In Candida glabrata, ERMES Component GEM1 Controls Mitochondrial Morphology, mtROS, and Drug Efflux Pump Expression, Resulting in Azole Susceptibility

Mitochondrial dysfunction or morphological abnormalities in human pathogenic fungi are known to contribute to azole resistance; however, the underlying molecular mechanisms are unknown. In this study, we investigated the link between mitochondrial morphology and azole resistance in Candida glabrata, which is the second most common cause of human candidiasis worldwide. The ER-mitochondrial encounter structure (ERMES) complex is thought to play an important role in the mitochondrial dynamics necessary for mitochondria to maintain their function. Of the five components of the ERMES complex, deletion of GEM1 increased azole resistance. Gem1 is a GTPase that regulates the ERMES complex activity. Point mutations in GEM1 GTPase domains were sufficient to confer azole resistance. The cells lacking GEM1 displayed abnormalities in mitochondrial morphology, increased mtROS levels, and increased expression of azole drug efflux pumps encoded by CDR1 and CDR2. Interestingly, treatment with N-acetylcysteine (NAC), an antioxidant, reduced ROS production and the expression of CDR1 in Δgem1 cells. Altogether, the absence of Gem1 activity caused an increase in mitochondrial ROS concentration, leading to Pdr1-dependent upregulation of the drug efflux pump Cdr1, resulting in azole resistance.


Introduction
Candida species can cause severe systemic infections in immunocompromised patients and are associated with high mortality rates [1,2]. Candida albicans is the most common cause of candidiasis. However, the frequent use of azole antifungals has led to the emergence of candidiasis caused by non-albicans Candida species, which display lower susceptibility to azoles [3,4]. Among these non-albicans Candida species, Candida glabrata is the first or second leading cause of candidemia in many countries [5][6][7]. The virulence mechanisms of C. glabrata are different from those of C. albicans; the most crucial difference is that C. glabrata does not cause significant damage to the host cell and does not provoke a strong response by the host immune system [8]. Treatment of C. glabrata infections is made difficult by the limited knowledge of its pathogenicity and its low susceptibility to azoles. Therefore, a better understanding of the mechanisms underlying azole resistance is critical for the treatment of C. glabrata.
Azole antifungals selectively inhibit 14α-lanosterol demethylase (Erg11) in ergosterol biosynthesis, leading to a depletion of ergosterol and an accumulation of the toxic sterol dimethylcholesta-8,24(28)-dien-3β,6α-diol [9]. In C. glabrata, azole resistance in clinical isolates is mainly caused by activating mutations in the transcription factor Pdr1, resulting in the overexpression of multi-drug transporters of the ATP-binding cassette (ABC) family, such as Cdr1 and Cdr2 [6,[10][11][12][13]. Upregulation of ABC transporters by Pdr1 has also been reported in mitochondrial dysfunction, leading to azole resistance [11,14]. Mutations in the mitochondrial genome or loss of mitochondria cause the so-called petite phenotype, observed in clinical isolates obtained from patients treated with azoles [15,16] or upon in vitro exposure to high concentrations of azoles [11]. Azole exposure induces a temporary loss of mitochondrial function [17]. However, it is unclear how azoles cause mitochondrial dysfunction and what molecular mechanisms of mitochondrial dysfunction contribute to the azole resistance mediated by ABC transporters. Mitochondrial fission and fusion are crucial for mitochondrial functioning. Mutants with defects in mitochondrial fission have been found to be azole-resistant in Aspergillus fumigatus [18], and deletion of CgSHE9, which is involved in mitochondrial inner membrane fission, has been reported to confer resistance to fluconazole in C. glabrata [17]. Therefore, it is important to clarify the mechanism through which mitochondrial dysfunction caused by defects in mitochondrial fission leads to azole resistance.
It has become clear that different organelles are in physical contact with each other and exchange substances and information through their contact sites [19,20]. In Saccharomyces cerevisiae, the outer mitochondrial membrane is connected to the endoplasmic reticulum (ER) through the contact site formed by the ER-mitochondria encounter structure (ERMES) complex [21]. The ERMES complex is composed of four core subunits: an ER-resident protein Mmm1, a cytosolic protein Mdm12, and mitochondrial outer membrane proteins Mdm10 and Mdm34 [21] ( Figure 1A). Gem1, a Rho GTPase of the outer mitochondrial membrane, was recently identified as a subunit of the ERMES complex and is suggested to be involved in the regulation of its function [22,23]. Mdm12, Mdm34, and Mmm1 have synaptotagmin-like mitochondrial-lipid-binding protein (SMP) domains [24], and the ERMES complex mediates lipid transport [25,26]. The ERMES complex is also involved in mitochondrial fission, the distribution of mtDNA, and mitophagy [27][28][29][30][31][32][33].
The ERMES complex was found to be involved in C. albicans and A. fumigatus pathogenicity, suggesting its potential as a target for new antifungal drugs [34][35][36]. Similar to S. cerevisiae, inactivation of the ERMES complex results in the disruption of mitochondrial tubular morphology in these pathogens. In C. albicans, the ERMES complex is involved in immune system evasion [35] and cell wall stress responses [34]. Recently, the yeast Gem1 homolog was isolated as GemA in A. fumigatus and was shown to be required for azole susceptibility, hyphal growth, virulence, and cell wall integrity [37].
In the present study, the link between the ERMES complex and azole resistance was investigated in C. glabrata. Upon the discovery that GEM1 deletion leads to azole resistance, its impact on mitochondrial morphology, Reactive Oxygen Species (ROS) accumulation, and the activation of azole drug efflux pumps was evaluated, providing new clues on the mechanisms of azole resistance associated with mitochondrial dysfunction. Growth of strains lacking ERMES components (∆mdm12, ∆mdm34, ∆mdm10, ∆mmm1, and ∆gem1) in the presence or absence of fluconazole. The cells were diluted to OD600 (optical density at 600 nm) of 0.5 in water and spotted in 4-fold serial dilutions (indicated by triangles) on agar plates of minimal SD medium containing the indicated concentration of fluconazole. All cells except for ∆mmm1 cells were incubated for 2 days (upper panel), and ∆mmm1 cells were incubated for 3 days at 30 °C or 37 °C. (C) Growth of strains lacking ERMES components on the plate containing glycerol as the carbon source. The cells were spotted in 4-fold serial dilutions on the minimum medium containing glucose or 3% glycerol and incubated for 3 days at 30 °C. Growth of strains lacking ERMES components (∆mdm12, ∆mdm34, ∆mdm10, ∆mmm1, and ∆gem1) in the presence or absence of fluconazole. The cells were diluted to OD 600 (optical density at 600 nm) of 0.5 in water and spotted in 4-fold serial dilutions (indicated by triangles) on agar plates of minimal SD medium containing the indicated concentration of fluconazole. All cells except for ∆mmm1 cells were incubated for 2 days (upper panel), and ∆mmm1 cells were incubated for 3 days at 30 • C or 37 • C. (C) Growth of strains lacking ERMES components on the plate containing glycerol as the carbon source. The cells were spotted in 4-fold serial dilutions on the minimum medium containing glucose or 3% glycerol and incubated for 3 days at 30 • C.

Strains and Media
The yeast strains used in the present study are listed in Table S1. Yeast cells were grown in rich medium (YPD; 2% peptone, 1% yeast extract, and 2% glucose) or minimal medium (SD; 0.17% yeast nitrogen base without amino acids and ammonium sulfate, 2% glucose, 5% ammonium sulfate, and appropriate amino acids) at 30 • C or 37 • C.

Quantitative RT-PCR
Cells were cultivated in SD minimal medium at 37 • C until the exponential phase was reached, collected by centrifugation, and washed twice with sterile distilled water at 4 • C. Total RNA was extracted using ISOGEN (Nippon Gene, Tokyo, Japan), and cDNA was synthesized using ReverTra Ace and random primers (Toyobo, Osaka, Japan). The amount of RNA for CDR1, CDR2, or PDR1 was determined by quantitative real-time PCR (qRT-PCR) on a LightCycler ® 96 System (Roche Diagnostics, Mannheim, Germany) with SYBR Green detection using the Thunderbird SYBR qPCR mix (Toyobo). The transcription levels were normalized to those of TEF1, a housekeeping gene that encodes an elongation factor 1. The following primer pairs were used; PDR1-F/PDR1-R, CDR1-F/CDR1-R, CDR2-F/CDR2-R, and TEF1-F/TEF1-R (Table S2). PCR conditions were as follows: pre-denaturation at 95 • C for 1 min, followed by 40 cycles of denaturation at 95 • C for 15 s, and annealing/extension at 60 • C for 1 min.

Mitochondrial Reactive Oxygen Species Quantification
Mitochondrial Reactive Oxygen Species (mtROS) production was monitored by staining with the MitoTracker Red CM-H2XROS (Thermo Fisher Scientific, Waltham, MA, USA).
The cells were incubated with 0.5 µM MitoTracker Red CM-H2XROS for 20 min in the dark, washed twice with fresh medium, and then observed under a BZ-9000 microscope (Keyence, Osaka, Japan) equipped with a 100× oil-immersion objective lens.

CgGem1 Is Involved in Azole Susceptibility
In C. glabrata, the predicted components of the ERMES complex, based on their homology to their S. cerevisiae counterparts, are encoded by CAGL0C02695g (MDM10), CAGL0E02365g (MDM12), CAGL0I07007g (MDM34), CAGL0D05698g (MMM1), and CAGL0M12276g (GEM1). To examine the involvement of the ERMES complex in the susceptibility of C. glabrata to azole drugs, each ERMES gene was individually deleted, and the resulting strains were tested for fluconazole susceptibility. The deletion of MDM12, MDM34, or MMM1 led to growth defects in the SD medium at 30 • C and was more pronounced at 37 • C ( Figure 1B). Deletion of GEM1 or MDM10 genes also led to a slight reduction in growth compared to the wild-type strains. Notably, ∆gem1 cells exhibited reduced susceptibility to fluconazole compared to wild-type cells at 30 • C and 37 • C, suggesting that Gem1p is involved in fluconazole-induced growth inhibition. The deletion of the remaining ERMES components resulted in decreased growth by fluconazole, similar to that of wild-type strains. Cells lacking GEM1, MDM12, or MDM34 grew more slowly on the minimum. The medium contains the nonfermentable carbon source glycerol, suggesting a mitochondrial dysfunction ( Figure 1C).
To further examine whether fluconazole resistance induced by the deletion of GEM1 is dependent on other ERMES components, we constructed double deletion mutants of GEM1 and each of the genes encoding ERMES subunits, and investigated their growth in the presence of fluconazole. Deletion of MDM10, MDM12, or MDM34 in ∆gem1 cells led to the abrogation of the azole resistance phenotype observed in the ∆gem1 single mutant ( Figure 1B). The ∆gem1 cells were resistant to another azole, ketoconazole, and the resistance was dependent on other ERMES components (Supplemental Figure S1). These results indicate that the effect of Gem1 on azole susceptibility requires a functional ERMES complex.
To evaluate the effect of Gem1 and Mdm34 on mitochondrial morphology in C. glabrata, the mitochondria-specific dye MitoBright LT Red was used ( Figure 2). Under the selected conditions, the wild-type cells displayed a branched tubular mitochondrial network. However, in the absence of GEM1 mitochondrial morphology, the ∆gem1 cells displayed shortened or collapsed tubular mitochondrial networks. In ∆mdm34 cell mitochondria, morphological abnormalities were even clearer, and these cells displayed mostly globular mitochondrial morphology. Furthermore, ∆mdm34 cells displaying no mitochondria-specific signals upon staining were often observed. Similar mitochondrial morphological abnormalities have been reported in ∆gem1 and ∆mdm34 cells of S. cerevisiae [21,27]. These results suggested that the ERMES complex is required for normal mitochondrial morphology in C. glabrata. Furthermore, mitochondrial morphological abnormalities in the ∆gem1∆mdm34 double deletion mutant were found to be more similar to ∆mdm34 cells than to ∆gem1 cells, indicating that the ERMES complex might still be partially functional in ∆gem1 cells but not in the absence of its core components, such as Mdm34.

The Azole Resistance Phenotype of ∆gem1 Cells Depends on Pdr1-Mediated CDR1 Upregulation
In C. glabrata, petite mutants with mitochondrial DNA (mtDNA) deficiency display increased resistance to azoles, which is associated with the upregulation of drug efflux pumps of the ABC superfamily [14]. Since abnormal mitochondrial morphology was observed in ∆gem1 cells, we investigated whether the azole resistance phenotype exhibited by ∆gem1 cells was mediated by the upregulation of azole drug efflux pumps. The mRNA levels of CDR1 and CDR2, which encode the major azole resistance drug efflux pumps, were assessed using quantitative RT-PCR in wild-type, ∆gem1, and ∆mdm12 cells cultivated in an SD medium. Interestingly, the expression of CDR1 was upregulated in both ∆gem1 and ∆mdm12 cells but was approximately 4-fold higher in ∆gem1 cells than in ∆mdm12 cells ( Figure 3A). The expression of CDR2 was also upregulated in both ∆gem1 and ∆mdm12 cells but was slightly higher in ∆gem1 cells than in ∆mdm12 cells. In addition, the mRNA levels of the transcription factor encoding PDR1 were also evaluated and found to be virtually unaffected by the deletion of GEM1 or MDM12. These results suggest that in ∆gem1 cells, the upregulation of CDR1 or CDR2 is not caused by increased transcription of PDR1 but likely through the activation of the encoding transcription factor, even in the absence of azole drugs.

The Azole Resistance Phenotype of ∆gem1 Cells Depends on Pdr1-Mediated CDR1 Upregulation
In C. glabrata, petite mutants with mitochondrial DNA (mtDNA) deficiency display increased resistance to azoles, which is associated with the upregulation of drug efflux pumps of the ABC superfamily [14]. Since abnormal mitochondrial morphology was observed in ∆gem1 cells, we investigated whether the azole resistance phenotype exhibited by ∆gem1 cells was mediated by the upregulation of azole drug efflux pumps. The mRNA levels of CDR1 and CDR2, which encode the major azole resistance drug efflux pumps, were assessed using quantitative RT-PCR in wild-type, ∆gem1, and ∆mdm12 cells cultivated in an SD medium. Interestingly, the expression of CDR1 was upregulated in both ∆gem1 and ∆mdm12 cells but was approximately 4-fold higher in ∆gem1 cells than in ∆mdm12 cells ( Figure 3A). The expression of CDR2 was also upregulated in both ∆gem1 and ∆mdm12 cells but was slightly higher in ∆gem1 cells than in ∆mdm12 cells. In addition, the mRNA levels of the transcription factor encoding PDR1 were also evaluated and found to be virtually unaffected by the deletion of GEM1 or MDM12. These results suggest that in ∆gem1 cells, the upregulation of CDR1 or CDR2 is not caused by increased transcription of PDR1 but likely through the activation of the encoding transcription factor, even in the absence of azole drugs.
Furthermore, to assess the eventual impact of the PDR genes on the azole resistance phenotype exhibited by ∆gem1 cells, double deletion mutants devoid of GEM1 and each PDR gene were constructed. Under the selected conditions, the single deletion of CDR1 and PDR1, but not of CDR2, led to fluconazole susceptibility. However, the azole resistance of ∆gem1 cells was suppressed by the deletion of CDR1 or PDR1, indicating that the azole-resistant phenotype in ∆gem1 cells is dependent on the PDR1-mediated upregulation of CDR1. Furthermore, the deletion of GEM1 led to increased azole sensitivity in ∆cdr1 or ∆pdr1 cells.

The GTPase Activity of Gem1 Is Required for Its Interaction with the ERMES Complex and Azole Susceptibility
C. glabrata Gem1 was predicted to contain two GTPase domains, two calciumbinding EF-hands, and a transmembrane domain ( Figure 4A). In S. cerevisiae, GTP hydrolysis by Gem1 is decreased by the amino acid substitution of serine S19 of the first GTPase domain or S462 of the second domain; and by the amino acid substitution of glutamic acid E225 of the first EF-hand domain or E354 of the second EF-hand domain [42]. To investigate whether these Gem1 domains are required in C. glabrata, site-directed mutagenesis was used to construct T19A, E216A, E344A, or S452A gem1 point mutants ( Figure 4B). These mutated gem1 genes were fused with a green fluorescent protein (GFP)encoding gene and introduced into a CEN/ARS-based low-copy plasmid. The plasmid expressing GFP-Gem1 restored the growth defect of ∆gem1 cells, suggesting that this fusion protein is functional (Supplemental Figure S2). To evaluate whether the Furthermore, to assess the eventual impact of the PDR genes on the azole resistance phenotype exhibited by ∆gem1 cells, double deletion mutants devoid of GEM1 and each PDR gene were constructed. Under the selected conditions, the single deletion of CDR1 and PDR1, but not of CDR2, led to fluconazole susceptibility. However, the azole resistance of ∆gem1 cells was suppressed by the deletion of CDR1 or PDR1, indicating that the azoleresistant phenotype in ∆gem1 cells is dependent on the PDR1-mediated upregulation of CDR1. Furthermore, the deletion of GEM1 led to increased azole sensitivity in ∆cdr1 or ∆pdr1 cells.
3.3. The GTPase Activity of Gem1 Is Required for Its Interaction with the ERMES Complex and Azole Susceptibility C. glabrata Gem1 was predicted to contain two GTPase domains, two calcium-binding EF-hands, and a transmembrane domain ( Figure 4A). In S. cerevisiae, GTP hydrolysis by Gem1 is decreased by the amino acid substitution of serine S19 of the first GTPase domain or S462 of the second domain; and by the amino acid substitution of glutamic acid E225 of the first EF-hand domain or E354 of the second EF-hand domain [42]. To investigate whether these Gem1 domains are required in C. glabrata, site-directed mutagenesis was used to construct T19A, E216A, E344A, or S452A gem1 point mutants ( Figure 4B). These mutated gem1 genes were fused with a green fluorescent protein (GFP)-encoding gene and introduced into a CEN/ARS-based low-copy plasmid. The plasmid expressing GFP-Gem1 restored the growth defect of ∆gem1 cells, suggesting that this fusion protein is functional (Supplemental Figure S2). To evaluate whether the constructed Gem1 mutated proteins interacted with the core components of the ERMES complex, we observed, by confocal microscopy, their intracellular localization in cells expressing mCherry-tagged Mdm34, used as an ERMES marker. The wild-type GFP-Gem1 always co-localizes to the foci containing Mdm34-mCherry, consistent with their co-existence as subunits of the ERMES complex ( Figure 4C). GFP-Gem1(S452A) showed stable dot-like colocalization with Mdm34-mCherry, suggesting that the S452A mutation did not affect the Gem1-ERMES interaction ( Figure 4C). The fluorescence of GFP-Gem1 (E344A) was detected predominantly in the region where Mdm34 was localized. In contrast, GFP-Gem1(T19A) and GFP-Gem1 (E216A) were not always co-localized with Mdm34-mCherry ( Figure 4C, black arrowhead), suggesting that these mutations hamper, but do not fully prevent, the interaction between Gem1 and the ERMES complex.
Next, we examined the effect of each mutation on the azole resistance in ∆gem1 cells. As expected, the transformation of GEM1 into ∆gem1 cells complemented the azolesusceptible phenotype ( Figure 4D). Similarly, the expression of gem1 (E216A) and gem1 (A344A), which are GEM1-bearing mutations in the EF-hand motifs, restored fluconazole susceptibility in ∆gem1 cells ( Figure 4D). In contrast, the expression of gem1 (T19A) and gem1 (S452A), which are GEM1-bearing mutations in the GTPase domains, did not complement the azole resistance phenotype exhibited by ∆gem1 cells. Taken together, these results suggest that the GTPase activity of Gem1 is required for ERMES-mediated azole susceptibility in C. glabrata.

Gem1-Dependent Mitochondrial ROS Concentration Affects Azole Susceptibility
Mitochondria are the main generators of reactive oxide species (ROS). ROS are toxic to cells but are also important signaling molecules. Therefore, we investigated whether mitochondrial morphological abnormalities induced by the deletion of GEM1 affected ROS concentration in the mitochondria. To detect mitochondrial ROS (mtROS), we used MitoTracker Red CM-H 2 XROS, a reduced dye that fluoresces upon oxidation. Interestingly, stronger fluorescent signals are observed in ∆gem1 cells than in wild-type cells ( Figure 5A). We also noticed that some ∆gem1 cells were not stained with MitoTracker Red CM-H 2 XROS ( Figure 5A). To confirm that the stronger signals in ∆gem1 cells were due to increased mtROS production, we treated ∆gem1 cells with N-acetylcysteine (NAC), an antioxidant known to reduce ROS generation. Treatment with NAC for 24 or 40 h resulted in a significant reduction of fluorescent signals in ∆gem1 cells compared to untreated cells ( Figure 5B). These results suggest that the absence of GEM1 contributes to increasing ROS concentrations, likely as a consequence of mitochondrial dysfunction.
did not affect the Gem1-ERMES interaction ( Figure 4C). The fluorescence of GFP-Gem1 (E344A) was detected predominantly in the region where Mdm34 was localized. In contrast, GFP-Gem1(T19A) and GFP-Gem1 (E216A) were not always co-localized with Mdm34-mCherry ( Figure 4C, black arrowhead), suggesting that these mutations hamper, but do not fully prevent, the interaction between Gem1 and the ERMES complex.

Discussion
Mitochondrial dysfunction is an important factor underlying azole resistance in C. glabrata [11,14,15]. Recent studies have shown that azole drugs accumulate in To investigate whether ROS production in ∆gem1 cells is responsible for the elevated expression of CDR1, we examined mRNA levels of CDR1 using quantitative RT-PCR. We found that treatment with NAC, which reduces ROS production in ∆gem1 cells, clearly decreased the expression of CDR1 compared to untreated cells ( Figure 5C). This result suggested that ROS production in ∆gem1 cells plays a role in the increased expression of CDR1.

Discussion
Mitochondrial dysfunction is an important factor underlying azole resistance in C. glabrata [11,14,15]. Recent studies have shown that azole drugs accumulate in mitochondria, at least upon contact with C. albicans cells [43]. However, the molecular mechanisms connecting the azole mode of action, mitochondrial activity and/or dysfunction, and azole resistance are unclear. In a study on A. fumigatus, mutants with defects in mitochondrial fission were reported to be azole-resistant [18]; therefore, we focused on the association between mitochondrial dynamics and azole resistance. As the ERMES plays an important role in mitochondrial fission [33], the impact of each subunit on azole resistance was evaluated. Interestingly, only a single deletion of GEM1 affected azole resistance; however, as azole resistance was suppressed in mutants with double deletions of GEM1 and additional ERMES components ( Figure 1B), the effect of GEM1-deletion on azole resistance depended on the remaining components. ∆gem1 cells showed abnormal mitochondrial morphology, increased mtROS, and upregulation of the drug efflux pump encoded by CDR1.
Mitochondria are dynamic organelles, and their continuous fission and fusion are necessary for the maintenance of mitochondrial function. The fission site of mitochondria is wrapped around by ER tubules [44], where the ERMES complex localizes [33]. At the fission site of mitochondria, the ERMES complex localizes adjacent to replicating mtDNA and is involved in the distribution of newly replicated mtDNA to the dividing mitochondria [33]. After mitochondrial fission, one of the dividing mitochondria is detached from the ER, a process that requires Gem1 in S. cerevisiae [33]. In particular, deletion mutants of the ERMES core complex display defects in the segregation of mtDNA into daughter cells, resulting in the loss of mitochondria over several generations of growth in S. cerevisiae [27,29,30,45]. Similarly, in C. glabrata, some cells losing mitochondria were observed in ∆mdm34 (Figure 2A), suggesting that the deletion of MDM34 mainly induces abnormal mtDNA distribution. The deletion of the ERMES core complex did not result in azole resistance ( Figure 1B); defects in mtDNA distribution might not be associated with azole resistance. In contrast, the deletion of GEM1 conferred resistance to azoles ( Figure 1B). C. glabrata mitochondria in ∆gem1 cells had a morphology similar to that observed in S. cerevisiae ∆gem1 cells (Figure 2A), suggesting that CgGem1 also contributes to the disruption of contact between the ER and mitochondria after mitochondrial fission. We found that azole resistance in ∆gem1 cells was suppressed by the deletion of the ERMES core complex, Mdm10, Mdm12, or Mdm34 ( Figure 1B). Deletion of any four proteins of the ERMES core complex results in the impaired formation of the ERMES complex and instability of ERmitochondrial contact [21]. Therefore, the ERMES complex-mediated contact between the ER and mitochondria is a crucial factor in the acquisition of azole resistance in C. glabrata.
Gem1 contains two GTPase domains and two Ca 2+ -binding EF-hand domains. In S. cerevisiae, mutations leading to decreased GTPase activity, Gem1(S19N) or Gem1(S462N), impair the maintenance of mitochondrial morphology and dissociation of ERMES from the fission site [33]. Significantly, we showed that similar amino acid substitutions in the GTPase domains of C. glabrata Gem1 were also sufficient to confer azole resistance ( Figure 4D). The fact that single mutations in GEM1 are sufficient to increase azole resistance suggests that GEM1 mutations may confer azole resistance in C. glabrata clinical isolates.
In ∆gem1 cells, the mRNA expression of the efflux pump-encoding genes CDR1 and CDR2 was found to be upregulated ( Figure 3A). Furthermore, the deletion of CDR1 or PDR1 suppressed the azole-resistant phenotype of ∆gem1 cells. (Figure 3B). These results indicate that the deletion of GEM1 leads to azole resistance by upregulating the expression of CDR1 via Pdr1 activation. The deletion of GEM increased the sensitivity of ∆cdr1 or ∆pdr1 to azole. It supposes that Pdr1 and Cdr1 are required for mtROS processing and that excess mtROS induced by GEM1 deletion reduces cell proliferation. Consequently, the deletion of GEM1 in ∆pdr1 or ∆cdr1 cells indicated an additive effect on fluconazole sensitivity. Double deletion of CDR2 and GEM1 indicated greater azole resistance than deletion of CDR2 alone; thus, CDR2 seems to have little effect on the mtROS processing. We also found that mtROS production is enhanced in ∆gem1 cells ( Figure 5A). The treatment with NAC, an antioxidant, reduced the mtROS production in ∆gem1 cells ( Figure 5B) and inhibited the upregulating of the expression of CDR1 (Figure 5C), implying that mtROS may be involved in the activation of Pdr1. Therefore, we propose that in ∆gem1 cells, increased mtROS production caused by mitochondrial dysfunction directly or indirectly activates Pdr1 and upregulates the drug efflux pump Cdr1, leading to azole resistance. An increase in intracellular ROS levels causes oxidative stress. In Kluyveromyces lactis, KIUpc2, whose homolog activates Pdr1 in response to fluconazole in C. glabrata [46], has been reported to be involved in the oxidative stress response [47]. In C. glabrata, Upc2 may activate Pdr1 in response to oxidative stress. However, it is not known whether oxidative stress or mtROS regulates the activation of Pdr1, and further analysis of ∆gem1 cells is needed.
As excess ROS is toxic to cells, cells must remove excess ROS and maintain a low concentration of ROS. We observed some cells that were not stained or very weakly stained with MitoTracker Red CM-H 2 XROS in ∆gem1 cells ( Figure 5A). No such cells were observed in the wild-type cells. The cells with reduced mtROS in ∆gem1 cells imply that the antioxidant system, such as scavenging excess mtROS, may work actively in ∆gem1 cells. Mitochondrial fission has been reported to be involved in the release of cytochrome C from mitochondria, which acts as a scavenger of ROS in the cytosol [48]. Analysis of ∆gem1 cells may reveal the mechanism by which excess mtROS is scavenged.
Altogether, the results described herein suggest that Gem1 is required to maintain mitochondrial dynamics and redox state in C. glabrata. In the absence of Gem1 activity, caused either by gene deletion or by point mutations in its GTPase domains, the mitochondrial ROS concentration increases, leading to Pdr1-dependent upregulation of the drug efflux pumps Cdr1 and Cdr2, resulting in azole resistance. This knowledge is expected to contribute to the development of much-needed antifungal molecules that target mitochondrial-related azole drug resistance.