The SET-domain protein CgSet4 negatively regulates antifungal drug resistance via the ergosterol biosynthesis transcriptional regulator CgUpc2a

Invasive fungal infections, which pose a serious threat to human health, are increasingly associated with a high mortality rate and elevated health care costs, owing to rising resistance to current antifungals and emergence of multidrug-resistant fungal species. Candida glabrata is the second to fourth common cause of Candida bloodstream infections. Its high propensity to acquire resistance toward two mainstream drugs, azoles (inhibit ergosterol biosynthesis) and echinocandins (target cell wall), in clinical settings, and its inherent low azole susceptibility render antifungal therapy unsuccessful in many cases. Here, we demonstrate a pivotal role for the SET {suppressor of variegation 3 to 9 [Su(var)3–9], enhancer of zeste [E(z)], and trithorax (Trx)} domain–containing protein, CgSet4, in azole and echinocandin resistance via negative regulation of multidrug transporter–encoding and ergosterol biosynthesis (ERG) genes through the master transcriptional factors CgPdr1 and CgUpc2A, respectively. RNA-Seq analysis revealed that C. glabrata responds to caspofungin (CSP; echinocandin antifungal) stress by downregulation and upregulation of ERG and cell wall organization genes, respectively. Although CgSet4 acts as a repressor of the ergosterol biosynthesis pathway via CgUPC2A transcriptional downregulation, the CSP-induced ERG gene repression is not dependent on CgSet4, as CgSet4 showed diminished abundance on the CgUPC2A promoter in CSP-treated cells. Furthermore, we show a role for the last three enzymes of the ergosterol biosynthesis pathway, CgErg3, CgErg5, and CgErg4, in antifungal susceptibility and virulence in C. glabrata. Altogether, our results unveil the link between ergosterol biosynthesis and echinocandin resistance and have implications for combination antifungal therapy.

Invasive fungal infections, which pose a serious threat to human health, are increasingly associated with a high mortality rate and elevated health care costs, owing to rising resistance to current antifungals and emergence of multidrugresistant fungal species. Candida glabrata is the second to fourth common cause of Candida bloodstream infections. Its high propensity to acquire resistance toward two mainstream drugs, azoles (inhibit ergosterol biosynthesis) and echinocandins (target cell wall), in clinical settings, and its inherent low azole susceptibility render antifungal therapy unsuccessful in many cases. Here, we demonstrate a pivotal role for the SET {suppressor of variegation 3 to 9 [Su(var) [3][4][5][6][7][8][9], enhancer of zeste [E(z)], and trithorax (Trx)} domain-containing protein, CgSet4, in azole and echinocandin resistance via negative regulation of multidrug transporter-encoding and ergosterol biosynthesis (ERG) genes through the master transcriptional factors CgPdr1 and CgUpc2A, respectively. RNA-Seq analysis revealed that C. glabrata responds to caspofungin (CSP; echinocandin antifungal) stress by downregulation and upregulation of ERG and cell wall organization genes, respectively. Although CgSet4 acts as a repressor of the ergosterol biosynthesis pathway via CgUPC2A transcriptional downregulation, the CSP-induced ERG gene repression is not dependent on CgSet4, as CgSet4 showed diminished abundance on the CgUPC2A promoter in CSP-treated cells. Furthermore, we show a role for the last three enzymes of the ergosterol biosynthesis pathway, CgErg3, CgErg5, and CgErg4, in antifungal susceptibility and virulence in C. glabrata. Altogether, our results unveil the link between ergosterol biosynthesis and echinocandin resistance and have implications for combination antifungal therapy.
C. glabrata is an asexual haploid budding yeast, which belongs to the Nakaseomyces clade (7,8). It shares a common ancestor with the non-pathogenic yeast Saccharomyces cerevisiae and is distinct in its virulence traits from other prevalent pathogenic species of Candida (7,8). C. glabrata lacks secreted proteolytic activity, does not form true hyphae, and displays elevated levels of stress resistance (7,8). Based on its position in the phylogenetic tree and close genetic relatedness with S. cerevisiae, pathogenesis mechanisms in C. glabrata are postulated to have arisen independently of other Candida pathogens (7,8). The ability to survive and replicate in human macrophages, suppress host innate immune response, form biofilms, adhere to host tissue, metabolic flexibility and high resistance to diverse stresses including antifungals, assist C. glabrata establish successful mucosal and invasive infections (1,7,8).
The incidence of C. glabrata infections has been on rise since last 2 decades, and alarmingly, this surge is also associated with an increased prevalence of co-resistance to two mainstream antifungal agents, azoles and echinocandins, thereby, hampering the success of antifungal therapy (3)(4)(5)(9)(10)(11)(12). Polyene, azole, and echinocandin drugs are largely being used in hospitals worldwide to treat bloodstream fungal infections (11)(12)(13). The target of azole antifungals, which are fungistatic, is an enzyme of the ergosterol biosynthesis pathway, lanosterol 14-alpha-demethylase enzyme, that is encoded by the ERG11 gene (13,14). Azole drug-induced growth inhibition is attributed to ergosterol depletion from the plasma membrane and accumulation of the toxic sterol intermediates (13,14). Contrary to azoles, the echinocandin antifungals target the β-glucan synthase enzyme, which is encoded by the family of FKS genes, and is involved in the synthesis of the cell wall structural polymer 1,3-β-D-glucan (11)(12)(13)15). The polyene class of antifungals bind to ergosterol and disrupt plasma membrane integrity as well as extract ergosterol from the cell membrane (12,13).
C. glabrata is inherently less susceptible to azole drugs, and C. glabrata infections are increasingly being associated with azole and/or echinocandin resistance (4,5,(9)(10)(11)(12). The elevated expression of ATP-binding cassette class of multidrug transporters is the primary mode of azole resistance in C. glabrata in hospitals worldwide (11,12,14). Genes coding for multidrug transporters are under the regulation of a Zn 2 Cys 6 binuclear zinc cluster domain-containing transcription factor, which is encoded by the CgPDR1 gene (14,16). A large number of single amino acid substitution mutations in CgPDR1 have been reported in azole-resistant clinical isolates of C. glabrata (14,17,18). CgCdr1 and CgCdr2 represent two major multidrug transporters in C. glabrata, whose elevated expression is associated with increased azole efflux, and the resultant decreased intracellular accumulation of azole drugs (14,19,20).
Another key regulator of azole antifungal response is the Zn 2 Cys 6 transcription factor CgUpc2A (Cagl0c01199p), which is required for basal and induced expression of the ergosterol biosynthesis (ERG) genes, with several ERG genes showing significant upregulation upon azole exposure (21)(22)(23). Intriguingly, CgUpc2A has recently been shown to bind to CgPDR1 and CgCDR1 promoter sequences, thereby linking ergosterol biosynthesis with the CgPdr1-mediated transcriptional regulatory network that controls expression of multidrug resistance (MDR) genes (24).
C. glabrata possesses three CgFKS genes, and point mutations in the hot spot regions of CgFKS1 and CgFKS2 genes largely account for echinocandin resistance in clinical isolates of C. glabrata (15,25). These hot spot regions represent echinocandin-binding regions, with mutations decreasing the binding affinity of echinocandins for the β-1,3-D-glucan synthase enzyme (12,15). Besides elevated CgFKS gene transcription, the echinocandin exposure also leads to increased expression of the chitin synthase genes, thereby resulting in compensatory activation of chitin synthesis in the cell wall (15,26). Mutations in the CgMSH2 gene, that codes for a component of the DNA mismatch repair pathway, have been associated with MDR in some isolates of C. glabrata (11,15).
However, the epigenetic regulation of MDR genes is largely unstudied in C. glabrata.
Here, we report that of six SET domain proteins (CgSet1-CgSet6) in C. glabrata, CgSet4 uniquely acts as a repressor of CgPdr1-dependent MDR and CgUpc2a-dependent ergosterol biosynthesis pathways. Besides showing that CgSET4 deletion results in decreased susceptibility to fluconazole (FLC) and caspofungin (CSP) drugs, we also report that ergosterol biosynthesis is downregulated in response to CSP stress. We further show that CgSet4-dependent negative regulation of CgPDR1 and CgERG genes is mediated through CgUpc2a. Finally, our animal infection studies reveal a role for the master transcriptional factor CgUpc2a, and two enzymes (CgErg3 and CgErg4) of the ergosterol biosynthesis pathway, as well as, for all CgSet proteins but for CgSet6, in survival of C. glabrata in the mouse systemic candidiasis model. Besides uncovering a link between ergosterol biosynthesis and echinocandin resistance, our findings yield key insights into the intertwined transcriptional networks that regulate cellular response to two seemingly distinct stresses, cell wall impairment, and ergosterol synthesis inhibition.

Results
The SET domain-containing protein CgSet4 negatively regulates azole and CSP resistance We have previously shown a role for putative histone chaperones CgFpr3 and CgFpr4, histone demethylase CgRph1, and histone H3K36-specific methyltransferase CgSet2 in regulating CgPDR1-dependent expression of multidrug transporter-encoding genes, and/or resistance to azole antifungals (28). Extending our results further, we, here, have examined the contribution of C. glabrata proteins, which possess the evolutionarily conserved SET domain, to resistance toward azole and echinocandin drugs. The SET domain consists of about 130 conserved amino acids (35). Several SET domain-containing proteins are known to methylate both histone and nonhistone proteins, with ε-amino group of the lysine residue in histones being able to undergo monomethylation, dimethylation, and trimethylation (35)(36)(37). Through in silico analysis, we first identified six genes, CgSET1-SET6, in the C. glabrata genome, which code for proteins containing the SET domain. The key features of these six SET domain-containing proteins CgSet1-CgSet6 and functions of their S. cerevisiae orthologs are listed in Table S1. The amino acid sequence alignment of the SET domain, and the predicted domain organization of CgSet1-Set6 proteins are shown in Fig. S1, A and B, respectively. Notably, CgSet1 and CgSet2 enzymes have recently been shown to be required for monomethylation, dimethylation, and trimethlyation of lysine-4, and trimethylation of lysine-36 residues, in histone H3, respectively (28,38), while this work was underway.
Next, we generated and characterized deletion strains for CgSET1, and CgSET3-6 genes, as we had recently shown that the loss of CgSET2 gene resulted in decreased sensitivity to the azole drug FLC (28). In consistence with the recent studies (28, 38), phenotypic analysis showed increased sensitivity and a moderate level of resistance toward FLC in Cgset1Δ and Cgset2Δ mutants, respectively, compared with wild-type (wt) cells (Fig. 1A). Contrarily, the Cgset4Δ mutant exhibited a high level of FLC resistance (Fig. 1A). In addition, an increased susceptibility of Cgset1Δ and Cgset2Δ mutants toward hydrogen peroxide (oxidative stressor) and elevated susceptibility of Cgset1Δ, Cgset2Δ, and Cgset3Δ mutants toward methyl methanesulfonate (DNA-damaging agent) were observed (Fig. 1B). Interestingly, Cgset1Δ, Cgset2Δ, Cgset3Δ, and Cgset6Δ mutants also displayed mild sensitivity to calcofluor white and congo red (cell wall stressors), compared with wt cells (Fig. 1C). No other notable phenotype was found for mutants lacking CgSet1-6 proteins. Besides implicating CgSet1-3 and CgSet6 proteins in survival of different stresses, these results indicate that CgSet1, and CgSet2 and CgSet4, modulate azole resistance albeit via opposite mode.
Furthermore, since we had previously reported that Tn7 insertion in the CgSET4 gene is associated with resistance to both FLC and CSP (echinocandin drug) (27), we next checked the role of SET domain-containing proteins in CSP tolerance in C. glabrata. We found the CgSET4 gene loss to be uniquely associated with CSP resistance, whereas Cgset1Δ, Cgset2Δ, Cgset3Δ, and Cgset6Δ mutants exhibited varied levels of increased susceptibility to CSP, compared with wt cells (Fig. 1D). The Cgset5Δ mutant exhibited wt-like growth in the presence of CSP (Fig. 1D). Besides showing differential requirement of CgSet1-CgSet6 proteins in antifungal tolerance, these data collectively implicate CgSet4 in resistance to both cell membrane-(azole) and cell wall-targeting (echinocandin) antifungal drugs.
To ascribe histone H3 lysine methyltransferase activity to SET domain-containing proteins in C. glabrata, we performed Western analysis and examined trimethylation of lysine 4 and lysine 36 residues in histone H3 in Cgset1Δ-Cgset6Δ mutants. As reported recently (28,38), the Cgset1Δ mutant exhibited no H3K4me3, whereas the Cgset2Δ mutant displayed very low levels of H3K36me3 (Fig. 1E). Notably, CgSET3-SET6 gene loss neither had any effect on trimethylation of the H3K4 residue nor on the H3K36 residue (Fig. 1E). Our attempts to determine methylation at other lysine residues in H3 were not successful because of nonavailability and/or nonspecificity of commercially available antibodies. However, these data for CgSet3-CgSet6 proteins are in agreement with known functions of their counterparts in S. cerevisiae, as ScSet3, ScSet4, and ScSet6 proteins have not yet been associated with histone lysine methyltransferase activity, whereas ScSet5 has been shown to monomethylate lysines 5, 8, and 12 in histone H4 (30,31,39,40).
Since we had recently reported a requirement for CgSet2 for survival in mice (28), we next assessed the survival of Cgset1Δ and Cgset3Δ-Cgset6Δ mutants in the murine model of systemic candidiasis. We found CgSet1 and CgSet3 to be required for survival of C. glabrata in kidneys, liver, and spleen, as 5-to 20-fold lower yeast colony-forming units were recovered from organs of the mice infected with Cgset1Δ and Cgset3Δ mutants, compared with the wt-infected mice (Fig. 1F). Contrarily, the Cgset4Δ-infected mice displayed about three-fold lower fungal burden in liver, spleen, and brain, as compared with the wtinfected mice (Fig. 1F). Of note, the Cgset1Δ-infected mice exhibited four-fold lower fungal burden in brain, whereas the Cgset5Δ-infected mice had eight-fold and three-fold lower fungal burden in kidneys and brain, respectively, of infected mice, as compared with the wt-infected mice (Fig. 1F). These data suggest that CgSet6 is not required for survival in the mouse systemic candidiasis model, whereas CgSet1, CgSet3, CgSet4, and CgSet5 are required in an organ-dependent manner, thereby underscoring the role of SET domaincontaining proteins in virulence of C. glabrata.
CgSET4 expression is downregulated in response to antifungal exposure The high level of FLC and CSP resistance in the Cgset4Δ mutant prompted us to characterize this mutant further. Therefore, we first verified, through mutant complementation analysis, that the FLC and CSP resistance in the mutant is due to the lack of CgSET4 gene. For this, we first cloned CgSET4 under its own promoter and expressed it ectopically in the Cgset4Δ mutant. Next, we performed serial dilution spotting assay and found the reversal of FLC ( Fig. 2A) and CSP (Fig. 2B) resistance in the Cgset4Δ mutant, as the Cgset4Δ/CgSET4 strain displayed wt-like susceptibility to both antifungals (Fig. 2, A and B). Consistent with the Cgset4Δ mutant phenotype, overexpression of the CgSET4 gene led to elevated sensitivity to FLC ( Fig. 2A) and CSP (Fig. 2B), suggesting that CgSet4 negatively regulates resistance toward FLC and CSP drugs in C. glabrata.
CgSet4 is an uncharacterized 350 amino acid protein that showed 40% identity with its S. cerevisiae ortholog and contains a 132 amino acid (146-277 amino acids) SET domain (Fig. S1B). In S. cerevisiae, Set4 has a paralog, Set3; however, neither Set3 nor Set4 has been reported to possess lysine methyltransferase activity (31,39). Similarly, we found normal H3K4me3 and H3K36me3 levels in the Cgset4Δ mutant (Fig. 1E). The amino acid sequence similarity and the lack of histone H3K4 and H3K36 methyltransferase activity in both CgSet4 and ScSet4 proteins prompted us to examine if ScSET4 expression could reverse antifungal resistance in the Cgset4Δ mutant. For this, we ectopically expressed the S. cerevisiae SET4 from the endogenous CgSET4 promoter in the Cgset4Δ mutant and compared growth profiles of the Cgset4Δ mutant expressing either ScSET4 or CgSET4, in the presence of antifungals. We found that unlike CgSET4, ectopic expression of the S. cerevisiae SET4 could not rescue FLC and CSP resistance phenotype of the Cgset4Δ mutant (Fig. 2, A and B), indicating functional differences between CgSet4 and ScSet4 proteins in modulation of antifungal resistance. Of note, while the SET4 gene loss led to azole resistance, it had no effect on CSP susceptibility in S. cerevisiae (33).
To examine the role of CgSet4 in other clinical strains of C. glabrata, we deleted CgSET4 gene in the reference strain CBS138 and two Indian clinical isolates YRK2289 and YRK2291 (27). The CgSET4 loss resulted in decreased susceptibility to FLC (Fig. S2A) and CSP (Fig. S2B) in all three Overnight-grown C. glabrata cultures were 10-fold serially diluted, and 3 μl of each dilution was spotted on the CAA medium lacking (CAA) or containing 16 μg/ml FLC (FLC-16), 32 μg/ml FLC (FLC-32), or 64 μg/ml FLC (FLC-64). Images were captured after 2 days of growth at 30 C. "wt" denotes the wild-type (wt) strain. B, serial dilution cell spotting analysis illustrating oxidative stress and DNA damage stress susceptibility of indicated C. glabrata strains. Hydrogen peroxide (H 2 O 2 ) and methyl methane sulfonate (MMS) were used at a concentration of 35 mM and 0.04%, respectively, in YPD medium. C, serial dilution cell spotting analysis illustrating cell wall stress susceptibility of indicated C. glabrata strains. Both calcofluor white (CFW) and Congo Red were used at a concentration of 2 mg/ml in CAA medium. D, liquid medium-based growth analysis illustrating caspofungin (CSP) susceptibility of indicated C. glabrata strains. C. glabrata strains were cultured at 30 ο C in CAA medium lacking (CAA) or containing 75 ng/ml CSP (CSP-75) or 150 ng/ml CSP (CSP-150) for 16 h. After incubation, cultures were diluted in PBS, and 3 μl of 5-, 25-, 125-, and 625fold diluted cultures were spotted on CAA medium, and growth was recorded after 1 day of growth at 30 C. E, representative immunoblots showing trimethylation on 4th and 36th lysine residues in histone H3 in indicated C. glabrata strains. Log-phase wt and CgsetΔ cells were collected, washed with PBS, strain backgrounds. Furthermore, CgSET4 overexpression led to FLC (Fig. S2A) and CSP (Fig. S2B) sensitivity in CBS138 and Indian clinical strains, albeit to different extent. These results suggest that CgSet4-dependent control of antifungal resistance is common among clinical strains of C. glabrata.
Next, to delineate the role of CgSet4 in antifungal resistance, we decided to first check its cellular localization. For this, we generated CgSet4-GFP fusion protein and found it to be functional, as it complemented FLC resistance phenotype of the Cgset4Δ mutant (Fig. S3A). Next, the confocal imaging analysis revealed that CgSet4 is primarily localized to the nucleus under all (regular, FLC-treated, and CSP-treated) conditions (Fig. 2C), indicating a predominant nuclear role for CgSet4 in C. glabrata. Furthermore, CgSet4-GFP was found to be highly enriched in the insoluble chromatin fraction, whereas being totally absent in the soluble cytosolic fraction (Fig. S3B), indicating that CgSet4 is a chromatin-associated protein.
Since CgSET4 gene loss was associated with drug resistance, we next checked the effect of antifungal treatment on CgSET4 transcript and protein levels. For this, we compared CgSET4 transcript and protein levels, via quantitative RT-PCR (qRT-PCR) and Western analysis, respectively, between untreated and FLC-or CSP-treated wt cells (Fig. 2D). We found that CgSET4 transcript (Fig. 2D) and CgSet4-GFP (58 kDa band) protein ( Fig. 2E) levels were about two-to five-fold lower in FLC-and CSP-treated wt cells, compared with untreated wt cells, indicating that C. glabrata responds to azole and echinocandin exposure by downregulating CgSET4 expression. Of note, this antifungal-induced CgSET4 repression is in accordance with the FLC and CSP resistance observed in the Cgset4Δ mutant (Fig. 1, A and D). Collectively, our data suggest that CgSet4 is a nuclear protein that may be a pivotal component of the azole and echinocandin antifungal resistance regulatory system in C. glabrata.

CgSet4 negatively regulates the CgPDR1 regulon
Since the azole resistance in C. glabrata is primarily associated with overexpression of the multidrug transporterencoding genes CgCDR1 and CgCDR2, owing to increased protein or activity levels of CgPdr1 (14), we next checked the transcript levels of CgPDR1, CgCDR1, and CgCDR2 genes in Cgset4Δ mutant. We found that 1.8-, 2.7-, and 2.7-fold increased expression of CgPDR1, CgCDR1, and CgCDR2 genes, respectively, in the Cgset4Δ mutant (Fig. 3A). Furthermore, while FLC exposure resulted in elevated CgPDR1, CgCDR1, and CgCDR2 transcript levels in wt cells, no such increase was observed in FLC-treated Cgset4Δ cells, compared with untreated Cgset4Δ cells (Fig. 3A). These data suggest that FLC resistance in the Cgset4Δ mutant could be due to high basal level expression of CgPDR1 regulon genes.
To verify this notion, we deleted CgCDR1 gene from the genome of both wt and Cgset4Δ strains. We found that while CgCDR1 gene loss rendered wt cells highly susceptible to FLC, CgCDR1 gene loss in the Cgset4Δ mutant reversed the azole resistance phenotype of the mutant, as the double mutant Cgset4Δcdr1Δ exhibited wt-like sensitivity to FLC (Fig. 3B). Importantly, expression of CgSET4 and CgCDR1 genes in the Cgset4Δcdr1Δ mutant led to elevated and diminished FLC susceptibility, similar to Cgcdr1Δ and Cgset4Δ mutant, respectively (Fig. 3B). Of note, the wt-like and not the Cgcdr1Δ-like FLC susceptibility of the double mutant Cgset4Δcdr1Δ raises the possibility that genes other than CgCDR1 may also contribute to FLC resistance in the Cgset4Δ mutant. Despite multiple attempts, we could not generate the double deletion strain Cgset4Δpdr1Δ lacking both CgSET4 and CgPDR1 genes, the reason underlying this is yet to be determined. Collectively, besides showing that FLC resistance in the Cgset4Δ mutant is predominantly dependent upon CgCdr1, these results implicate CgSet4 in the transcriptional repression of CgPDR1 in C. glabrata.

Cell wall composition is not altered in the Cgset4Δ mutant
After elucidating the molecular basis of azole resistance, we next focused on the echinocandin resistance in the Cgset4Δ mutant. CSP inhibits synthesis of the cell wall component, β-1,3 D-glucan, and a compensatory increase in cell wall chitin levels has been reported upon CSP exposure (15,26,41). Therefore, we next checked if CSP resistance in Cgset4Δ mutant could be due to an altered cell wall composition in the mutant. For this, we determined levels of all three components of the cell wall, viz., β-glucan, mannan, and chitin, via aniline blue, concanavalin A, and calcofluor white-based cell staining assays, respectively. Fluorescence-activated cell sorting analysis showed similar mannan and chitin content in cell walls of the wt and Cgset4Δ strains, and a very mild increase in β-glucan levels in the mutant cell wall, as compared with the wt cell wall (Fig. S4), thereby ruling out any significant contribution of the cell wall components to CSP resistance in the Cgset4Δ mutant.

Ergosterol biosynthesis gene expression is downregulated in response to CSP exposure
CgSet4 was found to be associated with chromatin ( Fig. S3B). Therefore, to uncover the molecular basis of CSP resistance in the Cgset4Δ mutant, we next profiled the global transcriptomes, using RNA-Seq approach, of logarithmic phase wt and Cgset4Δ cells in the presence and absence of CSP. For this, C. glabrata cells were treated with CSP at a sublethal concentration (0.25 μg/ml) for 1 h, and both wt and Cgset4Δ cells were found to retain viability during this treatment period (Fig. S5). and lysed with glass beads. Whole-cell lysates (50 μg protein) were resolved on 15% SDS-PAGE and probed with anti-H3, anti-H3K4me3, anti-H3K36me3, and anti-GAPDH antibodies. Bands representing histone H3 and CgGapdh proteins corresponded to about 17 and 36 kDa, respectively. CgGapdh was used as a loading control. F, organ fungal load in 6-to 8-week-old female BALB/c mice was determined 7 days postintravenous infection with indicated C. glabrata strains (4 × 10 7 cells). Diamonds and bars denote CFUs recovered from target organs of the individual mouse, and the CFU geometric mean (n = 7-10), respectively, for each organ. *p < 0.05; **p < 0.01; ***p < 0.001; Mann-Whitney test. CAA, casamino acid; CFU, colony-forming unit; SET, {suppressor of variegation 3 to 9 [Su(var)3-9], enhancer of zeste [E(z)], and trithorax (Trx)}; YPD, yeast extract-peptone-dextrose. , and 64 μg/ml (FLC-64) in CAA medium. "V," CgSET4, and ScSET4 denote empty vector, C. glabrata SET4, and Saccharomyces cerevisiae SET4 gene, respectively. B, liquid medium-based growth analysis illustrating caspofungin (CSP) susceptibility of indicated C. glabrata strains. CSP was used at a concentration of 75 ng/ml (CSP-75) and 150 ng/ml (CSP-150) in CAA medium. C, representative confocal microscopy images of CAA medium-grown logarithmic-phase wt cells expressing CgSet4-GFP ectopically showing colocalization of CgSet4-GFP with the Hoechst 33258-stained nuclei. C. glabrata strains were grown in CAA medium lacking (CAA) or containing 64 μg/ml FLC or 150 ng/ml CSP for 1 h and stained with 1 μg/ml Hoechst 33258 for 15 min at 37 ο C. Cells were washed, suspended in PBS, and visualized using the Confocal microscope (Zeiss LSM 700 equipped with 63×/1.44 numerical aperture objective). DIC, differential interference contrast; Hoechst 33258, DNA-binding stain. Bar represents 2.0 μm. D, qRT-PCR analysis showing transcriptional downregulation of the CgSET4 gene in response to FLC and CSP exposure. Log-phase wt cultures were left untreated or treated either with 64 μg/ml FLC or 250 ng/ml CSP for 1 h. RNA was isolated using the acid-phenol extraction method, and CgSET4 transcript levels were measured by qRT-PCR using the 2 −ΔΔCt method. Data (mean ± SEM, n = 3) were normalized against the CgTDH3 mRNA as control and represent fold change in CgSET4 gene expression in FLC-or CSP-treated wt cells, as compared with the CAA medium-grown wt cells (taken as 1.0). *p < 0.05; ***p < 0.001, paired two-tailed Student's t test. E, representative immunoblots showing RNA-Seq analysis revealed 1077 genes to be differentially expressed (≥1.5-fold change in expression, and a q value of ≤0.05) in wt cells upon exposure to CSP ( Fig. 4A and Table S2). Among these differentially expressed genes (DEGs), 475 and 602 genes were upregulated and downregulated, respectively ( Fig. 4A and Table S2). Gene Ontology enrichment analysis for biological process by the DAVID (Database for Annotation, Visualization, and Integrated Discovery; https://david.ncifcrf.gov/) tool revealed upregulation of genes involved in fungal-type cell wall organization, response to oxidative stress, fatty acid (FA) beta-oxidation, and the catabolism of 5-carbon sugars, xylose and arabinose ( Fig. 4B and Table S3A). Importantly, genes belonging to ergosterol biosynthesis, amino acid transport, zinc ion homeostasis, and amino acid biosynthetic process were found to be downregulated in CSP-treated wt cells, compared with untreated wt cells ( Fig. 4B and Table S3B). Consistent with our qRT-PCR data (Fig. 2D), the CgSET4 gene was downregulated in response to CSP exposure (Table S3B). Of note, CSP treatment also led to the repression of master regulators of the ergosterol biosynthesis genes, CAGL0C01199g (CgUPC2A) and CAGL0F07865g (CgUPC2B), and the plasma membrane sterol transporter gene (CgAUS1), and upregulation of the negative regulator of ERG gene biosynthesis CAGL0D05434g (CgROX1), and glycosylphosphatidylinositol-anchored aspartyl protease genes (CgYPS1, CgYPS7, CgYPS6, and CgYPS10) (Table S2). These proteases have previously been implicated in maintenance of the cell wall integrity, with CgYps1 also contributing to CSP tolerance (42,43), whereas CgRox1, a heme-dependent repressor of the hypoxic genes, has recently been implicated in negative regulation of the CgERG genes (44).
Overall, our transcriptome data show an overlap with the previously reported transcriptomes of CSP-treated C. albicans and S. cerevisiae cells (41,45,46). For example, genes related to cell wall biogenesis and stress response, viz., BAG7, CHS1, reduced CgSet4-GFP levels upon FLC and CSP exposure. Log-phase wt cultures were left untreated or treated either with 64 μg/ml FLC or 250 ng/ml CSP for 1 h. Cultures were spun down, cells were collected, and whole-cell extracts were prepared by glass bead lysis method. Samples containing 60 μg protein were resolved on 12% SDS-PAGE and probed with anti-GFP and anti-GAPDH antibodies. CgGapdh was used as a loading control. The intensity of individual bands in three independent Western blots was quantified using the ImageJ densitometry software, and the CgSet4-GFP signal was normalized to the corresponding CgGapdh signal. Data (mean ± SEM, n = 3) represent fold change in CgSet4-GFP levels in FLC or CSP-treated wt cells, compared with CAA medium-grown wt cells (considered as 1.0) and are plotted as a bar graph underneath the blot images. *p < 0.05; **p < 0.01, paired two-tailed Student's t test. CAA, casamino acid; qRT-PCR, quantitative RT-PCR. CWP1, FKS2, RLM1, and SLT2, were upregulated, and genes involved in amino acid biosynthesis, viz., ARG3, ILV6, and LEU1, were downregulated in both C. glabrata and S. cerevisiae upon CSP exposure (Table S2 and (41, 45)). Similarly, genes involved in ergosterol biosynthesis, ERG3, ERG4, ERG25, and ERG26, and high-affinity iron transport, FET3, FTR1, and FTH1, were downregulated in both C. glabrata and C. albicans (Table S2 and (46). Furthermore, despite CgSet4 being a chromatin-associated protein, its loss did not have a big impact on the transcriptome of C. glabrata, and only 48 genes were found to be differentially expressed in the Cgset4Δ mutant, compared with strains. D-F, qRT-PCR analysis validating the RNA-Seq data. Log-phase wt and Cgset4Δ cells were left untreated or treated with 250 ng/ml CSP in CAA medium for 1 h, and transcript levels of indicated genes were determined by qRT-PCR. Data (mean ± SEM, n = 3-4) were normalized against the CgTDH3 mRNA control, and represent fold change in expression of indicated genes in the CAA medium-grown Cgset4Δ mutant as compared with the CAA mediumgrown wt strain (taken as 1.0) (D), CSP-treated wt as compared with CAA medium-grown wt cells (taken as 1.0) (E), and CSP-treated Cgset4Δ as compared with CAA medium-grown Cgset4Δ cells (taken as 1.0) (F). *p < 0.05; ***p < 0.001, paired two-tailed Student's t test. CAA, casamino acid; DAVID, Database for Annotation, Visualization, and Integrated Discovery; qRT-PCR, quantitative RT-PCR.
wt cells (Fig. 4C and Table S4). Of 48 genes, 24 were upregulated and 24 were downregulated in the Cgset4Δ mutant (Table S4). Of these gene sets, 7 and 10 genes were upregulated and downregulated, respectively, upon both deletion of CgSET4 gene and CSP treatment of wt cells (Tables S2 and S4). Furthermore, three genes were upregulated upon CSP treatment of wt cells, whereas these were downregulated upon CgSET4 deletion (Tables S2 and S4). These genes included CgARO10, CgHBT1, and CgHOR7 (Tables S2 and S4). Similarly, 12 genes were downregulated in response to CSP treatment of wt cells, whereas these were upregulated upon CgSET4 deletion, which also included four ergosterol biosynthesis genes, CgERG2, CgERG3, CgERG5, and CgERG11 (Tables S2  and S4). DAVID analysis revealed genes involved in ergosterol biosynthesis, sterol import, and carbohydrate transport to be upregulated (Table S5A), whereas genes involved in the inositol biosynthetic process were found to be downregulated, upon CgSET4 deletion in C. glabrata (Table S5B). Of note, CgERG2, CgERG3, CgERG5, CgERG11, CgAUS1, and CgROX1 genes were upregulated in Cgset4Δ cells, compared with wt cells (Table S4).
Notably, CSP exposure led to differential regulation of 1159 genes in the Cgset4Δ mutant, compared with untreated Cgset4Δ cells, with 568 and 591 displaying upregulation and downregulation, respectively ( Fig. 4A and Table S6). Of these, 414 upregulated genes and 483 downregulated genes were common between CSP-treated wt and Cgset4Δ cells ( Fig. 4A and Table S6). Gene Ontology analysis revealed a substantial overlap (898 genes) between CSP-treated wt and Cgset4Δ cells (Fig. 4C, Tables S3 and S7), with fungal-type cell wall organization, FA beta-oxidation, endocytosis, and protein unfolding genes displaying upregulation in CSP-treated Cgset4Δ cells ( Fig. 4B and Table S7A), and amino acid transport, zinc ion homeostasis, ribosomal subunit export from nucleus, and cellular amino acid biosynthetic process genes showing downregulation in CSP-treated Cgset4Δ cells ( Fig. 4B and Table S7B). Of note, while CSP exposure led to wholesale downregulation of 14 ergosterol biosynthesis genes in wt cells (Table S2), its negative regulatory effect in Cgset4Δ cells was seen for five ergosterol biosynthesis genes, CgERG3, CgERG11, CgERG26, CgERG28, and CgHMG1, which were downregulated in the mutant (Table S4). Expression of the ergosterol biosynthesis pathway genes in untreated-and CSPtreated wt and Cgset4Δ strains is schematically presented in Fig. S6. Furthermore, similar to wt cells, CSP treatment led to upregulation of three glycosylphosphatidylinositol-anchored aspartyl protease genes, CgYPS1, CgYPS7, and CgYPS10, in Cgset4Δ cells (Table S6). Altogether, the transcriptional profiling data suggest that C. glabrata responds to echinocandin stress by transcriptional reprogramming of the cell wall organization, ergosterol biosynthesis, sugar metabolism, amino acid biosynthesis and transport, zinc ion homeostasis, and oxidative stress response genes.
Next, we verified the RNA-Seq data for many genes by qRT-PCR analysis and found good agreement between the RNA-Seq and qRT-PCR datasets. First, we showed that CgSET4 disruption resulted in transcriptional activation of CgYHB1 and CgTIR1 genes and repression of CgHOR7 and CgVID24 genes (Fig. 4D and Table S4). Next, we showed that CgSLT2 and CgFKS2 genes, and CgFET3 and CgTIR1 genes, were upregulated and downregulated, respectively, in CSPtreated wt (Fig. 4E) and Cgset4Δ (Fig. 4F) cells. Of note, while CgSlt2 is the terminal mitogen-activated protein kinase of the cell wall integrity pathway, and has previously been implicated in CSP tolerance (47,48), CgFks2 is the 1,3-β-Dglucan synthase enzyme, which is required for β-glucan synthesis, and mutations in the CgFKS2 gene have been associated with CSP resistance in C. glabrata (15,25).
Collectively, our transcriptome data suggest that expression of the ergosterol biosynthesis genes is altered in response to both CgSET4 deletion and CSP exposure, underscoring a hitherto unknown effect of CSP treatment on ergosterol biosynthesis in C. glabrata. Of note, the CSP-induced transcriptional downregulation of the CgSET4 gene in our RNA-Seq data is consistent with the CSP resistance phenotype of the Cgset4Δ mutant.

CgSet4 negatively regulates basal ERG gene expression
Ergosterol is a major sterol in the plasma membrane and represents an integral cell membrane component, whose synthesis in the cell is regulated by oxygen and iron abundance (49,50). Notably, ergosterol biosynthesis genes were upregulated in Cgset4Δ mutant (Table S4), with the mutant also exhibiting high level of resistance to both azole and echinocandin drugs (Fig. 1, A and B). In addition, CSP exposure resulted in downregulation of the ERG genes in wt cells (Table S2). Therefore, we hypothesized that besides impacting FLC resistance, ergosterol levels may also control resistance to the cell wall-targeting drug, CSP, in the Cgset4Δ mutant.
To test this hypothesis, we performed four experiments. First, we examined the expression of different ERG genes in the Cgset4Δ mutant. Ergosterol biosynthesis is a multistep process that consists of three modules (Fig. S7) (50). The mevalonate biosynthesis module takes place in the vacuole and the mitochondria, the farnesyl pyrophosphate biosynthesis pathway occurs in the vacuole, and the late stages of the pathway predominantly occur in the endoplasmic reticulum membrane (50,51). The late pathway starts and ends with the formation of squalene and ergosterol, respectively (50,51). Since late steps of the ergosterol biosynthesis pathway have been implicated in stress tolerance (50-52), we checked expression of the seven ERG genes, CgERG1, CgERG3, CgERG4, CgERG5, CgERG6, CgERG11, and CgERG25, which code for enzymes carrying out various reactions in the ergosterol biosynthesis pathway (Fig. S7). qRT-PCR analysis revealed about two-to five-fold higher expression of CgERG1, CgERG3, CgERG4, CgERG5, CgERG6, CgERG11, and CgERG25 genes in the Cgset4Δ mutant, compared with wt cells (Fig. 5A), suggesting that CgSET4 is required for repression of the late ERG genes under regular growth conditions. Of note, CgERG3, CgERG5, and CgERG11 transcript levels were also higher in Cgset4Δ mutant, compared with wt cells, in our RNA-Seq analysis (Table S4 and Fig. S6).
Second, we measured total ergosterol levels in the Cgset4Δ mutant. Consistent with increased ERG gene expression in the Cgset4Δ mutant, we found two-fold higher ergosterol content in the mutant, compared with wt cells (Fig. 5B). Expression of the CgSET4 gene led to wt-like ergosterol levels in the Cgset4Δ mutant (Fig. 5B), indicating that increased ergosterol content in the Cgset4Δ mutant was due to the lack of the CgSET4 gene. Collectively, these results suggest that CgSet4 inhibits the Data (mean ± SEM, n = 3-4) were normalized against the CgACT1 mRNA control, and represent fold change in gene expression in the CAA medium-grown Cgset4Δ mutant as compared with the CAA medium-grown wt strain (taken as 1.0). *p < 0.05; **p < 0.01, paired two-tailed Student's t test. B, UV spectrophotometry-based measurement of ergosterol content of indicated log-phase Candida glabrata strains. Cultures were grown in CAA medium to log phase and collected. After PBS washes, sterols were extracted using n-heptane, and spectral profiles were recorded between 200 and 350 nm. The typical four-peak curve, indicating the presence of ergosterol and dehydroergosterol (DHE) in samples, was used to measure ergosterol content. Data (mean ± SEM, n = 3) indicate the percentage of ergosterol per gram of the wet weight of the cell pellet. C and D, qRT-PCR analysis showing expression of indicated ERG genes in caspofungin (CSP)-treated wt (C) and CSP-treated Cgset4Δ cells (D). Log-phase wt and Cgset4Δ cells were left untreated or treated with 250 ng/ml CSP in CAA medium for 1 h, and transcript levels of indicated genes were determined by qRT-PCR. Data (mean ± SEM, n = 3-4) were normalized against the CgACT1 mRNA control, and represent fold change in gene expression in CSP-treated wt cells as compared with CAA medium-grown wt cells (taken as 1.0) (C), and CSP-treated Cgset4Δ cells as compared with CAA medium-grown Cgset4Δ cells (taken as 1.0) (D). *p < 0.05; **p < 0.01; ***p < 0.001, paired two-tailed Student's t test. E, ergosterol content in log-phase cells of indicated C. glabrata strains, which were grown in the CAA medium lacking (CAA) or containing 250 ng/ml CSP for 1 h, was measured. Data (mean ± SEM, n = 3) represent fold change in ergosterol levels in indicated strains as compared with CAA medium-grown wt cells (considered as 1.0). *p < 0.05, paired two-tailed Student's t test. CAA, casamino acid; qRT-PCR, quantitative RT-PCR. ergosterol biosynthesis process via downregulation of the CgERG genes in C. glabrata.
Third, we verified the effect of CgSET4 deletion on CSPmediated downregulation of the ergosterol biosynthesis pathway, as RNA-Seq analysis had revealed ERG genes to be repressed in CSP-treated wt cells (Table S2 and Fig. S6). We observed that CSP treatment led to transcriptional repression of CgERG3, CgERG4, CgERG5, CgERG6, CgERG11, and CgERG25 genes in both wt (Fig. 5C) and Cgset4Δ (Fig. 5D) strains. Importantly, however, CgERG1 was found to be downregulated and upregulated in CSP-treated wt and Cgset4Δ cells, respectively (Fig. 5, C and D), indicating CgSET4-dependent transcriptional downregulation of the CgERG1 gene in response to CSP exposure. Of note, CgERG1 transcript levels were found to be increased upon CgSET4 deletion (Fig. 5A), suggesting a negative regulatory role for CgSet4 in CgERG1 expression. The molecular basis underlying CgSet4-dependent CSP-induced upregulation of CgERG1, which is the converse of that of the other CgERG genes, is yet to be determined. Altogether, our results suggest that CgSet4 is indispensable for the regulation of basal expression of ERG genes, but it has no prominent role in global downregulation of the ergosterol biosynthesis pathway under CSP-treated conditions.
Fourth, we measured ergosterol content in CSP-treated wt and Cgset4Δ cells. Contrary to expectations from the transcript profiling data, the CSP exposure did not result in a significant decrease in total ergosterol levels in wt cells (Fig. 5E). Instead, the ergosterol content was similar between untreated and CSP-treated wt cells (Fig. 5E). Similarly, ergosterol levels were similar between untreated and CSP-treated Cgset4Δ cells, with CgSET4 loss resulting in 2.0-fold to 2.5-fold higher ergosterol in both the presence and the absence of CSP (Fig. 5E). These results suggest a complex multifaceted regulation of the ergosterol biosynthesis process in response to the antifungal CSP, wherein the decreased transcript levels of ERG genes do not translate into lower cellular ergosterol levels. Alternatively, it is possible that while transcriptional downregulation of the ERG genes is observed within 1 h CSP treatment, changes in the cellular ergosterol levels may become conspicuous after prolonged CSP treatment when cells had undergone a few divisions. This possibility warrants further investigations. In this context, it is noteworthy that deletion of the CgUPC2A gene, which codes for a major transcriptional activator of ERG genes, was found to have no effect on the ergosterol content (44).

Deletion of CgERG3 and CgERG4 genes reverses FLC resistance in the Cgset4Δ mutant
The late ergosterol biosynthesis pathway enzymes have previously been implicated in stress tolerance in S. cerevisiae (50)(51)(52). CgERG3, CgERG5, and CgERG4 genes code for C-5 sterol desaturase, C-22 sterol desaturase, and C-24(28) sterol reductase enzymes, respectively, which catalyze last three steps of the ergosterol biosynthesis pathway (Fig. S7) (50). CgErg3, CgErg5, and CgErg4 produce ergosta-5,7,24(28)-trienol, ergosta-5,7,22,24(28)-trienol, and ergosterol, respectively (Fig. S7) (50,51). Of note, the fungal Erg4 and Erg5 enzymes are not conserved in mammals (50), making them good antifungal targets. Importantly, mutations in the CgERG3 gene have recently been found in both FLC and anidulafungin (an echinocandin drug)-resistant isolates of C. glabrata (53). Since the Cgset4Δ mutant displayed resistance to FLC as well as to CSP (Fig. 1, A and D), we next undertook a genetic approach to decipher CSP-dependent modulation of the ergosterol biosynthesis pathway, which may have an impact on azole response. For this, we generated double deletion strains, which lacked both CgSET4, and CgERG3, CgERG5, or CgERG4 gene. As a control, the single deletion strains for CgERG3, CgERG4, and CgERG5 genes were also created. Growth analysis in the presence of FLC revealed that CgErg3 and CgErg4 are required for FLC tolerance, as Cgerg3Δ and Cgerg4Δ mutants displayed attenuated growth on FLC-supplemented medium (Fig. 6A). Of note, CgERG3 deletion has previously been associated with increased azole susceptibility (54). In contrast, CgERG5 gene loss had no effect on FLC susceptibility of C. glabrata (Fig. 6A). Ectopic expression of CgERG3 and CgERG4 genes complemented the elevated FLC susceptibility of Cgerg3Δ and Cgerg4Δ mutants, respectively (Fig. 6A), suggesting that the azole sensitivity is due to the lack of the corresponding CgERG gene. Furthermore, the deletion of CgERG3 and CgERG4 genes in the Cgset4Δ mutant background reversed the FLC resistance phenotype of the Cgset4Δ mutant (Fig. 6A), raising the possibility of an essential requirement for these genes as the downstream target/effector of CgSet4-dependent FLC response. Notably, similar to the Cgset4Δ mutant, the Cgset4Δerg5Δ double mutant was found to be FLC resistant (Fig. 6A), indicating a dispensable role for CgErg5 in CgSet4dependent modulation of the azole response. Moreover, since the Cgerg5Δ mutant did not exhibit elevated FLC susceptibility, it is possible that CgErg5 functions in ergosterol biosynthesis can either be bypassed and/or performed by another enzyme of the pathway. Alternatively, the differential FLC susceptibility of Cgerg3Δ, Cgerg4Δ, and Cgerg5Δ mutants could be due to impairment in both ergosterol biosynthesis and functioning of multidrug efflux pumps. In this context, it is noteworthy that the activity of multidrug transporters is known to be affected by the membrane lipid environment, and the plasma membranes of S. cerevisiae erg3Δ, erg4Δ, and erg5Δ mutants have been reported to be hyperpolarized, hyperpolarized, and nonhyperpolarized, respectively (55).

Deletion of CgERG5 reverses CSP resistance in the Cgset4Δ mutant
We next examined mutants' growth in the presence of CSP. We found that while deletion of CgERG4 and CgERG5 genes rendered C. glabrata cells susceptible to CSP, CgERG3 gene loss led to CSP resistance (Fig. 6B). These results indicate opposite roles for CgERG3 and CgERG4 and CgERG5 genes in CSP susceptibility in C. glabrata. Of note, CgERG5 deletion has earlier been reported to result in CSP sensitivity (56).
Interestingly, CSP resistance in the Cgset4Δ mutant was not observed upon deletion of the CgERG5 gene, with the Cgset4Δerg5Δ mutant in fact displaying increased CSP susceptibility, compared with wt cells (Fig. 6B). Contrarily, the Cgset4Δerg3Δ and Cgset4Δerg4Δ mutants exhibited increased growth in the presence of CSP, similar to that of the Cgset4Δ mutant (Fig. 6B). These data point toward CgSet4-dependent and CgSet4-independent requirement of CgERG genes in regulating antifungal resistance, with CgERG4 loss resulting in increased susceptibility to both azole and echinocandin tolerance, and CgERG3 and CgERG5 loss leading to increased FLC and decreased CSP, and increased CSP sensitivity, respectively. These data suggest that the ergosterol biosynthesis pathway may be more closely linked with the action of cell walltargeting echinocandin drugs, than it is currently being considered. In this context, it is worth noting that Cgerg1Δ, Cgerg3Δ, and Cgerg11Δ mutants have been shown to exhibit altered expression of ERG genes (23,54). Therefore, it is possible that the differential expression of other ERG genes may contribute to distinct antifungal susceptibilities of double deletion mutants. However, the precise basis underlying the differential CSP susceptibilities of CgergΔ and Cgset4ΔergΔ mutants is yet to be determined.
Altogether, we infer from these data that the upregulated ergosterol biosynthesis process in the Cgset4Δ mutant contributes to elevated ergosterol content and may also modulate the response to CSP stress. Our data also raise a possibility of a possible nexus between CgSet4-dependent azole and echinocandin resistance and CgSet4-dependent negative regulation of CgERG gene expression.

CgSet4 regulates ergosterol biosynthesis through CgUpc2a
To investigate the molecular link between CgSet4-mediated repression of CgERG genes and CSP resistance, we focused on two genes, CgUPC2A and CgUPC2B, that code for two Zn 2 -Cys 6 transcriptional activators of the ergosterol biosynthesis pathway in C. glabrata (21), and performed four experiments. First, we checked transcript levels of CgUPC2A and CgUPC2B genes in the Cgset4Δ mutant and found 2.3-fold higher and similar CgUPC2A and CgUPC2B gene expression, respectively, in the Cgset4Δ mutant, compared with wt cells (Fig. 7A). Furthermore, CSP exposure led to 2.4-fold and 1.5-fold downregulation of CgUPC2A and CgUPC2B genes, respectively, in wt cells (Fig. 7A). Contrarily, CgUPC2A expression remained the same between CSP-treated and untreated Cgset4Δ cells (Fig. 7A), suggesting that CgSet4 may be required to control CgUPC2A gene expression in response to CSP exposure. Of note, a small 1.4-fold decrease in CgUPC2B transcript levels was observed in CSP-treated Cgset4Δ cells, compared with untreated Cgset4Δ cells (difference not conspicuous in Figure 7A, as it depicts comparison with the untreated wt cells), thereby ruling out a major role of CgSet4 in CSP-induced CgUPC2B downregulation.
Second, we created deletion strains for CgUPC2A and CgUPC2B genes in both wt and Cgset4Δ strain backgrounds. Deletion of the CgUPC2B gene had no impact on the susceptibility toward FLC (Fig. 7B) or CSP (Fig. 7C). Contrarily, CgUPC2A gene loss rendered cells highly and moderately sensitive to FLC (Fig. 7B) and CSP (Figs. 7C and S8), respectively). These FLC susceptibility data are consistent with the published reports of CgUpc2a being a major regulator of ergosterol biosynthesis genes in C. glabrata under normal laboratory-growth conditions (21,22). CgUpc2a has also been shown to be required for basal and FLC-induced transcription of ERG genes, with CgUPC2A deletion in an azole-susceptible dose-dependent and an azole-resistant clinical isolate leading to decreased ergosterol content (22). Of note, CgUPC2A gene loss has recently also been associated with CSP sensitivity (57).
Notably, the double mutant Cgset4Δupc2bΔ exhibited resistance to both FLC and CSP drugs, similar to the Cgset4Δ mutant (Fig. 7, B and C). Contrary to this, the Cgset4Δupc2aΔ double mutant was found to be sensitive to both antifungals, compared with the Cgset4Δ mutant (Fig. 7, B and C). Importantly, the Cgset4Δupc2aΔ double mutant displayed highly attenuated growth in rich yeast extract-peptone-dextrose (YPD) medium (Fig. S9), underscoring the stress that the simultaneous loss of both genes poses to the cellular machinery. Importantly, ectopic expression of CgSET4, CgUPC2A, and CgUPC2B genes in single and double mutants complemented the mutants' altered drug susceptibility phenotypes (Fig. 7, B and C); however, the phenotypes were not rescued fully in Cgset4Δupc2aΔ and Cgset4Δupc2bΔ mutants expressing CgUPC2A and CgUPC2B, respectively (Fig. 7, B and C). The basis underlying this observation is yet to be determined.
Based on these data, which suggest that CgSet4 is likely to control ergosterol biosynthesis process by modulating the CgUpc2A-dependent ERG gene regulation, we hypothesized that the elevated expression of CgUPC2A and its target genes contribute to antifungal resistance in the Cgset4Δ mutant. If this is true, the CgUPC2A gene deletion is likely to impact the antifungal resistance gene expression negatively in the Cgset4Δ mutant. Therefore, to test this hypothesis, we checked the expression of CgPDR1, CgERG3, and CgERG11 genes in the Cgset4Δupc2aΔ double mutant. We found two-fold lower, similar, and similar transcript levels of CgPDR1, CgERG3, and CgERG11 genes, respectively, in the double mutant, compared with wt cells (Fig. 7D), indicating that CgUpc2A is a major activator of the CgPDR1 gene in the Cgset4Δ mutant. Furthermore, the wt-like expression of CgERG3 and CgERG11 genes in the Cgset4Δupc2aΔ double mutant, in comparison to elevated CgERG3 and CgERG11 transcript levels in the Cgset4Δ mutant (Fig. 5A), implicates CgUpc2A in the activation of CgERG genes in the Cgset4Δ mutant.
Finally, to examine if CgSet4 directly regulates CgUPC2A expression, we performed the chromatin immunoprecipitation (ChIP) analysis using CgSet4-GFP, and found CgSet4 bound to the CgUPC2A promoter. Furthermore, CSP exposure led to a decrease in the occupancy of CgSet4 on the CgUPC2A promoter (Fig. 7E), which is consistent with a reduction in CgSET4 transcript and protein levels in response to CSP exposure (Fig. 2, C and D). These data indicate multifactorial regulation of CgUPC2A gene expression, with probably another repressor accounting for the CSP-induced CgUPC2A downregulation. Notably, CgRox1 has recently been reported to be a negative regulator of ERG genes, with Cgrox1Δupc2aΔ mutant containing 1.5-fold higher ergosterol levels, indicating perturbed sterol homeostasis (44). Based on this report and the following three findings of our RNA-Seq analysis, we speculate that CgRox1 could play a role in CgUPC2A downregulation upon CSP exposure. First, the Cgset4Δ mutant was found to be  (Table S6 and Fig. S6); Second, CgROX1 was upregulated in the Cgset4Δ mutant (Table S4 and Fig. S6). Third, CgROX1 transcription is activated in CSP-treated wt cells (Table S2 and Fig. S6). However, the role of CgRox1 in cellular response to CSP is yet to be illustrated.
Collectively, we draw five major conclusions from our data. First, CgErg3, CgErg5, and CgErg4 enzymes modulate antifungal tolerance in C. glabrata. Second, CgSet4 is a negative regulator of CgUPC2A and CgERG gene expression under normal growth conditions. Third, CgUPC2A largely accounts for elevated levels of CgPDR1 and CgERG transcripts in the Cgset4Δ mutant. Fourth, CSP exposure leads to transcriptional repression of CgSET4, CgUPC2A, and CgUPC2B genes, which may in turn probably contribute to the downregulated ergosterol biosynthesis pathway in CSP-treated C. glabrata cells. Finally, CgSet4 may not be a key regulator of the CgUpc2a gene regulatory circuit in the presence of CSP. Based on these findings, we propose that the complex transcriptional loops probably drive CgERG gene expression in C. glabrata.
Altogether, we show for the first time that the SET domaincontaining protein CgSet4 is pivotal to antifungal resistance, as it acts as an upstream regulator of the CgUpc2a-dependent ergosterol biosynthesis gene expression system. In addition, our data suggest that ergosterol content in the fungal cell membrane may play a role in determining the efficacy of the cell wall-targeting antifungal drugs.

Discussion
C. glabrata is a common cause of Candida bloodstream infections in patients with a weakened immune system (1). The successful treatment of C. glabrata infections is becoming increasingly difficult as two mainstream antifungal drugs, azoles and echinocandins, are proving to be less effective because of emerging resistance in C. glabrata isolates in hospitals (4,5,11,12). Not only C. glabrata is intrinsically less susceptible to frequently used, cost-effective, and relatively safe azole antifungals, but it also acquires high levels of resistance to azole and echinocandin drugs, with the echinocandin class representing the last line of antifungal therapy in many hospitalized severely ill patients (4,5,9,11,12,14,15). The azole resistance in clinical settings has predominantly been attributed to gain-of-function mutations in the master transcriptional regulator-encoding gene, CgPDR1, whereas mutations in the hot-spot regions of β-1,3 glucan synthase enzyme-encoding genes CgFKS1 and CgFKS2 primarily account for echinocandin resistance in hospitals worldwide (14,15). Chromatin architecture is pivotal to gene expression regulation, with histone post-translational modifications playing a key role in maintenance of chromatin homeostasis (58). The epigenetic control of antifungal resistance mechanisms in C. glabrata is beginning to be elucidated (28,38,56).
Toward this end, we report an essential role for the SET domain-containing protein CgSet4 in the transcriptional downregulation of CgPDR1 and CgUPC2A genes, which code for the master regulator of MDR and ergosterol biosynthesis genes, respectively, in C. glabrata. In addition, we present the first systematic analysis unveiling functions of CgSet1-CgSet6 and CgErg3-CgErg5 proteins in C. glabrata. We show that the last enzyme of the ergosterol synthesis pathway CgErg4 is required for both azole and echinocandin tolerance, and, of six SET domain-containing proteins, only CgSet4 acts as a negative regulator of both azole and echinocandin antifungal resistance.
The SET domain, that consists of strongly conserved sequence motif of about 130 amino acids, was initially identified in the Drosophila proteins, Su(var)3 to 9, Enhancer-ofzeste and Trithorax, which regulate gene expression during development (35,59). The SET domain now has been reported in several proteins that perform diverse functions (30,36,37). The lysine methyltransferase activity has largely been associated with the SET domain-containing proteins, which modulate chromatin structure, function, and the consequent gene expression regulation (30,36,37). In addition to the SET domain, these proteins also possess cysteine-rich regions flanking N-terminal (pre-SET) and posterior to C-terminal (post-SET) region of the SET domain, which are pivotal to target protein recognition and lysine methyltransferase activity (36,37).
The National Center for Biotechnology Information conserved domain search analysis (https://www.ncbi.nlm.nih. gov/Structure/cdd/wrpsb.cgi) revealed that the SET domain in CgSet4 belongs to the SET Superfamily cl40432. The C. glabrata Set4 protein is 350 amino acid-long, whereas Set4 in S. cerevisiae consists of 560 amino acids. Despite sharing 40% identity with CgSet4 protein, the S. cerevisiae Set4 could not complement the decreased FLC and CSP susceptibility of the Cgset4Δ mutant (Fig. 2, A and B). In addition, while CgSET4 expression was downregulated (Fig. 2D), ScSET4 transcription was found to be activated upon azole exposure (33). These results highlight functional differences between the two proteins and are in accordance with other reported functions of ScSet4 protein. For example, ScSet4 plays an essential role in protection against oxidative stress by activating stress response genes (31), and 196 genes were found to be differentially regulated upon ScSET4 disruption (34). Contrarily, CgSET4 deletion had a minor effect on the C. glabrata transcriptome, with 48 genes displaying deregulation (Table S4). Moreover, CgSET4 deletion did not lead to an increased susceptibility toward hydrogen peroxide-induced oxidative stress (Fig. 1B), which could in part be due to elevated expression of the CgCTA1 gene in the Cgset4Δ mutant (Table S4). Furthermore, unlike ScSET4 (31), ectopic expression of CgSET4 was not found to be detrimental for cell growth in C. glabrata (Fig. 2A).
Despite these differences, Set4 in both C. glabrata and S. cerevisiae acts as a negative regulator of ergosterol biosynthesis (ERG) genes. SET4 disruption led to azole resistance, with Set4 regulating ERG3 and ERG11 expression through direct binding to their promoters in S. cerevisiae (33). Ergosterol biosynthesis in fungi is a multistep energy-consuming process, with oxygen and heme acting as cofactors for many enzymes of the ergosterol synthesis pathway (Fig. S7) (50,51). Furthermore, because of the centrality of ergosterol in maintaining fluidity, integrity, and functions of the plasma membrane, ergosterol biosynthesis is regulated at multiple levels, including transcriptional regulation, transport, and sterol feedback inhibition, and subcellular localization of enzymes (50,51). The late steps of the ergosterol biosynthesis pathway involve many demethylation, reduction, and desaturation reactions (Fig. S7), with ergosterol being transported to the plasma membrane from the site of its synthesis, endoplasmic reticulum (50,51).
ERG gene expression is tightly regulated at the transcriptional level, with hypoxia and iron depletion resulting in downregulation of ergosterol synthesis in S. cerevisiae (31,33,49). Azole antifungals target a rate-limiting step of the ergosterol biosynthesis pathway that is mediated by the cytochrome P450-dependent lanosterol 14 alpha-demethylase enzyme, which is encoded by the ERG11 gene (13,14). Azole exposure in C. glabrata is known to result in elevated expression of ERG genes including ERG11, and this gene induction is largely carried out by the Zn2-Cys6 binuclear cluster transcription factor, CgUpc2a (21)(22)(23)60). We show that CgSET4 deletion led to the increased expression of CgUPC2A, along with its target ERG genes, viz., CgERG2, CgERG3, CgERG4, CgERG6, and CgERG11 (Figs. 5A and 7A). In addition, elevated transcript levels of the zinc finger transcriptional activator gene CgPDR1 as well as the multidrug transporter genes CgCDR1 and CgCDR2 (CgPdr1 target genes) in the Cgset4Δ mutant (Fig. 3A) suggest that CgSet4 is a general repressor of two major azole response pathways in C. glabrata. Consistent with this, ChIP analysis revealed CgSet4-GFP to be present at the promoter region of CgPDR1 (Fig. S10). Of note, the CgSet1-dependent H3K4 methylation has recently been found to be increased on actively transcribing ERG genes in response to FLC (38), thereby highlighting the epigenetic regulation of CgERG genes in C. glabrata.
In S. cerevisiae, the major sterol regulator Upc2, a homodimer, is known to bind to the 7 bp sterol regulatory element (SRE) sequence TATACGA that is present in promoters of the ERG, sterol uptake, the DAN/TIR genes (50,51). The C-terminal domain of Upc2 has been shown to act as an ergosterolbinding and sensing domain, with the ergosterol-bound Upc2 residing in the cytosol (50). Ergosterol depletion leads to the release of ergosterol, and the translocation of Upc2a to the nucleus, resulting in the transcriptional activation of its targets including ERG genes (50). The C. glabrata ortholog of S. cerevisiae Upc2, CgUpc2a, has recently been postulated to act like its S. cerevisiae counterpart and shown to regulate the expression of a vast array of genes including ERG and CgPDR1 regulon genes, which contain SREs in their promoter regions (57).
Through ChIP-Seq analysis, Vu et al. (57) showed that CgUpc2a binds to about 1000 genes in C. glabrata. In addition, 64 genes including CgERG4 were identified as indirect target genes of CgUpc2a (57). CgUPC2A itself was found to be upregulated upon FLC treatment, and its disruption rendered cells susceptible to both FLC and CSP drugs (57). In addition, although FLC-induced upregulation of CgCDR1 and CgPDR1 genes was lower and similar between wt and Cgupc2aΔ strains, respectively (57), a pivotal role for CgUpc2a in CgPdr1dependent gene network was unveiled by the reduced binding of CgUpc2a to the mutated SRE in the CgPDR1 promoter, as well as, by the diminished FLC-induced induction of the SRE-lacking CgPDR1 gene (57). Our results of reduced CgPDR1 gene expression in the double mutant Cgset4Δupc2aΔ (Fig. 7D) suggest that the elevated CgUPC2A transcript levels contribute to the increased CgPDR1 gene expression in the Cgset4Δ mutant. Of note, further detailed investigations are required to determine if CSP susceptibility of the Cgset4Δup-c2aΔ mutant is due to an imbalance of sterol species or cell wall components or both.
In addition to the SET proteins, we have also investigated the role of three enzymes catalyzing late stages of ergosterol synthesis, CgErg3, CgErg5, and CgErg4, in antifungal tolerance and virulence. Intriguingly, while the Cgerg3Δ and Cgerg4Δ mutants exhibited elevated FLC sensitivity, the Cgerg5Δ mutant displayed FLC susceptibility, similar to that of the wt strain (Fig. 6, A and C). Furthermore, deletion of the CgERG3, CgERG5, and CgERG4 genes led to decreased, increased, and increased susceptibility to CSP, respectively (Fig. 6B). Notably, the echinocandin-resistant isolates in a microevolution study, which exhibited crossresistance to FLC, have recently been shown to carry mutations in the CgERG3 gene (53). Our mice infection studies revealed that in line with their distinct role in antifungal resistance, CgErg3 and CgErg4 were required for survival of C. glabrata in the murine model of systemic candidiasis in an organ-dependent manner, whereas CgErg5 was dispensable in this model (Fig. S11). In addition, whereas CgUPC2A deletion led to significantly attenuated survival of C. glabrata in various organs, CgUPC2B deletion affected survival adversely only in kidneys (Fig. S11), indicating that CgUpc2a plays a major role in virulence of C. glabrata. These results together also highlight the importance of ergosterol synthesis for C. glabrata-mammalian host interaction.
Furthermore, our data underscore that CgErg enzymes differ from one another in their requirement in cellular response to FLC and CSP. This difference among CgErg enzymes is likely to be determined by additional regulatory factors and/or other functions of CgErg proteins. These results are consistent with varied regulatory mechanisms of different ERG genes in S. cerevisiae and C. glabrata (21,23,33,50,57,60). In this context, it is worth noting that despite CgUpc2a showing strong binding to the CgERG1 gene promoter, FLCinduced activation of CgERG1 gene was found to be similar between wt and Cgupc2aΔ mutant (57), thereby pointing toward the complex multifactorial environmental cue-dependent regulation of individual CgERG genes. Our data provide further support to this notion, as CgSet4 was found to be indispensable and dispensable for the CSP-induced repression of CgERG1, and other CgERG (CgERG4, CgERG5, CgERG6, CgERG11, and CgERG25) genes, respectively (Fig. 5, C and D).
CSP exposure has recently been reported to result in elevated reactive oxygen species production (61). Consistent with this, our transcriptional profiling analysis revealed upregulation of oxidative stress response genes in CSP-treated cells. Furthermore, a multifactorial role of mitochondria has recently been reported in CSP tolerance in C. glabrata (61). Given that a reduction in ergosterol content is known to adversely affect mitochondrial DNA maintenance (62), it is possible that elevated ergosterol in Cgset4Δ mutant modulates mitochondrial functions, which may contribute to CSP resistance in the mutant. Elucidation of the nexus among sterol metabolism, CSP resistance, and mitochondrial functions will shed light on the underlying molecular mechanism.
Collectively, our data suggest that the transcriptional regulation of ERG genes is not necessarily reflected in cellular ergosterol levels. Consistently, despite CgUpc2a regulating ERG gene expression, its disruption led to no significant decrease in ergosterol levels (44,57). Similarly, cellular ergosterol levels were found to be decreased upon loss of both UPC2 and ECM22 (paralog of UPC2) genes, which code for activators of the sterol biosynthetic pathway in S. cerevisiae (50,63). Based on our data, we propose that CgSet4 is a negative regulator of the basal-level expression of the CgUPC2A gene (Fig. 8), whereas the cellular response to CSP involves a wholesale downregulation of ergosterol biosynthesis pathway that is probably initiated by the transcriptional downregulation of CgUPC2A (Fig. 8). CSP-induced repression of CgUPC2A is probably primarily carried out by another repressor protein (Fig. 8), with CgRox1 to be a likely candidate whose expression was found to be upregulated in response to CSP exposure (Table S2). Of note, CgRox1 has recently been shown to be a negative regulator of ERG genes, with loss-offunction mutations in CgROX1 rescuing the increased FLC susceptibility of the Cgupc2aΔ mutant (44). Furthermore, since the azole and CSP resistance of the Cgset4Δ mutant is reversed by CgUPC2A deletion (Fig. 7, B and C), CgUpc2a is likely to be the main effector protein of CgSet4, with CgUpc2a also Figure 8. A schematic summarizing key findings of the study. Candida glabrata cells maintain cellular ergosterol levels via tight regulation of ergosterol biosynthesis (CgERG) genes. CgSet4 acts as a repressor of CgPdr1-dependent multidrug resistance, and ergosterol biosynthesis pathways, as CgSET4 deletion leads to increased basal expression of CgPDR1 regulon and CgERG genes, elevated ergosterol content, and resistance to fluconazole and CSP antifungals. Under regular growth conditions, CgSet4 keeps CgUPC2A expression in check via binding to the CgUPC2A promoter. Since CgUpc2a is an activator of CgERG and CgPDR1 genes, the lower levels of CgUpc2a result in restrained expression of CgERG and CgPDR1 genes, thereby maintaining sterol homeostasis and susceptibility to azole antifungals. CSP exposure results in transcriptional and post-transcriptional downregulation of CgSET4, reduction in CgSet4 abundance on the CgUPC2A promoter, transcriptional activation and repression of CgROX1 and CgUPC2A, respectively, and the consequent downregulation of CgERG genes. Two lines of evidence, the Cgset4Δ mutant's proficiency in downregulating CgERG genes upon CSP exposure, and the decreased CgSet4 occupancy on CgUPC2A promoter in CSP-treated wt cells, point toward another repressor protein (repressor X) contributing to CgUPC2A and CgERG gene downregulation in response to CSP. This repressor is yet to be identified, although CgRox1, being an inhibitor of CgERG gene expression, appears to be a strong candidate for the same.
Role of CgSet4 in antifungal drug resistance controlling expression of the CgPdr1 regulon genes. Although the molecular basis for CgSet4-dependent regulation of CgPdr1 and CgUpc2a regulon genes is yet to be deciphered, our ChIP data suggest that it is likely to be through direct association of CgSet4 with the regulatory regions of CgUPC2A. However, the regulatory region of CgUpc2a, where CgSet4 could bind to, is yet to be identified.
Importantly, owing to a general repressive effect of CSP on ERG gene expression, our data underscore the need to revisit the combinatorial therapeutic regimen, which involves treatment of fungal infections with echinocandins along with other ergosterol-targeting drugs, polyenes (bind to ergosterol in the cell membrane), allylamines (inhibit squalene epoxidase, encoded by ERG1), or azoles (64,65).
Altogether, our data demonstrate CgSet4 to be a key regulator of CgPdr1-dependent MDR and CgUpc2a-dependent ergosterol biosynthesis pathways, thereby making CgSet4 as a major component of the cell wall and cell membrane homeostasis systems in C. glabrata.

Experimental procedures
Strains, media, and growth C. glabrata strains used in the study were derivatives of the BG2 strain and maintained in the rich YPD or minimal casamino acid (CAA) medium at 30 C. The Escherichia coli DH5α strain was used for plasmid propagation and maintained in LB medium at 37 C. Overnight cultures of C. glabrata strains were grown for 3 to 4 h at 30 C to obtain log-phase cultures. Antifungal drug and stress susceptibility was examined by serial dilution spotting assay and liquid medium-based growth analysis.

Mice infection assay
Mice infection studies were performed at the Animal House Facility of Centre for DNA Fingerprinting and Diagnostics (CDFD), Hyderabad, India in accordance with guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals, Government of India, and were approved by the Institutional Animal Ethics Committee [EAF/ RK/CDFD/15]. Briefly, C. glabrata strains were grown overnight in YPD or CAA medium, washed with PBS, and suspended in PBS. C. glabrata cells (4 × 10 7 ; 100 μl PBS cell suspension) were injected into the tail vein of 6-to 8-week-old female BALB/c mice. Mice were monitored for 7 days, sacrificed, and four organs, kidneys, liver, brain, and spleen, were collected. Organs were homogenized in PBS, and appropriate dilutions were plated on penicillin-and streptomycincontaining YPD medium. Mouse organ fungal burden was determined by counting colonies manually.

qRT-PCR
Total RNA was extracted from appropriate strains using the acid-phenol extraction method and digested with DNase I to eliminate any DNA contamination. DNase I-digested RNA (500 ng) was used to synthesize complementary DNA using the SuperScript III First-Strand Synthesis System for RT-PCR. The qPCR was performed using the SYBR green real-time PCR mastermix, and the sample C T (cycle threshold) values were normalized against the C T value of the control house-keeping CgACT1 or CgTDH3 gene. The fold change in expression under different conditions was determined using the comparative C T (2 −ΔΔC T ) method.

Microscopy analysis
The overnight grown Cgset4Δ/CgSET4-GFP strain was grown to log phase in CAA medium, followed by growth either in the absence (CAA) or the presence of FLC (64 μg/ml) or CSP (150 ng/ml), for 1 h. After incubation, cells corresponding to absorbance of 1.0 at 600 nm were collected, washed with PBS, and suspended in 100 μl of PBS containing 1 μg/ml Hoechst 33258 (Sigma; catalog no.: 94403) stain. After incubation at 37 C for 15 min, 3 to 5 μl of stained cell suspension was mounted on a slide and imaged with the Confocal microscope (Zeiss LSM 700) equipped with 63×/1.44 numerical aperture objective.

RNA-Seq analysis
Log-phase grown wt and Cgset4Δ mutant strains were grown with and without CSP (250 ng/ml) for 1 h. Total RNA was extracted using acid-phenol method followed by DNase digestion to remove any DNA contamination. The RNA samples were sent on dry ice to the AgriGenome Labs (a subsidiary of SciGenom Labs), Kakkanad, Kochi, Kerala, India (http://agrigenomelabs.com/), where these were further processed for sequencing. RNA quality was assured by taking RNA samples with RNA integrity number values ≥8. The complementary DNA library was prepared using TruSeq RNA Sample Prep Kits (Illumina), and 2 × 100 bp paired-end sequencing was performed on the Hiseq 2500 Illumina platform. The 40 to 60 million high-quality reads were obtained for each sample. The sequences were trimmed to remove unwanted sequences and aligned pairwise with the C. glabrata CBS138 reference genome (http://www.candidagenome.org). The DESeq analysis package was used, and genes that exhibited at ≥1.5-fold change in expression (q value = ≤0.05) were considered as DEGs.

Ergosterol estimation
The ergosterol content in C. glabrata strains was measured, as described previously (67). Briefly, strains were grown to the log phase in the CAA medium (150 ml) at 30 C, and absorbance at 600 nm of cultures was measured. After an absorbance at 600 nm normalization, cells were pelleted down and the wet weight of the cell pellet was determined. The cell pellets were suspended in freshly prepared alcoholic potassium hydroxide in sterile borosilicate glass screw-cap tubes and incubated at 85 C in a water bath for 3 h. After cooling to room temperature, the water and n-heptane (1:3 ratio) mixture was added to each tube, and tubes were vortexed vigorously for 3 min. The samples were kept static for about 15 min for phase separation, and the upper transparent layer of n-heptane was collected. The UV spectrophotometric profiles were recorded between 220 and 300 nm, with both ergosterol and 24 (28)dehydroergosterol (DHE) absorbing at 281.5 nm, whereas 24(28)-DHE alone absorbing at 230 nm. The ergosterol content was determined by subtracting the amount of 24 (28)-DHE (calculated from the absorbance at 230 nm) from the total ergosterol plus 24 (28)-DHE content (calculated from the absorbance at 281.5 nm), and calculated as a percentage of the wet weight of the cells using the following equation:

ChIP assay
Log-phase C. glabrata strains were grown in the presence or the absence of CSP (250 ng/μl) for 1 h. Cells were crosslinked with 1% formaldehyde for 20 min, followed by quenching with 125 mM glycine for 10 min. Cell pellets were collected and lysed in FA lysis buffer (1 mM EDTA [pH 8.0], 50 mM Hepes [pH 7.5], 140 mM NaCl, 0.1% sodium deoxycholate [w/v], 1× protease inhibitor, and 1% Triton X-100) by bead-beating. After removal of the cell debris by centrifugation, the supernatant was subjected to sonication for 40 min, with 30 s pulses of on and off at highest amplitude. The samples were centrifuged at 15,000 rpm for 10 min, and the soluble fraction was collected. The 1/10th volume of the soluble fraction was saved as "input" fraction, and the remaining soluble fraction was precleared with protein-A Sepharose beads, prior to immunoprecipitation with anti-GFP antibody for 4 h at 4 C. After incubation, beads were given four consecutive washes, with FA lysis buffer, lysis buffer containing 0.5 M NaCl, wash buffer (100 mM Tris-HCl [pH 8.0], 0.25 M LiCl, 0.5% NP-40, 0.5% deoxycholic acid [w/v], and 1 mM EDTA) and TE buffer. The beads were suspended in the elution buffer (50 mM Tris-HCl, 10 mM EDTA, and 1% SDS) and incubated overnight at 65 C for decrosslinking, followed by proteinase K treatment for 1 h. DNA was precipitated using phenol:chloroform:isoamyl alcohol, suspended in TE, treated with RNAse at 37 C for 1 h, and used as template for qRT-PCR using appropriate set of primers.

Other procedures
The cell wall analysis was performed, as described previously (66).

Statistical and functional analysis
The statistical significance was determined using the GraphPad Prism software (GraphPad Software, Inc). The twotailed Student's t test and the non-parametric Mann-Whitney test were used for intergroup comparisons and mouse organ fungal burden analysis, respectively. DEGs were analyzed and functionally annotated using the Candida Genome Database (http://www.candidagenome.org) and DAVID (https://david. ncifcrf.gov/) tools.

Data availability
The raw RNA-Seq data have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/geo/), with Gene Expression Omnibus accession number GSE202654.