Ctt1 catalase activity potentiates antifungal azoles in the emerging opportunistic pathogen Saccharomyces cerevisiae

Fungi respond to antifungal drugs by increasing their antioxidant stress response. How this impacts antifungal efficacy remains controversial and not well understood. Here we examine the role of catalase activity in the resistance of Saccharomyces cerevisiae to the common antifungals, fluconazole and miconazole, for which we report minimum inhibitory concentrations (MICs) of 104 and 19 μM, respectively. At sub-MIC concentrations, fluconazole and miconazole stimulate catalase activity 2-3-fold but, unexpectedly, deletion of cytosolic catalase (ctt1) makes cells more resistant to these azoles and to clotrimazole, itraconazole and posaconazole. On the other hand, upregulating Ctt1 activity by preconditioning with 0.2 mM H2O2 potentiates miconazole 32-fold and fluconazole 4-fold. Since H2O2 preconditioning does not alter the resistance of ctt1Δ cells, which possess negligible catalase activity, we link azole potentiation with Ctt1 upregulation. In contrast, sod2Δ cells deleted for mitochondrial superoxide dismutase are 4–8-fold more azole sensitive than wild-type cells, revealing that Sod2 activity protects cells against azole toxicity. In fact, the ctt1Δ mutant has double the Sod2 activity of wild-type cells so ctt1 deletion increases azole resistance in part by Sod2 upregulation. Notably, deletion of peroxisomal/mitochondrial cta1 or cytosolic sod1 does not alter fluconazole or miconazole potency.


Results
MICs of azoles for S. cerevisiae and their classification as fungicidal vs. fungistatic. Starting at an initial cell density of 10 6 cfu/ml and based on cell growth at different drug concentrations (Fig. 2), we determined the minimum inhibitory concentration (MIC µg/mL; µM) for our S. cerevisiae strain (BY4741) of six medically relevant azoles: itraconazole (32; 45), fluconazole (32; 105), posaconazole (32; 46), voriconazole (>256; >730), miconazole (8; 19) and clotrimazole (4; 12) ( Table S1). The structures of the azoles, shown as a footnote to Table S1, reveal that the drugs examined can be classified as triazoles (itraconazole, fluconazole, posaconazole and voriconazole) and imidazoles (miconazole and clotrimazole). The imidazoles are more potent antifungals than the triazoles and, in fact, cells are refractory to voriconazole (Table S1). An azole is classified as fungicidal if 1xMIC or 2xMIC promotes a ≥10 3 -fold reduction in the viable cfu/mL and Table S2 shows that the imidazoles are fungicidal under the present experimental conditions, whereas the triazoles are fungistatic with the exception of voriconazole.
Cultures of C. albicans (strain SC5314) at the same initial cell density (10 6 cfu/mL) exhibit MICs of >1 mM for fluconazole 42 and 60 µM for miconazole 43 . Thus, under our culture conditions, S. cerevisiae strain BY4741 is more albicans. Cytosolic Ctt1 is found in the cytoplasm of S. cerevisiae, whereas Cta1 is co-targeted to peroxisomes and mitochondria in respiring S. cerevisiae 48 and inferred to be associated with these two compartments in C. albicans. CuZnSod1 is localized in the cytoplasm and the mitochondrial intermembrane space of both yeasts, while MnSod2 is present in the mitochondrial matrix. C. albicans possesses an extra MnSod3 in the cytosol and cell-membrane-associated CuSod4-6, which are absent in S. cerevisiae. Note that C. glabrata possess only Cta1 5

Ctt1 catalase activity weakly combats miconazole-induced H 2 o 2 accumulation in S. cerevisiae cells.
Miconazole induces a rise in intracellular H 2 O 2 (Fig. 3C,F) despite also inducing catalase activity in wild-type cells. This led us to examine catalase activity and H 2 O 2 levels in the cta1Δ and ctt1Δ strains, which lack peroxisomal/mitochondrial and cytosolic catalase, respectively ( Fig. 1). Catalase activity (Fig. 3A,B) and H 2 O 2 levels ( Fig. 3C) are the same in wild-type and cta1Δ cells, which reflects the strong repression of Cta1 by glucose in the medium 34,47,48 . In contrast, Ctt1 is not repressed by glucose 34,49,50 and confers most of the catalase activity in cells grown in YPD since ctt1Δ cells are virtually devoid of catalase activity (Fig. 3A,B). Moreover, the azoles fail to induce catalase activity in the ctt1Δ strain (Fig. 3A,B) although ~4 times more H 2 O 2 accumulates in miconazole-treated vs. control over 24 h (Fig. 3F). It also is remarkable that the H 2 O 2 levels in wild-type and cta1Δ cells are ~75% those of ctt1Δ cells with negligible catalase activity (Fig. 3C,F). Thus, Ctt1 appears to be an ineffective scavenger of miconazole-induced H 2 O 2 .

Deletion of ctt1 or inhibition of catalase activity increases azole resistance in S. cerevisiae.
Peroxide-metabolizing enzymes have been associated with protection against cidal antimicrobials 4,5,8,14,15,45 . However, our observation that Ctt1 does not inhibit miconazole-induced H 2 O 2 accumulation (Fig. 3C,F) led us to ask whether Ctt1 actually protects cells against azole toxicity. As shown in Table 2 and Fig. S1, ctt1Δ cells display 4-and 8-fold higher MICs for fluconazole and miconazole, respectively, than the two strains with Ctt1 activity (Tables 2, S1). Given this surprising observation, we additionally determined the fold-change in MIC when ctt1 was deleted for the four other azoles. Both wild-type and ctt1Δ cells are refractory to voriconazole (Table S1) but the ctt1Δ strain is 8-fold less sensitive to posaconazole and 2-fold less sensitive to clotrimazole and itraconazole than wild-type cells (Table S1). Thus, Ctt1 appears to potentiate both fungistatic and fungicidal azoles in S. cerevisiae. Aminotriazole is a well-documented inhibitor of catalase activity in S. cerevisiae 51 . Thus, to directly probe the effect of inhibition of catalase activity on miconazole resistance we added aminotriazole to the cells. This compound did not inhibit the growth of any strain at concentrations as high as 100 mM (data not shown) but    Table 2). Relative intracellular H 2 O 2 levels measured by flow cytometry at (C) 8 h and (F) 24 h after DHR-stained wild-type , cta1Δ and ctt1Δ cells were exposed to 0.05xMIC miconazole or ethanol only (control). Experimental conditions: Cells at an initial OD 600 of 0.15 were grown in YPD at a medium-to-flask ratio of 1:5 at 30 °C/225 rpm. Catalase activity was assayed (see Materials and Methods) at 24 h after 3-mL cultures were challenged with azole in 12 μL of ethanol. For preconditioning, cultures were grown to OD 600 0.50 (12 h), 0.2 mM H 2 O 2 was added to the medium, cells were diluted 30 min later to OD 600 0.15 (10 6 cfu/mL) and challenged with azole in 3-mL cultures. To determine relative H 2 O 2 levels, cells grown in 3-mL cultures at initial OD 600 of 0.15 were was stained with 5 µM DHR in 1-mL aliquots at 30 °C, pelleted after 120 min, diluted to 10 6 cells/mL in PBS, fixed with 2% formalin (v/v) and analyzed by flow cytometry (see Materials and Methods). Relative fluorescence units (RFU; ex/em 490/520 nm) of individual cells were measured and the median RFU of 10,000 cells estimates a sample's relative H 2 O 2 level. All results represent the avg ± SEM of six independent experiments (n = 6). Statistical analyses performed using Student's t-test compare each sample with the wildtype untreated control. *p < 0.05 and **p < 0.01.

Deletion of ctt1 elevates MnSod2 activity in early log phase and increases miconazole resistance.
Although fungicide-dependent ROS production reportedly leads to fungal cell death 2,13,20 , we find no link here between elevated H 2 O 2 levels and miconazole sensitivity. In fact, ctt1Δ cells, which are the most miconazole resistant (Table 2), accumulate more H 2 O 2 on challenge with this azole (Fig. 3C,F). However, the miconazole resistance of C. albicans biofilms is dependent on the ROS-detoxifying activity of Sods 22 , and we 37 and others 36 have shown previously that suppressing or deleting catalase activity in S. cerevisiae upregulates mitochondrial MnSod2. Thus, we hypothesized that increased MnSod2 activity contributes to the enhanced azole resistance of ctt1Δ cells (Table 2). There are two Sod isoforms in S. cerevisiae (Fig. 1), and we find that the three strains exhibit similar total Sod activity, which doubles between 8 and 24 h but does not increase upon miconazole challenge (Fig. 4A,D). Since MnSod2 accounts for only 10-20% of the total Sod activity in cells growing on glucose 52 , to unmask any variation in this activity, we selectively inhibited CuZnSod1 with KCN 53 . This revealed 1.7-fold higher MnSod2 activity in untreated ctt1Δ cells vs. wild-type or cta1Δ cells (Fig. 4B,E).
We next compared the relative levels of O 2 •− in the three strains. Staining cells with the profluorescent dye, DHE, which is preferentially oxidized by O 2 •− 54 , we uncovered 2-fold less O 2 •− in the ctt1Δ strain relative to wild-type or cta1Δ cells (Fig. 4C). O 2 •− levels were a factor of ~1.3 higher in the cultures challenged with miconazole but remained significantly lower in ctt1Δ cells (Fig. 4C). The O 2 •− levels tripled between 8 and 24 h such that the 24-h miconazole-challenged cells contained > 10-fold more O 2 •− than the untreated 8-h cells (Fig. 4F vs. 4C). Also, the 24-h cultures have comparable O 2 •− levels and MnSod2 activity (Fig. 4E,F) so azole resistance must be associated with the O 2 •detoxifying activity of MnSod2 during exponential growth. Thus, we conclude that ctt1Δ cells are more azole resistant (Table 2) because they possess the higher MnSod2 activity in early log phase (Fig. 4B,C).
We additionally examined if the Sod mimetics, TEMPO • or mito-TEMPO • , protect wild-type S. cerevisiae against miconazole toxicity. These radicals are well-established O 2 •− scavengers and mito-TEMPO • is targeted to mitochondria 55 but not TEMPO • 55,56 . Addition of 1 mM TEMPO • did not change the MIC of miconazole (Table 3) and 1 mM mito-TEMPO • afforded only modest protection, doubling the MIC of miconazole in four of the six independent cultures examined ( Table 3). It is possible that O 2 •− scavenging by mito-TEMPO • is offset by the miconazole-induced increase in respiration in wild-type cells (Fig. S6). Increased respiration is not detected in the ctt1Δ mutant (Fig. S6) so its elevated MnSod2 activity may provide better protection against miconazole-dependent O 2 •− production than MnSod2 plus mito-TEMPO • in wild-type cells. Also, the efficacy of mito-TEMPO • may be lowered by its reaction with mitochondrial reductases 55 . To further explore the importance of Sod activity in azole resistance, we measured the fluconazole and miconazole MICs for sod1Δ and sod2Δ cells. MICs are the same for sod1Δ and wild-type cells, revealing that CuZnSod1 deletion does not impact miconazole resistance (Table 2), which is consistent with 1 mM TEMPO • having no protective effect (Table 3). However, the sod2Δ strain possesses fluconazole and miconazole MICs that are 4-and 8-fold lower, respectively ( Table 2). These results confirm that MnSod2 activity protects cells from azole toxicity and upregulation of MnSod2 activity in the ctt1Δ strain (Fig. 4B,E) increases its azole resistance ( Table 2).
Inhibiting catalase activity in the sod2Δ strain enhances miconazole resistance less than in wild-type cells. If Ctt1 activity potentiates the azoles by suppressing MnSod2, then inhibiting catalase activity in the sod2Δ strain should not enhance resistance. Treatment of sod2Δ cells with 25 mM aminotriazole resulted in undetectable catalase activity as seen for wild-type cells (Fig. S5). The MIC for miconazole increased from 1 to 4 µg/mL vs. the increase to 32 µg/mL seen on aminotriazole treatment of wild-type cells (Table 2). Hence, Ctt1 activity potentiates miconazole in large part by depressing MnSod2 activity or in other words, the O 2 •− detoxifying activity of MnSod2 combats azole toxicity and its deletion or suppression by Ctt1 activity lowers azole resistance in S. cerevisiae.

Cytosolic Ctt1 catalase activity, not elevated intracellular H 2 o 2 , potentiates azole toxicity. De
Nollin et al. found that fungistatic doses of miconazole stimulate catalase activity in S. cerevisiae 14 and proposed that this rescues cells from H 2 O 2 intoxication. We report here that sub-MIC concentrations of miconazole induce Ctt1 catalase activity up to 3-fold in our wild-type S. cerevisiae strain (BY4741) (Fig. 3) but this does prevent cells from accumulating ~4-fold more H 2 O 2 over 24 h than untreated cells (Fig. 3F). Furthermore, ctt1Δ cells with negligible catalase activity, accumulate more miconazole-induced H 2 O 2 than wild-type or cta1Δ cells (Fig. 3) but are more resistant to the azole (Table 2). Therefore, contrary to expectation 14  Cytosolic Ctt1 activity potentiates the azoles partly by depressing MnSod2 activity. Since ctt1Δ cells with the highest MnSod2 activity of the strains examined here (Fig. 4) are 4-8-fold more azole resistant than wild-type cells, we conclude that depression of MnSod2 activity on Ctt1 stimulation potentiates the azoles. MnSod2 is not essential for fermenting S. cerevisiae 57 but sod2Δ cells exhibit 4-8-fold lower azole resistance than wild-type cells (Table 2). However, we note that H 2 O 2 preconditioned wild-type cells are 4-fold more miconazole sensitive than sod2Δ cells (Table 2). Thus, strong Ctt1 induction may potentiate miconazole by additional mechanisms. For example, miconazole may bind to the heme of Ctt1 as reported for CYP51 58 , the 14α-demethylase in the ergosterol biosynthetic pathway 33 . This could promote heme-catalyzed azole autoxidation with the formation of reactive, cytotoxic species via mechanisms analogous to those we reported for hydrazides 59 . Elevated MnSod2 activity in early log phase increases azole resistance. The ctt1Δ cells from 8-h cultures possess higher MnSod2 activity and less O 2 •− than wild-type and cta1Δ cells (Fig. 4). Respiration is a major source of O 2 •− , and induction of respiration by miconazole reportedly increases its toxicity in S. cerevisiae whereas genetic blockage of respiration (by deleting TCA-cycle and ETC components) has the opposite effect 2 .

Respiration-derived O 2
•− inactivates aconitase 60,61 with the release of free iron, which catalyzes the production of highly toxic hydroxyl radicals via Fenton chemistry 1,37,62 .
We have previously reported on the positive biochemical and physiological effects of elevated MnSod2 activity in young cells deleted for cytochrome c peroxidase (ccp1Δ) 48 . Like ctt1Δ, the ccp1Δ mutant exhibits low O 2 •− and high H 2 O 2 levels plus it possesses stable aconitase activity, accumulates low amounts of free iron and hydroxyl radicals, amasses mitochondrial damage more slowly and lives longer than wild-type cells 48 . These traits arise from the beneficial mitochondrial H 2 O 2 stress response known as mitohormesis, which requires MnSod2 upregulation 17,36,37,63 . Presumably, the advantages of elevated MnSod2 activity in early log phase contribute to the increased miconazole resistance of ctt1Δ cells. At 24 h after miconazole treatment, the three strains possess comparable MnSod2 activity and O 2 •− levels (Fig. 4). Nonetheless, based on Rhod123 staining 64 , miconazole does not increase respiration in ctt1Δ cells (Fig. S6), suggesting that mitohormesis protects mitochondrial function 65 . Catalase and azole resistance in S. cerevisiae vs. C. albicans and C. glabrata. Given their different catalase and Sod isozymes (Fig. 1), it is informative to compare azole sensitivity in S. cerevisiae and C. albicans. It was reported in the 1970s that fungistatic doses of miconazole induce catalase activity in S. cerevisiae and C. albicans whereas fungicidal doses inhibit this activity 14 . We confirm these results for S. cerevisiae but show that only Ctt1 activity is induced (Fig. 3B) since peroxisomal/mitochondrial Cta1 is repressed by glucose 66 29 (Fig. 1), and synergistic killing of C. albicans biofilms by fluconazole and H 2 O 2 has been reported but no molecular mechanism was suggested 67 .
Like S. cerevisiae, the opportunistic yeast C. glabrata possesses Cta1, cytosolic CuZnSod1 and mitochondrial MnSod2. Antifungals also induce ROS production and stimulate catalase, Sod and glutathione peroxidase activities in C. glabrata 21,45 . Azole resistance is associated with increased catalase activity 5 and increased protein levels of thiol peroxidases 68 , but whether deletion of these antioxidant enzymes alters azole resistance in C. glabrata remains to be seen.

Conclusions
Although high catalase activity has been linked to azole resistance in C. albicans, C. glabrata and S. cerevisiae, the present study reveals that azole-induced upregulation of Ctt1 activity potentiates azole toxicity by depressing MnSod2 activity in S. cerevisiae, Hence, MnSod2 is an interesting antifungal target in this yeast but target antioxidant enzymes are likely to be species dependent. Therefore, to expand our knowledge of the role of a given antioxidant activity in fungal survival strategies, we need to establish the potency of antifungal drugs in yeasts singly deleted for the antioxidant enzyme of interest as performed here for S. cerevisiae.
Yeast strains. The Saccharomyces cerevisiae wild-type and mutant BY4741 strains used in this work are listed in Growth conditions and H 2 o 2 preconditioning. Precultures (10 mL) were obtained by growing single colonies of each strain in YPD (1% yeast extract, 2% peptone and 2% dextrose) for 24 h at 30 °C with high aeration (medium-to-flask ratio of 1:5 and shaking at 225 rpm). These cultures were used to inoculate 25 mL of fresh YPD in 125-mL flasks to give the experimental cultures at an initial OD 600 of 0.01 (OD 600 was measured at a 1.0-cm pathlength unless otherwise indicated). Cells (3 mL) were grown under the same conditions to mid-log phase (OD 600 0.50; 12 h) and preconditioned with 0.2 mM H 2 O 2 for 30 min at 30 °C/225 rpm where indicated.

Determination of azole minimum inhibitory concentration (MIC).
The solid azoles were dissolved in 100% ethanol to give stocks of 50 mg/mL fluconazole, 10 mg/mL voriconazole, 1 mg/mL miconazole and clotrimazole; and in 100% dimethyl sulfoxide (DMSO) to give stocks of 1 mg/mL posaconazole and itraconazole. Since H 2 O 2 preconditioning causes a 25-30% reduction in viable ctt1Δ cells 34 , the liquid cultures were diluted to OD 600 0.15 (10 6 cfu/mL) in fresh YPD before MIC determination. Our initial cell density is higher than suggested by the Clinical and Laboratory Standards Institute (10 3 cfu/m) 44 to provide sufficient cells for the biochemical analyses. Cells were exposed to different azole concentrations in 96-well plates (final volume of 200 μL per well) and MICs were determined as described 44 . Briefly, cells were mixed with the drug and OD 600 was measured on a SpectraFluor Plus Tecan plate reader at t = 0 and t = 24 h after growth at 30 °C without shaking. The MIC for each azole was determined from a plot of OD 600 at t = 24 h minus that t = 0 vs. [azole]. The MIC is the lowest antifungal concentration that results in no detectable growth after 24 h incubation 44 . MICs for cultures simultaneously treated with the azole and 1 mM TEMPO • , 1 mM mito-TEMPO • (Sod mimetics) 55,56 or 25 mM aminotriazole (catalase inhibitor) 51 were determined the same way in 96-well plates. To establish if an azole was fungicidal or fungistatic, wells containing 1xMIC and 2xMIC of the drug were serially diluted 10x after 24 h at 30 °C, plated onto YPD agar and grown for 2 days at 30 °C to measure the viable cfu/mL. A drug was considered fungicidal if 1xMIC or 2xMIC promoted a ≥10 3 -fold reduction in viable cfu 44 . Soluble protein extracts. Cells ±H 2 O 2 preconditioning and ± aminotriazole exposure were diluted to OD 600 0.15 in 3 mL of fresh YPD ± azole in a 15-mL Falcon tube, grown at 30 °C/225 rpm for 24 h, OD 600 values were measured, and soluble proteins were extracted as described previously 17,34 . Briefly, after centrifugation at 2000 × g, cells were washed 2x with 100 mM potassium phosphate buffer at pH 7.0 (KPi) containing 0.1 mM PMSF, the pellets were diluted into KPi/PMSF, and mixed with an equal volume of acid-washed glass beads (400-600 µm). Cells were disrupted by vortexing 4 × 15 s, the homogenates were spun at 13000 × g for 10 min at 4 °C, and the total protein concentration in the supernatants was determined by the Bradford assay with BSA as a standard 71 . Catalase and Sod activity assays. Cells exposed to azole concentrations below the MIC (sub-MIC) were used in the biochemical analyses to avoid the general metabolic collapse and down regulation of multiple enzyme activities seen at lethal drug concentrations 14,34 . To assay for catalase activity, 25-150 µL aliquots of soluble protein extract 72 containing 20-100 μg protein were added to 1.0 mL of 20 mM H 2 O 2 in 50 mM KPi in a cuvette. H 2 O 2 decomposition was monitored at 240 nm (ε 240 = 43.6 M −1 cm −1 ) 72 . One unit of catalase activity catalyzes the degradation of 1 µmol of H 2 O 2 per min 34,72 . Sod activity was assayed using the Superoxide Dismutase Detection Kit (Cell Technologies, CSOD100), where O 2 •− is generated by xanthine/xanthine oxidase and oxidized by XTT