signaling when disrupted (31).The experiment was repeated in an independently generated pool of insertion mutants in the BG14 background and also in an independently generated pool of insertion mutants in a pdr1∆ derivative of BG14 (32).This pdr1∆ pool was profiled because FK506 may regulate Pdr1 transcription factor independent of calcineurin (33) and because calcineurin can physically interact with Pdr1 (34).For each experiment, z-scores were calculated for all annotated genes and compared.FKS1 exhibited strongly negative z-scores in all three experiments (−8.9, -17.2, −9.6).Overall, the z-scores from the two BG14 experiments were highly correlated with each other (PCC = 0.41) and both correlated well with z-scores from the pdr1∆ experiment (PCC = 0.36 and 0.47).When the z-scores from pdr1∆ experiments were charted against the average z-scores of the replicate BG14 experiments, few outliers were evident (Fig. 1B).These findings suggest that dozens of genes reproducibly impact FK506 sensitivity in these conditions independent of Pdr1.
After averaging all three datasets, a total of 87 genes were depleted more than twofold and with z-scores less than −2.0 in response to FK506 (Table 1; Table S1).The SSD1 gene was not significant in these data sets even though ssd1∆ mutants were shown previously to be hypersensitive to FK506 (35).Closer inspection of SSD1 revealed that the gene was poorly covered with transposon insertions, suggesting that it was essential for viability in these conditions.Gene Ontology analyses revealed >4-fold enrichment of 10 processes, 9 of which involve vesicular trafficking (Table S2).When z-scores were limited to less than −4.0 (34 genes), 6 processes were enriched >4-fold and all involved vesicular trafficking.After manual curation, over 79% of the significant genes were found to contribute in some way to vesicular trafficking with the remainder contributing to cell wall biogenesis or unidentified functions.

Calcineurin improves fitness during stresses in vesicular trafficking
Defects in vesicular trafficking and endoplasmic reticulum functions were predicted for 69 of the 87 different mutants identified above as FK506-hypersensitive (Table 1; Table S1).The mutant genes include INP52, ARF1, a number of genes encoding Arf1-interacting proteins in S. cerevisiae such as VPS13, GGA1, and CHS6 (36).The screens also found all of the core subunits of the AP-1 and AP-1R complexes (APM1, APM2, APL2, APL4, APS1, LAA1, MIL1) that bind clathrin and Arf1 and promote vesicular transport from the Golgi complex to endosomes were among this group.Furthermore, we found that multiple subunits of the ESCRT complexes (VPS25, SNF8, STP22, DID4, VPS28, SNF7, VPS4) that promote endosomal trafficking, multivesicular body formation, and vacuolar delivery were also FK506-sensitive.A inp52∆ knockout mutant previously constructed in the CBS138-HTL strain background (37) exhibited strong hypersensitivity to the calcineurin inhibitors FK506 and cyclosporin A (Fig. 2).We knocked out ARF1, two genes encoding AP-1 subunits (APL2, APS1), and one subunit of the ESCRT-II complex (SNF8) in the BG14 strain of C. glabrata and tested for calcineurin dependence.The arf1∆, apl2∆, and aps1∆ mutants were strongly hypersensitive to the calcineurin inhibitors though both FIG 2 Responses of individual knockout mutants of C. glabrata to calcineurin inhibitors.Individual gene knockout mutants were constructed in wild-type strain BG14 and in the crz1∆ derivative as indicated.The inp52∆ and las21∆ knockout mutants of strain CBS138-HTL were obtained from the C. glabrata gene knockout collection (37).The indicated C. glabrata strains were grown to saturation, diluted 5-fold serially, frogged onto SCD agar medium containing calcineurin inhibitors, incubated for 1 day at 30°C, and then photographed.The CBS138-HTL and inp52∆ strains were grown on the same agar plates, but two intervening strains (shown later in Fig. 7A) were cropped out of the images.
formed slightly smaller colonies than the parent strain in the absence of calcineurin inhibitors (Fig. 2).The snf8∆ mutant exhibited very weak sensitivity to calcineurin inhibitors in these conditions (Fig. 2), which contrasts with the large effects observed in the transposon pools (average z-score = −8.2).SNF8 is a large gene with a high density of transposon insertions in our pools, which causes even small phenotypic effects to be highly significant in our z-score calculations.Three of these genes were also knocked out in a crz1∆ mutant background and tested similarly.The resulting double mutants behaved indistinguishable from the apl2∆, aps1∆, and snf8∆ single mutants (Fig. 2), suggesting that calcineurin promotes fitness during these stresses through Crz1-inde pendent effects.The relevant targets of calcineurin have not yet been identified.

ER stresses activate calcineurin, which promotes growth and cell survival
Several genes whose products promote secretory protein modifications in the ER were on the list of FK506-sensitive mutants with vesicular trafficking defects.Among these were ALG3, ALG5, ALG6, and ALG8 whose products function sequentially in the N-glycan biosynthetic pathway, CNE1 that binds glucosylated N-glycosylated secretory proteins, and GTB1 that promotes their deglucosylation.A cne1∆ mutant of C. glabrata was previously shown to be hypersensitive to FK506 (38).The alg5∆, alg6∆, and alg8∆ mutants of the BG14 background all exhibited mild sensitivity to calcineurin inhibitors (Fig. 3A).To determine whether calcineurin can become activated in response to deficiencies in non-essential components of N-glycosylation, we quantified expression of RCN2, a well-studied target of calcineurin and Crz1 in C. glabrata (7,39), using quantitative real-time PCR.Elevated levels of RCN2 expression were observed in alg6∆ and alg8∆ mutants (Fig. 3B).Additionally, several non-essential genes involved in GPI-anchor biosynthesis in the ER (LAS21, PER1, BST1) and trafficking of GPI anchored glycoproteins from the ER to the Golgi complex (EMP24) were also identified as hypersensitive to FK506 when disrupted with transposons.A las21∆ mutant in the CBS138-HTL background exhibited hypersensitivity to FK506 and cyclosporin A (Fig. 2).Insertions in the 354 bp gene CAGL0G03993g were highly significant in all three pools (average z-score = −5.2).This small gene is not transcribed (40), and its product is not conserved in any other species, suggesting that it is misannotated as a gene.Furthermore, the segment begins only 38 bp upstream of the essential gene GPI13, which encodes an enzyme critical for GPI-anchor biosynthesis in the ER.Therefore, insertions in the CAGL0G03993g segment likely lower the expression of GPI13, potentially causing stresses similar to insertions in the non-essential GPI-anchor genes and N-glycosylation genes.
Essential genes are poorly represented with transposon insertions, which limit their detectability in the genetic screens.However, potent inhibitors of essential compo nents of N-glycosylation and GPI-anchoring have become available.Manogepix is an experimental antifungal that blocks an intermediate step in GPI-anchor biosynthesis in the ER encoded by essential GWT1 (41).Tunicamycin is a natural product that blocks the first step of N-glycan biosynthesis in the ER encoded by essential ALG7 (42).The addition of manogepix and tunicamycin to exponentially growing cells at high concentrations caused rapid induction of RCN2 expression in wild-type cells but not in crz1∆ mutant cells (Fig. 4A), suggesting that acute and severe deficiencies in these ER-associated processes also activate calcineurin signaling.Interestingly, the degree of RCN2 induction was much higher for these ER stressors than for micafungin (Fig. 4A), a cell wall stressor that was previously shown to activate calcineurin signaling (6).
In S. cerevisiae, calcineurin signaling promotes cell survival during exposure to tunicamycin through Crz1-independent effects (20).To test whether calcineurin performs similar roles in C. glabrata, crz1∆ and cnb1∆ mutants were exposed to manogepix and tunicamycin, and then cell death was quantified by staining with propidium iodide.The crz1∆ mutant and BG14 control cells largely survived the stresses caused by manogepix and tunicamycin, whereas the cnb1∆ mutant rapidly died (Fig. 4B).The rate of cnb1∆ mutant cell death appeared somewhat slower in response to manogepix when compared to tunicamycin (Fig. 4B).In contrast, micafungin exposure caused very rapid cell death in all three C. glabrata strains (Fig. 4B).For these three antifungals, the rates of cell death in cnb1∆ cultures were inversely correlated with the magnitudes of calcineurin activation and RCN2 expression in wild-type cultures.This correlation between longer cell survival and higher calcineurin signaling is consistent with calcineurin functioning as a driver of tolerance to ER stressors.

YPS7-deficient mutants depend on calcineurin
Transposon insertions suggest that YPS7 deficiency causes hypersensitivity to FK506 (average z-score = −2.9).The YPS7 gene encodes 1 of 11 secreted aspartyl proteases (SAPs, or yapsins) that are N-glycosylated and GPI-anchored in the ER and trafficked to the plasma membrane where they function in cell wall remodeling (43).Yapsin deficiencies are not lethal but exhibit diminished shedding of Epa1 and other adhesins that are anchored to the cell wall by the actions of Dcw1 and Dfg5 (43,44).Two of the yapsins (YPS1, YPS5) have been shown to depend on Crz1 and calcineurin for maximum expression (7,45).In drop tests, we confirmed that yps7∆ knockout mutants in the BG14 strain background exhibited strong hypersensitivity to FK506 (Fig. 5).Similar results were obtained for the yps7∆ mutation introduced into a yps1∆ strain background and a yps5c∆ background that also lacked YPS2 and a cluster of eight YPS genes including YPS5 (Fig. 5).A strain bearing only YPS7 and lacking 10 other yapsins retained an ability to grow in FK506, while the strain lacking all 11 yapsins exhibited FK506 hypersensitivity (Fig. 5).These findings suggest that calcineurin promotes growth or survival of yps7∆ cells through compensatory effects on non-yapsin targets that remain to be identified.This compensatory effect of calcineurin could be important for virulence, as yapsins already have been shown to be required for the virulence of C. glabrata in mouse models of invasive candidiasis (43).

Co-essential paralogs regulated by calcineurin
The FKS1 gene encodes a catalytic subunit of beta-1,3-glucan synthase, the target of micafungin, and other echinocandin-class antifungals.FKS1-deficient mutants of C. glabrata are viable but hypersensitive to FK506 because the expression of FKS2, a functionally redundant co-essential paralog, depends on the activation of calcineurin and Crz1 (6,31).To determine whether other FK506-hypersensitive mutants might depend on calcineurin and Crz1 for the expression of co-essential paralogs, we scanned the top 87 genes found in our screen for paralogs and then searched published genetic FIG 5 Calcineurin increases the fitness of yps7∆ mutants independent of 10 yapsins.The indicated strains were serially diluted and frogged onto SCD agar medium with or without FK506 as described in Fig. 2. The slower growing BG1434 and BG1603 strains were photographed after 2 days of incubation at 30°C, while the others were photographed after 1 day.interaction data sets from S. cerevisiae for evidence of co-essentiality (46).Of the 87 genes, 15 have paralogs in C. glabrata and 9 of which are conserved and co-essential in S. cerevisiae (Table S1).We then quantified expression patterns of all nine co-essential paralogs relative to control genes using RT-PCR in wild-type and crz1∆ mutant cells with and without exposure to micafungin (see the Materials and Methods).These conditions activated calcineurin and induced expression of RCN2 in a Crz1-dependent fashion (Fig. 6).The same pattern of expression was observed for FKS2, as expected from earlier studies (6,31).Five co-essential paralogs (ARF2, INP53, IRA2, PMT3, SEC12) and seven other non-essential paralogs (BOI2, CCW12, FKS3, PEX31, UPC2, VPS501, YEH1) exhibited no significant response to micafungin and/or Crz1 deficiency (Fig. 6).One co-essential paralog (MYO3) responded to micafungin independent of Crz1 and was not further studied.In S. cerevisiae, calcineurin directly dephosphorylates and regulates the product of INP53 (47), the co-essential paralog of INP52 that exhibits calcineurin dependence when disrupted in C. glabrata (Fig. 2).It is possible that calcineurin regulates additional paralogs through processes independent of Crz1.
Interestingly, DFG5 and FLC2, the co-essential paralogs of DCW1 and FLC1, were expressed similar to FKS2 and RCN2 (Fig. 6), suggesting co-regulation by calcineurin and Crz1.Previous studies in S. cerevisiae (48) showed that calcineurin and Crz1 can similarly induce expression of DFG5, FLC2, FKS2, and RCN2 orthologs, indicating evolutionary conservation of the regulatory network.Drop tests using the dcw1∆ and flc1∆ knockout mutants generated previously in the CBS138-HTL background (37) exhibited strong hypersensitivity to FK506 and cyclosporin A similar to fks1∆ mutants in the BG14 background (Fig. 7A).The dcw1∆, flc1∆, and fks1∆ mutants also exhibited significantly elevated expression of RCN2 relative to the control strains, and this effect was blocked by FK506 in most cases (Fig. 7B).Thus, the genetic losses of FKS1, DCW1, and FLC1 seemed to cause stresses that activated calcineurin and Crz1 similar to micafungin exposure, which, in turn, increased the expression of the co-essential paralogs (FKS2, DFG5, FLC2) that serve to bolster cell wall biosynthesis and remodeling.
DCW1 and its paralog DFG5 encode enzymes that are N-glycosylated and GPI-anch ored to the plasma membrane (49) and catalyze the covalent attachment of numerous adhesins and other GPI-anchored proteins to the beta-glucan in the cell wall (44,50).FLC1 and FLC2 encode transmembrane proteins of the ER with incompletely established functions in S. cerevisiae (51,52).Through analysis of published data sets, we obtained evidence suggesting that FLC1 and DCW1 gene products form a functional partnership and that FLC2 and DFG5 gene products form a similar redundant partnership.First, the protein products of FLC2 and DFG5 physically interacted in S. cerevisiae (53).Second, flc2∆ and dfg5∆ mutants exhibited strikingly similar chemical interaction profiles (54) and cluster together when analyzed alongside thousands of other knockout mutants (55).Third, as illustrated (Fig. 8), strong fitness defects (synthetic lethalities) were observed in double mutants lacking both DCW1 and FLC2 and both DCW1 and DFG5, while no fitness defects were observed in double mutants lacking both DCW1 and FLC1 or DFG5 and FLC2 (46).These findings support a hypothesis that the calcineurin-responsive genes DFG5 and FLC2 function together in the same complex or pathway that is functionally redundant with the DCW1 and FLC1 complex or pathway.A major role of calcineurin and Crz1 during the response to cell wall damage, therefore, involves up-regulation of a paralogous set of genes involved in cell wall biosynthesis.

DISCUSSION
This study explores the genetic stresses under which calcineurin signaling becomes important for the growth and survival of C. glabrata in laboratory conditions.Strikingly, nearly 80% of the genetic stresses where calcineurin became crucial for fitness involved deficiencies in the endoplasmic reticulum and vesicular trafficking systems.Overall, these findings of FK506 sensitivity in C. glabrata appear more similar to the findings in the evolutionarily distant fission yeast Schizosaccharomyces pombe (56) than to those in the much closer relative Saccharomyces cerevisiae (57)(58)(59).Some of the differences may be attributed to very different experimental methods.The previous studies utilized large collections of gene knockout mutants that were tested individually.In contrast, this study employs large pools of transposon insertion mutants that were competing against one another and profiled en masse by deep sequencing.This method yielded a z-score for each gene.The z-score is sensitive to both the magnitude of the phenotype and the intrinsic noise, which can be very large for small genes and essential genes that contain few transposon insertions relative to other genes in the pool.For large transposon-rich genes such as IRA1, highly significant z-scores can be obtained with low effect size.Conversely, insignificant z-scores could arise for genes with high phenotypic sensitivity to FK506 but very low transposon coverage.This effect may explain why SSD1 was not identified in our data sets in spite of its previously established hypersensitivity to FK506 in a different C. glabrata strain (35).When S. cerevisiae gene knockouts were screened individually for elevated Ca 2+ uptake and calcineurin signaling rather than FK506 sensitivity (60), numerous genes involved in ER and vesicular trafficking processes were identified.An emerging theme from all these studies is that calcineurin activation and signaling play a broadly conserved role in the compensatory responses to stresses in vesicular trafficking.
The ER stressor tunicamycin has been shown previously to trigger calcineurin dependency and FK506 hypersensitivity in C. glabrata as well as C. albicans and S. cerevisiae (20).The genetic screens performed here reveal several additional non-essen tial genes of the N-glycosylation process in the ER as well as several non-essential genes of the GPI-anchoring process in the ER that modifies about 135 secretory proteins including the yapsins, adhesins, Dfg5, Dcw1, and others (61).Exploiting manogepix as an acute inhibitor of GPI-anchoring, we showed that calcineurin became activated rapidly, and this activation was critical for C. glabrata cell survival independent of Crz1.C. albicans may also rely on calcineurin for survival in response to manogepix and other GPI-anchoring deficiencies (62).The pro-survival effects of calcineurin during these forms of ER stress may overlap with the pro-survival effects of calcineurin in yeasts exposed to clinical azoles, which inhibit the ERG11 gene product required for ergosterol biosynthesis in the ER (63).If the pro-survival substrates of calcineurin can be identified, alternatives to FK506 that promote the conversion of these fungistats into fungicides may become possible.
Genetic deficiencies in most major steps of the vesicular trafficking network also resulted in calcineurin dependence in C. glabrata.Assuming the genes function similar to orthologs in S. cerevisiae, these steps include packaging of N-glycosylated and GPI-anch ored proteins in the ER, their trafficking to and further modification in the Golgi complex, and subsequent sorting and trafficking to the plasma membrane.Mutants deficient in endocytosis and trafficking to the vacuole also demonstrated calcineurin dependence.Dozens of additional essential and non-essential gene products accomplish all these processes.In S. cerevisiae, several proteins involved in vesicular trafficking have been identified as direct substrates of calcineurin (3,64).While orthologs of these substrates are conserved in C. glabrata, the motifs required for recognition by calcineurin often are not conserved (64).Some calcineurin substrates, such as the products of LAC1 and LAG1, completely lack the canonical docking motifs for calcineurin (65).These genes encode redundant ceramide synthases in the ER and are excellent candidates for involvement in pro-survival functions of calcineurin.Ceramide accumulation is highly toxic to fungal cells, and its detoxification depends on effective transport to the Golgi complex and enzymatic conversion to sphingolipids (66,67).Calcineurin signaling can inhibit ceramide biosynthesis observed during ER stress (68) and potentially mitigate toxicity when vesicular trafficking has been stressed/impaired.More research will be necessary to test this hypothesis and others in order to determine how calcineurin promotes the fitness of C. glabrata cells experiencing stresses in the ER and vesicular trafficking system.
Deficiencies in several cell wall biogenesis genes (FKS1, DCW1, FLC1, YPS7, and CCW22) were strongly dependent on calcineurin for proliferation in our screens.Unlike the gene deficiencies that stress the ER and vesicular trafficking systems, all these cell wall genes have paralogs or homologs in C. glabrata, many of which (FKS2, DFG5, FLC2, YPS1, YPS5) require calcineurin and Crz1 for maximal expression.Although its expression in C. albicans was dependent on calcineurin and Crz1 (69), CCW12 did not appear to be inducible by micafungin in a calcineurin-dependent manner in C. glabrata.Simultaneous deletion of YPS7 and all 10 of its paralogs were not lethal in C. glabrata (43), and the resulting undecuple mutant lacking all 11 yapsins still exhibited hypersensitivity to FK506.This finding suggests that calcineurin compensated for YPS7 and general yapsin deficiencies through some other mechanism.In S. cerevisiae, yps7∆ mutants depended on CNB1 for fitness but not CRZ1 (46), suggesting compensatory effects of calcineurin independent of Crz1 in that species.Crz1 was clearly important for the expression of the paralogs of FKS1, DCW1, and FLC1, and those paralogs exhibited synthetic lethality when disrupted in S. cerevisiae.Some unexpected synthetic lethal interactions were also observed, such as dcw1∆ flc2∆ and flc1∆ dfg5∆ (46).These findings implicate FLC1 and FLC2 in cell wall biogenesis as exclusive partners of DCW1 and DFG5, respectively.Supporting this idea, knockout mutants of FLC2 and DFG5 exhibited highly similar chemical interaction profiles (54), and the gene products exhibited physical interac tions (53) in S. cerevisiae.Though knockout mutants of FLC1 and FLC2 exhibit cell wall deficiencies in S. cerevisiae, the products localized to ER and were necessary for FAD import, which does not have any obvious role in cell wall biogenesis (51).These findings suggest that calcineurin and Crz1 play major roles in the expression of "reserve" cell wall biogenesis genes in response to deficiencies or inhibition of the primary paralogs.
Inhibitors of FKS1 and FKS2 gene products (echinocandins) are utilized clinically for the treatment of diverse fungal diseases.The DCW1 and DFG5 gene products represent excellent targets for the development of novel antifungals due to their broad conser vation in fungi as well as the extracellular location of their active sites (50).Chemicalgenetic screening has revealed at least two compounds that may inhibit DCW1 (or FLC1) based on the high susceptibility of both dfg5∆ and flc2∆ mutants of S. cerevisiae (54).Our findings predict that calcineurin signaling will promote resistance to such inhibitors by up-regulating the expression of the targets and/or paralogs similar to the action of calcineurin on micafungin resistance.Non-immunosuppressive compounds that specifically block fungal calcineurin would likely augment the potency of those antifungals while suppressing overall virulence even in their absence (17).Targeting the factors upstream and downstream of calcineurin that promote cell survival and proliferation also could be effective at controlling fungal infections.This study provides new insights into those factors in C. glabrata.

Strains and culture conditions
A complete list of C. glabrata strains used in this study is shown in Table S3.For individual gene knockouts, the coding sequences between start and stop codons were replaced with coding sequences of ScURA3 and ScHIS3 as described (70).Knockout mutants were authenticated by PCR using external primers (Table S4).Cells were cultured in synthetic complete 2% dextrose (SCD) medium at 30°C.

Genome-wide screens
Large pools of Hermes transposon insertion mutants in strains BG14 (wild-type) and CGM1094 (pdr1∆::HYGr) were thawed from frozen stocks (28), grown to stationary phase in synthetic complete 2% dextrose (SCD) medium, diluted 100-fold into fresh medium containing or lacking 1 µg/mL FK506 (SelleckChem), and shaken for 1 day at 30°C.Cells were then pelleted, washed once in SCD medium, resuspended in an equal volume of fresh SCD medium, and shaken for 1 day at 30°C.Cells were then pelleted, resuspended in 30 mL of 15% glycerol, and frozen in aliquots at −80°C.Genomic DNA was extracted from the aliquots, sheared by sonication, A-tailed, ligated to splinkerette adapters, PCR amplified, and sequenced using a MiSeq (Illumina) instrument as described previously (30).Sequence reads were demultiplexed, mapped to the BG2 reference genome, filtered for quality, and then tabulated gene-wise (30).The tabulated data were normalized, and then a z-score was calculated for each gene using the log 2 ratio of transposon insertions in FK506 versus control divided by the local standard deviation, which was estimated from the data as described previously (30).

Spot tests of drug susceptibility
Single colonies were picked and grown to saturation in SCD medium, serially diluted in 5-fold increments, and frogged to agar plates containing SCD medium with or without supplements of FK506 (1 µg/mL) or cyclosporin A (100 µg/mL; SelleckChem).Strains were grown at 30°C for 24 h.Images were taken on a Gel Doc XR+ (Image Lab, BioRad).

RT-PCR experiments
Single colonies were picked and grown overnight at 30°C to mid-log phase in SCD medium.For each sample, cells were diluted to OD 600 = 0.1 in fresh SCD medium with or without the stressor (see below) and shaken at 30°C.At the appropriate time points, 1.5 mL of the culture was harvested by centrifugation (14 k, 60 s), and the supernatant was aspirated.Cell pellets were flash frozen in liquid nitrogen and stored at −80°C until RNA extraction.Total RNA was extracted from cells using a hot acid phenol-chloroform extraction protocol (71).Briefly, cells were lysed in an RNA lysis buffer (6 mM NaOAc, 8.4 mM EDTA, 1% SDS), and then RNA was purified through two phenol extractions and a final chloroform extraction.RNA was precipitated with isopropanol and resuspended in TE buffer.RNA extracts were treated with DNAse (New England Biolabs) to ensure no genomic DNA contamination.One microgram of RNA was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Thermo).Real-time PCR was performed using the CFX96 Touch Real-Time PCR Detection System (BioRad) using the ABsolute Blue QPCR Mix SYBR Green Kit (ThermoFisher) with the following parameters: 15 min at 95°C, 40× (15 min at 95°C, 30 min at 58°C, 30 min at 72°C).Target gene transcript levels were normalized to averaged TEF1 and PGK1 transcript levels in each sample, and this ratio from each sample was normalized to that of untreated BG14 cells.Target primers are identified in Table S4.

Cell death assay
Single colonies were picked and grown to log phase at 30°C in SCD medium.Cells were back-diluted to an OD 600 = 0.1 and dosed with 0.6 µg/mL manogepix (SeleckChem), 20 µg/mL tunicamycin (Tocris Bioscience), or 0.12 µg/mL micafungin (Cayman Chemicals) and continued to grow at 30°C.Samples were taken at time-points, spun down, stained with propidium iodide (100 µg/mL) in PBS, and manually counted on a fluorescence microscope (Zeiss Axioscope).PI-positive cells were tallied out of 200 cells counted per sample.

FIG 1
FIG 1 Genome-wide analyses of FK506 susceptibility in C. glabrata.(A) A diverse pool of Hermes transposon insertion mutants in the BG14 strain background was diluted 100-fold into fresh SCD medium containing or lacking FK506 (1 µg/mL), shaken for 1 day, and then the insertion sites were sequenced and tabulated gene-wise.Genes falling 2 to 4 (blue) and greater than 4 (orange) standard deviations below the main diagonal are highlighted.(B) z-scores from panel A and an independent pool of insertion mutants in BG14 were averaged and plotted against z-scores obtained similarly from a pool of insertion mutants in a pdr1∆ derivative of BG14.

FIG 4
FIG4 Acute inhibition of N-glycosylation and GPI-anchoring in the ER trigger calcineurin signaling which promotes cell survival.BG14 (black symbols), crz1∆ (white symbols), and cnb1∆ (gray symbols) cells were grown to log phase in SCD medium at 30°C and then exposed to manogepix (0.6 µg/mL), tunicamycin (20 µg/mL), or micafungin (0.12 µg/mL).At the indicated times, samples were removed and analyzed by qRT-PCR to measure RCN2 expression (A) and by propidium iodide staining to measure cell death in the population (B).Data points indicate the averages of three biological replicates (±SD).

FIG 6
FIG6 Calcineurin and Crz1 regulate the expression of FKS2, DFG5, and FLC2 but not other co-essential paralogs in response to cell wall stressor (micafungin).The expression of the indicated genes was monitored by qRT-PCR in wild-type BG14 cells and a crz1∆ derivative with and without exposure to micafungin for 80 min.Columns indicate averages of four biological replicates (±SD).

FIG 7
FIG 7 Genetic deficiencies in cell wall biogenesis increase calcineurin signaling.(A) Drop tests were performed as described in Fig. 2 using the knockout mutants and wild-type parent strains as indicated.(B) The expression of the calcineurin-and Crz1-dependent gene RCN2 was monitored by qRT-PCR in the indicated strains during exponential growth in SCD medium at 30°C with or without exposure to FK506.Columns indicate averages of four replicates (±SD).Statistical significance was assessed using a Welch's t-test with Bonferroni correction for multiple comparisons (*P < 0.05; **P < 0.01; ***P < 0.005; ****P < 0.001; ns, not significant).

FIG 8
FIG 8 Genetic interactions of selected double mutants of S. cerevisiae.Genetic interaction scores of double knockout mutants in S. cerevisiae were obtained from Costanzo et al. (46).Strong negative interactions are indicated in black, while insignificant interactions are labeled "ns." Knockouts of most paralogs on the left resulted in the activation of calcineurin and increased expression or function of paralogs on the right.The top three functional groups all involve anchoring and shedding of GPI-anchored cell wall proteins (cwp) to cell wall glucan.