|Home | About | Journals | Submit | Contact Us | Français|
Candida albicans ERG3 encodes a sterol C5,6-desaturase which is essential for synthesis of ergosterol. Defective sterol C5,6 desaturation has been considered to be one of the azole resistance mechanisms in this species. However, the clinical relevance of this resistance mechanism is still unclear. In this study, we created a C. albicans erg3/erg3 mutant by the “Ura-blaster” method and confirmed the expected azole resistance using standard in vitro testing and the presence of ergosta-7,22-dien-3β-ol instead of ergosterol. For in vivo studies, a wild-type URA3 was placed back into its native locus in the erg3 homozygote to avoid positional effects on URA3 expression. Defective hyphal formation of the erg3 homozygote was observed not only in vitro but in kidney tissues. A marked attenuation of virulence was shown by the longer survival and the lower kidney burdens of mice inoculated with the reconstituted Ura+ erg3 homozygote relative to the control. To assess fluconazole efficacy in a murine model of disseminated candidiasis, inoculum sizes of the control and the erg3 homozygote were chosen which provided a similar organ burden. Under these conditions, fluconazole was highly effective in reducing the organ burden in both groups. This study demonstrates that an ERG3 mutation causing inactivation of sterol C5,6-desaturase cannot confer fluconazole resistance in vivo by itself regardless of resistance measured by standard in vitro testing. The finding questions the clinical significance of this resistance mechanism.
Candida albicans is the most common cause of deep mycoses in humans. Azole therapy has been well tolerated and effective for many forms of candidiasis. Of concern is that long-term azole treatment of oropharyngeal candidiasis in human immunodeficiency virus-infected patients has encountered progressive azole resistance (14, 40). Azole antifungals inhibit the biosynthesis of ergosterol, the major sterol of cell membrane, by targeting lanosterol 14α-demethylase encoded by ERG11 (38). Alteration of amino acid composition of lanosterol 14α-demethylase (37), increased drug efflux (32, 34), and altered ergosterol synthetic pathways due to blockage of sterol C5,6-desaturase encoded by ERG3 (17, 33) have been known as factors contributing to azole resistance in C. albicans and Saccharomyces cerevisiae (for reviews, see references 1 and 35). However, no relation between defective sterol C5,6-desaturase and azole resistance was found in Candida glabrata (9). The relevance of azole resistance in C. albicans erg3 mutants is still unclear. Although it has been reported that a few azole-resistant clinical isolates of C. albicans exhibited a sterol profile indicative of defective sterol C5,6 desaturation (4, 18, 26), the possibility remains that another mechanism(s) of azole resistance might have been present in those isolates. For instance, the Darlington strain, an erg3/erg3 mutant isolated from the oral cavity, was also azole resistant due to mutations in ERG11 (15, 24).
A mechanism by which erg3 mutations cause azole resistance has been proposed but is in part counterintuitive. The sterol composition of these mutants is largely ergosta-7,22-dien-3β-ol rather than ergosterol. The only difference between these molecules is the saturation of the C5-6 bond in ergosta-7,22-dien-3β-ol. Substitution of ergosta-7,22-dien-3β-ol for ergosterol in the cell membrane leads to increased, not decreased, sensitivity to a large number of toxic chemicals, detergent, ions, and low pH (11, 33). The contrary effect of increased azole resistance has been hypothesized to be due to the ability of the cell to circumvent the azole inhibition of C14 demethylation by successfully utilizing C14-methylated C5,6-saturated sterols (17, 18). What is not clear from these studies is whether this azole resistance in vitro translates into a decreased therapeutic response to azoles in vivo, particularly considering the increased fragility of the erg3 mutants.
Very recently, it has been reported that two clinical C. albicans isolates exhibiting defective activity of sterol C5,6-desaturase in their sterol compositions showed reduced virulence in mice and impaired hyphal formation in vitro compared to azole-susceptible clinical isolates (4). In addition, attenuated virulence of a laboratory strain (erg3Δ::hisG/erg3Δ::hisG-URA3-hisG erg11Δ::hisG/ERG11) generated by the “Ura-blaster” technique was shown (4). Although “Ura-blaster” is a useful method for gene disruption in C. albicans, a positional change of URA3 affects the expression level and activity of Ura3p, orotidine 5′-monophosphate decarboxylase (3, 19, 36). Because reduced URA3 expression itself attenuates virulence of C. albicans (3, 19, 36), effects of a gene disruption on virulence should be evaluated under the same conditions for the URA3 locus. To avoid positional effects on URA3 expression, reintroduction of URA3 into its original locus or an appropriate expression locus such as the RPS10 locus has been suggested (3, 36). For our in vivo studies, therefore, a wild-type URA3 was placed back into its native locus in the erg3 homozygote, and this allowed us to use a well-known control strain, CAF2-1 (8), which is derived from the wild-type isolate SC5314 (10) and is also a ura3/URA3 heterozygote. Here, we present a detailed evaluation of the erg3 mutant phenotype in C. albicans and cast doubt on the clinical relevance of this mechanism of resistance.
The C. albicans strains used in this study are listed in Table Table1.1. The C. albicans strains were routinely propagated in yeast peptone dextrose (YPD) medium (1% yeast extract, 2% peptone, 2% dextrose). The URA3 transformants were selected on minimal (MIN) (0.7% yeast nitrogen base without amino acid, 2% glucose) agar plates. The ura3 auxotrophs were obtained on MIN agar plates containing 0.1% 5-fluoroorotic acid (5-FOA; Lancaster, Pelham, NH) and 50 μg/ml uridine (8). Agar (1.5%) was added for solid media. RPMI 1640 medium was buffered with 0.165 M morpholinepropanesulfonic acid and was adjusted to pH 7.0. When needed, 50 μg/ml uridine was added. Escherichia coli strains were grown in Luria-Bertani medium containing 100 μg/ml ampicillin at 37°C.
All PCR products used in plasmid construction were sequenced before use. Transformation in C. albicans was performed by electroporation (Gene Pulser; Bio-Rad Laboratories, Richmond, CA) as described previously (39).
The 5′ end (0.3 kb) of the ERG3 open reading frame (ORF) was amplified with primers Tg1 (5′-ATGGATATCGTACTAGAAATTTGTG-3′) and Tg4 (5′-GCTGGGAAAAATTTAGGAGC-3′) from genomic DNA of strain CAI-4 (8). The PCR product was inserted into the BglII site of a plasmid containing the hisG-URA3-hisG cassette, p5921 (8), to yield pE3DC1. The 3′ end (0.4 kb) of the ERG3 ORF was obtained with primers Tg2 (5′-TCATTGTTCAACATATTCTCTATCG-3′) and Tg3 (5′-TCCAGTTGATGGGTTCTTCC-3′) and inserted into the BamHI site of pE3DC1 to yield pE3DC2. Amplified DNA products and digested fragments of p5921 and pE3DC1 were blunt ended (DNA Blunting kit; Takara) before ligation. The orientation of the inserted PCR products of ERG3 ORF 5′ and 3′ regions was verified at each step by PCR with primer pairs Tg1 and K11 (URA3 specific) (5′-GCTAACATCAATAACCCTCTTGGC-3′) for pE3DC1 and K10 (URA3 specific) (5′-CTGAGCAACAACCCCATACACAC-3′) and Tg2 for pE3DC2, respectively. Ten million CAI-4 cells were transformed with 2 μg of a 5-kb SacI-PstI fragment excised from pE3DC2. Ura+ transformants were obtained on MIN agar plates, and then Ura− isolates resulting from cis recombination between the hisG repeats were selected using 5-FOA (8). We performed sequential disruption of the C. albicans ERG3 gene by using the Ura-blaster technique again to yield erg3/erg3 strains.
A 5-kb BglII-PstI fragment containing the IRO1-URA3 locus (5, 20) was obtained from pLUBP, a kind gift from William A. Fonzi. Plasmid pLUBP consists of a pLITMUS28 backbone with a 5-kb BglII-PstI insert obtained from pUR3 (16). The Ura− erg3/erg3 strain, CAE3D, was transformed with 1 μg of this 5-kb fragment to place wild-type URA3 back into its original locus as described previously (5, 20). Transformants were selected by Ura prototrophy. Homologous recombination and no ectopic integration of the transforming DNA were confirmed by Southern blotting.
The growth rates of C. albicans strains were examined by the optical density at 600 nm (OD600) every hour. Tested media included YPD, MIN, yeast peptone glycerol (1% yeast extract, 2% peptone, 3% glycerol, 1% ethanol), and RPMI 1640, and tested growth temperatures included 25, 30, 37, 40, and 42°C. An overnight culture grown at 30°C was diluted 1 to 500 into each medium, and then the cultures were incubated in 250-ml flasks with shaking at 200 rpm.
Logarithmic-phase cultures were obtained by preculture in YPD medium. Cells were harvested, washed, and adjusted to the desired concentrations by counting the number of cells with a hemocytometer. Antifungal susceptibility assay was performed according to the M27-A2 standard protocol approved by the National Committee of Clinical Laboratory Standards (NCCLS) (25). Tested antifungal agents were fluconazole (Pfizer, Inc.), itraconazole (Janssen Pharmaceuticals), miconazole (Mochida, Inc.), and voriconazole (Pfizer, Inc.). RPMI 1640 medium adjusted to pH 7.0 was used. Cells were incubated in 96-well U-bottom microtiter plates at 35°C, and the OD600 was measured by a microplate spectrophotometer (Benchmark Plus; Bio-Rad Laboratories) at 24 and 48 h. The MIC50 was defined as the drug concentration required for 50% growth inhibition compared to that in the drug-free culture. Fluconazole susceptibility was also evaluated by Etest (AB Biodisk, Solna, Sweden) according to the manufacturer's instructions.
Sterol identification was made by gas chromatography-mass spectrometry (Hewlett Packard 6890/5973) using a DB5 capillary column (15 m by 0.25 mm; J&W Scientific), essentially as described previously (2, 13).
Southern blot analysis was performed following the standard protocol (30). The genomic DNA was digested with SalI and PstI. The 0.4-kb PCR product of the 3′ end of the ERG3 ORF (described above) was used as an ERG3 probe to monitor the recombination events. Both pre- and post-5-FOA isolates were also verified using an URA3 probe, which was obtained by PCR with primers K10 and K11 from p5921. The genomic DNA of the reconstituted Ura+ erg3/erg3 strain, CAE3DU3, was digested with HindIII and hybridized with the URA3 probe.
Northern blot analysis was performed following the methods described previously (41). Briefly, logarithmic-phase cultures at an OD600 of 0.75 were reincubated at 35°C in the absence and the presence of fluconazole at a concentration of 0.25 μg/ml. Total RNA was extracted when the culture reached an OD600 of 1.0 (approximately 90 min of incubation). An ERG3 probe for Northern blotting was amplified with primers designed in the deleted region of ERG3 ORF, Tg10 (5′-GGAAGAACCCATCAACTGGATGG-3′) and Tg11 (5′-GTGCCACTACTGCCATTCCA-3′). Gene probes for ERG11, CDR1, and MDR1 were amplified with primers described previously (12). Autoradiography was analyzed with a Fujix BAS-5000 image analyzer (Fuji Photo Film, Tokyo, Japan).
To induce hyphal growth, stationary-phase cells grown in YPD medium at 30°C were plated at approximately 100 cells/plate on spider agar (21), on 10% human serum agar, and on RPMI 1640 medium with 10% human serum agar. YPD agar was used as a control. Plates were incubated at 37°C. The cells were also grown in liquid RPMI 1640 medium in the absence and the presence of 10% human serum under the same conditions as those of the MIC assay. All tested media were adjusted to pH 7.0. Cell morphology was examined after 18-, 48-, and 72-h incubations.
Female, 8-week-old, BALB/c mice (Charles River Laboratories, Danvers, MA) were used in all experiments. Mice were maintained according to National Institutes of Health guidelines for animal care and in fulfillment of American Association for Accreditation of Laboratory Animal Care criteria (6). C. albicans strains for inoculation were grown in YPD medium at 30°C. Logarithmic-phase cells were harvested, washed, resuspended in sterile saline, and adjusted to the desired concentrations by counting the number of cells with a hemocytometer. Actual CFU in the inocula were determined by culturing serial dilutions of each preparation onto YPD plates. Mice were inoculated with a volume of 0.2 ml via the lateral tail vein.
Forty mice were divided into four groups. Ten mice of each group were injected with a higher or a lower inoculum of either CAF2-1 (ERG3/ERG3 ura3/URA3) or CAE3DU3 (erg3/erg3 ura3/URA3) on day 0 of the experiment. The mice were observed twice daily until day 24.
Twenty mice per group were injected with either CAF2-1 or CAE3DU3 on day 0 of the experiment. In each group, kidneys were removed from three mice euthanized on days 2 and 7 and from four mice on day 4. To assess fungal burden in tissue, the excised kidneys were weighed individually and homogenized in sterile saline by using a Precision Tissue Grinder (Kendall, Mansfield, MA). Aliquots of 100 μl from kidney homogenates and their dilutions of 10−1 and 10−2 were plated onto YPD agar. Colonies were counted after 3 days of incubation at 30°C, and CFU per gram of kidney were calculated. The remaining 10 mice in each group were monitored for survival until day 24 of the experiment.
Three mice per group were injected with CAF2-1 or CAE3DU3 on day 0 of the experiment. Both kidneys were excised on day 4 and fixed in 10% neutral buffered formalin. Paraffin-embedded tissue sections were stained with Grocott-Gomori methenamine silver stain. Tissues were microscopically examined for morphology of C. albicans cells.
Twenty mice per group were injected with either CAF2-1 or CAE3DU3 on day 0 of the experiment. The mice were treated with fluconazole (Diflucan; Pfizer. Inc.) given by gavage at 40 mg/kg of body weight once a day for 4 days, starting at 3 h after inoculation. As a control, mice were treated with the equivalent volume (0.2 ml/gavage) of sterile saline. Kidneys were excised from all mice on day 4 of the experiment, and kidney CFU were determined as described above.
Multivariate regression analyses with log CFU as the dependent variable were used to assess the difference between the two groups in the in vivo virulence assay. The estimated group difference in log CFU and its associated 95% confidence intervals are presented. F tests were used to derive P values for assessing the significance of the group. Log-rank tests were used to compare the survival rates of mice. In the [3H]fluconazole accumulation assay and the in vivo fluconazole treatment experiment, Student's t test was used to analyze differences between mean values of groups of data. A significance level of 0.05 was used to determine statistical significance. All analyses were conducted using STATA version 8.2 (STATA Corp., College Station, TX).
Both copies of ERG3 in C. albicans strain CAI-4 (8) were disrupted sequentially by means of the Ura-blaster technique and 5-FOA selection, yielding the following strains: the Ura+ erg3/ERG3 strain (CAD1U), Ura− erg3/ERG3 strain (CAD1), Ura+ erg3/erg3 strain (CAE3DU), and Ura− erg3/erg3 strain (CAE3D) (Table (Table1).1). Each strain construction was confirmed by Southern blotting with the ERG3 or the URA3 probe (data not shown). The IRO1-URA3 locus of CAE3D was reconstituted by transformation with a 5-kb BglII-PstI fragment of pLUBP. Southern blotting with the URA3 probe confirmed that a copy of URA3 was placed back to its native locus in the reconstituted Ura+ erg3/erg3 strain, CAE3DU3 (data not shown).
We examined the effect of ERG3 disruption on the growth rate of C. albicans before performing susceptibility assays. As representative data, doubling times of each strain in YPD medium at 30°C were as follows: 100 min for CAF2-1 and CAE3DU3, 123 min for CAD1U and CAE3DU, and 145 min for CAI-4, CAD1, and CAE3D. The increases in doubling times were due to an ectopic expression of URA3 at the ERG3 locus (CAD1U and CAE3DU) and uracil auxotrophy (CAI-4, CAD1, and CAE3D). The growth ability of C. albicans was not affected by ERG3 disruption under the various conditions, including different media and growth temperatures (25, 37, 40, and 42°C) as described in Materials and Methods.
Antifungal susceptibilities of the erg3 disruptants were determined by broth dilution tests following the NCCLS M27-A2 protocol (25) (Table (Table2).2). Although heterozygous disruption of ERG3 did not affect antifungal susceptibilities, the erg3 homozygotes CAE3D and CAE3DU3 were found to be resistant to fluconazole in RPMI 1640 medium with a MIC of >64 μg/ml and were also resistant to other azoles, such as itraconazole (MIC, >16 μg/ml), miconazole (MIC, >16 μg/ml), and voriconazole (MIC, >16 μg/ml). The absence of URA3 did not affect the antifungal susceptibilities.
Some fluconazole-resistant C. albicans isolates have been shown to exhibit significant trailing growth, which is also known as a low-high MIC phenotype (23, 29). Fluconazole MICs for strains having this type of growth appear to be low at 24 h but are much higher at 48 h. To examine whether the low-high MIC phenotype can be seen in the erg3 homozygote, the fluconazole MIC was measured at 24 and 48 h (Fig. (Fig.1).1). Although the optical density of the growth control was relatively low at 24 h, the erg3 homozygote, CAE3DU3, consistently showed fluconazole resistance with a MIC of >64 μg/ml at both 24 and 48 h. The fluconazole MIC of CAE3DU3 was also measured by Etest and was >256 μg/ml at both 24 and 48 h; meanwhile, that of the control strain, CAF2-1, was 0.25 μg/ml at 24 h and 0.38 μg/ml at 48 h.
We confirmed the phenotypes of the erg3 hetero- and homozygotes by sterol assay (Table (Table3).3). Sterol contents of the erg3 heterozygote, CAD1, were similar to those of CAI-4. On the other hand, in the erg3 homozygote CAE3D, ergosta-7,22-dien-3β-ol accumulated in place of ergosterol and no C5,6-desaturated sterols were detected. The presence of URA3 did not affect sterol composition (data not shown). The sterol profile of CAE3D was consistent with that of the previously reported C. albicans erg3 mutant (erg11/ERG11 erg3/erg3) (33).
Northern blot analysis of CAE3DU confirmed the lack of ERG3 transcript and the increased ERG11 expression compared to that of CAF2-1 and CAD1U (Fig. (Fig.2).2). We sequenced ERG11 in the erg3 homozygote, CAE3D, and found no difference from published sequence data. In addition, there was no difference in expression levels of CDR1, an ATP-binding cassette transporter gene (34), in the presence of fluconazole among CAF2-1, CAD1U, and CAE3DU (data not shown). MDR1, a gene encoding a membrane transport protein of the major facilitator superfamily (34), was not expressed detectably in any of the strains (data not shown).
The effect of ERG3 disruption on the filamentous growth of C. albicans was monitored under known hyphae-inducing conditions (7, 21). CAF2-1 and the Ura+ erg3 heterozygote, CAD1U, formed abundant hyphae under all tested conditions except YPD agar and RPMI 1640 broth (see Materials and Methods). In contrast, the Ura+ erg3 homozygote, CAE3DU3, did not show a filamentous form including germ tube, pseudohypha, and hypha under any of the tested conditions (data not shown).
To clearly assess the effects of defective ERG3 on virulence, survival and C. albicans burden in kidney tissue were monitored in mice intravenously inoculated with CAE3DU3, an erg3/erg3 strain containing a copy of URA3 at the native locus, versus CAF2-1 (ERG3/ERG3 ura3/URA3). Again, these paired strains showed the same growth rate at a variety of temperatures in vitro. Actual CFU (CFU/mouse) inoculated into mice for monitoring their survival were 0.904 × 106 and 4.52 × 105 for CAF2-1 and 0.922 × 106 and 4.61 × 105 for CAE3DU3. At both of higher and lower inoculum sizes, mice injected with CAE3DU3 survived significantly longer (P < 0.001 each) than those injected with CAF2-1 (Fig. (Fig.33).
To assess C. albicans burden in kidney tissue, both kidneys were excised on days 2, 4, and 7 of the experiment from mice infected with CAF2-1 or CAE3DU3 (Table (Table4).4). Inocula were 4.65 × 105 for CAF2-1 and 4.79 × 105 CFU/mouse for CAE3DU3. Concurrently, 10 mice of each inoculum group were monitored for survival. Survival curves of both groups were consistent with the results shown in Fig. Fig.3,3, and no mice died before day 8 of this experiment. Adjusted for the date of sacrifice, the mean log CFU of CAF2-1 was 1.86 (95% confidence interval, 0.91 and 2.81) higher (F test; P = 0.001) than that of CAE3DU3. The reduced virulence of CAE3DU3 was shown by both the longer survival (Fig. (Fig.3)3) and the lower kidney burdens (Table (Table4)4) of mice inoculated with this strain relative to CAF2-1.
For histopathologic analysis, three mice per group were injected with CAF2-1 and CAE3DU3. Actual CFU (CFU/mouse) of each inoculum were 4.91 × 105 for CAF2-1 and 4.85 × 106 for CAE3DU3. Kidneys were excised 4 days after injection, and tissue sections were stained with Grocott-Gomori methenamine silver stain (Fig. (Fig.4).4). Kidney histopathology revealed that CAF2-1 cells formed abundant and intact hyphae, but almost all CAE3DU3 cells were blastospores. No intact hypha was detected, but aborted hyphal formation was observed in kidney tissues infected with CAE3DU3 (Fig. (Fig.4,4, arrows).
Murine candidiasis caused by CAE3DU3 was treated with fluconazole to examine whether the erg3 homozygote shows fluconazole resistance in vivo as observed in vitro. It is difficult to assess drug efficacy between two strains having different virulence levels. Normally, it is easier to treat strains with low levels of virulence than those with high levels of virulence. In this study, therefore, mice were injected with CAE3DU3 (4.85 × 106 CFU/mouse) at a 10-fold higher concentration than CAF2-1 (4.91 × 105 CFU/mouse), and kidney CFU of both strains obtained from untreated groups were comparable (P = 0.3) (Fig. (Fig.5).5). No mice died until euthanasia for kidney removal on day 4. Fluconazole treatment by gavage at 40 mg/kg/day for 4 days after injection effectively reduced kidney CFU of CAF2-1 compared to that of the saline-treated group (P < 0.0001) (Fig. (Fig.5).5). Despite the fluconazole resistance measured by standard in vitro testing and a 10-fold higher inoculum compared to that of CAF2-1, kidney CFU of CAE3DU3 in the fluconazole-treated group was significantly less than that in the saline-treated group (P < 0.0001).
In this study, we interrupted both copies of ERG3 in C. albicans and confirmed the known phenotypes, including in vitro azole resistance and the altered sterol profile. Novel findings obtained from our C. albicans erg3 homozygote were as follows: attenuated virulence in mice was accompanied by the reduction of kidney fungal burden and defective hyphal formation was observed in kidney tissues. Lastly, the erg3 homozygote was susceptible to fluconazole in vivo.
The erg3 homozygote showed no overexpression of CDR1 and MDR1 compared to the wild-type strain and contained no mutation in ERG11. The increased ERG11 expression level observed in the erg3 homozygote was consistent with previous reports (28, 33), except the report by Chau et al. (4). A feedback mechanism caused by the homozygous disruption of ERG3, which acts at the late phase (downstream of ERG11) in the ergosterol biosynthesis pathway, may account for the ERG11 overexpression observed in the erg3 homozygote. However, overexpression of ERG11 is thought to have only a modest impact on azole resistance (31, 35).
To our knowledge, there is so far only one report addressing effects of defective C5,6-desaturase on morphology and virulence of C. albicans (4). In that report, no congenic strain was used as a control but defective filamentous growth of the erg3 mutant in the presence of serum was shown, as confirmed here. The longer survival of mice infected with the mutant was not accompanied by a reduction of kidney fungal burdens, in contrast to our findings. Their failure to find a reduced fungal burden in the kidneys is probably due to the selection of a single and early time point 24 h after injection. In our study, a marked attenuation of virulence by ERG3 disruption in C. albicans was equally evident from both the longer survival (Fig. (Fig.3)3) and the lower kidney burdens (Table (Table4)4) of mice inoculated with CAE3DU3 relative to the control strain, CAF2-1. In addition, we found a decreased ability of CAE3DU3 to form intact hyphae not only in vitro but in kidney tissues (Fig. (Fig.4).4). Our results obtained from in vivo experiments using female BALB/c mice and the control strain CAF2-1 were consistent with the data previously reported with that strain (22), indicating that an appropriate internal control had been selected.
Our experiment in mice also found that the fluconazole resistance of the C. albicans erg3 homozygote could not be demonstrated in vivo (Fig. (Fig.4).4). To answer the question about the effect of fluconazole in the experimentally infected mouse, inoculum sizes of CAF2-1 and CAE3DU3 were chosen which provided a similar kidney burden. Under these conditions, kidney burdens of both strains were significantly decreased in the fluconazole-treated groups compared to the saline-treated control groups.
Of all our findings with the erg3 homozygote, the most unexpected was the efficacy of fluconazole in a murine model of disseminated candidiasis. The inactivation of sterol C5,6-desaturase induced fluconazole resistance in vitro, consistent with the previous report (33). However, clinical significance of this resistance mechanism is still controversial, because only a few azole-resistant clinical isolates have exhibited a sterol profile indicative of defective sterol C5,6 desaturation (4, 18, 26). Information is limited because that mutation has not often been sought in clinical isolates. Another reason that such mutants may be rare in the infected host is the decreased virulence of such mutants. What is unclear is whether the erg3 mutation alone contributes to clinically relevant resistance. Several studies in C. albicans have confirmed that multiple mechanisms are often involved in high-level resistance to fluconazole in an individual isolate. Both mutations in the ERG11 gene and increased drug efflux are quite common (27, 35). Although a possibility remains that an erg3 mutation spontaneously occurred in clinical settings and may have a role in azole resistance when combined with other mechanisms, this study suggests that an erg3 mutation causing inactivation of sterol C5,6-desaturase is unlikely to confer in vivo fluconazole resistance by itself.
We thank Yoko Kawamura for sterol assay, Chiung-Yu Huang and Dean Follmann for statistical analyses, Katsunori Yanagihara, Yoichi Hirakata, and Kazunori Tomono for helpful discussion, and William A. Fonzi for pLUBP.
This research was supported by grants from the Japanese Ministry of Education (Grant-in-Aid for Scientific Research) and the Japanese Ministry of Health and Welfare and by the Intramural Research Program of the NIH, National Institute of Allergy and Infectious Diseases.