Differential Azole Antifungal Efficacies Contrasted Using a Saccharomyces cerevisiae Strain Humanized for Sterol 14α-Demethylase at the Homologous Locus

Abstract

Inhibition of sterol-14α-demethylase, a cytochrome P450 (CYP51, Erg11p), is the mode of action of azole antifungal drugs, and with high frequencies of fungal infections new agents are required. New drugs that target fungal CYP51 should not inhibit human CYP51, although selective inhibitors of the human target are also of interest as anticholesterol agents. A strain of Saccharomyces cerevisiae that was humanized with respect to the amino acids encoded at the CYP51 (ERG11) yeast locus (BY4741:huCYP51) was produced. The strain was validated with respect to gene expression, protein localization, growth characteristics, and sterol content. The MIC was determined and compared to that for the wild-type parental strain (BY4741), using clotrimazole, econazole, fluconazole, itraconazole, ketoconazole, miconazole, and voriconazole. The humanized strain showed up to >1,000-fold-reduced susceptibility to the orally active azole drugs, while the topical agents showed no difference. Data from growth kinetic measurements substantiated this finding but also revealed reduced effectiveness against the humanized strain for the topical drugs. Cellular sterol profiles reflected the decreased susceptibility of BY4741:huCYP51 and showed a smaller depletion of ergosterol and accumulation of 14α-methyl-ergosta-8, 24(28)-dien-3β-6α-diol than the parental strain under the same treatment conditions. This strain provides a useful tool for initial specificity testing for new drugs targeting CYP51 and clearly differentiates azole antifungals in a side-by-side comparison.

Azoles inhibit the synthesis of ergosterol in fungi through direct binding to the cytochrome P450 (CYP) sterol 14α-demethylase (CYP51, Erg11p), a mono-oxygenase that undertakes sequential reactions to remove the methyl group (26). A nitrogen of the azole ring binds to the iron atom in the heme prosthetic group of CYP51, located in the active site of the enzyme, while the N-1 substituent group interacts with amino acids in the folded enzyme structure above the heme (10). The inhibition of sterol 14α-demethylation results in a depletion of ergosterol and an accumulation of 14α-methylated sterols, including 14α-methyl-ergosta-8,24(28)-dien-3β-6α-diol, which inhibit fungal growth (17). Mutations which affect the binding of azoles to CYP51 are commonly found in azole-resistant fungal strains (15, 23), and new drugs are required to address this problem as well as the spectrum of fungal diseases.

One obvious potential side effect of new CYP51 inhibitors is the inhibition of human CYP51, which may result in a reduction in the synthesis of cholesterol and affect the endocrine system (29). Human and fungal CYP51 sequences are around 36% identical (human CYP51 and Candida albicans CYP51 are 35.9% identical; human CYP51 and Saccharomyces cerevisiae CYP51 are 35.8% identical) and catalyze the sterol 14α-demethylase reaction. Despite their relatively low sequence similarities, CYP51s are likely to be topologically similar in function and appear to contain structurally conserved regions (13). The similarities in the structures of human and fungal CYP51s have posed queries and problems for the treatment of fungal infections with azoles, in addition to the potential for drug-drug interactions due to the inhibition in humans of other CYPs that alter the half-lives of other agents that may be administered to patients (18). Azole antifungals are divided into the imidazoles (e.g., miconazole, clotrimazole, econazole, and ketoconazole) and the triazoles (e.g., itraconazole, fluconazole, and voriconazole). The earliest clinical imidazole-based azole antifungals (clotrimazole, econazole, and miconazole) were originally used as topical treatments. Ketoconazole was the first oral treatment for systemic fungal infections (7) but was limited by its toxicity/adverse effects (14, 21, 22). The triazoles were developed in an aim to produce more-specific, less-toxic, and more-potent antifungal drugs. Fluconazole and itraconazole have good antifungal activity and are less toxic than ketoconazole (16). However, the emergence of fluconazole resistance and absorption problems with itraconazole have led to the development of a second generation of triazoles, including voriconazole, which can be used to treat fluconazole-resistant strains (5) and aspergillosis (24). The adverse effects of azole drugs could be due to interactions with human CYPs, including CYP51. It is therefore important that antifungal drugs for systemic use are selective for fungal CYP51.

Previous experimental studies on the specificities of azoles with respect to human and fungal enzymes have relied upon the expression of recombinant proteins, assaying their activities in reconstituted systems and determining the 50% inhibitory concentrations (IC₅₀s) of drugs (3, 11, 12, 27). Eukaryotic CYP51 is a membrane-bound protein, and although recombinant proteins enable a direct comparison of drug binding and protein activity in a cell-free system, functional CYP51 is extremely difficult and time consuming to express, purify, and reconstitute with a highly lipophilic substrate and reductase partner. Therefore, these techniques do not present a convenient test for the specificity of new CYP51 inhibitors. In this study, we report on a strain of S. cerevisiae containing human CYP51 (huCYP51) targeted to replace the native S. cerevisiae CYP51 (ScCYP51) at the chromosomal locus and under the control of the yeast CYP51 promoter, which may be used as a tool for testing the specificity of azoles and for general chemical screen technology based on the assessment of growth.

MATERIALS AND METHODS

Strains and growth conditions.

We used S. cerevisiae BY4741 (ATCC 201388) MATa his3Δ1; leu2Δ0; met15Δ0 ura3Δ0 (EUROSCARF, Frankfurt, Germany). Unless stated otherwise, both the wild-type and humanized strains were grown on yeast minimal media (YM), pH 7.0, containing a 0.67% (wt/vol) yeast nitrogen base without amino acids (Difco, Detroit, Michigan) and 2% (wt/vol) glucose (Fisher, Loughborough, United Kingdom), with 50 μg/ml of histidine and 100 μg/ml each of leucine, methionine, and uracil added (Sigma, Poole, United Kingdom).

Escherichia coli DH5α (Stratagene, La Jolla, CA) was used for plasmid construction and DNA amplification.

Construction of strain BY4741:huCYP51.

The human CYP51A1 cDNA sequence (accession number Q16850) was obtained from Swiss-Prot (http://expasy.org/sprot/). The 5′-upstream- and 3′-downstream-flanking sequences of the S. cerevisiae CYP51 open reading frame (accession number YHR007C) were obtained from the Saccharomyces genome database (http://www.yeastgenome.org/). A DNA sequence consisting of the 5′-flanking sequence from S. cerevisiae, the human CYP51 sequence, and the 3′-flanking sequence from S. cerevisiae was designed. Restriction enzyme sites were added to facilitate cloning and insertion of the URA3 marker gene, flanked by loxP sites, for selection in yeast. The nucleotide sequence was optimized for expression in S. cerevisiae (codon adaptation index of 0.571, compared to a codon adaptation index of 0.0859 for the unoptimized human CYP51 sequence [25]; synthesized by GeneCust [Evry, France]) (Fig. 1). The synthesized gene was cloned into a NotI site in vector pUC57 (pUC57:huCYP51). The URA3-loxP marker was excised from pUG72 (accession number P30117; EUROSCARF), using SpeI and BglII, and was ligated (T4 ligase; Promega, Madison, WI) into the pUC57:huCYP51 vector cut with SpeI and BglII (Promega). The replacement huCYP51 cassette was amplified in E. coli and excised from pUC57:huCYP51 by using NotI (Promega). S. cerevisiae strain BY4741 was transformed with the huCYP51 cassette by electroporation. A 10-ml culture containing 1 × 10⁷ cells/ml grown in YPD medium (1% [wt/vol] yeast extract [Duchefa, Haarlem, The Netherlands], 2% [wt/vol] peptone [Duchefa], 2% [wt/vol] glucose [Fisher]) was harvested, washed twice with sterile water, washed sequentially with 1 M sorbitol (Fisher) and 1 M sorbitol-20 mM HEPES (Sigma), and resuspended in 80 μl of 1 M sorbitol-20 mM HEPES. Electroporation was performed in 2-mm gap cuvettes at 1.5 kV, 25 μF, and 200 Ω in a Gene Pulser II system (Bio-Rad). One milliliter of ice-cold YPD medium was added, and cells were incubated at room temperature for 1 h before being plated on YM-uracil containing 2% (wt/vol) agar (Duchefa). Replacement of the native ScCYP51 by huCYP51 at its chromosomal locus was confirmed by PCR and DNA sequencing (MWG, London, United Kingdom) (data not shown). Excision of the URA3-loxP marker was achieved by induction of Cre recombinase from BY4741:huCYP51:pSH62 transformants (EUROSCARF) (6).

FIG. 1.

Diagram and sequences of the huCYP51 replacement cassette. The upstream and downstream flanking regions of S. cerevisiae CYP51 are shown in uppercase. The huCYP51 sequence with S. cerevisiae codon bias (codon adaptation index = 0.571 [25]) and conflicting restriction enzyme sites removed is shown in lowercase. NotI sites for excision from the pUC57 vector are shown underlined. BglII and SpeI sites for insertion of the URA3 marker are shown in boldface. Transcription initiation (AAA) and transcription termination (polyadenylation, TAAATAAA) sequences are in boldface and underlined. The sequence between the stop codon (taa) and the polyadenylation site contains the sequence for a histidine tag not utilized in this study.

Transcript analysis.

Quantitative reverse transcriptase PCR (RT-PCR) was used to analyze the transcription levels of the native CYP51 and the huCYP51 gene in the two strains. Total RNA was isolated from 3 × 10⁸ cells from overnight cultures of BY4741 and BY4741:huCYP51 by hot phenol (4). RNA samples were adjusted to 100 ng/μl by using a NanoDrop spectrophotometer (Thermo Scientific), and cDNA was synthesized by using an Ambion RETROscript kit. An internal control, the S. cerevisiae actin gene (accession number YFL039C), was amplified by using primers actin F (5′-TCCCAGGTATTGCCGAAAGAAT-3′), actin R (5′-AGCCAAGATAGAACCACCAATC-3′; amplicon size, 125 bp), native ScCYP51, ScCYP51F (5′-CATTCACCGTTGGAACAAAG-3′) and ScCYP51R (5′-GAGAGCTGACACCCTTAGA-3′; 96 bp), and huCYP51, huCYP51F (5′-GTGCTGGTAGACATAGATGTATTG-3′) and huCYP51R (5′-CCATCAATCAAATCAAATTCATAC-3′; 103 bp). The PCR mixtures contained 10 μl of Sybr green master mix (Thermo Scientific), 2.5 μl of a 4-μM concentration of the forward and reverse primer, and 5 μl of the cDNA template (1 in 100 dilution) or RNA (1 in 200 dilution). The process was performed with negative controls containing water and no template, and a four-step, 10-fold dilution series was used to calculate the efficiency of the primer sets (determined as ≥94%) and quantify the mean starting quantity of each transcript (according to Bio-Rad iCycler software, version IQ5 2.0).

Amplification was performed by using an iCycler (Bio-Rad) with the following conditions: 15 min at 95°C, 50 cycles of 15 s at 94°C, 30 s at 56°C, and 30 s at 72°C. Melt curve analysis and gel electrophoresis verified the specificity of each amplicon. All reactions were performed in triplicate on each plate, and each plate was duplicated. Starting quantities of the target genes (huCYP51, ScCYP51) in the RNA samples were subtracted from the cDNA samples and were normalized against actin levels.

Immunoblotting and protein localization.

An antibody against the huCYP51 peptide, W₈₀AKSPPESENKEQLE₉₄C (GeneCust), was made. Yeast cytosolic and microsomal protein fractions were separated by sodium dodecyl sulfate from both BY4741 and BY4741:huCYP51 and were transferred to Hybond-P membranes (Amersham). A primary antibody against huCYP51p (anti-huCYP51) was diluted 1:1,000 in blocking buffer (50 g of low fat milk powder in 0.1 M potassium phosphate [pH 7.5] containing 0.05% [vol/vol] Tween 20 and 0.15 M NaCl). Binding of the antibody to antigens was detected with alkaline phosphatase-conjugated rabbit immunoglobulin G (Sigma) and detected by using a Sigmafast BCIP (5-bromo-4-chloro-3-indolylphosphate)/Nitro Blue Tetrazolium alkaline phosphatase substrate tablet (Sigma).

Identification of sterols.

Both BY4741 and BY4741:huCYP51 were grown in 20 ml of YM (pH 7.0) with an initial inoculum of 2.25 × 10³ cells/ml for 48 h at 30°C at 200 rpm. All azole antifungals were dissolved in dimethyl sulfoxide (adjusted to 1% [vol/vol] for azole solutions and negative controls). Cells were harvested and washed twice with sterile water. Nonsaponifiable lipids were extracted as reported previously (9). Samples were dried in a vacuum centrifuge (Heto). Sterol samples were derivatized by using bistrimethylsilyltrifluoroacetamide; Sigma), and gas chromatography mass spectrometry was performed by using a VG12-250 mass spectrometer (VG Biotech) with a split ratio of 20:1. Individual sterols were identified by reference to relative retention times, mass ions, and fragmentation patterns.

Susceptibility testing.

The susceptibilities of BY4741 and BY4741:huCYP51 to clotrimazole (0 to 16 μg/ml), econazole (0 to 16 μg/ml), fluconazole (0 to 128 μg/ml), itraconazole (0 to 128 μg/ml), ketoconazole (0 to 16 μg/ml), miconazole (0 to 16 μg/ml), and voriconazole (0 to 128 μg/ml) were determined by an adaptation of the CLSI (formerly NCCLS) M-27A broth microdilution method (19) for use with S. cerevisiae strains, using microtiter plates in triplicate. Overnight cultures were diluted in YM, pH 7.0, to give an inoculum density of 2.5 ×10³ cells/ml in a final volume of 200 μl per well with 20 μl of drug diluted in YM. Cells were grown at 30°C for a period of 48 h. MICs were determined as the concentrations of the drugs causing an 80% inhibition of growth.

Growth rates of azole-treated cells.

The growth of BY4741 and BY4741:huCYP51 was monitored by using a Bioscreen C (Oy Growth Curves Ab Ltd.). Three hundred microliters from overnight cultures was diluted to 1 × 10⁵ cells/ml in YM, pH 7.0. Azoles were dissolved in dimethyl sulfoxide, and controls were added as before to each well of a honeycomb 2 plate to final concentrations ranging from 0.0625 μg/ml to 16 μg/ml. Cultures were incubated at 30°C, and the optical density at 600 nanometers (OD₆₀₀) was measured every 15 min for 48 h. All strain/drug combinations were replicated on a second plate and were repeated on each plate, giving a total of four measurements for each. Maximum growth rates for each strain/drug treatment were determined from all four measurements, using the Excel (Microsoft) trend line function. The maximum OD was determined by calculating the mean maximum OD reached by each strain/drug combination.

RESULTS

Characterization of the BY4741:huCYP51 strain.

CYP51 (ERG11) gene deletion in S. cerevisiae is lethal, producing an obligate anaerobic ergosterol auxotroph (8), and functional complementation by huCYP51 is evidenced by growth under aerobic conditions without the requirement for exogenous ergosterol. For comparative studies and to avoid overexpression of huCYP51, we placed the gene under the control of the yeast CYP51 (ScCYP51) promoter. Expression was monitored by RT-PCR and Western blot analysis. RT-PCR showed that the yeast and huCYP51 transcripts were equally abundant (Fig. 2A). Western blot analysis of cytosolic and microsomal extracts from both BY4741 and BY4741:huCYP51 with anti-huCYP51 showed the human enzyme to be correctly targeted to the microsomal membranes (Fig. 2B, lane 4).

FIG. 2.

(A) Transcription levels of native ERG11 and huCYP51. RT-PCR was performed, mean starting quantities were determined, and data were normalized to the values for yeast actin. Error bars show SEs for different quantities. (B) Western blot analysis of huCYP51p localization from cytosolic fractions (lanes 1 and 3) and microsomal fractions (lanes 2 and 4). Molecular mass markers (M) are shown in kilodaltons. An arrow indicates the band corresponding to huCYP51p (56.8 kDa), present in a microsomal fraction from the BY4741:huCYP51 strain.

To confirm that complementation was based on in vivo sterol 14α-demethylation, we analyzed the sterol content for both strains (Table 1). Similar percentages of ergosterol were observed in BY4741:huCYP51 and BY4741: 54% (± 0.82 standard error [SE]) versus 57% (± 1.99 SE), respectively. However, the huCYP51 gene product may be slightly less effective than native ScCYP51, since a greater percentage of the total sterol in repeat measurements was lanosterol (23% [± 1.08 SE] versus 17% [± 1.22 SE] in BY4741).

TABLE 1.

The composition of major sterols in both untreated and azole-treated BY4741 and BY4741:huCYP51

Strain	Sterolb	% (± SE) Total sterols with indicated treatment:a
		Untreated	16 μg/ml of fluconazole	8 μg/ml of itraconazole	4 μg/ml of ketoconazole	0.25 μg/ml of voriconazole
BY4741	Erg	57 ± 1.00	NDc	6 ± 1.63	12 ± 3.89	10 ± 1.41
	Lan	17 ± 1.22	42 ± 3.19	42 ± 5.66	35 ± 0.41	35 ± 1.08
	3,6 diol	NDc	29 ± 3.54	24 ± 7.07	19 ± 5.66	24 ± 0.41
BY4741:huCYP51	Erg	54 ± 0.82	46 ± 1.78	52 ± 1.41	51 ± 1.41	49 ± 1.47
	Lan	23 ± 1.08	24 ± 1.08	24 ± 2.12	27 ± 4.60	27 ± 0.41
	3,6 diol	NDc	6 ± 0.4	5 ± 1.08	7 ± 3.19	4 ± 0.82

^aValues shown are the means of two replicate samples.

^bErg, ergosterol; Lan, lanosterol; 3,6 diol, 14α-methyl-ergosta-8,24(28)-dien-3β-6α-diol.

^cND, not detected.

Prior to the drug treatments, it was important to determine whether the growth of the humanized strain was impaired or altered in comparison to that of the wild-type strain. Figure 3 shows the growth curves obtained for the untreated BY4741 and BY4741:huCYP51 strains. Both strains grew to a similar final OD₆₀₀ of around 1.4, and the maximal growth rates were of the same order (Table 2); in fact, the humanized strain appeared to grow slightly faster (an increase of an OD of 0.157 per hour) than the wild-type strain (an increase of an OD of 0.144 per hour).

FIG. 3.

Growth curves of BY4741 (gray lines) and BY4741:huCYP51 (black lines). The growth curves of untreated cultures and the lowest concentration of a treatment that resulted in significantly slower growth of a strain are shown.

TABLE 2.

Maximum growth rates and ODs of BY4741 and BY4741:huCYP51 treated with azoles

Treatment and drug amt (μg/ml)	BY4741		BY4741:huCYP51
	Mean maximum growth rate	Mean maximum OD	Mean maximum growth rate	Mean maximum OD
Untreated	0.144	1.46	0.157	1.39
Clotrimazole
0.5	0.110	1.16	0.162	1.43
1	0.123	1.29	0.173	1.41
2	0.088	1.08	0.169	1.33
Econazole
1	0.084	0.87	0.157	1.22
Fluconazole
8	0.036	0.47	0.160	1.34
16	0.030	0.55	0.171	1.42
Itraconazole
0.5	0.042	0.55	0.166	1.40
1	0.042	0.55	0.176	1.42
2	0.044	0.56	0.170	1.36
4	0.039	0.56	0.164	1.33
8	0.036	0.52	0.136	1.28
16	0.035	0.54	0.136	1.25
Ketoconazole
2	0.088	0.98	0.188	1.35
4	0.060	0.73	0.204	1.18
8	0.039	0.62	0.119	1.19
Miconazole
0.5	0.024	0.76	0.174	1.41
1	0.034	0.72	0.141	1.25
2	0.023	0.30	0.167	1.34
Voriconazole
0.125	0.057	0.89	0.157	1.45
0.25	0.046	0.75	0.159	1.44
0.5	0.046	0.66	0.153	1.45
1	0.039	0.54	0.156	1.42
2	0.051	0.62	0.154	1.35
4	0.043	0.62	0.154	1.36
8	0.043	0.67	0.168	1.43
16	0.039	0.67	0.159	1.33

^aAll drugs were tested between 0.0625 μg/ml and 16 μg/ml. Values are the means of four replicate samples. The SEs of growth rate measurements do not exceed ±10 % of the means except the measurements for BY4741 treated with 0.125 μg/ml of voriconazole, 2 and 4 μg/ml of ketoconazole, 0.5 μg/ml of clotrimazole, and 0.5 μg/ml of miconazole, which did not exceed 20%. The significance of the difference between the values for untreated growth and treated growth was measured by Student's t test, using a confidence interval of 99%. Growth rates that are significantly less than those for the untreated strains and OD maximums below 1.0 are shown in boldface. Only treatments resulting in significantly slower growth of at least one of the strains are shown; treatments resulting in no growth are not shown.

Susceptibility testing.

The susceptibility of both strains to azole antifungal drugs was tested. Of the drugs tested, the triazoles (fluconazole, itraconazole, and voriconazole) and ketoconazole (an imidazole used for oral treatment) showed specificity in their effects on the strains, with much-reduced efficacy for inhibiting the humanized strain. The earlier drugs used topically exhibited no specificity in this comparison (Table 3).

TABLE 3.

MICs of seven azole drugs for BY4741 and BY4741:huCYP51

Drug	MIC (μg/ml) for indicated strain:
	BY4741	BY4741:huCYP51
Clotrimazole	8	8
Econazole	8	8
Fluconazole	8	>128
Itraconazole	1	>128
Ketoconazole	1	8
Miconazole	8	8
Voriconazole	0.125	>128

^aResults were consistent for three experiments.

In parallel to the determination of MICs, sensitivities to the drugs, as assessed by their effects on growth rates and maximum ODs, were also measured, using a Bioscreen C instrument (Fig. 3 and Table 2). The data showed the same trend with regard to the specificities of the drugs, with a higher affinity for the yeast CYP51 than for the human CYP51 observed for all drugs tested. This extended to clotrimazole, econazole, and miconazole, with which there was clearly reduced growth in the humanized strain compared to the treated parent strain, although the MIC was not altered in the humanized strain.

Voriconazole displayed the greatest specificity of the drugs tested, and the growth data match the MICs determined in this study. A significant reduction in growth rate and the restriction of maximum OD₆₀₀ to <1 was seen with BY4741 at 0.125 μg/ml of voriconazole, while no significant reduction in growth was seen at 16 μg/ml with BY4741:huCYP51. The MIC of 0.125 μg/ml of voriconazole determined for BY4741 versus >128 μg/ml for BY4741:huCYP51 is in agreement with these data. Fluconazole significantly reduced the growth rate of the wild-type strain and reduced the maximum OD₆₀₀ to less than 1 at 8 μg/ml, whereas growth of the humanized CYP51 strain was unaffected at 16 μg/ml. The growth rate measurements obtained with fluconazole agree with our results obtained in tests for MICs. Itraconazole significantly reduced the growth rate and restricted the maximum OD₆₀₀ to <1 at 0.5 μg/ml for the wild-type strain, and although a significantly slower growth rate was observed for the humanized CYP51 strain at 8 and 16 μg/ml, the maximum OD₆₀₀ was not greatly affected. These results concur with results for MICs, although the concentration required to inhibit the growth of the wild-type strain is lower when tested in this way. Treatment with ketoconazole resulted in a significantly lower growth rate at 2 μg/ml, higher than the MIC of 1 μg/ml. However, growth of the humanized strain with ketoconazole was observed at 8 μg/ml, which is also above the MIC we estimated under different conditions. In the cases of clotrimazole, econazole, and miconazole, the concentrations required to inhibit the wild-type strain were lower (0.5, 1, and 0.5 μg/ml, respectively), and indeed, no growth could be detected at all with concentrations above, 2, 1, or 2 μg/ml, respectively.

Sterol composition of azole-treated strains.

The sterol composition of strains treated with azoles for 48 h was determined. Azole drugs inhibit CYP51 (sterol 14α-demethylase); therefore, it was expected that an increase in the concentration of the substrate, lanosterol, would be observed upon treatment with azole. An increase in the sterol, 14α-methyl-ergosta-8,24(28)-dien-3β-6α-diol, was also associated with growth arrest and the mode of action in previous studies (17).

The concentrations of fluconazole (16 μg/ml), itraconazole (8 μg/ml), ketoconazole (4 μg/ml), and voriconazole (0.25 μg/ml) used were chosen by determining the highest concentrations which still resulted in sufficient growth (biomass) for sterol extraction. Table 1 shows the percentages of ergosterol, lanosterol, and 14α-methyl-ergosta-8,24(28)-dien-3β-6α-diol present in the strains. As expected, the wild-type strain (BY4741) showed a greater accumulation of lanosterol and 14α-methyl-ergosta-8,24(28)-dien-3β-6α-diol than the BY4741:huCYP51 strain when treated with azoles. Importantly, the percentage of ergosterol in BY4741 was greatly reduced by treatment, whereas in the BY4741:huCYP51 strain, a level of around 50%, which is typical of this strain when untreated, was maintained. For example, treatment of BY4741 with fluconazole (16 μg/ml) resulted in no detectable ergosterol, whereas the ergosterol level remained at 46% (±1.76 SE) in BY4741:huCYP51.

DISCUSSION

Azole antifungals target CYP51, causing an accumulation of 14 methyl sterol intermediates and a depletion of ergosterol. The effect of perturbed sterol content on the fungal membrane gives rise to growth inhibition (17, 28). CYP51 is also present in humans, and when designing new drugs to target fungal CYP51, the specificity of the new compounds must be considered. We have constructed and validated a strain that may be utilized for preliminary specificity screenings of new drugs. The yeast strain contains the human CYP51 gene, but not the fungal CYP51 gene, through integration at the chromosomal locus of yeast CYP51, and this strain has been shown to be less susceptible to some azole drugs than the wild-type yeast strain. The results obtained are complementary to those of previous enzyme inhibition-based studies on the specificities of azole antifungals (3, 12, 27).

Experimental studies of the specificities of azole drugs have utilized recombinant protein to determine the K_m and V_max of the enzyme in reconstituted systems and the subsequent IC₅₀s (the inhibition of enzyme activity by drugs). Three studies have used these techniques to investigate the relative susceptibilities of C. albicans CYP51 and human CYP51 (3, 12, 27). These studies found the human enzyme to be less susceptible to fluconazole, itraconazole, and ketoconazole. Our study used MICs determined for an 80% inhibition of growth and resulted in an efficacy ratio comparing the parent strain to the humanized one of 16 for fluconazole, 128 for itraconazole, and 8 for ketoconazole, where the humanized strain was least inhibited in growth. Although our MIC data cannot be directly compared to in vitro enzyme inhibition data using C. albicans CYP51, our results are in reasonable agreement with those previously determined for ketoconazole (12, 27) and fluconazole (3) and with the ratio for itraconazole determined by Lamb et al. (12). Our MIC data show some correlation of trends with the results for IC₉₀ determined by Barchiesi et al. (2) and Pfaller et al. for S. cerevisiae (20). Pfaller et al., however, found ketoconazole to be far less effective than indicated by the data presented here and by Barichiesi et al. The results from these different susceptibility tests vary greatly in their numerical values, which may be due to differences in methodology and/or the use of different strains of S. cerevisiae.

The data on the growth rates of drug-treated strains in this study agree with the MICs determined and enabled identification of growth inhibition at lower concentrations not detectable through MIC testing. One additional observation was that both strains exhibited significantly faster growth rates at some subinhibitory azole concentrations. At present we cannot explain this phenomenon. However, subinhibitory concentrations of azoles have been shown to affect the transcription levels of a large number of genes in S. cerevisiae, including genes involved in the control of the cell cycle (1), which could explain this increase in growth rate.

Importantly, the method we employed here gave reproducible results, which are required to detect significant differences in the sensitivities of the enzyme target. The humanized strain, BY4741:huCYP51, therefore provides a useful tool for the rapid and high-throughput initial specificity testing of new drugs designed to inhibit fungal CYP51. This is an in vivo system that allows direct comparisons of the specificities of the drugs and could also be used to further investigate inhibitors of CYP51 not only for antifungals but also for anticholesterol chemotherapies.