A Crucial Role for Ergosterol in Plasma Membrane Composition, Localisation, and Activity of Cdr1p and H+-ATPase in Candida albicans

Candida albicans is an opportunistic fungal pathogen of humans. Treatment of C. albicans infections relies on azoles, which target the lanosterol 14α-demethylase (Erg11p) encoded by the ERG11 gene. Our results show that targeted gene disruption of ERG11 can result in resistance to ergosterol-dependent drugs (azoles and amphotericin B), auxotrophy and aerobically viable erg11Δ/Δ cells. Abnormal sterol deposition and lack of ergosterol in the erg11Δ/Δ strain leads to reduced plasma membrane (PM) fluidity, as well as dysfunction of the vacuolar and mitochondrial membranes, resulting respectively in defects in vacuole fusion and a reduced intracellular ATP level. The altered PM structure of the erg11Δ/Δ strain contributes to delocalisation of H+-ATPase and the Cdr1 efflux pump from the PM to vacuoles and, resulting in a decrease in PM potential (Δψ) and increased sensitivity to ergosterol-independent xenobiotics. This new insight into intracellular processes under Erg11p inhibition may lead to a better understanding of the indirect effects of azoles on C. albicans cells and the development of new treatment strategies for resistant infections.


Introduction
The fungal microorganism Candida albicans is a part of the healthy mucosal surface microbiota of around 50% of the human population [1]. It is also the cause of fungal infections in millions of immunologically deficient individuals worldwide, causing candidiasis of oral, gastrointestinal, and vaginal surfaces, as well as candidemia or invasive candidiasis [2]. Current treatment of candidiasis involves the use of antifungal agents, such as polyenes and azoles. Both types of antifungals cause disturbances in the structure and functioning of the plasma membrane, either by binding ergosterol (polyenes, e.g., amphotericin B) or by inhibiting the enzyme lanosterol 14α-demethylase, which is involved in ergosterol biosynthesis (azoles, e.g., fluconazole) [3].
Ergosterol is a neutral lipid of fungal membranes, and is critical for many cellular processes [4], hence disruption of its synthesis has become a focus of antifungal therapies, particularly as the prevalence of Candida resistance to many antifungal drugs increases. The lanosterol 14α-demethylase gene (ERG11) encodes a member of the cytochrome P450 family of enzymes that converts lanosterol to ergosterol and is therefore essential for the synthesis of ergosterol [5]. Alterations to the target enzyme, either via overexpression or point mutations of the ERG11 gene, are a cause of growing azole resistance in C. albicans [6]. Another mechanism for the development of azole resistance involves overexpression of genes encoding multidrug resistance (MDR) transporters, which pump out drugs from the cell. There are three MDR transporters involved in azole efflux in C. albicans: Cdr1p and For specific experiments, cells were grown until they reached either early (8 h), late (14 h) logarithmic or stationary (24 h) phase. Cells were centrifuged (4500 rcf, 5 min.), washed twice (4500 rcf, 5 min.) with either phosphate-buffered saline (PBS), H 2 O dd or citrate-phosphate (CP) buffer (pH 6.0), and resuspended in either PBS, H 2 O dd or CP to the indicated OD 600 .

Plasmids and Strains Construction
Plasmid pSFS5 was a generous gift from Prof. J. Morschhäuser (Würzburg, Germany).

Real Time PCR
The RNA isolation, cDNA synthesis and calculations of gene expression levels were prepared as previously described [10].
The following gene-specific primers were used:

Sterol Analysis
Sterol analysis was performed according to the method proposed by [15]. To the evaporated lipid samples, 0.5 mL CHCl 3 , 0.5 mL MetOH-KOH (0.6 M) and 20 µg of cholesterol were added. After 1 h of incubation (23 • C) 0.325 mL 1M HCl and 0.125 mL H 2 O dd were added and centrifuged (5000 rcf; 10 • C; 5 min). The lower layer was transferred to fresh tubes, dried and 100 µL BSTFA/TMCS was added. Samples were heated at 85 • C for 90 min, then 50 µL hexane was introduced to tubes and vortexed. The analysis was performed with a gas chromatograph (Agilent 7890) equipped with column HP 5 MS (30 m × 0.25 mm inner diameter, i.d. × 0.25 mm film thickness, f.t.) and a 5975C Mass Detector. The column was maintained at 100 • C for 0.5 min −1 , then increased to 240 • C at a rate of 25 • C min −1 , and finally to 300 • C at a rate of 3 • C min −1 (for 5 min) with helium as a carrier gas at a flow rate of 1 mL·min −1 [15]. The injection port temperature was 250 • C. Cholesterol was used as an internal standard. Tetramethylsilane (TMS)-derived ergosterol and lanosterol were analysed with reference to retention times and fragmentation spectra for standards. Other sterol TMS ethers were identified by comparison with the NIST database or literature data and quantitated using a standard curve for lanosterol.

Membrane Fluidity Assessment
The assay was based on a Saccharomyces cerevisiae protocol [16], with modifications. KS028 and CAF2-1 suspensions (PBS, OD 600 = 0.1, 3 mL) were labelled with laurdan (final conc. = 5 × 10 −6 M; Microorganisms 2019, 7, 378 5 of 17 20 min.; 25 • C; in darkness). The probes were excited at 366 nm (Ex slit = 10 nm), and fluorescence spectra were recorded at 400-550 nm (Em slit = 2.5 nm) (PMT voltage = 400 V) using a fluorescence spectrophotometer equipped with a xenon lamp (HITACHI F-4500; manufacturer: Hitachi, Tokyo, Japan). For the data analysis, modified general polarisation (GP) was calculated as follows: the difference of the sum of fluorescence intensities (IFs) from 425 to 450 nm and the sum from 475 to 525 nm, divided by the sum of IFs from 425 to 450 nm and from 475 to 525 nm.

Proton Extrusion Assay
The method was based on a S. cerevisiae protocol [18], with modifications. Real time acidification of KS028 and CAF2-1 strain suspensions (H 2 O dd ; OD 600 = 1.0; 20 mL) was monitored every 10 s for 12 min using a pH-meter (Eutech Instruments CyberScan PH 5500, ThermoFisher Scientific, Warsaw, Poland), equipped with MiniTrode electrode (manufacturer: Hamilton; distributor: Sigma-Aldrich; Poznań, Poland). Due to exposure to hypotonic conditions, cell suspensions under identical conditions were checked for plasma membrane permeabilisation (using PI, according to the protocol of [13]). In each experiment, pH values at t 0 were equal to 5.7 ± 0.1. For clearer presentation these have been normalised to 5.7.

Di-4-ANEPPS Assay
KS028 and CAF2-1 suspensions (CP; OD 600 = 0.1; 3 mL) were labelled with di-4-ANEPPS according to our protocol [19]. For data analysis, red-blue signal ratio (R-B ratio) was calculated by dividing the sum of IFs between 580 and 620 nm by the sum of IFs between 540 and 580 nm as described previously [19]. All results were normalised to 1 for the plasma membrane potential of 8 h CAF2-1 cells

Western Blotting
Crude protein extracts from CAF2-1, KS028, YHXW11 and KS045 strains were isolated according to [21], with the following modifications: after TCA precipitation, proteins were washed with ice-cold acetone, dried and resuspended in SB followed by 45 min denaturation at 37 • C. Electrophoretic separation, transfer and immunodetection of Cdr1p (CAF2-1, KS028) was performed following [21]. For Pma1p detection, the following modifications were applied: crude proteins from YHXW11 and KS045 strains were separated on 10% SDS-polyacrylamide gels and primary mouse anti-GFP antibodies were used, followed by HRP-conjugated rabbit antimouse antibodies. The remaining steps were performed as in [21].

R6G Efflux Assay
The R6G efflux assay was performed as described previously [21]. IFs were collected 15 min after R6G efflux initiation and normalised to 1 for the efflux activity of 8 h CAF2-1 cells.

Statistical Analysis
For each experiment we performed at least three independent replicates. Statistical significance was determined using Student's t-test (binomial, unpaired).
Cultured in complex YPD medium, the KS028 (erg11∆/∆) strain was viable under aerobic conditions with no differences in the temporal distribution of growth phases from the WT strain ( Figure 1A). However, the growth rate of the C. albicans K028 strain was significantly (p = 8.7 × 10 −5 ) lower than that of the WT strain (t d s = 179.4 ± 4 min and 207.4 ± 9.4 min for WT and KS028 strains, respectively).  The inhibitory effect of azoles on Erg11p is dependent on C. albicans growth phase [24,25]. Thus Cultured on solid complex medium (YPD), the KS028 strain developed visually sparser groupings of colonies than the WT strain ( Figure 1B). The C. albicans KS028 strain displayed an auxotrophic phenotype due to the lack of growth on solid minimal medium (YNB) ( Figure 1B). Therefore, the complex solid medium was chosen for the phenotypic tests with antimycotics. We used complex broth for all other experiments.
The C. albicans strain KS028 showed no growth inhibition within the tested range of amphotericin B (0.125 to 10 µg/mL), fluconazole (0.5 to 128 µg/mL), and ketoconazole (0.05 to 8 µg/mL) ( Figure 1B). Similarly, the resistance of KS028 strain towards azoles and amphotericin B was expressed as no detectable MIC 50 values of these drugs. The growth of the WT strain was, however, fully inhibited by amphotericin B from levels of 0.5 µg/mL (MIC 50 = 0.25 µg/mL). The residual growth of the WT strain was present but limited at concentrations from 1 µg/mL fluconazole and 0.05 µg/mL ketoconazole upwards (MIC 50 = 1 and 0.015 µg/mL for fluconazole and ketoconazole, respectively). According to literature, this can be attributed to the fungistatic properties of azoles [23]. The growth of the WT strain treated with azoles (≥1 µg/mL fluconazole and ≥0.05 µg/mL ketoconazole, Figure 1B and data not shown) was slower than that of KS028. On YPD with and without a 0.5 µg/mL fluconazole addition, the growth of both strains was comparable ( Figure 1B).

The C. albicans Strain erg11∆/∆ Accumulates Lanosterol at a Constant Level during Growth, but 14α Methylergosta-8-24(28)dienol at Increasing Level due to High Expression of ERG3 Gene
The inhibitory effect of azoles on Erg11p is dependent on C. albicans growth phase [24,25]. Thus we examined the physiology of erg11∆/∆ knock-outs throughout the process of cell aging. The following time points were chosen: early and late logarithmic phase (8 and 14 h, respectively), and stationary phase (24 h).
ERG11 and ERG3 gene expression in C. albicans WT after 14 and 24 h culture was considerably lower than after 8 h culture ( Figure 2). After 8 h culture, the ERG11 knockout strain showed four-fold higher expression of ERG3 than the WT strain. After 14 h culture, in C. albicans KS028, the expression of ERG3 was less than half that of early logarithmic phase (8 h), and remained at a comparable level after 24 h ( Figure 2B). In the WT strain only, we compared 2 −∆∆CT values for the expression of the ERG3 and ERG11 genes. After 8, 14, and 24 h growth, expression of the ERG3 gene was 2, 4, and 1.5-fold lower than the expression of the ERG11 gene, respectively. Blocking the synthesis of ergosterol by inhibition of the Erg11p results in the accumulation of lanosterol, which is converted by ∆(24)-sterol C-methyltransferase (Erg6p) into eburicol, then by Δ5,6-desaturase (Erg3p) into 14α methylergosta-8-24(28)dienol [26]. In the strain with uninterrupted ergosterol synthesis, we observed the highest ergosterol accumulation during logarithmic growth (14 h). After 24 h, we observed almost a 10-fold decrease in ergosterol from 14 h levels ( Table 2). As expected, we did not detect 14α methylergosta-8-24(28)dienol in C. albicans WT. The level of lanosterol decreased with the aging of the culture, and after 24 h it was less than half the level of ergosterol (Table Blocking the synthesis of ergosterol by inhibition of the Erg11p results in the accumulation of lanosterol, which is converted by ∆(24)-sterol C-methyltransferase (Erg6p) into eburicol, then by ∆5,6-desaturase (Erg3p) into 14α methylergosta-8-24(28)dienol [26]. In the strain with uninterrupted ergosterol synthesis, we observed the highest ergosterol accumulation during logarithmic growth (14 h). After 24 h, we observed almost a 10-fold decrease in ergosterol from 14 h levels ( Table 2). As expected, we did not detect 14α methylergosta-8-24(28)dienol in C. albicans WT. The level of lanosterol decreased with the aging of the culture, and after 24 h it was less than half the level of ergosterol (Table 2). In the C. albicans KS028 strain lacking the ERG11 gene, the level of lanosterol was stable, regardless of the age of culture, whilst the level of 14α methylergosta-8-24(28)dienol increased with the aging of the culture and after 24 h culture was the highest ( Table 2). In addition, small levels of eburicol, 14α methylfecosterol and 4α methylfecosterol were determined in the strain KS028 (Table S1). Table 2. Sterols (µg/mg dry mass of isolated lipids, means ± SD, n = 3) in C. albicans CAF2-1 (WT) and KS028 (erg11∆/∆) strains. Statistical analysis was performed in accordance to µg/mg values after 8 h of culture (**, p < 0.01; ***, p < 0.001); General polarization values (GP; means ± SD, n = 6) of Laurdan incorporated into plasma membranes of C. albicans CAF2-1 (WT) and KS028 (erg11∆/∆) strains. Statistical analysis was performed in accordance to GP values after 8 h for the WT strain (**, p < 0.01; ***, p < 0.001). The fluidity of the plasma membrane (PM) depends in part on the concentration and type of sterols expressed [27]. Comparing C. albicans WT with KS028 strains, we observed a significantly lower PM fluidity in the strain without ergosterol (Table 2), which indicates the significant role of ergosterol in the construction and functionality of the membrane. In both strains, membrane fluidity decreased with the aging of the culture, and in stationary phase (24 h), was accordingly 4.5-and 2.5-fold lower for the WT and KS028 strains than in late logarithmic phase (14 h) ( Table 2). Ergosterol is the most abundant sterol in cell membranes, being present not only in the plasma membrane (PM), but also in the membranes of other organelles, such as vacuoles or mitochondria [4,28]. In S. cerevisiae, ergosterol deficient mutants have defects, such as in vacuole fusion and mitochondrial morphogenesis [4]. We noticed many small vacuoles in C. albicans KS028 cells by the 14 h of culture in contrast to the WT strain in which this effect was observed only in stationary phase (Figures 3  and 4). Ergosterol deficiency, caused by the action of azoles on C. albicans, contributes to the abnormal functioning of these organelles and a reduced level of intracellular ATP [29]. Comparing the amount of ATP in C. albicans WT and KS028, we found a slightly lower level of ATP in the non-ergosterol strain at the late logarithmic phase (14 h) and almost five-fold lower levels of ATP in this strain in stationary phase (24 h) ( Table 3).
The lack of ergosterol in the PM as well as a reduced level of ATP in the cell may cause, as demonstrated previously in Candida glabrata, the abnormal functioning of H + -ATPase [30], an enzyme responsible for maintaining membrane potential [31]. In C. albicans KS045 strain, which lacks ergosterol, and with GFP-labelled H + -ATPase, we observed the delocalisation of Pma1p (H + -ATPase) from the PM to the vacuole after 8 h of culture in contrast to the WT strain, in which Pma1p after 8 h of strain growth was still localised in the PM ( Figure 3A). In both tested strains, Pma1p was also present in the correct place of its functioning, i.e., in the PM after up to 24 h of culturing ( Figure 3A).   (8,14, and 24 h). The samples were resolved on 10% SDS-PAGE gel and probed by an anti-GFP antibody. Ponceau S staining was used as the loading control. Experiment is a representative of three independent assays and the presented conditions were resolved in the same gel and cut out into separate lines ( Figure S3A). (E) Relative PMA1 gene expression in WT (CAF2-1) and KS028 (erg11∆/∆) strains during growth (8,14, and 24 h). Gene expression levels as means of 2 −∆∆CT values (n = 6) ± SD; normalised to 1 for 8 h WT gene expression level. Statistical analysis in (C) and (E) compared data during different growth phases, and between C. albicans WT and KS028 (erg11∆/∆) strains (*, p < 0.05; **, p < 0.01; ***, p < 0.001). expression in KS028, in contrast to the WT strain in which we did not observe such processes ( Figure  4B,C). Despite the additional level of Cdr1p in the KS028 strain, the activity of this transporter measured by the standard R6G efflux test was more than half of that of the WT strain ( Figure 4D). In contrast to the WT strain, C. albicans KS028 showed high sensitivity to Brefeldin A and Fluphenazine, which are substrates for Cdr1p and have toxicity mechanisms not related to membrane interaction [34] ( Figure 4E). This result indicates a disturbed Cdr1p activity and confirms the reduced activity of the   The activity of H + -ATPase in real time can be measured by acidification of the cell environment [18]. We observed, together with the aging of the culture, a decrease in the pH of the surroundings of the cells. The medium containing the C. albicans WT strain after 12 min of measurement was about 0.3 pH units lower than the medium of the KS028 strain in the early logarithmic phase of growth (8 h) and respectively about 0.2 pH units lower in the late logarithmic phase of growth (14 h) ( Figure 3B). Differences between strains in acidification of their external environments indicate a significantly reduced H + -ATPase activity in the strain without ergosterol compared to the WT strain. After 14 h of culture, in the KS028 strain, H + -ATPase showed very weak activity ( Figure 3B). This effect might have also contributed to the lower level of Pma1p in the KS028 strain than in the WT strain, regardless of the phase of growth ( Figure 3D). This low level of Pma1p results from decreased PMA1 gene expression ( Figure 3E). In both strains, PMA1 gene expression decreased with cell aging, however after 14rs and 24 h of growth, gene expression was sufficient to maintain a stable level of Pma1p.
The plasma membrane potential in the KS028 strain was lower compared to the WT strain, but it was not as reduced, as indicated by the weak H + -ATPase activity measured by acidification of the environment outside the cells ( Figure 3C).

In C. albicans Strain erg11∆/∆, Cdr1p Is Still Synthesised during Growth but Has Reduced Activity and Is Rapidly Delocalised to the Vacuole
Ergosterol, as a sterol present in fungi and absent in mammalian cells, is a target for antifungal drugs such as azoles [32]. Azoles, especially fluconazole, are removed from C. albicans cells by PM-localised ABC and MFS transporters [33]. Any lack or deficiency of ergosterol in PM may result in dysfunction of membrane proteins, including ABC transporters through their delocalisation to the inside of the cell. As in the case of Pma1 (Figure 3A), in a ergosterol knockout strain we observed the delocalisation of Cdr1p from the PM to the vacuole by the early logarithmic phase of growth (8 h) in contrast to the WT strain, which maintained the correct localisation of this protein in the PM after 8 h culture ( Figure 4A). Results obtained in Western blot analysis and Real-time PCR showed that Cdr1p was still synthesised during the growth of KS028. Additionally, the CDR1 gene underwent increased expression in KS028, in contrast to the WT strain in which we did not observe such processes ( Figure 4B,C). Despite the additional level of Cdr1p in the KS028 strain, the activity of this transporter measured by the standard R6G efflux test was more than half of that of the WT strain ( Figure 4D). In contrast to the WT strain, C. albicans KS028 showed high sensitivity to Brefeldin A and Fluphenazine, which are substrates for Cdr1p and have toxicity mechanisms not related to membrane interaction [34] ( Figure 4E). This result indicates a disturbed Cdr1p activity and confirms the reduced activity of the Cdr1 transporter in the strain lacking ergosterol, as measured with the R6G test ( Figure 4D,E).

Discussion and Conclusions
The ERG11 gene is considered essential to C. albicans viability [35], despite Erg11p blockage by azoles having a fungistatic rather than fungicidal effect [23]. Sanglard et al. [36] showed that C. albicans erg11∆/∆ mutants could be produced from either an erg3∆/∆ background or as a result of mitotic recombination after culturing ERG11/erg11∆ heterozygotes in the presence of amphotericin B.
In this work, we show that the targeted gene deletion of ERG11 can be achieved using selection on complex YPD medium (here YPD with nourseothricin). We identified the auxotrophic phenotype of the erg11∆/∆ strain ( Figure 1B, YNB), which is suppressed after adding the combination of adenine and uracil, but not amino acids ( Figure S1). Thus the potential for selecting erg11 transformants on minimal media could have previously been restricted.
In contrast to reported erg11∆ mutants of S. cerevisiae and C. glabrata [35,37,38], C. albicans erg11∆/∆ was viable under aerobic conditions ( Figure 1A), most likely due to intraspecies differences in sterol metabolism under aerobic and anaerobic conditions [39]. S. cerevisiae and C. glabrata import exogenous sterols under hypoxia and are characterised by aerobic sterol exclusion, whereas, C. albicans is able to assimilate sterols only under aerobic conditions, and therefore relies mostly on endogenous ergosterol [39].
Due to the lack of target enzyme (Erg11p) and the final product of the pathway (ergosterol), the C. albicans erg11∆/∆ strain was invulnerable towards azoles and amphotericin B (Figure 1B), which is in agreement with data previously published by Sanglard et al. [36]. We observed differences between C. albicans erg11∆/∆ and WT strains in type and level of sterol expression ( Table 2). We observed a four-fold increase in the expression of ERG3 in erg11∆/∆ compared with the WT strain, which is consistent with the conversion of 14α methylfecosterol to 14α methylergosta-8-24(28)dienol [40].
Sterols, especially ergosterol, are necessary for proper PM construction and fluidity, however, our results indicate that PM fluidity also depends on other factors, especially in cells in stationary phase ( Table 2). Qi Y. and co-workers found that, in C. glabrata, the fluidity of the PM as well as the activity of H + -ATPase decreases with decreasing extracellular pH [30]. When cells reach stationary phase, the environment is rich in metabolites, including organic acids, and the pH is lower than at the start of the culture, which may explain the strong decrease in PM fluidity we observed following 24-h culture of both strains (Table 2). Neither the replacement of ergosterol with 14α methylergosta-8-24(28)dienol, nor increased lanosterol in the KS028 strain, restored for the fluidity of the PM to that of the WT strain, therefore the PM in the strain without ergosterol was more rigid (Table 2). Sterols accumulate mostly in the inner leaflet of the PM [42], whereas FM4-64 dye is inserted into the outer leaflet of the PM and, from there, is passed on to vacuolar membranes by endocytosis [43]. We observed partial accumulation of the dye in the PM in the erg11∆/∆ strain, contrary to the WT strain, where the dye was entirely accumulated in the vacuolar membranes ( Figures 3A and 4A). Increased rigidity of the PM in erg11∆/∆ strain most likely increases the duration of outer-to-inner leaflet transport, as shown for the FM4-64 dye.
Despite a lack of significant differences in the level of intracellular ATP between the tested strains during logarithmic growth, the activity of H + -ATPase was much lower in strain KS028 in comparison to the WT strain (Table 3, Figure 3). In the WT strain, PMA1 expression was maintained at a higher level than in a mutant without ergosterol, especially in the early phase of growth ( Figure 3E). Within 24 h of culture, however, the Pma1p level was lower in the KS028 mutant than in the WT strain ( Figure 3D). Despite the lower level of Pma1p in the ERG11 knockout mutant, and the delocalisation of the protein from the PM to the vacuole within 8 h of culture, we recorded a persistent presence of H + -ATPase for 24 h in the PM ( Figure 3A). However, the H + -ATPase activity in late logarithmic phase is very low compared to the WT strain ( Figure 3B). This indicates that the reason for reduced H + -ATPase activity is rather a lack of ergosterol in the PM, and subsequent reduced fluidity of the membrane, thus, the inability to maintain the correct location of the transporter, rather than the lack of energy for H + -ATPase function or a lower intracellular H + -ATPase level.
The upregulation of the ERG11 gene is one of the established mechanisms of C. albicans' resistance to azoles [44]. Another type of azole resistance is their active efflux from cells in which ABC transporters are involved, mainly Cdr1p [45]. The decreased activity of Cdr1p in the ergosterol-free strain may result from a similar H + -ATPase lack of right localization of Cdr1p in the PM (Figure 4A), as well as from decreased PM potential caused by a decrease in H + -ATPase activity.