Disruption of the Candida albicans CYB5 Gene Results in Increased Azole Sensitivity

Abstract

Sterol synthesis in fungi is an aerobic process requiring molecular oxygen and, for several cytochrome-mediated reactions, aerobically synthesized heme. Cytochrome b₅ is required for sterol C5-6 desaturation and the encoding gene, CYB5, is nonessential in Saccharomyces cerevisiae. Cyb5p and Ncp1p (cytochrome P-450 reductase) appear to have overlapping functions in this organism, with disruptions of each alone being viable. The cytochrome P-450 reductase phenotype has also been shown to demonstrate increased sensitivity to azole antifungals. Based on this phenotype, the CYB5 gene in the human pathogen Candida albicans was investigated to determine whether the cyb5 genotype was viable and would also demonstrate azole sensitivity. Sequential disruption of the CYB5 alleles by direct transformation resulted in viability, presumably conferred by the presence of a third copy of the CYB5 gene. Subsequent disruption procedures with a pMAL2-CYB5 rescue cassette and a CYB5-URA3 blaster cassette resulted in viable cyb5 strains with no third copy. The C. albicans CYB5 gene is concluded to be nonessential. Thus, the essentiality of this gene and whether we observed two or three alleles was dependent upon the gene disruption protocol. The C. albicans cyb5 strains produced a sterol profile containing low ergosterol levels and sterol intermediates similar to that reported for the S. cerevisiae cyb5. The C. albicans cyb5 shows increased sensitivity to azoles and terbinafine, an inhibitor of squalene epoxidase, and, unexpectedly, increased resistance to morpholines, which inhibit the ERG2 and ERG24 gene products. These results indicate that an inhibitor of Cyb5p would not be lethal but would make the cell significantly more sensitive to azole treatment.

The fungal end product sterol, ergosterol, and its biosynthetic pathway are the targets of the majority of the antifungal compounds currently in use for human infections and agricultural applications. The primary class of compounds used in human infections is the azoles, drugs that inhibit the C-14 demethylation of the pathway intermediate lanosterol. Overuse of the azoles, especially in immunocompromised patients, has led to increases in the incidence of antifungal resistance (5). As resistance levels to these compounds rise, the need to identify new targets for antifungal therapy expands. Several reactions in the ergosterol biosynthetic pathway of the pathogenic fungus Candida albicans have been identified (1, 18, 26) as required for viability and, thus, might be subject to exploration as new antifungal sites. Inhibition of these steps could prove to be effective antifungal interventions because this organism cannot import exogenous sterol (19, 29), making sterol biosynthesis obligatory.

Sterol biosynthesis in fungi is an exclusively aerobic process because several steps in the ergosterol biosynthetic pathway require molecular oxygen or heme, which are also synthesized only under aerobic conditions and used primarily in cytochrome-mediated reactions. Among the latter are the cytochrome P-450-dependent step in lanosterol demethylation (encoded by CYP51, also referred to as ERG11 and ERG16), the cytochrome P-450-dependent desaturation step at C-22 (encoded by CYP61, also referred to as ERG5), and the cytochrome b₅ requiring C-5 desaturation (encoded by ERG3). In the case of the cytochrome P-450 steps, two distinct enzymes are involved (11, 12), but each requires the same electron donor, the NADPH cytochrome P-450 oxidoreductase (encoded by NCP1). Ncp1p also donates electrons in the presterol, oxygen-requiring reaction catalyzed by squalene epoxidase (32). The cytochrome b₅ step in the pathway utilizes NADH-cytochrome b₅ reductase or NADPH-cytochrome c reductase as electron carriers in the creation of the double bond at C5-6 (25).

Sutter and Loper (27) reported that disruption of the Saccharomyces cerevisiae NCP1 gene is not lethal despite the fact that two steps in the pathway utilize this electron donor and that one of the steps, the C-14 demethylation of lanosterol, has proven to be essential based on the fact that disruption of the cytochrome P-450 demethylase gene (CYP51) results in nonviability (17). The presence of an alternative electron carrier, perhaps cytochrome b₅, was postulated to explain the viability of the ncp1 mutant (13). This substitution of function by cytochrome b₅ was confirmed in a follow-up study (20). Similarly, the gene encoding cytochrome b₅ in S. cerevisiae was found to be nonessential (28). Disruption of CYB5 generates no growth phenotype in a wild-type background but results in lethality when present in a ncp1 background (28). This suggests that NCP1 might provide a reciprocal function for the missing cytochrome b₅ protein.

Ncp1p and Cyb5p appear to have overlapping functions in S. cerevisiae. One of the phenotypes of ncp1 is an increased sensitivity to azole antifungals (27). This is postulated to be due to the increased sensitivity of the substitute electron carrier in the C-14 demethylation reaction and suggests a mechanism of action for azoles beyond interacting with Cyp51p. A more recent study (28) has confirmed the azole sensitivity of the ncp1 phenotype and has demonstrated that ncp1 cells still produce about 25% of the ergosterol produced by NCP1 strains. In addition, elevated levels of lanosterol were not detected, indicating that the cells were able to complete the C-14 demethylation reaction as well as the other reactions where Ncp1p normally functions. Based on an in vitro assay (13) and on suppression of the ncp1 phenotype by CYB5 (28), Cyb5p could be the replacement electron donor.

Based on the increased azole sensitivity of ncp1 strains coupled with the fact that required reactions in sterol biosynthesis in S. cerevisiae and C. albicans can differ (16), it is worthwhile to investigate the characteristics of the cyb5 and the ncp1 phenotypes in this pathogenic fungus. This report investigated whether CYB5 is an essential gene and whether Cyb5p may be a potential drug target. This report also employed various methods of gene disruption due to difficulties in isolating homozygous mutants for the determination of gene essentiality.

MATERIALS AND METHODS

Strains and plasmids.

Escherichia coli strain DH5-α [φ80dlacZΔM15 recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 deoR Δ(lacZYA-argF)U169] was used for routine bacterial transformations. C. albicans strain BWP17 (ura3Δ::λimm-434/ura3Δ::λimm-434 his1::hisG/his1::hisG arg4::hisG/arg4::hisG) was obtained from A. Mitchell (31). C. albicans strain CAI-4 (ura3Δ::λimm-434/ura3Δ::λimm-434) was obtained from W. Fonzi (9). S. cerevisiae strain Wb5Δ was obtained from G. Truan (28). The C. albicans strains generated in this study are listed in Table 1. Plasmids pGEM-URA3, pGEM-HIS1, and pRS-ARG4 were obtained from A. Mitchell (31). Plasmid pDBI52 was obtained from C. Kumamoto (6).

TABLE 1.

C. albicans strains derived from BWP17 generated in this study

Strain	Disruption protocol	Genotype
KRC2	Direct transformation	CYB5/cyb5::ARG4
KRC3	Direct transformation	cyb5::URA3/cyb5::ARG4/CYB5
KRC6	pMAL-CYB5 cassette	CYB5/cyb5::ARG4 ADE2/ade2::pMAL-CYB5
KRC7	pMAL-CYB5 cassette	cyb5::URA3/cyb5::ARG4 ADE2/ade2::pMAL-CYB5
KRC9	URA3-CYB5 blaster	CYB5/cyb5::HIS1
KRC10	URA3-CYB5 blaster	CYB5/cyb5::HIS1 ERG25/erg25::hisG-URA3-CYB5-hisG
KRC11	URA3-CYB5 blaster	cyb5::ARG4/cyb5::HIS1 ERG25/erg25::hisG-URA3-CYB5-hisG
KRC12	URA3-CYB5 blaster	cyb5::ARG4/cyb5::HIS1 ERG25/erg25::hisG
KRC13	URA3-CYB5 blaster	cyb5::ARG4/cyb5::HIS1 ERG25/erg25::hisG ADE2/ade2::pMAL-CYB5

Media and growth.

E. coli strains were grown at 37°C in Luria broth (LB) (Bio101) containing 10 g of tryptone, 5 g of yeast extract, and 10 g of NaCl per liter. For solid media, 20 g of agar (Difco) per liter was added. Ampicillin (Sigma) was added after autoclaving at a concentration of 60 μg/ml for plasmid selection.

S. cerevisiae and C. albicans strains were grown on complete synthetic medium (CSM) or YPD medium. CSM medium contained 0.17% yeast nitrogen base without amino acids (Difco), 0.5% ammonium sulfate, 2% d-glucose, and a mixture of amino acids (Bio 101). CSM lacking specific nutrients was used in selection. YPD was comprised of 1% Bacto yeast extract, 2% Bacto peptone, and 2% d-glucose. Adenine was added at 360 mg/liter to C. albicans cultures to ensure optimal growth (YPAD). Uridine was added at 80 mg/liter to media used to cultivate C. albicans ura3 strains. Maltose was used at 2% to replace glucose in some media (YPAM). S. cerevisiae cultures were grown at 30°C, while C. albicans cultures were grown at 30°C and 37°C.

Transformations.

E. coli DH5α transformations were accomplished by standard protocols (2). C. albicans transformations were carried out as previously described (31).

DNA sequencing.

DNA sequencing was performed in the Biochemistry Biotechnology Facility at the Indiana University School of Medicine. Initial sequence data obtained with the T3 and T7 primers was used to construct new primers. Primers were obtained from Sigma-Genosys, Ltd. (The Woodlands, Tex.).

PCR.

The standard PCR was composed of template DNA (5 to 10 ng of plasmid DNA or 10 to 50 ng of genomic DNA), sequence-specific oligonucleotide primers (10 to 60 pmol of each primer), 0.2 mM deoxynucleoside triphosphates (Sigma, St. Louis, Mo.), appropriate buffer, sterile water, and a thermostable DNA polymerase.

The two kinds of thermostable DNA polymerases used in this study were Taq (Promega) and Expand High Fidelity Taq polymerase (EHF Taq) (Roche, Indianapolis, Ind.). Taq polymerase was largely used for screening genetic constructions and genetic insertions or other PCRs where high DNA fidelity was not critical. The reaction buffer contained 10× buffer without MgCl₂ (50 mM KCl, 10 mM Tris-HCl [pH 9.0], 1% Triton X-100), 1.5 to 2.0 mM MgCl₂ and 1.25 U of enzyme. The other polymerase used in this study was EHF Taq (Roche). This polymerase was used for PCRs when DNA accuracy was more crucial. The reaction buffer contained 10× buffer with 15 mM MgCl₂ (20 mM Tris-HCl [pH 7.5], 100 mM KCl, 1 mM dithiothreitol, 0.1 mM EDTA, 0.5% [vol/vol] Tween 20, 0.5% [vol/vol] Nonidet P40, 50% [vol/vol] glycerol), and 2.6 U of polymerase was used per reaction.

PCRs were carried out on a Perkin Elmer GeneAmp2400 (Perkin Elmer, Inc., Foster City, Calif.) or the PTC-0200 DNA Engine Thermal Cycler Chassis (M.J. Research, Waltham, Mass.) with standard methods as described in Jia et al. (16). The PCR primers used in this study are listed in Table 2.

TABLE 2.

Primers used in this study

Primer	Sequence (5′-3′)a	Application
b5-5 drI	CCATCCTTTCGTTACATACATTTATACAATCATGTCAGAAGAAACCGTTACTATTTACGAGTTTTCCCAGTCACGACGTT	Disruption of CYB5 with ARG4
b5-3 drI	GCAAAGTTAGTTTTGTAATAATAAGCACCAAATGCAAGTAAAAAGACACCAACAGCAATCTGTGGAATTGTGAGCGGATA	Disruption of CYB5 with ARG4
b5-5 chI	AATTCAGCCCCTTTCAACCC	Check ARG4 and URA3 CYB5 disruptions
b5-3 chI	GTGGTAAATTGGGATTCTCC	Check ARG4 and URA3 CYB5 disruptions
b5-5 drIII	CCATCCTTTCGTTACATACATTTATACAATCATGTCAGAAGAAACCGTTACTATTTACGATTATGAAGAAGTTTCCAAACGTTTTCCCAGTCACGACGTT	Disruption of second CYB5 allele with URA3 and ARG4
b5-3 drIII	CCAACAGCAATCAATGGGAAATTAATACCACTATCTTCAGTAGCATAAGATTGAGCATGTTTAGCTTCAACAGGTTTAGCGTGGAATTGTGAGCGGATA	Disruption of second CYB5 allele with URA3 and ARG4
b5-5EcoRV	GGGATATCATGTCAGAAGAAACCGTTAC	Amplify CYB5 in BWP17
b5-3XbaI	GGTCTAGAAATTATCTAAGTTCTATTACCCTCTC	Amplify CYB5 in BWP17
HIS1-5	GGGGATCCGACTCGACAGGTACCTGG	Amplify HIS1 from CAI4
HIS1-3	GGTCTAGAAAACCGTACCAGGT	Amplify HIS1 from CAI4
ADE-3ck2	AGTATTCACGGATAGATCTGTTAGAG	ADE2 integration
pMAL-5	GCAGTTGAGAATGTTAGTTTTTG	ADE2 integration
b5-5 drII	ATACATATATACTACAAAAAGAAAATAGAAGCACAATAATCGAAGGGAGAGAGGGAGAGGGTTTTCCCAGTCACGACGTT	Disruption of CYB5 with HIS1
b5-3 drII	GTAAGAATATCAAATTGCTTGCAGATGGAAATATTCTCTACTACTCTCTTTCAAATTAATTGTGGAATTGTGAGCGGATA	Disruption of CYB5 with HIS1
b5-5 chII	GTACTGTTGTTGGAGTCTCG	Check CYB5 disrup-tions with ARG4 and HIS1
b5-3 chII	GATAATGTCAACGTGAGTCG	Check CYB5 disrup-tions with ARG4 and HIS1
b5-5 promoter	GGTCTAGATCTTAAACCTTTAGTGTGGG	Amplify CYB5 in BWP17
b5-3XbaI	GGTCTAGAAATTATCTAAGTTCTATTACCCTCTC	Amplify CYB5 in BWP17
hisG-3	GATAATACCGAGATCGAC	ERG25 integration
E25-b5-3chI	CGTTCAGCCATAGCCCAACC	ERG25 integration
E25-5chII	AGCACAACAGGATGACGCCC	ERG25 integration
b5-5probeI	CTTCTGAAGCTACTACCACC	Southern analysis for CYB5 disruptions
b5-3probeI	CCACTATGCTTTATCATGGG	Southern analysis for CYB5 disruptions
b5-5ProbeX	ATAGAACTCATGATGATTTATGGGTCG	Southern analysis for CYB5
b5-3ProbeX	ACCTTTCAAATTACCAATATATAATTTTTG	Southern analysis for CYB5
ARG4-5	CAACCTATTCGCTGGGCCC	Detection of ARG4 disruption in Southern analysis
ARG4-5R	AGCAGTACCACCAAGAGC	Detection of ARG4 disruption in Southern analysis
NJ-ura3-5	TCATGCCTCACCAGTAGC	Detection of URA3 disruption in Southern analysis 3
KR-ura3-5	GGTACCTGATTTCTAGAAGGACC	Detection of URA3 disruption in Southern analysis

^aPlasmid sequences are in bold.

Sterol analysis.

Sterols were isolated as previously described (22) and analyzed by gas chromatography. An HP5890 Series II equipped with a Hewlett-Packard Chemstation software package was used to analyze the sterols. The capillary column (DB-5) was 15 m by 0.25 mm by 0.25 μm (J&W Scientific, Folsom, Calif.). The gas chromatograph was programmed from 195°C (1 min at 195°C, then an increase at 20°C/min until 240°C and from 240 to 280°C at 2°C/min). The linear velocity was 30 cm/s, with nitrogen as the carrier gas. All injections were run in the splitless mode.

Gas chromatography/mass spectrometry analyses of sterols were done with a Thermoquest Trace 2000 gas chromatograph interfaced to a Thermoquest Voyager mass spectrometer. The gas chromatography separations were done on a fused silica column, DB-5MS, 20 m by 0.18 mm by 0.18 μm film thickness (J&W Scientific, Folsom, Calif.). The injector temperature was 190°C. The oven temperature was programmed to remain at 100°C for 1 min, followed by a temperature ramp of 6.0°C/min to a final temperature of 300°C. The final temperature was held for 25 min. Helium was the carrier gas with a linear velocity of 50 cm/s in the splitless mode. The mass spectrometer 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 850 atomic mass units at 0.6-s intervals.

Southern analysis.

The probes used were amplified from a C. albicans wild-type strain (BWP17) by PCR. Probe 1 was used for cytb5 fragments disrupted by selectable markers. The primers used, b₅-5probeI and b₅-3probeI (Table 2), were located 1,539 bp and 862 bp, respectively, upstream of the start codon and amplified a 677-bp DNA fragment. Probe 2 was designed to hybridize to the CYTB5 open reading frame. The primers used, b₅-5ProbeX and b₅-3ProbeX (Table 2), were located 46 bp downstream of the start codon and 129 bp upstream of the stop codon, respectively, and amplified a 200-bp DNA fragment. The probes were designed from sequences not present in cyb5 remnants of the disrupted alleles. Probes were labeled with the Rediprime II random prime labeling system (Amersham, Piscataway, N.J.) according to the manufacturer's protocol. Blotting was accomplished with a Nytran Supercharge nylon transfer membrane (Schleicher and Schuell, Keene, N.H.) according to the manufacturer's protocol.

Growth and drug susceptibility testing.

Cells were grown overnight at 30°C in 50-ml cultures. Cell concentration was determined by hemacytometer and diluted to yield concentrations of 2 × 10⁷, 2 × 10⁶, 2 × 10⁵, and 2 × 10⁴ cells/ml; 5 μl of each dilution was spotted on the medium to be tested. Cerulenin, clotrimazole, cycloheximide, and nystatin were obtained from Sigma. Fluconazole and ketoconazole were obtained from ICN Pharmaceuticals Inc. (Cosa Mesa, Calif.), and tridemorph and fenpropimoph were obtained from Sigma Aldrich. Terbinafine was a gift from I. Hapala (Slovak Academy of Science), and itraconazole was a gift from Janssen Pharmaceuticals. Stock solutions of cerulenin, terbinafine, and tridemorph were made in 95% ethanol, and stock solutions of cycloheximide and fluconazole were made in distilled H₂O. Dimethyl sulfoxide was used as the solvent for clotrimazole, itraconazole, and ketoconazole. All stocks were stored at −20°C.

RESULTS

The C. albicans CYTB5 gene was identified from the C. albicans genomic DNA sequence (19X contig) obtained from the Stanford University DNA Sequencing Project. The CYB5 gene is a 381-bp open reading frame encoding a 126-amino-acid protein. C. albicans Cyb5p shows 43% identity at the amino acid level with S. cerevisiae Cyb5p.

In order to determine whether Cyb5p is essential for growth in C. albicans, both copies of the CYB5 gene would have to be sequentially disrupted. Since this organism does not import exogenous sterol (19, 29), the second disruption would have to take place in the presence of a third copy of the gene if Cyb5p is essential. Since the CYB5 gene from S. cerevisiae was shown to be nonessential (28), the first attempt at C. albicans CYB5 disruption was carried out assuming that this gene was also nonessential.

Method I, disruption of CYB5 by direct transformation; disruption of the first allele.

Method I is diagrammatically represented in Fig. 1A. PCR primers b₅-5drI and b₅-3drI (Table 2) containing 60 bp of the CYB5 gene and 20 bp of plasmid sequence flanking a selectable marker were used to amplify the ARG4 gene on plasmid pRS-ARG4. The resulting 2.1-kb DNA fragment was used to transform C. albicans strain BWP17, and transformants were selected on CSM-Arg+Ade medium. PCR was used to verify that the disruption was correct. Checking primers b₅-5chI and b₅-3chI (Table 2) were designed to amplify a 0.5-kb fragment in the CYB5 allele and a 2.4-kb fragment in the ARG4-disrupted allele. Figure 2, lanes A and B, show the predicted bands of the wild type and the heterozygote KRC2, respectively. Sterol analysis (not shown) for KRC2 showed a wild-type profile.

FIG. 1.

Disruption methods used in this study. (A) Direct transformation disruption method. (B) Disruption method with an inducible rescue cassette. (C) Disruption method with the URA3 blaster system.

FIG. 2.

PCR confirmation of disruption of the C. albicans CYB5 gene by sequential transformation. Lane A shows the 0.5-kb fragment of the CYB5 wild-type allele; lane B shows the 0.5-kb CYB5 allele and the 2.4-kb fragment of the ARG4-disruptped cyb5 allele in the heterozygote KCR2. Lane C shows the 1.7-kb fragment of the HIS1-disrupted cyb5 allele, the 2.4-kb fragment of the ARG4-disrupted cyb5 allele, and the 0.5-kb fragment of the third copy CYB5 wild-type allele of KCR3.

Disruption of the second allele.

PCR primers b₅-5drIII and b₅-3drIII (Table 2) containing 60 bp of CYB5 sequence not represented in the ARG4-disrupted CYB5 allele and 20 bp of plasmid sequence were used to amplify the URA3 gene on pGEM-URA3. The PCR products were use to transform KRC2, and transformants were selected on CSM-Ura-Arg+Ade. Confirmation of disruption was accomplished with primers b₅-5chI and b₅-3chI (Table 2), which would amplify a 1.7-kb fragment in the URA3-disrupted allele. Figure 2, lane C, shows the results of the confirmation with one isolate (KRC3) of the cyb5 homozygote. The 1.7-kb fragment representing the URA3-disrupted allele and the 2.4-kb fragment representing the ARG4-disrupted allele are present. However, a 0.5-kb fragment reflecting a wild-type allele was also present. Sequencing of the 0.5-kb fragment isolated from the gel indicated 100% identity with the CYB5 sequence, indicating that it was an authentic copy of the wild-type allele. The same results were obtained with other isolates of KRC3. Sterol analysis (not shown) of strain KRC3 indicated a wild-type sterol profile, indicating that the third allele was functional. Despite successful disruption of both CYB5 alleles, the presence of the third wild-type allele may imply that this gene is essential. To confirm the essentiality of this gene, we employed an alternative method of disruption.

Method II, disruption of CYB5 with an inducible rescue cassette; preparation of the rescue cassette.

Method II is shown in Fig. 1B. The vector pKR01, containing a copy of the CYB5 gene under the control of the pMAL2 promoter, was constructed by inserting a CYB5 open reading frame into pDB152 (6). Primers b₅-5EcoRV and b₅-3XbaI (Table 2) were used to amplify the CYB5 open reading frame in strain BWP17, creating a 5′ EcoRV site and a 3′ XbaI site on the 0.6-kb fragment. The plasmid was confirmed to have the correct insertion and generated a 6.4-kb fragment when linearized.

Since the disruption system in place used ARG4 and URA3 as selectable markers for the CYB5 disruptions, the URA3 marker in pKR01 was replaced with HIS1. Primers HIS1-5 and HIS1-3 (Table 2) were designed with a 5′ BamHI site and a 3′ XbaI site to amplify a 1.6-kb HIS1 fragment from C. albicans strain CAI-4. HIS1 was ligated to replace URA3 to generate pKR02. This plasmid was linearized within the ade2 sequence with BamHI and used to transform the CYB5 heterozygote KRC2. Transformants were selected on CSM-Arg-His+Ade. Randomly selected colonies were checked by PCR for the proper integration at the ADE2 locus. Primers pMAL-5 and ADE-3ck2 were designed to amplify a 3.4-kb fragment representing the proper integration. Figure 3, lane A, shows the proper integration for one isolate, KRC6.

FIG. 3.

PCR confirmation of disruption of the C. albicans CYB5 gene with an inducible rescue cassette. Lane A shows the 3.4-kb fragment representing the integration of pKR02 containing the pMAL-CYB5 cassette at the ADE2 locus of KCR6. Lane B shows the 2.4-kb fragment of the ARG4-disrupted cyb5 allele and the 1.7-kb fragment of the HIS1-disrupted cyb5 allele of KCR7. Lane C confirms the presence (3.4-kb fragment) of the pMAL-CYB5 cassette in KCR7.

Disruption of the second CYB5 allele.

The remaining wild-type CYB5 allele in KRC6 was disrupted by transformation again with the 1.7-kb URA3 fragment. Selection was on CSM-Arg-His-Ura with maltose as the carbon and energy source. Several colonies were analyzed by PCR to determine the nature of the disruptions and integration. Primers b₅-5chI and b₅-3chI (Table 2) were used to check for the ARG4 and URA3 disruptions of the CYB5 alleles. Figure 3, lane B, shows the 2.4-kb fragment representing the ARG4-disrupted allele and the 1.7-kb fragment representing the URA3-disrupted allele in one isolate (KRC7). With primers pMAL-5 and ADE-3ck2 (Table 2), the plasmid integration at the ADE2 locus is shown to be correct by virtue of the presence of the 3.4-kb fragment (Fig. 3, lane C). Identical results were obtained with four additional isolates of KRC7.

The cyb5 homozygote (KRC7) was spot plated on maltose and glucose media to determine whether the CYB5 gene was essential. The results shown in Fig. 4 indicate that growth was equivalent on all media. If the CYB5 gene is essential, normal growth in maltose would be expected, but growth on glucose would be minimal due to lack of induction of the wild-type CYB5 allele under the control of pMAL2. While some leakage through the pMAL2 promoter would be expected (1, 3), the amount of product formed would not allow normal growth if the gene is essential. Since growth was normal on glucose, the CYB5 gene is concluded to be nonessential.

FIG. 4.

Growth of C. albicans cyb5 with the pMAL-CYB5 cassette (KRC7) in YPA and CSM media with glucose and maltose.

Sterol accumulation in strain KRC7 grown on maltose and glucose was determined. Figure 5 shows the sterol profiles along with those of BWP17 and Wb5Δ (an S. cerevisiae cyb5 mutant). The sterol profile of maltose-grown KRC7 resembles that of its wild-type strain, with ergosterol constituting over 85% of total sterol. When grown on glucose, however, KRC7 produced a profile similar to that of Wb5Δ, with decreased ergosterol and accumulation of the pathway intermediates ergosta-7,22-dien-ol, episterol, and fecosterol. The accumulation of substantial amounts of ergosta-7,22-dien-ol was expected because C-5 desaturation is a Cyb5p-dependent reaction.

FIG. 5.

Sterol profiles of the C. albicans wild-type (BWP17), the S. cerevisiae cyb5 mutant (Wb5Δ), the C. albicans cyb5 mutant (KCR7) grown on maltose and glucose, and the C. albicans cyb5 mutant (KCR12). Peaks: A, ergosterol; B, fecosterol; C, epsiterol; D, lanosterol; E, ergosta-7,22-dien-ol. RT, retention time.

Method III, disruption of CYB5 with the URA3 blaster system.

Since the direct transformation method suggested that CYB5 was essential and the rescue cassette method indicated that the gene was not essential, a third method (shown in Fig. 1C) was employed for final confirmation. In this method, the CYB5 gene was integrated adjacent to the URA3 gene and flanked by Salmonella hisG repeats (URA3 blaster cassette) (9). This third copy would serve as a rescue copy until plating on 5-fluoroorotic acid, which selects against URA3 and for cells in which hisG recombination excises the CYB5-URA3 fragment.

HIS1 disruption of the first CYB5 allele.

Primers b₅-5drII and b₅-3drII (Table 2) were designed to contain 60 bp of the CYB5-flanking sequence and located 769 bp upstream of the start codon and 289 bp downstream of the stop codon, respectively, and 20 bp of plasmid sequence. They were used to amplify the HIS1 marker on pGEM-HIS1. The resulting 2.7-kb fragment containing the HIS1 gene flanked by adjacent CYB5 sequence was used to transform strain BWP17, and transformants were selected on CSM-His+Ade. Ten of the resulting 93 transformants were tested by PCR for the proper configuration. Checking primers b₅-5chII and b₅-3chII (Table 2) were designed to amplify a 1.9-kb fragment representing the wild-type CYB5 allele and a 3.1-kb fragment representing the HIS1-disrupted allele. Figure 6, lane A, shows the presence of both fragments in the heterozygote KRC9.

FIG. 6.

PCR confirmation of disruption of the C. albicans CYB5 gene with the URA3 blaster cassette. Lane A shows the 3.1-kb fragment of the HIS1-disrupted cyb5 allele and the 1.9-kb fragment of the wild-type CYB5 allele of KCR9. Lane B shows the 1.7-kb fragment indicating integration of the URA3-CYB5 cassette at the ERG25 locus in KCR10. Lane C confirms the CYB5 heterozygosity of KCR10. Lane D shows the 3.7-kb fragment of the ARG4-disrupted cyb5 allele and the 3.1-kb fragment of the HIS1-disrupted cyb5 allele of KCR11. Lane E confirms (1.7-kb fragment) the continued presence of the URA3-CYB5 cassette at the ERG25 locus in KCR11. Lane F shows the presence of the ARG4- and HIS1-disrupted cyb5 alleles in KCR12. Lane G shows the confirmation of hisG recombination eliminating the URA3-CYB5 construct (2.9-kb fragment) and the wild-type ERG25 allele on the second chromosome.

Construction and integration of the CYB5-URA3 blaster cassette.

The vector pIU1300 was used as the basic structure for the URA3 blaster cassette. pIU1300 is pBluescript into which the C. albicans ERG25 gene (encoding the sterol C-4 methyloxidase) has been integrated. The URA3 blaster cassette has been integrated within the ERG25 sequence.

With primers b₅-5promoter and b₅-3XbaI (Table 2), the CYB5 sequence including 654 bp of promoter and 187 bp of terminal sequence was amplified from BWP17. The resulting 1.3-kb fragment contained XbaI sites at both ends. After removal of one of the XbaI sites in pIU1300, both the resulting plasmid, pKR04, and the CYB5 fragment were digested with XbaI, ligated, and used to transform E. coli DH5α. Forty resulting colonies out of 1,450 were screened for the appropriate integration of the CYB5 sequence within the URA3 blaster adjacent to the URA3 gene. Transformants that yielded the proper orientation were designated pKR05.

XhoI and ClaI were used to extract a 9.1-kb fragment from pKR05 containing the erg25-hisG-CYB5-URA3-hisG-erg25 construct. This fragment was used to transform KRC9, and transformants (KRC10) were selected on CSM-His-Ura+Ade. To confirm integration at the ERG25 chromosomal locus, primers hisG-3 and E25-b₅-3chI (Table 2) were designed to amplify a sequence within the hisG repeat to a chromosomal region flanking ERG25, respectively. The expected fragment would be 1.7 kb. Figure 6, lane B, shows the presence of the 1.7-kb fragment in KRC10. KRC10 was also rechecked for the continued presence of the chromosomal heterozygosity (Fig. 6, lane C) of CYB5.

ARG4 disruption of the second CYB5 allele.

Primers b₅-5drIII and b₅-3drIII (Table 2) were used to amplify the ARG4 gene on plasmid pRS-ARG4, as described in the disruption of the first CYB5 allele by direct transformation above. The resulting 2.1-kb fragment was used to transform strain KRC10, and transformants were selected on CSM-Arg-His-Ura. Twelve of 240 transformants were analyzed by PCR for the proper configuration. Primers b₅-5chII and b₅-3chII (Table 2) were used to check the disrupted CYB5 alleles, and primers hisG-3 and E25-b₅-3chI (Table 2) were used to confirm the CYB5-URA3 blaster cassette integration at the ERG25 locus. Figure 6, lane D, shows the 3.1-kb HIS1-disrupted allele and the 3.7-kb ARG4-disrupted allele fragments, and lane E shows the 1.7-kb fragment indicating cassette integration at ERG25 for KRC11.

Recombination of hisG fragments and the cyb5 null.

KCR11 was freshly grown and plated on CSM containing 5-fluoroorotic acid. After 3 days of incubation, a total of 40 colonies emerged. The colonies were patched onto CSM-Ura to confirm the loss of the URA3 gene. Twenty of the colonies were confirmed as Ura⁻. Five isolates were chosen for confirmation by PCR for continuing CYB5 disruptions, and hisG recombination expected in the resulting strain, KRC12. The ARG4- and HIS1-disrupted alleles were confirmed as still present (Fig. 6, lane F) in four of the five KRC12 isolates chosen. The fifth showed only the HIS1-disrupted allele and a wild-type allele and was set aside. Primers E25-b₅-3chI and e25-5chII (Table 2) were used to amplify a 1.9-kb fragment representing the wild-type ERG25 allele and a 2.9-kb fragment representing the hisG recombination product in the second allele into which the CYB5-URA3 blaster cassette was inserted. All four of the KRC12 isolates yielded 1.9-kb and 2.9-kb fragments, indicating the presence of a wild-type ERG25 allele and the hisG recombination at the other ERG25 locus, respectively (Fig. 6, lane G). Sterol analysis of KCR12 indicated that the predominant sterol was ergosta-7,22-dien-ol, consistent with the sterol profile in the authentic S. cerevisiae cyb5 mutant Wb5Δ (26). Other sterols found in lesser amounts, such as fecosterol, ergosterol, episterol, and lanosterol, were also found in both C. albicans and S. cerevisiae cyb5 mutants (Fig. 5). Cyb5p is nonessential by this procedure.

Confirmation of disruptions by Southern analysis.

Southern analysis was done to confirm the results generated by PCR for the CYB5 heterozygotes and cyb5 nulls created in this study. Two probes were employed. The first was located upstream of the CYB5 chromosomal alleles and designed to detect wild-type and disrupted versions of the CYB5 gene (primers b₅-5probeI and b₅-3probeI in Table 2). The second probe was designed to detect the CYB5 gene integrated at any position within the C. albicans genome. (primers b₅-5probeX and b₅-3probeX in Table 2). DNA was prepared for both probes with EcoRI.

Figure 7A shows the results with probe 1. The wild-type (BWP17) shows the predicted 5.6-kb fragment representing the wild-type CYB5 allele. Strain KRC3 containing the ARG4- and URA3-disrupted alleles shows a heavy band at approximately 4.5 kb and the presence of the wild-type allele (5.6 kb). Further experiments with the URA3 and ARG4 probes (primers in Table 2) confirmed that both the predicted 4.5-kb ARG4- and 4.3-kb URA3-disrupted bands are present in the heavy band shown, confirming the presence of the third copy of the CYB5 gene in KRC3. KRC7 shows the same doublet at 4.5 kb, and again both the ARG4- and URA3-disrupted alleles were found to be present. KRC9, a heterozygote, shows the presence of the 5.6-kb wild-type allele and the 6.9-kb fragment representing the HIS1-disrupted allele. KRC12 (and KRC11-not shown) shows the presence of the HIS1-disrupted allele (6.9 kb) and the ARG4-disrupted allele (4.5 kb). The results confirm those obtained with PCR.

FIG. 7.

Confirmation of the CYB5 disruption strategies by Southern analysis. The upper panel shows the results with probe 1 to detect wild-type and disrupted versions of CYB5. The wild-type band of 5.6 kb is shown in the wild-type strain (BWP17) and as a third copy in KCR3. KCR3 and KCR7 show the presence of a thick band of approximately 4.4 kb, comprising the 4.3-kb URA3-disrupted and 4.5-kb ARG4-disrupted cyb5 alleles. The 6.9-kb band representing the HIS1-disrupted cyb5 allele is shown in KCR9 and KCR12. The lower panel shows the results with probe 2. BWP17 shows the wild-type allele (5.6-kb band), while KCR12 shows no bands confirming that the CYB5 gene is not found at any location in the genome.

Probe 2 results are shown in Fig. 7B. The chromosomal wild-type CYB5 allele is displayed as a 5.6-kb band in BWP17. KRC12, as expected, produced no bands, indicating that no copies of the CYB5 gene remain after hisG recombination eliminated the cassette copy of CYB5.

Drug susceptibilities of the C. albicans cyb5 mutant KRC12.

The viable C. albicans cyb5 mutant KCR12 generates a novel sterol profile, and it was of interest to see if this array of sterol alters the cell's sensitivity to a variety of drugs. Of particular interest would be the sensitivities to commonly used antifungal agents. Figure 8 shows the growth patterns of the C. albicans wild-type, BWP17 (CYB5), and KRC12, the cyb5 null, on three azoles (ketoconazole, fluconazole, and itraconazole), inhibitors of Erg11p (30), two morpholines (tridemorph and fenpropimorph), inhibitors of Erg24p and Erg2p (4), and terbinafine, an inhibitor of squalene epoxidase (Erg1p) (14). In addition, cycloheximide, a protein synthesis inhibitor, cerulenin, a fatty acid synthase inhibitor (23), and fluphenazine, a calmodulin antagonist (10), were tested.

FIG. 8.

Sensitivity of C. albicans BWP17 (CYB5) and the cyb5 null (KRC12) to a variety of antifungal agents and cellular inhibitors.

KCR12 was shown to be significantly more sensitive to the azoles than the wild type. Terbinafine sensitivity also increased, but the cyb5 strain was considerably more resistant to morpholines than the CYB5 strain. There were no changes in sensitivities to general metabolic inhibitors. These patterns are shown in Table 3 in terms of the drug concentration at which no growth was evident after 72 h of incubation.

TABLE 3.

Effects of antifungal drugs and cellular inhibitors on wild-type (CYB5) and cyb5 null mutants of C. albicans

Inhibitor	Inhibitory concna (μg/ml)
	CYB5	cyb5
Ketoconazole	>1.5	0.5
Fluconazole	>10	0.5
Itraconazole	1.5	<0.5
Tridemorph	0.025	>5
Fenpropimorph	0.025	0.1
Terbinafine	3	1
Cycloheximide	>140	>140
Cerulenin	>1.5	1.5
Fluphenazine	>10	>10

^aConcentration at which there was no visible growth after 72 h of incubation.

In order to confirm that the KRC12 azole-sensitive phenotype was due to the loss of Cyb5p, a copy of the CYB5 gene was reintroduced into KCR12. Plasmid pKR02 was linearized within the ade2 sequence with AvaI and used to transform KRC12. Transformants (KRC13) were selected on CSM-Arg-His-Ura. Figure 9 shows the growth of KRC13 on glucose and maltose in the presence of the three tested azoles. Growth on glucose resulted in increased azole sensitivity, as expected with no functional Cyb5p present. Induction of the CYB5 gene on maltose resulted in azole sensitivity approaching the levels seen with the wild type (Fig. 8).

FIG. 9.

Azole sensitivity of C. albicans KCR13, a cyb5 null containing a CYB5 gene under pMAL2 control, grown on glucose and maltose.

DISCUSSION

The CYB5 gene of C. albicans was identified from a database based on homology to the CYB5 gene from S. cerevisiae. In order to determine whether Cyb5p was essential for growth in C. albicans, both copies of the gene were sequentially disrupted with CYB5 sequences flanking selectable markers. The direct transformation disruption protocol resulted in the successful disruption of both alleles in the presence of a third copy of the wild-type CYB5 allele. The third-copy phenomenon has been previously reported in C. albicans and was thought to occur only with essential genes (8). Based on this assumption, we would conclude that the CYB5 gene is essential. However, since the gene is not essential in S. cerevisiae, confirmation through other approaches was sought.

With a second protocol, a third copy of the CYB5 gene under the control of pMAL2, the maltose-inducible promoter, was inserted at the ADE2 locus in a CYB5 heterozygote. The second allele of the CYB5 gene was then disrupted. This was done in the presence of maltose to rescue cyb5 homozygotes should the gene be essential. When transferred to glucose, the cyb5 null (KRC7) grew normally, indicating that the gene is nonessential in C. albicans. Low-level leakage through the pMAL2 promoter in the absence of inducer might be anticipated, but previous studies (1, 3) with this promoter show that very slow growth results when the gene in question is essential. Sterol analysis of KRC7 indicated a profile that included low levels of ergosterol and an array of intermediates that closely resembled the sterol accumulation of the cyb5 (Wb5Δ) mutant of S. cerevisiae (28). This approach did not reveal a third copy of the CYB5 gene and indicated that the gene is not essential.

A final disruption procedure was used to confirm the results with the pMAL2 inducible rescue cassette. This approach used the URA3 blaster cassette into which a third copy of the CYB5 gene had been integrated adjacent to the URA3 marker. The cassette was integrated into an unrelated chromosomal site in the genome of a CYB5 heterozygote. Following disruption of the second allele, the resulting chromosomal cyb5 homozygote was plated on 5-fluoroorotic acid to eliminate the URA3-CYB5 fragment via hisG recombination. The resulting strain (KRC12) was viable, indicating, once again, that the CYB5 gene is not essential. This procedure has the advantage of having no remaining copies of the CYB5 gene, thus eliminating the possibility of recombination of the CYB5 allele, as is possible in the pMAL2 system (1, 3), to other sites not under the control of pMAL2.

The issue of gene disruption is complicated due to the third-allele phenomenon which occurs in C. albicans and also due to possible trisomies that occur in this organism. Whereas sequential gene disruption of essential genes results in two disrupted alleles, a third allele is observed. The origin of these trisomies is unclear because various C. albicans strains may have trisomic regions or entire chromosomes. A recent report by Chen et al. (7) finds that in some isolates of CAI-4, chromosome 1 is trisomic, explaining why genes such as GLY1, YAL36, CPP1, CPH1, and FKS1 are found to be triplicated but nonessential. However, our results suggest that we see trisomic CYB5 only when performing a transformation of the CYB5/cyb5 heterozygote and not when we introduced a rescue cassette before disruption of the second allele and not when we selected for deletion of the CYB5 allele with the URA3 blaster approach. Thus, direct transformation to determine essentiality was less reliable than rescue or deletion approaches. Importantly, the pMAL2 inducible CYB5 cassette and the URA3 blaster-CYB5 deletion cassettes were not associated with the CYB5 locus to avoid allelic and chromosomal ploidy interfering with the final test for gene essentiality.

Sterol analysis of KRC12 indicated the synthesis of some ergosterol and sterol intermediate accumulation similar to that of S. cerevisiae Wb5Δ. This indicates that another electron donor is able to partially substitute for Cyb5p and allows the synthesis of some ergosterol. Based on these data, accompanied by PCR and Southern analysis confirmations, we conclude that the CYB5 gene is not essential in C. albicans.

The results obtained here raise a question regarding the third gene copy phenomenon. Chromosome duplication and duplication of smaller genomic elements (8) have been suggested as mechanisms by which third copies of some genes arise. To this point, third copies have been shown to occur only with essential genes, and recently the use of the UAU1 disruption procedure was suggested as a rapid test for determining whether a given gene was essential (8). Unfortunately, this procedure could not be successfully implemented in this case, possibly because the CYB5 gene is so small. Since there was no evidence of a third copy in the latter two disruption protocols, the results also indicate the possibility that the third-copy phenomenon may occur as a spontaneous event in certain cases when the chromosomal alleles are being inactivated. Alternatively, the successful direct disruption procedure where the third copy was found may have occurred only in a small subpopulation in which the CYB5 allele existed in triplicate.

Since inhibition of Cyb5p would result in a viable cell with altered sterol composition, it was of interest to determine whether such a cell would show altered responses to common inhibitors and antifungal drugs. Surprisingly, we found a significant increase in azole sensitivity for KCR12. In order to confirm the linkage of azole sensitivity and the lack of Cyb5p, the azole experiment was repeated with KRC13, a derivative of KCR12. KCR13 is KRC12 into which the pMAL2-CYB5 cassette has been integrated at the ADE2 locus. When grown on glucose, KRC13 shows a pattern of increased inhibition by azoles similar to that seen with KCR12. The increased sensitivity is not quite as great, probably due to low-level leakage through the pMAL2 promoter (1, 3). Nearly normal azole sensitivity is restored when KCR13 is grown on maltose, indicating that the loss of Cyb5p is responsible for the increased azole sensitivity. Other C. albicans sterol mutants that also produced altered sterol profiles showed no such increases in azole sensitivity (15, 16). Coupled with no changes in inhibition levels with cellular inhibitors, this indicates that increased azole sensitivity is not likely due to increased permeability to these compounds. The cyb5 null was also more sensitive to terbinafine, an inhibitor of squalene epoxidase.

These results with the azoles are consistent with these drugs' having effects on substitute electron donors in reactions within the sterol biosynthetic pathway. Both the squalene epoxidase and the C-14 demethylase steps are dependent upon NCP1 as an electron carrier. The ncp1 genotype has also been shown to be more sensitive to azoles, indicating the possibility that Cyb5p and Cprp do have significant functional overlap and that both may be sensitive to azole inhibition. Further investigations are needed to understand why cyb5 mutants are more resistant to the morpholine class of sterol biosynthesis inhibitors. However, S. cerevisiae elo2 and elo3 sphingolipid mutants (24) are commonly found to be morpholine resistant, and these mutations are known to be suppressors of erg2 and erg24 (21).