Targeted gene disruption in Candida albicans wild-type strains: the role of the MDR1 gene in fluconazole resistance of clinical Candida albicans isolates



Resistance of the pathogenic yeast Candida albicans to the antifungal agent fluconazole is often caused by active drug efflux out of the cells. In clinical C. albicans strains, fluconazole resistance frequently correlates with constitutive activation of the MDR1 gene, encoding a membrane transport protein of the major facilitator superfamily that is not expressed detectably in fluconazole-susceptible isolates. However, the molecular changes causing MDR1 activation have not yet been elucidated, and direct proof for MDR1 expression being the cause of drug resistance in clinical C. albicans strains is lacking as a result of difficulties in the genetic manipulation of C. albicans wild-type strains. We have developed a new strategy for sequential gene disruption in C. albicans wild-type strains that is based on the repeated use of a dominant selection marker conferring resistance against mycophenolic acid upon transformants and its subsequent excision from the genome by FLP-mediated, site-specific recombination (MPAR-flipping). This mutagenesis strategy was used to generate homozygous mdr1/mdr1 mutants from two fluconazole-resistant clinical C. albicans isolates in which drug resistance correlated with stable, constitutive MDR1 activation. In both cases, disruption of the MDR1 gene resulted in enhanced susceptibility of the mutants against fluconazole, providing the first direct genetic proof that MDR1 mediates fluconazole resistance in clinical C. albicans strains. The new gene disruption strategy allows the generation of specific knock-out mutations in any C. albicans wild-type strain and therefore opens completely novel approaches for studying this most important human pathogenic fungus at the molecular level.


Candida albicans is an important opportunistic fungal pathogen of humans that causes superficial as well as disseminated infections in immunocompromised patients (Odds, 1988). The antifungal agent fluconazole, which inhibits the synthesis of ergosterol, the major sterol in the fungal plasma membrane, is a widely used compound for treating C. albicans infections. In recent years, however, fluconazole-resistant strains have emerged, especially in AIDS patients with recurrent oropharyngeal candidiasis (OPC) receiving prolonged fluconazole therapy (Rex et al., 1995; Ghannoum et al., 1996). Different mechanisms may be responsible for the development of drug resistance in previously susceptible C. albicans strains, for example mutations in the drug target, sterol 14α-demethylase (14DM), which lower the affinity of the enzyme for the drug, or enhanced expression of the ERG11 gene encoding 14DM (Vanden Bossche et al., 1994; White, 1997a, b; Franz et al., 1998; Lopez-Ribot et al., 1998; Sanglard et al., 1998;). Another common resistance mechanism is increased fluconazole efflux out of the cell, and this correlates with enhanced expression of certain multiple drug resistance genes, the ATP-binding cassette (ABC) transporters CDR1 and CDR2 and/or the major facilitator MDR1 (Sanglard et al., 1995, 1997; White, 1997a; Franz et al., 1998; 1999; Lopez-Ribot et al., 1998). The MDR1 gene is not usually expressed at detectable levels in fluconazole-susceptible C. albicans isolates, but is constitutively activated in many fluconazole-resistant isolates. When overexpressed in Saccharomyces cerevisiae, MDR1, like CDR1 and CDR2, conferred fluconazole resistance upon transformants (Sanglard et al., 1995). However, whereas inactivation of CDR1 in C. albicans rendered the cells hypersusceptible to fluconazole, homozygous mdr1/mdr1 mutants of C. albicans CAI4 or CAF4-2, auxotrophic strains that are frequently used for genetic manipulations, did not exhibit enhanced sensitivity to fluconazole (Sanglard et al., 1996; Morschhäuser et al., 1999). This observation might result from the fact that, in contrast to CDR1, the MDR1 gene is not or is barely expressed in strains CAI4 and CAF4-2.

In S. cerevisiae, pleiotropic drug resistance is also mediated by membrane transport proteins of the ABC transporter and major facilitator superfamilies. The expression of genes encoding these transport proteins is controlled by the regulatory proteins Pdr1p, Pdr3p and yAP1p (Meyers et al., 1992; Balzi et al., 1994; Katzmann et al., 1994; 1995; Wemmie et al., 1994; Mahéet al., 1996; Alarco et al., 1997; Coleman et al., 1997; Nourani et al., 1997a; Wolfger et al., 1997), and mutations in these regulators have been identified that result in upregulation of their respective target genes (Carvajal et al., 1997; Nourani et al., 1997b). Recently, functional homologues of these regulators, CAP1 and FCR1, have also been found in C. albicans (Alarco et al., 1997; Talibi and Raymond, 1999). Although a role for CAP1 and FCR1 in fluconazole resistance of clinical C. albicans strains has not yet been demonstrated, it is conceivable that mutations in such regulators might affect the expression of several, including unknown, target genes, each of which could contribute to drug resistance. Direct assessment of the role of one of the known efflux pumps in C. albicans drug resistance requires its inactivation in fluconazole-resistant clinical isolates in which it is overexpressed. However, because of the lack of a dominant selection marker, the genetic manipulation of C. albicans has been largely confined to auxotrophic strains that are tedious to generate from wild-type strains, and no such investigation has been performed to date.

In a previous study, we have described two series of C. albicans isolates from AIDS patients with recurrent episodes of OPC that developed fluconazole resistance during therapy (Franz et al., 1998). It was shown by DNA fingerprinting that, in both cases, fluconazole resistance had developed in a previously susceptible strain and that a combination of different mechanisms had contributed to a gradual increase in drug resistance. In both series of isolates, the observed reduced intracellular drug accumulation correlated with high levels of MDR1 mRNA that was undetectable in the corresponding fluconazole-susceptible isolates. As for other clinical C. albicans strains, the precise molecular changes that resulted in constitutive MDR1 expression in these two strains are still unknown, but we were able to demonstrate that MDR1 activation was caused by mutations in a trans-regulatory factor (Wirsching et al. 2000). This raised the possibility that, in addition to MDR1, other unknown genes might have been activated and be the actual cause of drug resistance.

We have reported recently a new strategy for sequential gene disruption in C. albicans that is based on the repeated use of the URA3 marker for integrative transformation in a ura3-negative host strain and its subsequent deletion by FLP-mediated, site-specific recombination (Morschhäuser et al., 1999). This mutagenesis procedure eliminates the need for a negative selection scheme for marker removal, which, after inactivation of the first copy of a target gene, is necessary for the second round of transformation to generate homozygous mutants in the permanently diploid fungus C. albicans. In the present study, we have combined this mutagenesis strategy with the use of the dominant selection marker MPAR, a mutated form of the C. albicans IMH3 gene encoding inosine monophosphate dehydrogenase, which confers resistance against mycophenolic acid (MPA) on C. albicans transformants (Köhler et al., 1997; S. Theiß and G. Köhler, in preparation). This new mutagenesis procedure (MPAR flipping) allowed us to disrupt the MDR1 gene in MDR1-overexpressing clinical C. albicans isolates from the two series mentioned above to assess its contribution to fluconazole resistance.


Sequential gene disruption by MPAR-flipping

The MPARflipping cassette contained in plasmid pSFI1 was constructed as described in Experimental procedures and is depicted in Fig. 1A. It consists of the MPAR selection marker and a genetically engineered FLP gene under the control of the inducible SAP2 promoter and is flanked by direct repeats of the 34 bp minimal FLP recombination target site. There are several unique restrictions sites on both sides of the cassette, which can be used to excise the MPAR flipper from pSFI1 and insert it into target genes by a single cloning step.

Figure 1.

A. Genetic structure of the MPAR flipper contained in plasmid pSFI1. The promoter of the C. albicans SAP2 gene (PSAP2) is indicated by the thin arrow; the transcription termination sequence of the C. albicans ACT1 gene (ACT1T) is indicated by the filled circle. FRT, FLP recombination target; caFLP, C. albicans-adapted FLP gene; MPAR, MPA resistance marker. Only relevant restriction sites are given; unique sites that can be used to cut out the MPAR flipper from pSFI1 are shown in bold. A, ApaI; N, NotI; P, PstI; Sl, SalI; ScI, SacI; ScII, SacII; Xh, XhoI.

B. Sequential gene disruption by MPAR flipping. The genomic structures of wild-type parent, transformants containing the MPAR flipper (MPARFLIP) and derivatives after deletion of the MPAR flipper are indicated. The open box represents the target gene.

The strategy for sequential gene disruption using the MPAR flipper is outlined in Fig. 1B. After insertion of the cassette between cloned sequences of a target gene, the corresponding linear DNA fragment is used to transform a C. albicans host strain to MPA resistance. The flanking sequences direct integration of the cassette into one allele of the target gene by homologous recombination. Transformants containing the correct insertion are then propagated in YCB-BSA medium, which contains protein as the sole nitrogen source. In this medium, the SAP2 promoter is induced (Hube et al., 1994; Staib et al., 1999), resulting in expression of the FLP gene and excision of the MPAR flipper by FLP-mediated, site-specific recombination. The MPA-sensitive derivatives generated after deletion of the MPAR flipper can easily be identified by their reduced growth on plates containing low amounts of MPA, leading to the formation of smaller colonies compared with MPA-resistant cells (Staib et al., 1999). The heterozygous mutants are then used for a second round of gene disruption that generates the desired homozygous mutants. The MPAR marker used for transformation is not present in the final mutants, which differ from the wild-type parent strain only by the two disrupted alleles of the target gene, both of which contain one copy of the FRT site.

Disruption of the MDR1 gene in fluconazole-resistant clinical C. albicans strains

From each of the two series of clinical C. albicans isolates described in a previous study (Franz et al., 1998), we selected the most resistant isolate that exhibited high MDR1 mRNA levels (isolates G5 and F5, Table 1) to disrupt the MDR1 gene and thereby assess its contribution to drug resistance. To allow for discrimination between all possible recombination events after the second round of transformation (see below), the MPAR flipper was inserted at two different positions within the MDR1 coding region (Fig. 2A). In plasmids pSFIM3 and pSFIM4, MDR1 sequences between positions 784 and 1138 or between positions 606 and 929, respectively, were replaced by the MPAR flipper, such that the flanking upstream and downstream MDR1 regions had different lengths in the two plasmids. In both plasmids, the deleted region includes the signature sequences WRW and PET that are conserved between all members of cluster I of the multidrug resistance subfamily of yeast major facilitators, to which MDR1 from C. albicans belongs (Goffeau et al., 1997).

Table 1. C. albicans strains used in this study.
  1. a . MPA R-FLIP denotes the FRT-PSAP2-FLP-ACT1T-MPAR-FRT cassette (MPAR flipper).

G2Clinical isolate from patient G, fluconazole-susceptible Franz et al. (1998)
G5Clinical isolate from patient G, fluconazole-resistant Franz et al. (1998)
G5M401Derivative of G5, MDR1/mdr1:: MPAR-FLIPaThis study
G5M402Derivative of G5M401, MDR1/mdr1:: FRTThis study
G5M431Derivative of G5M402, mdr1::FRT/mdr1:: MPAR-FLIPThis study
G5M432Derivative of G5M431, mdr1::FRT/mdr1:: FRTThis study
F2Clinical isolate from patient F, fluconazole-susceptible Franz et al. (1998)
F5Clinical isolate from patient F, fluconazole-resistant Franz et al. (1998)
F5M401Derivative of F5, MDR1/mdr1:: MPAR-FLIPThis study
F5M402Derivative of F5M401, MDR1/mdr1:: FRTThis study
F5M431Derivative of F5M402, mdr1::FRT/mdr1:: MPAR-FLIPThis study
F5M432Derivative of F5M431, mdr1::FRT/mdr1:: FRTThis study
Figure 2.

A. Insertion of the MPAR flipper within the MDR1 coding region (open arrow) in plasmids pSFIM3 (top) and pSFIM4 (bottom). The thick lines above the MDR1 coding region represent the probes used for Southern hybridization. The MPAR flipper is not drawn to scale. EI, EcoRI; ScI, SacI; ScII, SacII; Sp, SphI; Xh, XhoI.

B. Genomic structure of the MDR1 alleles in the C. albicans isolates G5 and F5 showing the EcoRI restriction site polymorphism that can be used to distinguish between the two alleles.

C. Possible FLP-mediated recombination events resulting in loss of the MPAR flipper from strain G5M431; the size of the resulting EcoRI fragments is given. (a) Specific excision of the MPAR flipper by intrachromosomal recombination between the flanking FRT sites, resulting in strain G5M432. (b) and (c) Loss of the MPAR flipper by interchromosomal recombination of FRT sites on homologous chromosomes after DNA replication if the centromere is located to the left, as in (b), or to the right, as in (c). Only the MPA-sensitive derivatives resulting from such mitotic recombination events are shown.

The Southern blot in Fig. 3A shows all the steps of MDR1 disruption in isolate G5. The insert from pSFIM4 was used for the first round of transformation. In the parent strain G5, the two MDR1 alleles (arbitrarily designated as MDR1-1 and MDR1-2) can be discriminated by an EcoRI restriction site polymorphism, as illustrated in Fig. 2B, and are located on a 4.9 kb and a 5.7 kb EcoRI fragment (Fig. 3A, lane 1). In the transformant G5M401, the MPAR flipper was inserted into the MDR1-1 allele, as can be seen from the disappearance of the 4.9 kb band and the generation of a new fragment of 3.0 kb ranging from the upstream EcoRI site to an EcoRI site within the MPAR flipper (lane 2). This band disappeared after excision of the MPAR flipper by FLP-mediated recombination of the flanking FRT sites and, in the resulting strain G5M402 (lane 3), a new 4.6 kb fragment was generated that was 254 bp smaller than the original wild-type fragment as a result of the insertion of 70 bp of heterologous sequences (the FRT site plus polylinker sequences) instead of the deleted part of the MDR1 coding region. A faint band of this size was also seen in strain G5M401 (lane 2), demonstrating that, during growth of this strain, some recombination had already occurred without induction of the SAP2 promoter (see Discussion). The heterozygous mutant G5M402 was then transformed with the insert from pSFIM3 to disrupt the remaining intact MDR1 allele. In strain G5M431, the 5.7 kb EcoRI fragment was replaced by a 3.2 kb fragment (lane 4). The MPAR flipper was excised again from the genome of G5M431 by FLP action, resulting in strain G5M432, in which the 3.2 kb fragment was replaced by a new fragment of 5.4 kb, 285 bp smaller than the original wild-type fragment (lane 5). In addition to the desired intrachromosomal recombination between the FRT sites flanking the disruption cassette (a in Fig. 2C), MPA-sensitive derivatives could also now have been generated by mitotic interchromosomal recombination between one of the FRT sites in MDR1-2 and the FRT site in the disrupted MDR1-1 allele on the homologous chromosome. In this case, the cells would have become homozygous for all chromosomal regions centromere-distal from the site of crossing over and would not represent true specific mutants. Depending on the (unknown) orientation of the MDR1 gene with respect to the centromere, such an interchromosomal recombination would have resulted in a new fragment of either 4.8 kb (centromere to the left, b in Fig. 2C) or of 5.2 kb (centromere to the right, c in Fig. 2C). The generation of the 5.4 kb EcoRI fragment in strain G5M432 demonstrates that specific excision of the MPAR flipper had occurred.

Figure 3.

Southern hybridization of EcoRI-digested genomic DNA of the clinical C. albicans isolates and mdr1 mutants with an MDR1-specific probe (the 5′MDR1 fragment, see Fig. 2). The identity of the fragments is shown on the right-hand side of the blots; the molecular sizes are given on the left-hand side of the blots.

A. 1, G5 (MDR1/MDR1); 2, G5M401 (MDR1/mdr1::MPAR-FLIP); 3, G5M402 (MDR1/mdr1::FRT); 4, G5M431 (mdr1::FRT/mdr1::MPAR-FLIP); 5, G5M432 (mdr1::FRT/mdr1::FRT).

B. 6, F5 (MDR1/MDR1); 7, F5M401 (MDR1/mdr1::MPAR-FLIP); 8, F5M402 (MDR1/mdr1::FRT); 9, F5M431 (mdr1::FRT/mdr1::MPAR-FLIP); 10, F5M432 (mdr1::FRT/mdr1::FRT).

In the same way as described for isolate G5, homozygous mdr1/mdr1 mutants were also generated from the clinical isolate F5. In this strain, the two MDR1 alleles can also be distinguished from each other by an EcoRI restriction site polymorphism (Fig. 2B) and are located on 4.9 kb and 9.0 kb fragments (Fig. 3B, lane 6). After transformation with the insert from pSFIM4, strain F5M401 was obtained, which contained the MPAR flipper inserted into the MDR1 allele located on the 9.0 kb fragment (lane 7). Excision of the MPAR flipper resulted in strain F5M402 (lane 8), which was subsequently transformed with the insert from pSFIM3. In the transformant F5M431, the MPAR flipper was inserted into the remaining wild-type allele (lane 9), from which it was excised to result in strain F5M432 (lane 10). The genomic structure of all mutant derivatives of strains G5 and F5 was also analysed by Southern hybridization with a probe from the 3′MDR1 flanking region (see Fig. 2A) and by digestion of the DNA with EcoRV, which produces smaller hybridizing fragments, confirming the correct integration and specific excision of the MPAR flipper (data not shown).

Phenotypic analysis of mdr1/mdr1 mutants

Transcription of the MDR1 gene in wild-type strains and heterozygous and homozygous mutants was analysed by Northern hybridization with an MDR1–ACT1 hybrid probe (Fig. 4). As described previously (Franz et al., 1998), MDR1 transcription could not be detected in the fluconazole-sensitive isolates G2 (lane 1) and F2 (lane 5), whereas the corresponding resistant isolates G5 (lane 2) and F5 (lane 6) exhibited high levels of MDR1 mRNA. MDR1 mRNA levels were reduced in the heterozygous mutants G5M402 (lane 3) and F5M402 (lane 7), and no intact MDR1 transcript was detected in the homozygous mutants G5M432 (lane 4) and F5M432 (lane 8).

Figure 4.

Northern hybridization of total RNA from the clinical C. albicans isolates and mdr1 mutants with an MDR1–ACT1 hybrid probe. The identity of the mRNAs is indicated; the ACT1 mRNA served as an internal control. 1, G2 (MDR1/MDR1); 2, G5 (MDR1/MDR1); 3, G5M402 (MDR1/mdr1); 4, G5M432 (mdr1/mdr1); 5, F2 (MDR1/MDR1); 6, F5 (MDR1/MDR1); 7, F5M402 (MDR1/mdr1); 8, F5M432 (mdr1/mdr1).

To assess the contribution to fluconazole resistance of MDR1 activation in isolates G5 and F5, the minimum inhibitory concentration (MIC) of fluconazole of the mdr1 mutants was determined and compared with that of the corresponding clinical isolates (Fig. 5A). In the homozygous mutant G5M432, the MIC of fluconazole was reduced fourfold compared with the fluconazole-resistant parent strain G5, but was still considerably higher than that of the matched fluconazole-susceptible isolate G2. The heterozygous mutant G5M402 displayed an intermediate MIC. Similarly, the homozygous mutant F5M432 exhibited a fourfold reduced MIC of fluconazole compared with the fluconazole-resistant parent strain F5, but the MIC was still somewhat higher than that of the matched fluconazole-susceptible isolate F2. Again, the heterozygous mutant F5M402 showed an intermediate MIC. These results demonstrate that activation of the MDR1 gene contributed to fluconazole resistance in the two clinical C. albicans isolates, but that additional mechanisms also played a role in the development of drug resistance in these strains.

Figure 5.

Susceptibility of the clinical C. albicans isolates and mdr1 mutants against (A) fluconazole, (B) ketoconazole and (C) 4-nitroquinoline-N-oxide. The MIC of each drug is given (in μg ml−1).

To assess the specificity of the effects of MDR1 disruption, the resistance of the mutants against two other drugs was also analysed and compared with that of the corresponding clinical isolates. Overexpression of the C. albicans MDR1 gene in S. cerevisiae confers resistance against the triazole fluconazole, but not against the imidazole ketoconazole (Sanglard et al., 1995), suggesting that ketoconazole is not a substrate for Mdr1p. In accordance with these findings, the mdr1 mutants of the two clinical isolates were not more susceptible against ketoconazole than their parent strains (Fig. 5B). MDR1 did not therefore contribute to the elevated resistance against ketoconazole of isolates G5 and F5 compared with the matched clinical isolates G2 and F2. In contrast, MDR1 confers resistance to 4-nitroquinoline-N-oxide (4-NQO) when overexpressed in S. cerevisiae (Ben-Yaacov et al., 1994). The fluconazole-resistant clinical isolates G5 and F5 also exhibited enhanced resistance against 4-NQO compared with isolates G2 and F2 (Fig. 5C). Resistance against 4-NQO was reduced in the heterozygous mutants G5M402 and F5M402, and disruption of both copies of the MDR1 gene in strains G5M432 and F5M432 completely abolished 4-NQO resistance; the mutants were even slightly more susceptible than the matched clinical isolates G2 and F2. Activation of the MDR1 gene was therefore the only mechanism responsible for 4-NQO resistance in the clinical isolates G5 and F5.


The results of this study provide the first direct genetic evidence that MDR1 is responsible for fluconazole resistance of clinical C. albicans strains that have activated the gene under conditions in which it is normally repressed, as MDR1 disruption in two unrelated C. albicans isolates resulted in diminished resistance against the drug. Heterozygous mutants displayed reduced MDR1 mRNA levels and an intermediate resistance phenotype, demonstrating that both alleles of the MDR1 gene were activated in the parent strains and contributed to fluconazole resistance. This finding is in agreement with our recent observation that MDR1 activation in these two clinical isolates was caused by mutations in a trans-regulatory factor and not by promoter mutations in one of the two alleles (Wirsching et al. 2000). Disruption of both MDR1 alleles, however, did not completely abrogate the enhanced fluconazole resistance of isolates G5 and F5 compared with the matched clinical isolates G2 and F2. This confirms our previous results showing that multiple molecular mechanisms contributed to a stepwise development of fluconazole resistance in these two series of isolates. In addition to an activated MDR1 gene, isolate G5 had a mutation in both alleles of the ERG11 gene encoding the azole target enzyme 14DM. This mutation was not present in the previous isolates of this series, including G2, and led to a reduced fluconazole susceptibility of 14DM, as seen by the 10-fold higher drug concentration that was necessary to inhibit enzyme activity in cell-free extracts (Franz et al., 1998). Isolate F5, in addition to MDR1 activation, expressed the ERG11 gene at strongly enhanced levels compared with the previous isolates of this series, including F2. The increased ERG11 expression also contributed to fluconazole resistance of isolate F5, although to a lesser degree, as suggested by the twofold increase in the drug concentration needed to inhibit 14DM activity in cell-free extracts (Franz et al., 1998). The difference in the fluconazole susceptibilities of the mdr1 mutant F5M432 and the matched clinical isolate F2 observed in the present study is in agreement with this conclusion.

Mutations in ERG11 or enhanced expression of the gene can cause increased azole resistance, but would not be expected to affect susceptibility against drugs that are not ergosterol biosynthesis inhibitors. Accordingly, these additional alterations did not contribute to the enhanced resistance of isolates G5 and F5 against 4-NQO, which is a supposed substrate of the Mdr1p efflux pump (Ben-Yaacov et al., 1994). In contrast to fluconazole, resistance to 4-NQO was completely abolished in the mdr1 mutant strains G5M432 and F5M432, suggesting that MDR1 activation was the only mechanism responsible for the increased resistance of isolates G5 and F5 against this compound. Ketoconazole, on the other hand, is not supposed to be a substrate for Mdr1p (Sanglard et al., 1995). In accordance with this, resistance of the isolates G5 and F5 against ketoconazole was not as strongly increased in comparison with the matched susceptible isolates G2 and F2, respectively, as that seen for fluconazole (see Fig. 5A and B), because only mutation or altered expression of the ERG11 gene, but not MDR1 activation, would contribute to ketoconazole resistance. Correspondingly, MDR1 disruption did not influence the susceptibility of the mutants against ketoconazole, confirming the specificity of the effect of MDR1 inactivation.

The assessment of the role of a specific gene in a certain phenotype requires its inactivation in an appropriate host strain. As drug resistance is a trait of only a subset of C. albicans strains, the usual approach using an auxotrophic model strain for targeted gene disruption was not feasible for evaluating the role of MDR1 in fluconazole resistance. Positive selection of transformants of C. albicans wild-type strains has only recently become possible with the development of the dominant selection marker MPAR (S. Theiß and G. Köhler, in preparation). However, for the construction of homozygous mutants, the selection marker has to be removed from the genome after the disruption of the first allele of a target gene to allow its use for inactivation of the remaining intact copy. Using the standard ‘blaster’ strategy that relies on the spontaneous loss of a marker by homologous recombination between direct repeats flanking the marker (Fonzi and Irwin, 1993), this would still be a laborious task because of the low frequency of the deletion event and because, in contrast to the URA3 marker (Boeke et al., 1984), loss of the MPAR marker cannot be positively selected for in the presence of the wild-type IMH3 gene. The ‘flipper’ strategy that relies on marker loss by induced, FLP-mediated, site-specific recombination circumvents this problem. Markerless derivatives are obtained with high frequency (Morschhäuser et al., 1999), and MPA-sensitive colonies can easily be detected on suitable indicator plates (Staib et al., 1999). As the final mutants do not contain a heterologous gene (or an original C. albicans gene either absent or integrated at a heterologous site in the genome, as in the case of the URA3 blaster and URA3 flipper strategies), they differ from the parent only by the inactivated target gene and are otherwise still prototrophic wild-type C. albicans, thus eliminating the risk of undesired effects not related to target gene disruption (Lay et al., 1998).

FLP-mediated loss of the MPAR marker occurred at a detectable frequency even before induction of the PSAP2FLP fusion (for example, see Fig. 3, lane 9), indicating that the SAP2 promoter was induced in some cells even in SAP2-repressing medium. This was not observed in the construction of mutants using the URA3 flipper (Morschhäuser et al., 1999), probably because ura3-negative derivatives would rapidly be overgrown by prototrophic cells in YPD medium not supplemented with uridine. In contrast, there is no selection pressure for the maintenance of the MPAR marker in the absence of MPA, and MPA-sensitive cells that have lost the marker can be recovered. Nevertheless, this finding came as a surprise, because we did not detect FLP-mediated excision of the MPAR marker under SAP2-repressing conditions when a PSAP2FLP fusion was integrated at the original SAP2 locus in the Candida genome (Staib et al., 1999). This observation suggests that there is some deregulation of the SAP2 promoter when it is integrated at an ectopic site. This does not really affect the construction of mutants using the MPAR-flipping strategy, but it should be kept in mind when one analyses the regulation of a promoter using ectopically integrated reporter constructs.

The direct assignment of a role for MDR1 in fluconazole resistance of C. albicans is important with respect to efforts to search for inhibitors of efflux pumps mediating drug resistance. However, the implications of the possibility of constructing specific mutants from any C. albicans strain extend far beyond the analysis of resistance mechanisms to the study of the biology and pathogenicity of this fungus at the molecular level. Strains within the species C. albicans differ not only in drug resistance but also with respect to virulence factors, such as adhesion to various host surfaces, secretion of hydrolytic enzymes or phenotypic switching (Kondoh et al., 1987; Calderone and Braun, 1991; Soll et al., 1993; Ibrahim et al., 1995). The widely used model strain SC5314, from which CAI4 and similar auxotrophic mutants are derived, certainly does not represent all the different C. albicans strains able to cause infection, and the relative importance of single virulence traits probably differs among the various strains, some being better adapted to colonization/infection of specific mucosal surfaces and others with a higher capacity to cause disseminated infection. Molecular genetic approaches such as targeted gene disruption can now be applied to C. albicans strains with special characteristics and thus serve to corroborate, modify or expand knowledge gained from the analysis of the model strain SC5314.

Experimental procedures

C. albicans strains and growth conditions

C. albicans strains used in this study are listed in Table 1. The clinical isolate pairs G2 and G5 and F2 and F5 are fluconazole-susceptible and resistant isolates of two different series of matched isolates (i.e. representing the same C. albicans strain) from AIDS patients with OPC and have been described previously (Franz et al., 1998). G2 and F2 are the last isolates in each of the two series that did not express the MDR1 gene detectably, whereas G5 and F5 were isolated from later OPC episodes in the same patients and are the isolates with the highest MIC of fluconazole, displaying high MDR1 mRNA levels. Strains were kept as frozen stocks at −80°C and were subcultured on YPD agar plates (10 g of yeast extract, 20 g of peptone, 20 g of glucose, 15 g of agar l−1) at 30°C. For routine growth of the strains, YPD liquid medium was used and, for induction of the SAP2 promoter, YCB-BSA (23.4 g of yeast carbon base, 4 g of BSA l−1, pH 4.0).

Construction of plasmids

Plasmid pSFI1 that contains the MPAR flipper (see Fig. 1A) was constructed by replacing the SalI–PstI fragment with the URA3 gene in plasmid pSFU1 (Morschhäuser et al., 1999) with a SalI–PstI fragment containing the MPAR marker from plasmid pAFI3 (Staib et al., 1999). Plasmids pSFIM3 and pSFIM4 contain the MPAR flipper inserted between MDR1 sequences and were constructed in the following way (nucleotide positions are with respect to the MDR1 start codon; Fling et al., 1991). For pSFIM3, MDR1 sequences from positions −568 to 793 and from positions 1131–2393 were amplified using the polymerase chain reaction (PCR) with the primer pairs MDR5 (5′-TTGAACCGCGGAATGGACCAAAACTAGGACC-3′) and MDR8 (5′-CCTTATAGAGCTCTACTGGTAACTATTGGCG-3′), and MDR4 (5′-TTTCGCTCGAGTTAAACATTTCACCCTCG-3′) and MDR9 (5′-CCGTAATGTAATTGCATGCAGTAGGCGCAGTC-3′), respectively, thereby introducing SacII, SacI, XhoI and SphI restriction sites (underlined). The SacI–SacII 5′MDR1 and the XhoI–SphI 3′MDR1 fragments were then ligated together with the XhoI–SacII fragment containing the MPAR flipper from pSFI1 into the SacI/SphI-digested vector pUC18, resulting in pSFIM3. Similarly, for pSFIM4, MDR1 sequences from positions −568 to 626 and from positions 918–2393 were amplified using PCR with the primer pairs MDR3 (5′-CCGGCAATATTATTTACCGCGGCAGTGGGG-3′) and MDR8, and MDR6 (5′-GGCTAAAAGCTCGAGAGCCATCACCGG-3′) and MDR9, respectively (the introduced SacII and XhoI restriction sites are underlined). The SacI–SacII 5′MDR1 and the XhoI–SphI 3′MDR1 fragments were then ligated together with the XhoI–SacII fragment containing the MPAR flipper into the SacI/SphI-digested vector pUC18, resulting in pSFIM4.

C. albicans transformation

C. albicans strains were transformed by electroporation (Köhler et al., 1997) with the gel-purified SacI–SphI fragments from plasmids pSFIM3 and pSFIM4. MPA-resistant transformants were selected on minimal agar plates [6.7 g of yeast nitrogen base without amino acids (YNB; BIO 101)], 2 g of glucose, 0.77 g of complete supplement medium (CSM-URA; BIO101), 15 g of agar l−1) containing 10 µg ml−1 MPA. Single colonies were picked after 5–7 days of growth, restreaked on plates containing 10 µg ml−1 MPA and then propagated further on YPD agar plates.

Isolation of genomic DNA and Southern hybridization

Genomic DNA from C. albicans strains was isolated as described previously (Millon et al., 1994). Southern hybridization with enhanced chemiluminescence-labelled probes was performed with the ECL labelling and detection kit provided by Amersham according to the instructions of the manufacturer.

RNA isolation and Northern hybridization

Total RNA from C. albicans was isolated by the hot acidic phenol method (Ausubel et al., 1989). The MDR1-ACT1 fragment from plasmid pMDR2 (Franz et al., 1998) that was used as a probe was labelled with a random-primed DNA labelling kit (Roche Diagnostics), and Northern hybridization was performed under stringent conditions using standard protocols (Ausubel et al., 1989).

Drug susceptibility tests

Stock solutions of the drugs were prepared by dissolving fluconazole (Pfizer UK) in water (1 mg ml), and ketoconazole and 4-NQO (Sigma-Aldrich Chemie) in DMSO (2 mg ml−1 and 0.2 mg ml−1 respectively). The stock solutions were diluted to 200 µg ml−1 for fluconazole, 10 µg ml−1 for ketoconazole and 1 µg ml−1 for 4-NQO in high-resolution medium [14.67 g of HR-medium (Oxoid), 1 g of NaHCO3, 0.2 M phosphate buffer, pH 7.2], and the MICs of the drugs were determined using a microdilution method described previously (Ruhnke et al., 1994), with the exception that readings were done after 24 h instead of 48 h.


This study was supported by the Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie (BMBF grant O1 K18906-0). We are indebted to Gerwald Köhler for the generous gift of the MPAR marker before publication. Clarissa Radecke is acknowledged for her advice with respect to the MIC determinations.