Sequencing, Disruption, and Characterization of the Candida albicans Sterol Methyltransferase (ERG6) Gene: Drug Susceptibility Studies in erg6 Mutants

Strains and plasmids.

C. albicans CAI4 (Δura3::imm434/Δura3::imm434), received from W. Fonzi (10), was used for disruption of both copies of ERG6. The S. cerevisiae erg6 deletion strain BKY48-5C (α leu2-3 ura3-52 erg6Δ::LEU2) was used as the recipient strain for transformation with the Candida genomic library (13). Escherichia coli DH5α was used as the host strain for all plasmid constructions. Plasmid pRS316 was obtained from P. Heiter, and Bluescript plasmid was obtained from Stratagene, La Jolla, Calif.

Media.

CAI4 was grown on YPD complete medium containing 1% yeast extract (Difco), 2% Bacto Peptone (Difco), and 2% glucose. Complete synthetic medium (CSM) was used for transformation experiments and contained 0.67% yeast nitrogen base (Difco), 2% glucose, and 0.8 g of a mixture of amino acids plus adenine and uracil (Bio 101) per liter. CSM dropout medium contained the same ingredients as CSM, but without uracil. Uridine was added at 80 mg per liter to ensure growth of CAI4. CSM containing uridine and 5-fluoroorotic acid (5-FOA) at 1 g/liter was used to regenerate the ura3 genetic marker as outlined by Fonzi and Irwin (10). All experiments were carried out at 30°C unless otherwise indicated.

Cloning of ERG6.

Transformation of S. cerevisiae BKY48-5C by using the Candida gene library was carried out by a lithium acetate-modified protocol developed by Gaber et al. (11) for erg6 transformations. The C. albicans ERG6 gene was cloned by transforming a S. cerevisiae erg6 deletion strain (BKY48-5C) with a Candida genomic DNA library obtained from S. Scherer at the University of Minnesota (13). Transformants containing putative Candida ERG6 DNA were subcloned into the Saccharomyces vector pRS316 for complementation analyses and DNA sequencing. All Candida transformations for disruption experiments were carried out essentially in accordance with the procedures of Sanglard et al. (30). Plasmid p5921, obtained from Fonzi (10), was the source of the URA3 blaster for Candida ERG6 disruption experiments.

Approximately 1,250 transformants were obtained by plating on a uracil dropout medium that ensured the presence of the plasmid. These transformants were then screened on medium containing 0.06 μg of cycloheximide per ml. S. cerevisiae erg6 strains are nystatin resistant and cycloheximide sensitive. Transformants that were resistant to this level of cycloheximide (Cyh^r) were further tested for the presence of intracellular ergosterol. Sterols extracted from the S. cerevisae erg6 strains and the transformants were analyzed by UV spectrophotometry and gas chromatography-mass spectrometry (GC-MS) to confirm the sterol profile.

DNA sequencing of the Candida ERG6 gene.

Both strands of the plasmid insert containing the ERG6 gene were sequenced by the Sanger dideoxy chain termination method. Initially, T3 and T7 primers were used, and as DNA sequence became available, primers were generated from sequenced DNA.

PCR.

PCR analyses were used to verify disruptions of both Candida ERG6 genes. Primers P1, P2, and P3 were used to distinguish disrupted ERG6 genes on the basis of size and are in the ERG6 gene itself. Primer 4 is in the hisG region of the URA3 blaster. P1 was 5′-CACATGGGTGAAATTAG-3′ and could be used with all other primers. P2 was 5′-CTCCAGTTCAATTAGCAG-3′, P3 was 5′-TGTGCGTGTACAAAGCAC-3′, and P4 was 5′ GATAATACCGAGATCGAC-3′. PCR buffers and Taq polymerase were obtained from Promega. The buffer composition was 10 mM Tris-HCl (pH 9) and 2 mM MgCl₂, and reactions mixtures contained 0.2 mM deoxynucleoside triphosphates and 0.5 U of polymerase. Conditions for amplification were as follows: the first cycle was denaturation at 94°C for 5 min; this was followed by 40 cycles of annealing at 50°C for 2 min, elongation at 72°C for 3 min, and denaturation at 94°C for 1 min. A final elongation step at 72°C for 20 min completed the reaction. The protocols used for preparation of the Candida template DNA described by Ausubel et al. (1).

Sterol analyses.

Nonsaponifiable sterols were isolated as described previously (23). UV analysis of sterols in extracts was accomplished by scanning wavelengths from 200 to 300 nm with a Beckman DU 640 spectrophotometer. GC analyses of nonsaponifiable sterols were conducted on a HP5890 series II equipped with the Hewlett-Packard Chemstation software package. The capillary column (HP-5) was 15 m by 0.25 mm by 0.25 mm (film thickness) and was programmed to increase from 195 to 300°C (3 min at 195°C and then increased at 5.5°C/min until the final temperature of 300°C was reached and held for 4 min). The linear velocity was 30 cm/s with nitrogen as the carrier gas, and all injections were run in the splitless mode. GC-MS analyses were done with a Varian 3400 GC interfaced to a Finnigan MAT SSQ 7000 MS. The GC separations were done on a DB-5 fused-silica column (15 m by 0.32 mm by 0.25 mm [film thickness]) programmed to increase from 50 to 250°C at 20°C/min after a 1-min hold at 50°C. The oven temperature was then held at 250°C for 10 min before the temperature was increased to 300°C at 20°C/min. Helium was the carrier gas, with a linear velocity of 50 cm/s in the splitless mode. The MS was in the electron impact ionization mode at an electron energy of 70 eV, an ion source temperature of 150°C, and scanning from 40 to 650 atomic mass units at 0.5-s intervals.

Drug susceptibility testing in C. albicans.

Drug susceptibilities of C. albicans wild-type and erg6 strains were conducted by using cells harvested from overnight YPD plates grown at 37°C. Cells were suspended in YPD medium to a concentration of 10⁷ (optical density at 660 nm of 0.5) cells per ml. Cells were plated by transferring 5 μl of the original suspension (10⁰) plus 10⁻¹ and 10⁻² dilutions onto YPD plates containing the drug to be tested. The plates were incubated for 48 h at 37°C and observed for growth. Clotrimazole, brefeldin A, cerulenin, cycloheximide, nystatin, and fluphenazine were obtained from Sigma, St. Louis, Mo. Fenpropiomorph and tridemorph were obtained from Crescent Chemical Co., Hauppage, N.Y. Ketoconazole was obtained from ICN, Costa Mesa, Calif. Terbinafine was a gift from D. Kirsch (American Cyanamid, Princeton, N.J.). Stock solutions of terbinafine, tridemorph, brefeldin A, and cerulenin were prepared in ethanol. Clotrimazole, ketoconazole, and fenpropiomorph stocks were prepared in dimethyl sulfoxide, and fluphenazine and cycloheximide stocks were prepared in water. Nystatin was dissolved in N,N-dimethyl formamide (Sigma).

Nucleotide sequence accession numbers.

The GenBank accession number for the C. albicans ERG6 gene is AF031941. GenBank accession numbers for the previously determined nucleotide sequences of ERG6 from S. cerevisiae, Arabidopsis thaliana, and Triticum ativum are X74249, U71400, and U60755, respectively.

Cloning of the C. albicans ERG6 gene.

Four Saccharomyces erg6 transformants which grew on cycloheximide were analyzed for sterol content. erg6 mutants which fail to synthesize ergosterol due to defects in the C-24 transmethylase gene accumulate principally zymosterol, cholesta-5,7,24-trien-3β-ol, and cholesta-5,7,22,24-tetraen-3β-ol (23). UV scans of the sterols obtained from a Saccharomyces erg6 strain as well as an erg6 transformant containing the Candida ERG6 gene are shown in Fig. 1. Sterols giving the erg6 spectrum contain absorption maxima at 262, 271, 282, and 293 nm as well as maxima at 230 and 238 nm. The latter two absorption maxima are due to conjugated double bonds which occur in the sterol side chain (cholesta-5,7,22,24-tetra-en-3β-ol). The ERG6 transformed strain does not have a conjugated double bond in the side chain and gives absorption maxima only at 262, 271, 282, and 293 nm. The remaining three transformants yielded similar profiles. Additionally, GC analysis of the erg6 mutant and the ERG6 transformants confirmed the presence of ergosterol in the latter strains (data not shown). These results were confirmed by MS.

Of the four transformants restoring the ability of the erg6 mutant to synthesize ergosterol, there were two different types, designated pCERG6-20 and pCERG6-9, with insert sizes of 8 and 14 kb, respectively (Fig. 2). The pCERG6-9 insert contained the entire 8-kb DNA fragment of pCERG6-20, suggesting that the ERG6 gene resided within the 8-kb fragment. Growth of ergosterol-producing transformants on media containing 5-FOA resulted in the loss of the transforming plasmid, which restored the BKY48-5c strain back to the erg6 phenotype; this indicated that ergosterol production of the pCERG6-20 and -9 transformants was plasmid mediated. To locate the ERG6 gene within the plasmid insert, an approximately 4-kb subclone of the left arm of pCERG6-20 was inserted into the Saccharomyces vector pRS316, yielding plasmid pIU880, which was able to complement erg6 (Fig. 2). Plasmid pIU882, which contains a 2.4-kb overlap with pIU880, also complemented erg6, suggesting that the Candida ERG6 gene lies within this 2.4-kb fragment. A 2.4-kb XbaI-EcoRI subclone of pIU880 inserted into pRS316 resulted in pIU885 containing the entire ERG6 gene.

DNA sequencing of the Candida ERG6 gene.

The 2.4-kb XbaI-EcoRI DNA insert of pIU885 (Fig. 2) was selected for sequencing. The DNA and amino acid sequences are presented in Fig. 3. The Candida ERG6 gene encodes the sterol methyltransferase, which contains 377 amino acids and is 66% identical to the Saccharomyces enzyme. Figure 4 shows the sequence alignment between the Candida, Saccharomyces, Arabidopsis, and Triticum sterol methyltransferases, and the levels of identity of Candida to the latter two are 40 and 49%, respectively. A 9-amino-acid region (Fig. 4; amino acids 127 to 135 in the C. albicans sequence) represents the highly conserved S-adenosylmethionine binding site (6).

Creation of a C. albicans ERG6 heterozygote.

Disruption of the Candida ERG6 gene to derive a sterol methyltransferase-deficient strain was made more difficult since Candida, unlike Saccharomyces, is diploid and, thus, both copies of the ERG6 gene must be disrupted. To accomplish this, the URA3 blaster system developed by Fonzi (10) was used. The URA3 blaster contains ∼3.8 kb comprised of repeat elements of hisG (derived from Salmonella) flanking the Candida URA3 gene. The plasmid pIU887-A containing the URA3 blaster inserted into the ERG6 gene is shown in Fig. 2. The 2.4-kb XbaI-EcoRI ERG6 DNA fragment was cloned into the pBluescript vector KS(+) in which a HindIII site was filled in with the Klenow fragment of DNA polymerase I (pIU886). pIU886-L was subsequently derived by deleting a 0.7-kb HindIII fragment within the ERG6 coding sequence, filling in this site with Klenow fragment, followed by the addition of BamHI linkers. Plasmid 5921, containing the URA3 blaster, was digested with SnaBI and StuI, both blunt-cutting enzymes, followed by religation. This resulted in a deletion of 6 bp in one of the hisG regions and destruction of these two sites. The modified 5921 plasmid was then digested with BamHI and BglII to release the 3.8-kb URA3 blaster, which was then ligated into pIU886-L that had been digested with BamHI to generate pIU887-A.

C. albicans CAI4 was transformed by using the 5.3-kb BglII-SnaBI fragment containing the URA3 blaster and ERG6 flanking recombinogenic ends of 0.8 and 0.9 kb. Transformants containing the single disrupted ERG6 allele resulting in heterozygosity for ERG6 were confirmed by using PCR after selection for loss of the URA3-hisG region. Intrachromosomal recombination between the linear hisG sequences resulted in the loss of one of these hisG repeats and the URA3, thus permitting reuse of the URA3 blaster for the subsequent disruption of the ERG6 gene on the homologous chromosome. Selection for colonies on medium containing 5-FOA resulted in growth of only uridine-requiring strains (5).

Creation of C. albicans erg6 strains.

The creation of a Candida erg6 mutant strain in which both alleles were disrupted was accomplished in two different ways. The ERG6 heterozygote was placed onto plates containing high concentrations of nystatin (15 μg/ml), and nystatin-resistant colonies appeared after 3 days. We surmised that mitotic recombination resulted in homozygous ERG6 and erg6 segregants and that these nystatin-resistant colonies might be the erg6 homozygotes. When colony purified, these resistant colonies indeed turned out to be erg6 homozygotes (see below). The second method used to generate erg6 homozygotes was to transform the ERG6 heterozygote with the URA3 blaster. Two kinds of transformants were obtained, wild-type and slow-growing colonies. Both types of colonies were tested for resistance to nystatin, and only the slower-growing colonies were nystatin resistant.

Confirmation of erg6 homozygosity by sterol analyses.

The sterols isolated from wild-type and putative erg6 homozygotes were analyzed by UV spectrophotometry and GC-MS. All of our putative erg6 homozygotes contained erg6-like UV scans similar to the S. cerevisiae erg6 scan shown in Fig. 1. Additionally, GC-MS of erg6 mutant sterols confirmed that only cholesterol-like (C-27) sterols accumulate since the side chain cannot be methylated. Figure 5 shows a GC profile demonstrating that the putative erg6 mutants accumulate C-27 sterols and are deficient in side chain transmethylation. Whereas the predominant sterol in the CAI4 wild type is ergosterol (peak B, 76%), the principal sterols in erg6 mutants are zymosterol (peak A, 43%), cholesta-5,7,24-trien-3β-ol (peak D, 6%), cholesta-7,24-dien-3β-ol (peak E, 9%), and cholesta-5,7,22,24-tetraen-3β-ol (peak F, 29%).

PCR confirmation of homozygous disruptions.

Confirmation of the disruption of both copies of the C. albicans ERG6 gene by mitotic recombination of the heterozygote and by a second transformation using the URA3 blaster was performed by using four PCR primers. The URA3 blaster containing a 3.8-kb region of hisG-URA3-hisG replaced 0.7 kb of ERG6 DNA (Fig. 6A). This was followed by deletion of the hisG-URA3 sequence such that, in effect, the remaining 1.2-kb hisG sequence replaces a 0.7-kb ERG6 deletion. The expected PCR amplifications of CAI4 using primer pair P1-P2 or P1-P3 are 1.5 and 2.15 kb, respectively (Fig. 6B, lanes 1 and 2). The expected products from P1-P2 amplification of the heterozygote CAI-4-6-5 are 1.5 kb (wild-type allele) and 2.01 kb (disrupted ERG6 allele), and the expected products from amplification using the P1-P3 primers are 2.15 kb (wild type) and 2.65 kb (disrupted ERG6); these products are visible in Fig. 6B, lanes 3 and 4. Primer pair P1-P4 gives a 1.1-kb band, demonstrating the presence of hisG within the ERG6 sequence (data not shown). The erg6 homozygotes 5AB-15, obtained by mitotic recombination, and HO11-A3, obtained by URA3 blaster disruption, yield identical amplification products with primer pairs P1-P2 (2.01 kb) and P1-P3 (2.65 kb), as shown in Fig. 6B, lanes 5 to 8.

Drug susceptibilities of C. albicans erg6 strains.

The susceptibilities of the erg6 strains as compared to that of wild-type C. albicans were determined by using a number of antifungal compounds and general cellular inhibitors (Fig. 7). The erg6 strains were shown to be more resistant to nystatin while showing nearly identical sensitivities to the azole antifungals clotrimazole and ketoconazole. Significantly increased susceptibilities of the erg6 strains were noted for tridemorph and fenpropiomorph, inhibitors of sterol Δ14-reductase and Δ8-Δ7 isomerase (2); terbinafine, an allylamine antifungal inhibiting squalene epoxidase (16); brefeldin A, an inhibitor of Golgi function (33); cycloheximide, a common protein synthesis inhibitor; cerulenin, an inhibitor of fatty acid synthesis (25); and fluphenazine, a compound which interferes with the function of calmodulin (14).

The determination of drug concentrations sufficient to completely inhibit growth on plates yielded the data shown in Table 1. The concentration of nystatin required for complete inhibition of the wild type (2.5 μg/ml) is within the normal range for a wild-type strain (23), while the erg6 mutants show a resistance level similar to that noted for erg6 mutants of S. cerevisiae (23). As demonstrated by growth on plates (Fig. 7), the azoles show equal efficacies against both wild-type and erg6 mutant strains. In contrast, the erg6 mutants show significantly increased susceptibilities to other antifungals and metabolic inhibitors. erg6 susceptibilities to cerulenin and fluphenazine were twofold greater, while those for terbinafine and brefeldin A were about 50 times greater, than those of the wild type. Cycloheximide susceptibility was increased about 11-fold in the erg6 mutants, while the greatest increases in susceptibility were shown for the morpholines fenpropiomorph (100-fold) and tridemorph (several thousandfold). The erg6 heterozygote showed essentially the same drug sensitivities as those of the wild type, CAI4, for all inhibitors tested.