Heterologous Expression of Full-Length Lanosterol 14α-Demethylases of Prominent Fungal Pathogens Candida albicans and Candida glabrata Provides Tools for Antifungal Discovery

Targeting lanosterol 14α-demethylase (LDM) with azole drugs provides prophylaxis and treatments for superficial and disseminated fungal infections, but cure rates are modest for immunocompromised patients and individuals with comorbidities. The efficacy of azole drugs has also been reduced due to the emergence of drug-resistant fungal pathogens.

ABSTRACT

Targeting lanosterol 14α-demethylase (LDM) with azole drugs provides prophylaxis and treatments for superficial and disseminated fungal infections, but cure rates are modest for immunocompromised patients and individuals with comorbidities. The efficacy of azole drugs has also been reduced due to the emergence of drug-resistant fungal pathogens. We have addressed these problems by expressing in Saccharomyces cerevisiae functional, hexahistidine-tagged, full-length Candida albicans LDM (CaLDM6×His) and Candida glabrata LDM (CgLDM6×His) for drug discovery purposes and determining their X-ray crystal structures. Compared with S. cerevisiae overexpressing LDM6×His (ScLDM6×His), the reduced susceptibility of CgLDM6×His to all azole drugs tested correlated with its level of overexpression. In contrast, the reduced susceptibility to short-tailed (fluconazole and voriconazole) but not medium-tailed (VT-1161) or long-tailed azoles (itraconazole and posaconazole) indicates CaLDM6×His works best when coexpressed with its cognate NADPH-cytochrome P450 reductase (CaNcp1A) rather than the host reductase (ScNcp1). Overexpression of LDM or Ncp1 modified the ergosterol content of yeast and affected growth inhibition by the polyene antibiotic amphotericin B. Affinity-purified recombinant Candida LDMs bind carbon monoxide and show tight type II binding of a range of azole drugs, including itraconazole, posaconazole, fluconazole, and voriconazole. This study provides a practical basis for the phenotype-, biochemistry-, and structure-directed discovery of novel antifungals that target LDMs of fungal pathogens.

INTRODUCTION

The eukaryotic CYP51 (sterol 14α-demethylase) proteins of the endoplasmic reticulum are bitopic, membrane-monospanning, cytochrome P450 monooxygenases that use substrates such as lanosterol, eburicol, 24,25-dihydrolanosterol, and/or obtusifoliol (1, 2). In fungi they are a key enzyme in the ergosterol biosynthetic pathway (3). In humans, cholesterol is the corresponding end product of lanosterol metabolism. Inhibition of lanosterol 14α-demethylase (LDM) depletes fungal cells of ergosterol required for the function of plasma membrane enzymes and ultimately for the sparking reaction needed to initiate growth (4, 5). Inhibition or deletion of fungal LDMs can also result in the accumulation of toxic fecosterol metabolites (6). In contrast, the uptake of sterols downstream of LDM in sterol biosynthesis pathways can render the enzyme nonessential. Humans can obtain these sterols in the diet, and some fungal pathogens may also utilize host sterols, e.g., Candida glabrata can use cholesterol instead of ergosterol under aerobic conditions (7).

LDM is the target of the azole drugs used widely to prevent or treat fungal infections in humans. Life-threatening invasive fungal infections (IFIs) are caused predominantly by opportunistic pathogens that include Candida albicans, Aspergillus fumigatus, Cryptococcus neoformans, Pneumocystis jirovecii, Candida parapsilosis, Candida tropicalis, and Ajellomyces capsulatus and by emerging pathogens such as C. glabrata and Candida krusei (8–10). Multiple generations of azole drugs have resulted in substantial increases in potency, broader spectrum of action, and, in some cases, reduction in side effects caused by interaction with drug-metabolizing cytochrome P450s in the human host (11, 12). Imidazoles such as ketoconazole and prochloraz appeared in the 1970s and were among the first azoles used in the clinic and as agrochemicals, respectively. The imidazoles were replaced in the clinic during the 1990s by the long-tailed triazole itraconazole (ITC) and the short-tailed triazole fluconazole (FLC). The short-tailed triazole voriconazole (VCZ) was introduced in the early 2000s because it is more potent and of broader spectrum than FLC, i.e., it inhibited the growth of organisms such as A. fumigatus that are intrinsically resistant to FLC. However, metabolism of VCZ by liver cytochrome P450 enzymes can require clinical monitoring of serum levels due to drug interactions (13). Azoles that have entered the clinic more recently include the long-tailed triazole posaconazole (PCZ) (14) and the medium-tailed triazole isavuconazole (IVZ) (15), which is administered as its prodrug, isavuconazonium sulfate (16). IVZ is metabolized by CYP3A4 and CYP3A5, and while appearing to have lower toxicity than VCZ, it can be problematic in some patients (12). The medium-tailed tetrazoles VT-1161 (17, 18) and VT-1129 (19, 20) have been designed to minimize interactions with host cytochrome P450s and are in clinical trials. VT-1161 is targeted against superficial infections such as those caused by dermatophytes. VT-1129 is slightly less hydrophobic, shows good activity against C. albicans, C. neoformans, and Cryptococcus gattii, and is more likely to be marketed for use against invasive fungal infections.

The repeated use of azole drugs can be problematic, as they are fungistatic rather than fungicidal, in part due to their ability to interact with transcription factors and induce overexpression of drug efflux pumps (21, 22). This gives rise to drug tolerance that can provide time for genetic change. Low to modest levels of resistance to azole drugs are conferred by mutations in the target LDM, its overexpression due to changes in promoters or aneuploidy, and as a result of mutations that inactivate the enzyme Δ5,6-sterol desaturase, thereby preventing biosynthesis of toxic fecosterols from the eburicol generated on inhibition of LDM (9, 23, 24). High-level resistance is conferred by mutations in transcriptional regulators that enable constitutive overexpression of drug efflux pumps, including members of the major facilitator superfamily (MFS) and ATP-binding cassette (ABC) transporters of the ABCG class in C. albicans and C. glabrata.

Despite improvements in efficacy of azole antifungals, 30 to 55% of patients with IFIs caused by C. albicans and up to 80% of patients with IFIs caused by A. fumigatus die from the infection itself or the sepsis that results (10, 25). Although still relatively uncommon, patients infected with azole-resistant strains of C. albicans or A. fumigatus are more difficult and expensive to treat, often require extended hospitalization, and are less likely to survive the infection (9). Furthermore, the widespread use of azole prophylaxis in intensive care settings over the last 2 decades has increased the incidence of infection with intrinsically azole-resistant strains of fungi, such as C. glabrata, C. krusei, Candida auris, and Rhizopus arrhizus. Azole prophylaxis may also have contributed to the development of cross-resistance by C. glabrata to the structurally unrelated echinocandins that can be due to mutation in a key DNA repair mechanism (26). These considerations highlight the unmet need to identify effective fungicides that selectively target fungal LDM and are not affected by mutations in the target enzyme or subject to drug efflux.

Until recently, the ongoing development of fungal LDM as a drug target was limited by a paucity of high-throughput screens. These include phenotypic and biochemical screens using the drug target overexpressed in a suitable host and computer-based techniques to identify compounds that bind to the target. We have made significant progress toward implementing these approaches by overexpressing constitutively full-length S. cerevisiae LDM (ScLDM) with a C-terminal hexahistidine tag (ScLDM6×His) from the PDR5 locus of hypersensitive yeast strains deleted of multiple drug efflux pumps (2). The functional overexpression of ScLDM6×His has provided robust phenotypic screens not confounded by the presence of drug efflux pumps. Purified ScLDM6×His has yielded high-resolution X-ray crystallographic structures of the full-length enzyme in complex with its substrate lanosterol, the pseudosubstrate estriol, several azole drugs, and, most recently, a range of azole agrochemicals (2, 27–29). It has also provided the first structures of full-length fungal mutant enzymes, including ScLDM6×His Y140F/H mutants in complex with FLC, VCZ, ITC, and PCZ (2, 28, 30). Informative crystal structures have also recently been obtained for the catalytic domains of LDMs from two pathogenic fungi, A. fumigatus and C. albicans. These structures used recombinant enzymes truncated of the N-terminal membrane-associated and transmembrane helices and have involved complexes with VCZ, PCZ, and the experimental drugs (R)-N-(1-(2,4-dichlorophenyl)-2-(1H-imidazol-1-yl)ethyl)-4-(5-phenyl-1,3,4-oxadiazol-2-yl)benzamide (VNI), VT-1161, and VT-1598 (31).

In an effort to develop additional drug discovery tools and as a prelude to protein crystallography, we report the physiological and biochemical characterization of recombinant full-length hexahistidine-tagged C. albicans LDM (CaLDM6×His) and C. glabrata LDM (CgLDM6×His) expressed from the PDR5 locus of S. cerevisiae. The function of CaLDM6×His and CgLDM6×His, detected as reduced susceptibility to short-, medium-, and long-tailed azoles, indicates that CaLDM6×His works best when coexpressed with its cognate NADPH-cytochrome P450 reductase (CaNcp1) rather than the host reductase (ScNcp1). Both affinity-purified recombinant LDMs bind carbon monoxide and show tight type II binding of a range of azole drugs, including ITC, PCZ, FLC, and VCZ.

RESULTS

Biochemical characterization of recombinant LDM6×His and CaNcp1p6×His expression using Western blotting and mass spectrometry.

Crude membranes were prepared from yeast strains that expressed ScLDM6×His, CaLDM6×His, CgLDM6×His, and/or ScNcp1-6×His or CaNcp1A6×His from the PDR5 or PDR15 locus. SDS-polyacrylamide gels of crude membrane preparations obtained from strains expressing putative recombinant LDMs from the PDR5 locus showed a dramatic increase in a protein band migrating at ∼60 kDa for ScLDM6×His (Fig. 1A, lane 2) and CgLDM6×His (lane 6) and ∼58 kDa for CaLDM6×His (lane 7) compared with control preparations from a representative host strain (lane 1). This result was consistent with the predicted size of these recombinant proteins. The overexpressed bands were excised, and the presence of recombinant C. glabrata or C. albicans LDM6×His was confirmed by tandem mass spectrometry (MS/MS) protein sequencing of tryptic fragments. High-level coverage (>60% primary sequence coverage; see Fig. S1A and B in the supplemental material) for each protein primary sequence was found. The GGR linker engineered at the C terminus was detected for CaLDM but not CgLDM. Identical amounts of ScLDM were expressed from either the PDR5 (lane 2) or PDR15 (lane 3) locus under the control of the Pdr1-3 transcriptional regulator acting on a PDR5 promoter. Compared with the host, strains expressing putative ScNcp1-6×His from the PDR5 locus (Fig. 1, lanes 4 and 5) or CaNcp1A6×His from the PDR15 locus (Fig. 1, lanes 8 and 9) gave intensely stained protein bands of ∼78 kDa, the size expected for the recombinant protein. The identities of these bands were confirmed by MS/MS protein sequencing of tryptic fragments. The overexpressed ScNcp1-6×His gave 69% coverage, while CaNcp1A6×His gave 70% primary sequence coverage, in both cases including the C-terminal engineered GGR linker (Fig. S1C and D).

FIG 1.

SDS-PAGE and Western blots of crude membrane preparations. (A) Coomassie-stained SDS-PAGE. Samples of crude membranes (15 μg) were separated by SDS-PAGE and stained with Coomassie blue as described in Materials and Methods. (B) Western blots decorated with a mouse anti-6×His antibody. Samples of crude membranes (15 μg) were separated by SDS-PAGE and Western blots decorated using an anti-His tag antibody detected by ECL as described in Materials and Methods. Lanes: 1, strain Y785, ADΔ Δpdr5::URA3; 2, strain Y941, ADΔ Δpdr5::ScERG11-6×HIS; 3, strain Y2461, AD2Δ Δpdr15::ScERG11-6×HIS; 4, strain Y2360, AD2Δ pdr5::ScNCP1-6×HIS; 5, strain Y2462, AD2Δ Δpdr15::ScERG11-6×HIS Δpdr5::ScNCP1-6×HIS; 6, strain 2374, ADΔ Δpdr5::CgERG11-6×HIS; 7, strain Y2373, AD2Δ ΔScERG11 Δpdr5::CaERG11-6×HIS; 8, strain 2376, AD2Δ Δpdr15::CaNCP1A-6×HIS; 9, strain 2378, AD2Δ Δpdr15::CaNCP1A-6×HIS Δpdr5::CaERG11-6×HIS. Asterisks show bands of interest. Solid asterisks show bands subject to tryptic digestion and mass spectrometry. Estimated molecular weights (kDa) are shown above bands.

Western blots of the crude membrane preparations were probed using an anti-6×His tag antibody to assess the relative amount of recombinant His-tagged protein produced and compared with an ADΔ host strain that contained a reconstituted empty PDR5 locus (Fig. 1B). Each strain with a 6×His-containing open reading frame (ORF) inserted at either the PDR5 or PDR15 locus overexpressed the expected recombinant protein construct, as shown by the intensity and relative migration of protein bands decorated with the anti-His tag antibody. Recombinant His-tagged protein was not detected in membranes from the control host strain. His-tagged fragments indicative of LDM degradation were not detected in the recombinant crude membrane preparations. In agreement with the Coomassie blue-stained profiles, Western blots showed that CgLDM6×His was expressed in slightly smaller amounts than ScLDM6×His and about twice the level of CaLDM6×His (Fig. 1A and B, lanes 2, 6, and 7). The Coomassie blue-stained profiles indicated that CaLDM6×His was expressed at 2- to 3-fold, CgLDM6×His at 4- to 5-fold, and ScLDM6×His at 6-fold higher levels than the native LDM (Fig. 1A, lanes 1, 2, 6, and 7). However, strains expressing CaNCP1A6×His from the PDR15 locus (strains Y2376 and Y2378) or ScNCP1-6×His from the PDR15 locus (strains Y2460 and Y2462) gave an additional His-tagged and Coomassie blue-stained protein band at ∼68 kDa of comparable (CaNcp1A6×His) or at about half the intensity (ScNcp1-6×His) of the ∼77-kDa band (Fig. 1B, lanes 4, 5, 8, and 9). We suspect, but have yet to prove, the ∼68-kDa band was a product of N-terminal proteolytic processing.

Strains Y2291 and Y2461 conferred identical azole resistance profiles (Table 1) by expressing comparable amounts of functional ScLDM6×His from the PDR5 or PDR15 locus, respectively (Fig. 1A and B, lanes 2 and 3). Strain Y2462 coexpressed equivalent amounts of ScNcp1-6×His from the PDR5 locus and ScLDM6×His from the PDR15 locus, respectively (Fig. 1A and B lane 5). Coexpression of CaLDM6×His and CaNcp1A6×His in strain Y2378 gave an excess of both forms of the cognate reductase protein over its target LDM (Fig. 1A and B, lane 9). The amount of CaLDM6×His expressed from the PDR5 locus was unaffected by coexpression of CaNcp1A6×His (Fig. 1A and B, lanes 7 and 9). We conclude that the pdr1-3 transcriptional regulator used the PDR5 promoter to drive constitutive functional expression from both the PDR5 and PDR15 loci at levels that depended on the ORF involved.

TABLE 1.

Susceptibilities to azole drugs conferred by expression of native and recombinant LDMs and a cognate Ncp1^a

Strain	Strain description	MIC80
		Ergosterol (% wet wt)	AMB (μM)	FLC (μM)	VCZ (nM)	VT-1161 (nM)	ITC (nM)	PCZ (nM)
Y785	Host 1	1.64 ± 0.17	2.1 ± 0.3	1.6 ± 0.1	43 ± 9	83 ± 12	136 ± 9	180 ± 27
Y2411	Host 2	2.14 ± 0.06	2.1 ± 0.2	1.5 ± 0.2	42 ± 3	55 ± 5	155 ± 20	166 ± 30
Y1031	Host 1 PDR5::CgERG11	2.58 ± 0.03	2.5 ± 0.3	5.3 ± 0.8	131 ± 8	186 ± 20	256 ± 46	390 ± 49
Y2374	Host 1 PDR5::CgERG11 ΔScERG11	1.81 ± 0.10	2.5 ± 0.3	4.5 ± 0.4	128 ± 7	165 ± 22	255 ± 45	371 ± 37
Y2372	Host 2 PDR5::CaERG11	2.06 ± 0.14	1.6 ± 0.2	3.8 ± 0.4	83 ± 4	137 ± 4	235 ± 26	352 ± 33
Y2458	Host 2 PDR5::CaERG11 ΔScERG11	2.50 ± 0.14	1.6 ± 0.2	2.8 ± 0.2	48 ± 2	103 ± 6	234 ± 26	321 ± 25
Y2376	Host 2 PDR15::CaNCP1A	1.70 ± 0.16	1.5 ± 0.2	1.4 ± 0.2	49 ± 5	43 ± 6	131 ± 7	120 ± 20
Y2378	Host 2 PDR5::CaERG11 PDR15::CaNCP1A	2.14 ± 0.09	1.4 ± 0.1	4.6 ± 0.2	116 ± 4	130 ± 7	235 ± 26	325 ± 50
Y2459	Host 2 PDR5::CaERG11 PDR15::CaNCP1A ΔScERG11	2.54 ± 0.10	1.6 ± 0.1	3.8 ± 0.1	76 ± 4	110 ± 11	226 ± 24	303 ± 41
Y2291	Host 2 PDR5::ScERG11	2.89 ± 0.17	2.7 ± 0.4	6.3 ± 0.9	166 ± 20	205 ± 17	314 ± 19	376 ± 26
Y2300	Host 2 PDR5::ScERG11 ΔScERG11	2.57 ± 0.16	2.4 ± 0.4	5.2 ± 0.8	143 ± 23	188 ± 16	246 ± 9	317 ± 32
Y2460	Host 2 PDR5::ScNCP1	0.86 ± 0.03	1.5 ± 0.2	2.2 ± 0.2	78 ± 5	86 ± 14	137 ± 12	141 ± 8
Y2461	Host 2 PDR5::ScERG11	3.09 ± 0.01	2.9 ± 0.2	5.4 ± 0.4	165 ± 20	182 ± 22	273 ± 13	340 ± 10
Y2462	Host 2 PDR5::ScNCP1 PDR15::ScERG11	1.32 ± 0.14	1.5 ± 0.2	5.5 ± 0.6	214 ± 22	173 ± 16	223 ± 13	289 ± 21

^aAll measurements were carried out in triplicate in 3 separate experiments. The errors are standard deviations.

Antifungal susceptibilities of strains expressing recombinant LDMs to antifungal drugs.

The physiological function of overexpressed recombinant LDMs was evaluated by determining differential susceptibilities of the host strains and recombinant S. cerevisiae strains overexpressing CaLDM6×His, CgLDM6×His, or ScLDM6×His to short-tailed (FLC and VCZ), medium-tailed (VT-1161), and long-tailed (ITC and PCZ) azole inhibitors of LDM. These properties were visualized using agarose diffusion drug susceptibility assays (Fig. S2) and measured as MIC₈₀ values in complete synthetic medium (Table 1). In some cases the LDM was coexpressed with its cognate NADPH-cytochrome P450 reductase.

Agarose diffusion drug susceptibility assays, using the glucan synthase inhibitor micafungin as an independent control, showed that the amphotericin B susceptibilities of the host and recombinant yeast strains overexpressing LDMs varied slightly. These changes appeared to correlate with the ergosterol content of cell membranes, i.e., the overexpression of ScERG11 in strains Y2291 and Y2300 (Fig. S2A) increased the amount of ergosterol detected in cells and diminished the size of zones of inhibition due to amphotericin B compared with those of host strain Y2411. In contrast, overexpression of ScNcp1-6×His in strain Y2460 reduced the cellular content of ergosterol and increased the size of zones of inhibition for both amphotericin B and micafungin. In strain Y2376, a barely detectable increase in susceptibility to amphotericin B (compared with the host strain Y2411) resulted when CaNcp1A6×His was overexpressed in amounts comparable to those of the expression of ScNcp1 in strain Y2460. As expected, strains Y2460 and Y2376 gave susceptibilities to VT-1161, ITC, and PCZ comparable to those of the control host strain but also had slightly reduced susceptibility to FLC or VCZ.

Compared with strains Y2300 and Y2374, which overexpress ScLDM6×His and CgLDM6×His, respectively, together with the native ERG11 deleted, the expression of CaLDM6×His in Y2458 gave weaker resistance to short- but not medium- or long-tailed azole drugs. The weaker resistance was corrected substantially in strain Y2459 by coexpressing the cognate NADPH-cytochrome P450 reductase of CaLDM from the PDR15 locus. Quantitation of these observations and more details on development of the coexpression system are described below.

As described previously, the MIC₈₀ values obtained for the ADΔ and AD2Δ host strains show comparable susceptibilities to amphotericin B or azole antifungals (32). Our observations using agarose diffusion assays were confirmed by demonstrating a strong linear correlation between the amount of ergosterol in host and recombinant yeast strains overexpressing CgLDM6×His or ScLDM6×His and/or ScNcp1-6×His and the MIC₈₀ values obtained with amphotericin B (Fig. S3A). Overexpression of ScLDM-6×His (Table 1) or CgLDM6×His (Table 1) reduced susceptibility to amphotericin B, consistent with more active ergosterol biosynthesis than that in their host strains. Overexpression of ScNcp1-6×His in strain Y2460 (Table 1) increased susceptibility to amphotericin B. Furthermore, coexpression of ScNcp1-6×His with ScLDM6×His in strain Y2462 reduced susceptibility to amphotericin B compared with that of strain Y2291, which overexpressed ScLDM6×His only. The increases in susceptibility to amphotericin B appear to be due to modest reductions in the normal ergosterol content of the cells, i.e., ScNcp1-6×His expression negatively affected the ergosterol biosynthesis pathway. These results demonstrate, for the first time, a dose-dependent relationship between polyene susceptibility and the ergosterol content of yeast cells.

Compared with the host strain Y2411, the overexpression of CaNcp1A6×His in strain Y2376 (Table 1) reduced ergosterol content and increased amphotericin B susceptibility significantly, while overexpression of the CaLDM6×His locus in strain Y2372 decreased ergosterol content slightly and increased amphotericin B susceptibility significantly. The coexpression of CaNcp1A6×His and CaLDM6×His in strains Y2378 and Y2459 did not further reduce amphotericin susceptibility. These results show that ScLDM6×His overexpression produced ergosterol more effectively than CaLDM6×His overexpression and that ScNcp1-6×His overexpression was more deleterious for the ergosterol biosynthetic pathway than CaNcp1A6×His overexpression. In contrast to the linear relationship between cellular ergosterol content and amphotericin B susceptibility in yeast, there was no correlation between ergosterol content and susceptibility to PCZ (Fig. S3A and B).

Compared to the control host strain Y2411, overexpression of ScNcp1-6×His in strain Y2460 did not affect growth (data not shown) and gave slightly reduced susceptibility to short- and medium-tailed azoles but not long-tailed azoles (Tables 1 and 2). Strains overexpressing ScLDM6×His, CgLDM6×His, or CaLDM6×His from the PDR5 locus, together with the native LDM expressed from the ERG11 locus, showed a 1.5- to 4-fold reduction in susceptibility to short-, medium-, and long-tailed azoles compared with those of the host strains (Table 2). When ScLDM6×His or CgLDM6×His was overexpressed, deletion of the native ScERG11 did not modify significantly susceptibility to ITC, PCZ, VT-1161, FLC, or VCZ (Tables 1 and 2). This result showed these recombinant LDMs to be functional. When CaLDM6×His was overexpressed, deletion of the native ScERG11 did not affect susceptibility to ITC and PCZ, but susceptibility to the short-tailed triazoles FLC and VCZ and to the medium-tailed tetrazole VT-1161 increased by 1.3-, 1.4-, and 1.3-fold, respectively (Table 1, strains Y2372 and Y2458). For each of the short- and medium-tailed azole drugs tested, the sum of the susceptibility value obtained for strain Y2458 and the host Y2411 closely approximated the value obtained for strain Y2372. These results confirmed that the overexpressed recombinant CaLDM6×His was functional. However, with the native ScERG11 deleted in strain Y2458, CaLDM6×His overexpression conferred reductions of susceptibility of 1.9-fold for VT-1161, 1.9-fold for FLC, and only 1.1-fold for VCZ compared to host strain Y2411 (Table 2). These were more modest reductions in susceptibility than those of strains deleted of native ScERG11 and expressing recombinant ScLDM6×His (strain Y2300; 3.3-fold for FLC, 3.4-fold for VT-1161, and 3.4-fold for VCZ compared with its AD2Δ host) or CgLDM6×His (strain Y2374; 2.8-fold for FLC, 2.0-fold for VT-1161, and 3.0-fold for VCZ compared with its ADΔ host; Table 2). These drug subclass-specific responses are unlikely to be due to the ∼2-fold lower expression of CaLDM6×His than CgLDM6×His.

TABLE 2.

Evaluation of susceptibilities to azole drugs conferred by expression of native and recombinant LDMs and a cognate Ncp1p^a

Strain	Strain description	Ratio of MIC80 values
		AMB	FLC	VCZ	VT-1161	ITC	PCZ
Y2300/Y2411	Host 2 PDR5::ScERG11 ΔScERG11/Host 2	1.1 ± 0.3	3.3 ± 1.2	3.4 ± 0.9	3.4 ± 0.7	1.6 ± 0.3	1.9 ± 0.6
Y2461/Y2411	Host 2 PDR15::ScERG11/Host 2	1.4 ± 0.2	3.6 ± 0.7	3.9 ± 0.8	3.3 ± 0.7	1.8 ± 0.3	2.0 ± 0.4
Y2460/Y2411	Host 2 PDR5::ScNCP1/Host 2	0.7 ± 0.2	1.5 ± 0.4	1.9 ± 0.5	1.6 ± 0.2	0.9 ± 0.2	0.8 ± 0.3
Y2462/Y2411	Host 2 PDR5::ScNCP1 PDR15::ScERG11/Host 2	2.1 ± 0.2	3.7 ± 0.9	5.1 ± 0.8	3.1 ± 0.6	1.4 ± 0.3	1.7 ± 0.4
Y2462/Y2460	Host 2 PDR5::ScNCP1 PDR15::ScERG11/Host 2 PDR5::ScNCP1	1.0 ± 0.3	2.5 ± 0.6	2.7 ± 0.5	2.0 ± 0.6	1.6 ± 0.3	2.0 ± 0.3
Y2374/Y785	Host 1 PDR5::CgERG11 ΔScERG11/Host 1	1.3 ± 0.4	2.8 ± 0.5	3.0 ± 1.0	2.0 ± 0.6	1.9 ± 0.4	2.1 ± 0.7
Y2376/Y2411	Host 2 PDR15::CaNCP1A/Host 2	0.7 ± 0.1	0.9 ± 0.3	1.1 ± 0.3	0.8 ± 0.2	0.8 ± 0.2	0.7 ± 0.3
Y2458/Y2411	Host 2 PDR5::CaErg11 ΔScERG11/Host 2	0.8 ± 0.2	1.9 ± 0.4	1.1 ± 0.2	1.9 ± 0.3	1.5 ± 0.4	1.9 ± 0.7
Y2378/Y2411	Host 2 PDR5::CaERG11 PDR15::CaNCP1A/Host 2	0.7 ± 0.1	3.1 ± 0.5	2.7 ± 0.3	2.4 ± 0.3	1.5 ± 0.4	2.0 ± 0.3
Y2459/Y2411	Host 2 PDR5::CaERG11 PDR15::CaNCP1A ΔScERG11/Host 2	0.8 ± 0.1	2.5 ± 0.5	1.8 ± 0.2	2.0 ± 0.4	1.5 ± 0.4	1.6 ± 0.6
Y2459/Y2376	Host 2 PDR5::CaERG11 PDR15::CaNCP1A ΔScERG11/ Host 2 PDR15::CaNCP1A	1.1 ± 0.2	2.7 ± 0.6	1.6 ± 0.2	2.6 ± 0.7	1.7 ± 0.3	2.6 ± 0.8
Y2378/Y2459	Host 2 PDR5::CaERG11 PDR15::CaNCP1A/ Host 2 PDR5::CaERG11 PDR15::CaNCP1A ΔScERG11	0.9 ± 0.1	1.2 ± 0.1	1.5 ± 0.1	1.2 ± 0.2	1.0 ± 0.2	1.1 ± 0.3
Y2459/Y2458	Host 2 PDR5::CaERG11 PDR15::CaNCP1A ΔScERG11/ Host 2 PDR5::CaERG11 ΔScERG11	1.0 ± 0.2	1.4 ± 0.1	1.6 ± 0.1	1.1 ± 0.2	1.0 ± 0.2	0.9 ± 0.2

^aHost 1 is Y785. Host 2 is Y2411. Results are derived from the data shown in Table 1.

Coordination with the heme is expected to dominate the binding of short-tailed triazoles. It was expected to be less important for the binding of the long-tailed triazoles because they have a wide range of contacts in the substrate entry channel (SEC) of the ligand binding pocket (LBP). Furthermore, the tetrazole of VT-1161 should be a weaker ligand of the heme than a triazole, with other contacts in the LBP contributing significantly to overall affinity. As efficient reduction of heme in heterologously expressed LDMs requires interaction with a cognate reductase, the effects of CaNCP1A expression and coexpression with LDM6×His on drug susceptibilities were determined. BLAST alignment shows the primary sequence of C. glabrata NADPH-cytochrome P450 reductase (CgNcp1) has 66% identity to ScNcp1 with 1% gaps, while CaNcp1A has 50% identity with 5% gaps. Thus, ScNcp1 was expected to recognize ScLDM and CgLDM more efficiently than CaLDM. In addition, CaNcp1A was expected to preferentially interact with its natural counterpart, CaLDM.

Strain Y2376, which retains the native ERG11 and overexpresses CaNcp1A6×His from the PDR15 locus, grew normally (data not shown) and was slightly more susceptible to all the azole drugs, except possibly VCZ, than its host (Tables 1 and 2). Strain Y2459, which has the native ERG11 deleted but coexpresses CaNcp1A6×His from the PDR15 locus and CaLDM6×His from the PDR5 locus, showed 2.7- and 1.6-fold reductions in susceptibility for FLC and VCZ, respectively, compared to those for strain Y2376 (Table 2). Compared to strain Y2459, retention of ERG11 in strain Y2378 reduced susceptibility to FLC 1.2-fold and VCZ 1.5-fold but without significant effect on susceptibility to the medium- and long-tailed azoles (Tables 1 and 2). This result shows that functional expression of the CaNcp1A6×His/CaLDM6×His recombinant enzyme pair dominated ergosterol metabolism and that the native ERG11 had no significant effect on the modest (1.5- to 2-fold) reduction in susceptibility to the long-tailed triazoles ITC and PCZ conferred by the overexpression of CaLDM6×His. Comparison of strains Y2458 and Y2459 showed that this reduction in susceptibility to ITC and PCZ was independent of the expression of the cognate reductase (Table 1). These results are consistent with the hypothesis that expression of the cognate reductase increased the content of reduced heme in coexpressed LDM, allowing more efficient interaction with short- but not long-tailed azoles. Susceptibility to VT-1161 gave a phenotype resembling that of the long-tailed azoles. Susceptibility to VT-1161 conferred by overexpression of CaLDM6×His in strain Y2458 was not changed significantly by coexpression of recombinant CaNcp1A6×His in strain Y2459. Furthermore, the retention of native ScERG11 in strains Y2372 and Y2378 reduced susceptibility to VT-1161 only 1.3-fold (Y2372/Y2458) and 1.2-fold (Y2378/Y2459), respectively. These results also showed that the recombinant CaLDM dominated the VT-1161 reduced susceptibility phenotype, while the overexpression of CaNcp1A conferred a barely detectable reduction in susceptibility to this compound.

The modest impact of NCP1 overexpression alone on azole susceptibility was taken into account by measuring the azole susceptibilities of the strains overexpressing the recombinant CaLDM plus its cognate recombinant Ncp1 but with the native ERG11 deleted compared with the strain overexpressing the cognate recombinant Ncp1A6×His but retaining the native LDM (Table 2, Y2459/Y2376). The 1.6- to 2.7-fold reduction in susceptibility to azole drugs reflects the functional contribution of the recombinant LDM relative to the native LDM. This was a significant improvement on the 1.1- to 1.9-fold reductions in azole susceptibility (Table 2, Y2458/Y2411) obtained when CaLDM was expressed without CaNcp1A. It was also comparable to the 1.6- to 2.7-fold increase in azole resistance (Table 2, Y2462/Y2460) achieved when recombinant ScErg11 and ScNcp1 were coexpressed and the native ERG11 retained.

Drug and substrate binding by affinity-purified enzyme.

The biochemical function of the recombinant LDMs was assessed by measurement of the binding of azole drugs. Nickel-nitrilotriacetic acid (Ni-NTA) affinity-purified LDMs from C. albicans and C. glabrata were quantitated using their carbon monoxide binding spectra at 450 nm. These affinity-purified enzymes showed carbon monoxide binding difference spectra for reduced enzyme with a peak at 448 nm and a minor shoulder at 420 nm, like ScLDM6×His (Fig. 2). The spectra indicated at least 90% of each enzyme was functional, as judged by the ratio of the absorbance at 448 nm and 420 nm. Samples of each enzyme at 1 μM showed type II binding of azole drugs comparable to that of ScLDM6×His (Fig. 3 and Table 3). Difference spectra were used to calculate affinity constants, in accord with either allosteric or multipoint binding, using the Hill equation (K_DH). All of the triazoles tested showed tight type II binding, with concentrations of drug giving half-maximal ΔA_max (K_1/2) of 0.32 to 0.61 μM, modest Hill numbers (1.7 to 2.3), and K_DHs of 0.0.07 to 0.37 μM (Table 3), apart from CgLDM6×His binding of FLC. This outlier gave a low Hill number (1.72), the highest K_1/2 (1.03 μM), and the highest calculated affinity constant (K_DH of 1.05 μM). Otherwise, the K_DH values obtained for the azole drugs with the S. cerevisiae and C. glabrata enzymes were comparable, as were about half of those obtained for the C. albicans enzyme. Use of the Morrison equation to take into account the amount of enzyme in the binding assays gave a poorer fit with the type II binding data than the Hill equation. Therefore, affinity constants calculated using the Morrison equation are not presented.

FIG 2.

Carbon monoxide of reduced recombinant LDMs. The difference spectra for the LDMs were obtained as described in Materials and Methods.

FIG 3.

Type II difference spectra for 1 μM ScLDM6×His, CgLDM6×His, or CaLDM6×His incubated with the representative azole drugs VCZ and ITC. Difference spectra obtained in the presence of saturating VCZ (2 to 3 μM) and ITC (2 to 3 μM) with 1 μM fungal LDM are shown on the right, and the peak-minus-trough curves for type II binding (ITC, circles; VCZ, triangles) are shown on the left.

TABLE 3.

Type II binding of triazole drugs by CaLDM, CgLDM and ScLDM^a

Sample and azole	ΔAmax	[Azole]0.5 (μM)	Hill number	KDH (μM)
CaLDM6×His
PCZ	0.0465 ± 0.0021	0.609 ± 0.045	2.00 ± 0.31	0.37 ± 0.09
ITC	0.0432 ± 0.0027	0.586 ± 0.045	1.95 ± 0.31	0.33 ± 0.09
VCZ	0.0447 ± 0.0015	0.510 ± 0.029	1.84 ± 0.17	0.29 ± 0.06
FLC	0.0338 ± 0.0028	0.456 ± 0.049	1.82 ± 0.27	0.24 ± 0.09
CgLDM6×His
PCZ	0.0447 ± 0.0016	0.444 ± 0.027	2.21 ± 0.29	0.17 ± 0.05
ITC	0.0381 ± 0.0008	0.389 ± 0.017	2.31 ± 0.24	0.11 ± 0.03
VCZ	0.0450 ± 0.0008	0.392 ± 0.013	2.17 ± 0.16	0.13 ± 0.02
FLC	0.0554 ± 0.0032	1.028 ± 0.086	1.72 ± 0.15	1.05 ± 0.15
ScLDM6×His
PCZ	0.044 ± 0.014	0.340 ± 0.042	1.90 ± 0.42	0.14 ± 0.09
ITC	0.037 ± 0.010	0.343 ± 0.056	1.70 ± 0.08	0.16 ± 0.03
VCZ	0.043 ± 0.011	0.315 ± 0.120	2.35 ± 0.49	0.07 ± 0.02
FLC	0.047 ± 0.003	0.387 ± 0.101	2.20 ± 0.76	0.14 ± 0.07

^aMeasurements of ΔA_max were between 429 ± 0.5 nm and 411.5 ± 1 nm. K_1/2, concentration of drug giving half-maximal ΔA_max; K_DH, affinity constant determined using the Hill equation. The errors are standard deviations from at least 2 separate experiments.

DISCUSSION

The cognate NADPH-cytochrome P450 reductase bolsters LDM function.

CgLDM6×His overexpressed in the yeast host ADΔ confers resistance to azole drugs at 50 to 70% of the level obtained with overexpressed ScLDM6×His, even after deletion of the native ERG11 (27, 28). This resistance is commensurate with the relative levels of constitutive expression from the PDR5 locus of CgLDM6×His or ScLDM6×His. It shows that the C. glabrata enzyme functions effectively using the endogenous ScNcp1, with gene deletion showing the native LDM makes a minor contribution to overall resistance when CgLDM6×His is overexpressed from the PDR5 locus.

Despite the CaLDM6×His protein being overexpressed constitutively from the PDR5 locus at levels about 3-fold lower than that of ScLDM6×His and only about 2-fold higher than that of the native enzyme, CaLDM6×His is functional. With ScERG11 deleted, overexpression of CaLDM6×His confers modest reductions in susceptibility to short-tailed azole drugs (1.9-fold for FLC and 1.1-fold for VCZ) compared to overexpressed ScLDM6×His (∼5-fold for FLC and VCZ). Coexpression of CaLDM6×His with the native ERG11 retained reduces susceptibility to short-tailed azoles a further 1.4-fold for FLC and 1.7-fold for VCZ but has no effect on susceptibilities to the long-tailed azoles. These differences in susceptibilities between short-tailed and long-tailed azoles may be due to incompatibility between the surfaces required for efficient electron transfer from ScNcp1 to CaLDM. For example, nonconservation of surfaces near the conserved C470 could make the interaction of ScNcp1 with CaLDM less efficient than that with ScLDM or CgLDM. Significant structural differences detected among relevant structurally aligned residues in the surfaces of ScLDM/CgLDM/CaLDM are the following: chemically different residue, G156/G157/F148, A451/A453/K451, and H474/L476/Q474; modified loop, N²⁷⁶DIQD²⁸⁰/N²⁷⁷DIQN²⁸¹/G²⁶⁹DIDPN²⁷⁴; conserved residue in a different conformation, K151/K153/K144.

We hypothesized that coexpression of CaLDM6×His with its cognate C. albicans NADPH-cytochrome P450 reductase would increase the efficiency of CaLDM heme reduction and thereby reduce the disparity in the target available for the binding of short- versus medium- and long-tailed azole drugs. Consistent with this hypothesis, coexpression of CaLDM from the PDR5 locus together with its cognate reductase from the PDR15 locus reduces susceptibility to short-tailed azoles but not medium- or long-tailed azoles. For example, compared with the overexpression of CaLDM with ScERG11 deleted in strain Y2458, the coexpression of CaNcp16×His and CaLDM6×His with ScERG11 deleted in strain Y2459 further reduces susceptibility to short-tailed azoles, i.e., a 1.4-fold reduction for FLC and a 1.6-fold reduction for VCZ (Table 2). Retention of the native ERG11 in the coexpressing strain further reduced susceptibility to short-tailed azoles (1.2-fold for FLC and 1.5-fold for VCZ, strain Y2378/Y2459). As coexpression of CaLDM6×His with CaNcp1A6×His (strain Y2459) did not further reduce the susceptibility to VT-1161, ITC, and PCZ conferred by overexpression of CaLDM6×His (strain Y2458), we conclude that the overexpressed recombinant CaLDM was functional and susceptible to these medium- and long-tailed azoles. Furthermore, the retention of native ScERG11 in strains Y2372 and Y2378 reduced susceptibility to VT-1161 by only 1.3-fold (Y2372/2458) and 1.2-fold (Y2378/Y2459), respectively, and did not significantly reduce that to ITC or PCZ. We conclude that coexpression of CaNcp16×His with CaLDM6×His confers reduced susceptibility to the short-tailed but not the medium- or long-tailed azoles. Thus, the interaction of LDM with the short-tailed triazoles FLC and VCZ appears to be modulated by interactions with the active-site heme, while those of the medium-tailed tetrazole VT-1161 and long-tailed azoles ITC and PCZ are more dependent on interaction with the SEC.

The function of the full-length CaLDM and CgLDM in vitro was demonstrated by spectrophotometric analysis of the Ni-NTA affinity-purified enzymes. Both recombinant enzymes bind carbon monoxide under reducing conditions to give a dominant absorbance peak at 445 nm comparable to that of full-length ScLDM6×His (Fig. 2). The in vitro function of the recombinant enzymes was confirmed by showing tight type II binding of the short-tailed azole drugs FLC and ITC and long-tailed azole drugs ITC and PCZ (Fig. 3). These studies, which pave the way for the structural analysis shown in the companion paper (33), provide the first physiological and in vitro characterization of full-length CaLDM and CgLDM expressed from the genome of S. cerevisiae.

Ergosterol content is affected by LDM and NCP1 expression.

Constitutive overexpression of ScERG11 or CgERG11 and ScNCP1 or CaNCP1 in the AD2Δ host affects ergosterol biosynthesis differentially (see Fig. S3 in the supplemental material). Overexpression of ScERG11 and CgERG11 increases the ergosterol content of cells and reduces amphotericin B susceptibility. Other key enzymes, such as Erg1p and Erg3p in the yeast ergosterol biosynthesis pathway, require expression of native ScNCP1. We found overexpression of a potentially noncognate, recombinant Ncp1 is deleterious to ergosterol metabolism. Thus, overexpression of ScNcp1-6×His or CaNcp1-6×His diminishes cellular ergosterol content and increases susceptibility to amphotericin B. The dose-dependent response obtained is consistent with the concept that the fungicidal effect of amphotericin B depends directly on its binding to ergosterol (34), provided there is sufficient ergosterol to enable disruption of membrane function on drug binding. Our finding is also consistent with previous studies where amphotericin B resistance is due to disruption of the ergosterol biosynthesis pathway by mutation or anaerobiosis (35) or when ergosterol is innately absent and/or replaced by host cholesterol, i.e., no ergosterol is available to bind amphotericin B. By taking the impact of Ncp1 expression into account, coexpression of ScLDM or CaLDM with its cognate reductase increases resistance to all azole drugs comparably, i.e., 1.7- to 2.7-fold (Table 2, strains Y2462/Y2460 and Y2459/Y2376). These findings provide a precedent for using coexpression of cognate NADPH-cytochrome P450 reductases to support functional expression in S. cerevisiae of diverse fungal pathogen, plant, and human host CYP51s. This approach will assist biochemical exploration of these targets and their application in drug discovery.

The value of overexpressed fungal LDM for drug discovery.

Heterologous overexpression in yeast of functional hexahistidine-tagged LDMs from the major fungal pathogens C. glabrata and C. albicans augments knowledge obtained from the overexpression of functional ScLDM6×His in S. cerevisiae (2, 27–31). The constitutive overexpression of CgLDM6×His and CaLDM6×His from the PDR5 locus in the genome of hypersensitive S. cerevisiae hosts deleted of 7 ABC transporters provides a stable platform not confounded by the vagaries of plasmid-based expression, e.g., the impact of variable plasmid number (32). This approach has enabled robust assessment of the response of LDMs from two major fungal pathogens of humans to both established and novel antifungals and expression of functional fungal LDMs for structural analysis. For example, we have used X-ray crystal structures for full-length S. cerevisiae (2), C. glabrata, and C. albicans LDMs (see the companion paper [33]) to define a deep LBP comprised of an active site, SEC, and putative product exit channel relevant to a wide range of fungal pathogens. In silico screens that identify and rank hits that dock with the crystal structures and derived homology models can now be combined with phenotypic screens to identify agents that target the fungal enzyme in biological contexts. For example, compounds that dock with fungal but not human LDM structures in silico can be screened in a phenotypic assay that uses an ultrasensitive AD2Δ yeast strain containing the essential native ERG11 gene. Compounds that block growth in that system can then be tested using agarose diffusion drug susceptibility assays and MIC measurements to determine their specificity with fungal LDM. This can be achieved by comparing the susceptibility of the AD2Δ host with derivative strains in which the native ScLDM has been deleted and ScLDM6×His or human LDM overexpressed from the PDR5 or PDR15 locus under the control of the pdr1-3 transcriptional regulator. The overexpression of functional CgLDM and CaLDM in S. cerevisiae, bolstered with coexpression of the cognate NADPH-cytochrome P450 reductase in the case of CaLDM, enables rapid screening of compounds targeting ScLDM for activity against the LDMs from pathogens in a physiologically equivalent S. cerevisiae host background. Furthermore, the high-level azole resistance conferred by overexpressing fungal drug efflux pumps from the major fungal pathogens in the same host allows identification of fungicidal LDM inhibitors that circumvent this prominent form of azole resistance.

Future prospects.

LDM is the target of the azole drugs used to treat fungal infections of humans, animals, and plants. While enjoying considerable success as preventative measures in medicine and agriculture as well as treatments for superficial and disseminated infections, cure rates with azole drugs are poor for immunocompromised patients undergoing invasive surgery or transplants as well as for individuals with comorbidities, such as AIDS, tuberculosis, and malaria (9). Furthermore, the efficacy of azole drugs in the clinic and the field has been reduced due to the emergence of drug-resistant pathogens. It therefore is important to identify highly potent, fungus-specific LDM inhibitors, preferably with low-nanomolar affinities, and especially compounds likely to avoid the impact of known mutations, such as those described in this report. Testing of the most highly ranked virtual hits in yeast-based phenotypic and biochemical screens, followed by measurements of the susceptibility of clinical isolates, provide the tools to identify scaffolds for broad-spectrum azole and non-azole antifungals that target the LDMs of fungal pathogens. By using S. cerevisiae strains expressing individual drug efflux pumps and highly resistant clinical isolates overexpressing combinations of drug efflux pumps, the problem of azole resistance mediated by the known drug efflux mechanisms can be addressed. None of the azole antifungals in clinical use or antifungals in clinical development, such as VT-1161, has overcome this increasingly significant problem. Different substituents from those found in current azole drugs may help circumvent azole resistance.

MATERIALS AND METHODS

Yeast strains and culture media.

The yeast strains used in the present study are shown in Table S1 in the supplemental material. Yeast strains were grown on YPD medium, containing 1% (wt/vol) Bacto-yeast extract (BD Difco Laboratories Inc., Franklin Lakes, NJ), 2% (wt/vol) Bacto-peptone (BD Difco), and 2% (wt/vol) glucose. Synthetic defined (SD) medium was used for selection of transformants. It contained 2% (wt/vol) glucose, 0.67% (wt/vol) yeast nitrogen base without amino acids (BD Difco), 1.8% (wt/vol) agar (Oxoid Ltd., Hampshire, UK), and either uracil dropout (Qbiogene, Irvine, CA) or histidine dropout (Formedium, Norfolk, UK) complete supplement mixture. Liquid SD medium with complete supplement mixture (Formedium) containing 10 mM morpholineethanesulfonic acid and 20 mM HEPES buffered with Tris to pH 6.8 was used for MIC₈₀ determinations.

Materials.

Desalted oligonucleotides (see Table S3 in the supplemental material), FLC, ITC, VCZ, PCZ, amphotericin B, and lanosterol were purchased from Sigma-Aldrich, Ltd. (St. Louis, MO). VT-1161 was prepared by MicroCombiChem (Wiesbaden, Germany) using the methods described by Hoekstra et al. (17). Micafungin was supplied by Astellas Pharma Inc. (Osaka, Japan). Colony PCRs were carried out using TaKaRa DNA polymerase (TaKaRa Bio Inc., Shiga, Japan). All other PCRs were performed using KOD Hot Start DNA polymerase (Novagen, Madison, WI). PCR clean-up and DNA gel extraction were carried out using kits from Qiagen, Pty. Ltd. (Limburg, Netherlands). Genomic DNA from yeast was isolated using the Y-DER kit from Thermo Fisher (Waltham, MA). Yeast DNA transformation was carried out using an Alkali Cation yeast transformation kit from Qbiogene (Irvine, CA). DNA transformation cassettes and genes inserted at the S. cerevisiae PDR5, PDR15, or ERG11 locus were confirmed by DNA sequence analysis performed at the Genetic Analysis Services facility (University of Otago, Dunedin, New Zealand). The presence of recombinant protein was verified by mass spectrometry using an LTQ Orbitrap XL hybrid ion trap-Orbitrap mass spectrometer (Thermo Fischer Scientific, New Zealand) at the Centre for Protein Research (University of Otago, Dunedin, New Zealand).

Construction of recombinant strains.

The S. cerevisiae ADΔ and AD2Δ hosts (2, 28, 32) were used to create the strains used in this study. The host ADΔ has the URA3 ORF deleted, while AD2Δ also has the HIS1 ORF deleted. The two host strains are deleted of 7 pleiotropic drug resistance (PDR) ABC transporters and the PDR3 transcriptional regulator, but they include the mutant pdr1-3 transcriptional regulator that drives constitutive expression from the PDR5 locus or from the PDR15 locus with the PDR5 promoter introduced (32). Strains of S. cerevisiae that express full-length recombinant LDM6×His (Table S1) were constructed using transformation cassettes that encode C-terminal hexahistidine (6×His)-tagged ORFs of C. albicans or C. glabrata ERG11 (alignment of primary sequences is shown in Fig. S1), together with a PGK transcription terminator and the URA3 selection marker downstream. The transformation cassettes were transferred by homologous recombination into the PDR5 locus of the yeast host strain ADΔ (32) or AD2Δ (28). In some cases the C-terminal 6×His ERG11 ORF, together with a PGK transcription terminator and a selection marker (URA3 or HIS1), was incorporated into the PDR15 locus and bordered upstream by a PDR5 promoter sequence. Two open reading frames of C. albicans ERG11 were tested. One was the native ERG11 ORF (strain Y2472), while the other (strain 2473) had the triplet 787CT789G encoding serine in C. albicans but leucine in S. cerevisiae to 787AGC, modified to encode serine in S. cerevisiae. Both ORFs gave the same physiological responses when expressed in S. cerevisiae, and results obtained using the latter ORF only are reported. In some cases the C. albicans NADPH-cytochrome P450 reductase A (CaNCP1A) allele ORF from the database strain SC5314 with a C-terminal 6×His tag was transformed into the S. cerevisiae PDR5 or PDR15 locus in order to determine the phenotype of this construct or to complement ERG11 inserted previously or subsequently at the other of these two loci. The PDR5 and the PDR15 loci were used to overexpress both an LDM and its cognate reductase at comparable levels due to the pdr1-3 transcriptional regulator in the host background acting constitutively on PDRE elements in PDR5 promoters at both loci. In some recombinant yeast strains the native ScERG11 ORF was deleted by replacement with a URA3 marker. In some other cases, the URA3 marker was deleted using homologous recombination by transformation with an oligonucleotide containing sequence of 39 nucleotides upstream and 43 nucleotides downstream of the ORF plus a spacer fragment of 2 nucleotides. Selection with 5-fluoroorotic acid was used to obtain the desired Ura⁻ strains. This removes 1,475 nucleotides, including URA3, the remainder of the PDR5 ORF, and 1,523 nucleotides downstream of PDR5. The ORFs inserted into the PDR5 and PDR15 loci or at the disrupted ERG11 locus, together with flanking sequences, were confirmed by DNA sequence analysis. For example, the genomic DNA sequence of each transformant used in the present study was confirmed from at least 720 nucleotides upstream to 4 nucleotides downstream of the recombinant ERG11 or NCP1 ORFs.

Agarose diffusion drug susceptibility assays and MIC determinations.

The susceptibilities of strains to antifungal drugs were observed as zones of growth inhibition in agarose diffusion drug susceptibility assays and quantitated as MIC₈₀ values using broth microdilution assays. The agarose diffusion assays were carried out as previously described by Keniya et al. (36), with micafungin serving as a control drug expected to act independently of ergosterol content or its biosynthesis. The MIC₈₀s were defined as 80% growth inhibition compared to that of the no-drug controls, because the triazole drugs are fungistatic rather than fungicidal and can give trailing growth. MIC₈₀s to FLC, VCZ, VT-1161, ITC, and PCZ were determined in 96-well microtiter plates using SD buffered to pH 6.8 (28) instead of RPMI 1640. Cells were seeded at an optical density at 600 nm (OD₆₀₀) of 0.01 (3 × 10⁴ CFU), and the plates were incubated at 30°C with shaking at 200 rpm for 48 h. Cell growth was assessed by measurement of the OD₆₀₀ of wells using a BioTek Synergy 2 multimode plate reader (BioTek Instruments, Vermont, USA). Each MIC₈₀ value was determined using triplicate measurements that pooled 4 clones of each strain in three separate experiments (n = 3, nine measurements in total).

Determination of the ergosterol content of yeast cells.

Sterols were extracted from stationary-phase yeast cultures and lanosterol content quantitated as described previously (37). Yeast cells, grown overnight at 30°C with shaking at 200 rpm in 50 ml of glucose-supplemented SD medium, were harvested by centrifugation at 3,000 rpm for 5 min and washed twice with sterile distilled water. Samples (100 to 120 mg wet weight) of cell pellets were saponified for 1 h at 85°C after vortexing for 1 min in 3 ml of 25% alcoholic potassium hydroxide solution. Nonsaponified sterols were extracted from the cooled sample by adding 1 ml of sterile distilled water and 3 ml of n-heptane and vortexing for 3 min. The clarified heptane layer obtained after standing was diluted 5-fold in absolute ethanol, and absorption spectra were recorded between 200 and 500 nm using an Ultrospec 6300 Pro UV-visible spectrophotometer. In the typical four-peak curve obtained, the peak at 281.5 nm corresponds to the presence of ergosterol plus the late sterol biosynthetic intermediate 24(28)-dehydroxyergosterol (DHE), with only DHE showing an intense spectral band at 230 nm. The percentage of ergosterol in samples was determined by subtracting the percent contribution of DHE determined at 230 nm from the total sterol detected at 281.5 nm. The equation used was the following: percent ergosterol/wet weight = [(A_281.5/290) − (A₂₃₀/518)]/pellet weight in mg, where 290 and 518 are the extinction coefficients (E 1%. cm⁻¹) at 281.5 and 230 nm, respectively (37).

Preparation of crude membranes and Western blot analysis of His-tagged recombinant protein.

Yeast cells were grown in multiple 1.0-liter liquid cultures in baffled 3-liter Erlenmeyer flasks or in 50-ml flasks for microscale experiments. The cultures were grown in YPD medium to an OD₆₀₀ of ∼8 at 30°C with shaking at 200 rpm. Harvested yeast cells were broken using a bead-beating protocol and crude membranes prepared by differential centrifugation (2). The protein concentrations of crude membrane fractions were estimated using the Lowry method (38), with bovine serum albumin (Thermo Fisher) as the standard. Samples containing 15 μg of crude membrane protein were separated by SDS-PAGE in 8% acrylamide gels at pH 8.5 using the method of Laemmli (39) and stained with Coomassie blue R250 or electrotransferred onto Amersham Hybond-ECL 0.45-μm nitrocellulose membrane (GE Healthcare, Auckland, New Zealand) using a standard protocol (100 V, 1.5 h). The membranes were blocked with Tween 80 and milk in phosphate-buffered saline (PBS), decorated for 2 h using 0.5 U/membrane of an anti-His6-peroxidase mouse monoclonal antibody (Roche Diagnostics, Auckland, New Zealand), and washed with blocking buffer. Enhanced chemiluminescence was detected by exposing the membranes to 25 ml 0.1 M Tris-HCl buffer, pH 8.6, containing 5.5 mg luminol (Sigma) dissolved in 60 μl dimethyl sulfoxide (DMSO), 0.28 mg p-coumaric acid (Sigma) dissolved in 10 μl DMSO, and 7.7 μl 30% hydrogen peroxide (Sigma). An Odyssey FC imaging system and Image Studio Lite v.5.2.5 software (LI-COR Biotechnology, Lincoln, NE) were used to record and analyze the data.

Drug binding studies.

The concentration of functional cytochrome P450 used for drug binding studies was determined using carbon monoxide binding spectra according to the protocol described by Guengerich et al. (40). Ni-NTA affinity-purified LDM6×His, used for spectroscopic assays, was eluted using affinity purification buffer containing 50 mM l-histidine instead of imidazole (41). l-Histidine was removed from the sample by washing the enzyme with solubilization buffer containing 16 mM n-decyl-β-d-maltoside using Amicon Ultra-4 centrifugal filters (50-kDa molecular weight cutoff; Merck Millipore, Ltd., Cork, Ireland). The removal of l-histidine was checked by taking the absolute spectra of the sample using an Ultrospec 6300 Pro UV-visible spectrophotometer. The heme peak for wild-type protein with no ligand was at ∼417 nm. With l-histidine bound, the peak was detected at ∼420 nm. Enzyme concentration was determined by saturating the sample cuvette with CO gas prior to the addition of sodium dithionite. The reference cuvette containing the same amount of enzyme was treated with sodium dithionite only. The P450 concentration was determined by measuring the difference in absorbance between 446 and 490 nm and using an extinction coefficient of 91 mM⁻¹ cm⁻¹ (42). Absorption spectra were recorded with a Cary 1 Bio UV-visible spectrophotometer using 10-mm-path UV transparent plastic cuvettes (GE Healthcare Life Sciences, UK). Difference spectra were measured using 1 μM LDM titrated with the triazole drugs ITC, PCZ, VCZ, and FLC. Triazole drugs dissolved in DMSO were added to the sample cuvette, with the same amount of DMSO added to the reference cuvette. The total amount of DMSO was <2% of the total volume in the cuvette. Difference spectra between 350 and 500 nm were recorded, and the trough-peak absorbance changes were used to plot binding curves. The dissociation constant, K_DH, for type II binding of triazole drugs was calculated using GraphPad Prism 6 software (GraphPad Prism, San Diego, CA) by applying the Hill equation using the following formula: ΔA = ΔA_max × [azole]ⁿ/([azole]ⁿ + K_DHⁿ), with ΔA_max being the maximum change in absorbance and [azole] the azole concentration.