The genetic manipulation of the human fungal pathogen Candida albicans is difficult because of its diploid genome, the lack of a known sexual phase and its unusual codon usage. We devised a new method for sequential gene disruption in C. albicans that is based on the repeated use of the URA3 marker for selection of transformants and its subsequent deletion by FLP-mediated, site-specific recombination. A cassette was constructed that, in addition to the URA3 selection marker, contained an inducible SAP2P–FLP fusion and was flanked by direct repeats of the minimal FLP recognition site (FRT). This URA3 flipper cassette was used to generate homozygous C. albicans mutants disrupted for both alleles of either the CDR4 gene, encoding an ABC transporter, or the MDR1 gene, encoding a membrane transport protein of the major facilitator superfamily. After insertion of the URA3 flipper into the first copy of the target gene, the whole cassette could be efficiently excised by induced FLP-mediated recombination, leaving one FRT site in the disrupted allele of the target gene. The URA3 flipper was then used for another round of mutagenesis to disrupt the second allele. Deletion of the cassette from primary and secondary transformants occurred exclusively by intrachromosomal recombination of the FRT sites flanking the URA3 flipper, whereas interchromosomal recombination between FRT sites on the homologous chromosomes was never observed. This new gene disruption strategy facilitates the generation of specific, homozygous C. albicans mutants as it eliminates the need for a negative selection scheme for marker deletion and minimizes the risk of mitotic recombination in sequential disruption experiments.
Candida albicans is the most important fungal pathogen in humans causing superficial as well as often fatal disseminated infections in immunocompromised patients (Odds, 1988). In recent years molecular biological methods have been developed for C. albicans to gain insight into its biology and mechanisms of pathogenicity (Pla et al., 1996). The diploid nature of the organism and the lack of a known sexual phase, however, have hampered the genetic manipulation of C. albicans. In addition, C. albicans has a non-canonical codon usage, translating the normally leucine-specific codon CTG as serine (Santos et al., 1997). This unusual decoding is a reason for the frequent failure to express heterologous genes in C. albicans, for example for use as reporter genes or as selection markers (Leuker et al., 1992; Morschhäuser et al., 1998). Genetic manipulation of C. albicans is therefore largely based on the use of auxotrophic host strains in conjunction with selection markers complementing the defect in the corresponding biosynthetic pathway. Integrative transformation occurs with high frequency by homologous recombination in C. albicans, allowing the targeted disruption of genes for the construction of specific mutants. However, owing to the diploid genome, both alleles of a given gene have to be inactivated in order to obtain homozygous mutants. Although C. albicans strains with defects in several biosynthetic pathways exist in which two different markers can be used for sequential disruption of the two alleles of the target gene (Kurtz and Marrinan, 1989), these host strains have in most cases suffered from chemical mutagenesis and/or UV irradiation and probably contain additional unknown, undesired mutations. The construction of a ura3-negative derivative of the clinical isolate SC5314 by targeted mutagenesis (Fonzi and Irwin, 1993) has circumvented the latter problem, and strain CAI4 and similar derivatives of SC5314 are now used by most researchers for molecular genetic studies of C. albicans. The adaptation of the method of sequential gene disruption in S. cerevisiae (Alani et al., 1987) for C. albicans allowed the construction of homozygous C. albicans mutants using only a single selection marker (Fonzi and Irwin, 1993). A cassette consisting of the C. albicans URA3 gene flanked by direct repeats of the S. typhymurium hisG gene is inserted between sequences of the target gene, and the linear DNA fragments are used in a first round of transformation to disrupt one of the two chromosomal copies of the gene by the hisG–URA3–hisG cassette. The URA3 gene has the advantage that it can repeatedly be used as a selection marker because ura3-negative strains are resistant against 5-fluoro-orotic acid (FOA) (Boeke et al., 1984), and derivatives of the first-round transformants in which the URA3 gene was deleted by rare, spontaneous recombination between the hisG repeats can be selected on FOA plates and be used for a second round of transformation, resulting in homozygous mutants. Although a similar mutagenesis procedure was also described with the GAL1 gene as a selection marker (Gorman et al., 1991), the URA3 blaster strategy is now the standard method for the construction of specific C. albicans mutants. One problem that arises when two or more genes are sequentially inactivated is the generation of ura3-negative derivatives from secondary transformants by mitotic recombination between homologous chromosomes, which occurs at about the same frequency as recombination between the hisG repeats (Fonzi and Irwin, 1993). Such mitotic recombinants became homozygous for all chromosomal sequences distal from the site of crossing-over including possible recessive mutations and, in the absence of restriction site polymorphisms, cannot be distinguished from true specific mutants.
The site-specific recombinase FLP from Saccharomyces cerevisiae is used as a tool for genetic manipulations in a wide variety of organisms from bacteria to mammalian cells (for a review see Kilby et al., 1993). We have genetically engineered the FLP gene for expression in C. albicans and used it as a reporter of C. albicans virulence gene expression (Staib et al., 1999). When the FLP gene was placed under the control of the C. albicans SAP2 promoter (SAP2P), FLP activity was only observed under SAP2-inducing conditions. The availability of an FLP gene whose expression in C. albicans can be regulated in vitro enabled us to design a new gene disruption strategy for this asexual, diploid fungus. The principle of the method relies on the repeated use of a mutagenesis cassette that contains the URA3 selection marker and an inducible SAP2P–FLP fusion and is bordered by direct repeats of the FLP recognition sequence. After insertion into a target gene the cassette is specifically deleted again with high frequency by induced, FLP-mediated excision (URA3-flipping), in contrast to the rare, spontaneous marker deletion in the URA3-blasting method. This new method was successfully applied for the construction of two different homozygous C. albicans mutants.
Sequential disruption of the CDR4 gene
The URA3 flipper cassette was constructed as described in Experimental procedures. It consists of the URA3 selection marker and the SAP2P–FLP fusion and is flanked by direct repeats of the 34 bp minimal target sequence of the FLP recombinase (FLP recognition target, FRT). For disruption of the C. albicans CDR4 gene, encoding a recently identified ABC transporter (Franz et al., 1998b), the cassette was inserted into the CDR4 coding region in plasmid pSFUC2, thereby deleting the region between nucleotide positions 305 and 2769. A KpnI–SacI fragment from pSFUC2 containing the URA3 flipper flanked by CDR4 sequences was then used to transform the ura3-auxotrophic C. albicans strain CAI4 (Fig. 1). URA3 transformants were selected on minimal agar plates without uridine and the correct insertion of the URA3 flipper into the CDR4 locus was confirmed by Southern hybridization. In the parent strain CAI4, both CDR4 alleles were located on a 3.2 kb BglII fragment that hybridized with the probe used (Fig. 2, lane 1). Replacement of 2.5 kb of CDR4 sequences in one of the alleles by the 4.2 kb URA3 flipper resulted in the generation of a new fragment of about 4.9 kb, as shown in Fig. 2 for two independent transformants, SFLUC1A and SFLUC1B (lanes 2 and 3). Single colonies of both transformants were subsequently inoculated into YCB–BSA liquid medium containing 100 μg ml−1 uridine to induce the SAP2P–FLP fusion and allow the growth of ura3-negative derivatives generated by FLP-mediated excision of the URA3 flipper. The SAP2 promoter is induced when low molecular nitrogen sources become limiting and the BSA in the medium is the only available nitrogen source that occurs after several hours of growth in YCB–BSA (Hube et al., 1994; Staib et al., 1999). A dilution of the YCB–BSA culture was spread on minimal agar plates containing 10 μg ml−1 uridine. On plates with this limiting uridine concentration ura-auxotrophic clones grew more slowly than the prototrophic parent strains and formed smaller colonies, a phenotype that served as a convenient marker for rapid screening. About 10% of the cells in the cultures were ura-auxotrophic. For each of the primary transformants, SFLUC1A and SFLUC1B, four auxotrophic clones were analysed by Southern hybridization and in all eight clones the 4.9 kb BglII fragment was replaced by a 0.8 kb fragment expected after FLP-mediated excision of the URA3 flipper. From each parent strain, one ura3-negative derivative (strains SFLUC2A and SFLUC2B, lanes 4 and 5 in Fig. 2) was subsequently used to disrupt the second copy of the CDR4 gene. To distinguish later between intrachromosomal recombination of the FRT sites flanking the cassette and interchromosomal recombination of FRT sites on homologous chromosomes (see below), the URA3 flipper was inserted into a different position in the CDR4 gene in plasmid pSFUC4, resulting in a deletion of the region between nucleotide positions 1444 and 2036. A KpnI–SacI fragment from pSFUC4 was then used to transform strains SFLUC2A and SFLUC2B (see Fig. 1). In addition, a similar fragment from plasmid pCDR4M, which contained only the URA3 gene instead of the URA3 flipper at the same position in the CDR4 gene, was used for transformation of the two strains to generate prototrophic homozygous mutants from which the URA3 gene could not be deleted again. As only the wild-type copy of the CDR4 gene but not the already disrupted allele in SFLUC2 contained sequences homologous to the flanking regions, integration was targeted to the remaining wild-type CDR4 allele. Insertion of the URA3 flipper resulted in the loss of the 3.2 kb fragment and generation of a new 6.8 kb fragment, whereas insertion of only the URA3 gene instead of the URA3 flipper generated a new fragment of 4.0 kb. For each parent strain, one transformant containing the correct insertion of the URA3 flipper (strains SFLUC3A and SFLUC3B, lanes 6 and 7) or the URA3 gene (strains SFLUC4A and SFLUC4B, lanes 8 and 9) is shown in Fig. 2. The URA3 flipper in strains SFLUC3A and SFLUC3B was subsequently deleted again. After passage in YCB–BSA medium, 17% and 27% of the cells were ura-auxotrophic. In addition to the desired intrachromosomal recombination between the FRT sites flanking the URA3 flipper, ura3-negative derivatives might also be generated by mitotic interchromosomal recombination between the centromere-proximal FRT site on the chromosome with the URA3 flipper and the single FRT site on the homologous chromosome (see Fig. 3). Such mitotic recombinants would become homozygous for the whole chromosomal region on the centromere-distal side of the FRT site, similar to mitotic recombinants arising by homologous recombination (Fonzi and Irwin, 1993). As illustrated in Fig. 3, possible ura3-negative clones arising by interchromosomal recombination could be distinguished from those produced by specific excision of the URA3 flipper. From each of the parent strains, 11 ura3-auxotrophic derivatives were analysed by Southern hybridization, and, as exemplified in Fig. 2 by strains SFLUC5A and SFLUC5B (lanes 10 and 11), all contained the 2.6 kb BglII fragment expected after excision of the URA3 flipper by recombination between the flanking FRT sites, demonstrating that interchromosomal recombination between FRT sites on homologous chromosomes had not occurred at a detectable frequency. For complementation analysis, a complete copy of the CDR4 gene was introduced into strains SFLUC5A and SFLUC5B by transformation with the BamHI-linearized plasmid pCDR4K. Integration was targeted to the CDR4 allele disrupted in the second round of mutagenesis, as this allele contained homologous sequences on both sides of the BamHI site, in contrast to the first disrupted allele in which a large portion of homologous sequences downstream of the BamHI site was absent (see Fig. 1 and Experimental procedures). For each parent strain, one transformant is shown in Fig. 2 (strains SFLUC6A and SFLUC6B, lanes 12 and 13). The replacement of the 2.6 kb BglII fragment by the 3.2 kb wild-type fragment and an additional 8.8 kb fragment containing the remaining plasmid sequences demonstrated the correct integration of a single copy of pCDR4K into the CDR4 locus.
Analysis of cdr4 mutants
Transcription of the CDR4 gene in the different mutants was analysed by Northern hybridization with a hybrid probe containing CDR4 sequences still present in the two disrupted alleles as well as sequences from the ACT1 gene as an internal control (Fig. 4). The heterozygous CDR4/cdr4 mutants SFLUC1A/B containing the URA3 flipper in one of the CDR4 alleles (lanes 2 and 3) contained about half as much CDR4 mRNA as the parent strain CAI4 (lane 1), indicating that both CDR4 alleles were expressed in the wild-type strain. Correspondingly, CDR4 mRNA levels in strains SFLUC2A/B in which the URA3 flipper is deleted (lanes 4 and 5) were also lower than in strain CAI4. In addition, strains SFLUC2A/B contained a shorter transcript (CDR4Δ1) of the size expected after deletion of 2.5 kb of the CDR4 coding region. No intact CDR4 mRNA was present in the homozygous mutants SFLUC3A/B and SFLUC4A/B in which the second CDR4 allele was disrupted by the URA3 flipper or the URA3 gene respectively (lanes 6–9). After excision of the URA3 flipper from the second disrupted allele, another truncated transcript (CDR4Δ2) of the size expected from the 0.6 kb deletion appeared in strains SFLUC5A/B (lanes 10 and 11). Because the CDR4 reading frame in the transcripts with the 2.5 kb and 0.6 kb deletions was interrupted upstream of the N-terminal nucleotide binding fold and upstream of the first transmembrane region, respectively, the two truncated transcripts could not direct synthesis of a functional Cdr4 protein. Reintroduction of a complete copy of the CDR4 gene in the second disrupted CDR4 allele resulted in the disappearance of the CDR4Δ2 transcript and production of a full-length CDR4 mRNA (lanes 12 and 13).
Because other Cdr proteins, Cdr1p and Cdr2p, mediate resistance of C. albicans against a variety of drugs by actively transporting them out of the cell, the susceptibility of the cdr4 mutants to different drugs was tested (see Experimental procedures). No enhanced susceptibility of the mutants to any of the drugs tested was found, neither when the ura-prototrophic homozygous mutants SFLUC3 and SFLUC4 were compared with the ura-prototrophic wild-type strain SC5314, nor when the ura-auxotrophic mutant SFLUC5 was compared with the auxotrophic parent strain CAI4 (data not shown). Therefore, Cdr4p either does not transport these drugs, as has been suggested for Cdr3p (Balan et al., 1997), or its contribution to drug efflux would only be observed in mutants in which other transporters such as Cdr1p are absent, as has been shown for Cdr2p (Sanglard et al., 1997).
Disruption of the MDR1 gene
To confirm the general applicability of the URA3-flipping method for the generation of specific homozygous C. albicans mutants, we disrupted a second gene, MDR1, encoding a membrane transport protein of the major facilitator superfamily (Fling et al., 1991). This gene is constitutively activated in many drug-resistant clinical C. albicans isolates, but not detectably expressed in strain CAI4 during in vitro growth in several media (Franz et al., 1998a). Because the frequency of chromosomal integration in C. albicans by homologous recombination correlates with the transcriptional activity of the target gene (Srikantha et al., 1995), it seemed possible that FLP-mediated recombination, too, might work less well in transcriptionally inactive chromosomal regions.
Two different insertions of the URA3 flipper in the MDR1 coding region were constructed. In plasmids pSFUM1 and pSFUM2, MDR1 sequences between positions 783 and 932 or between positions 606 and 1141, respectively, were replaced by the URA3 flipper, and the inserts from these two plasmids were used to disrupt both MDR1 alleles in strain CAI4 (see Fig. 5). As described above for the mutagenesis of the CDR4 gene, sequential disruption of the MDR1 alleles was performed two times independently and was confirmed by Southern hybridization (Fig. 6). Insertion of the URA3 flipper into the first MDR1 allele resulted in a new EcoRV fragment of 3.6 kb (lanes 2 and 3), which was reduced to 2.08 kb after excision of the cassette (lanes 4 and 5). Disruption of the second MDR1 allele resulted in the loss of the remaining 2.16 kb wild-type fragment and generation of new fragments of 3.4 kb or 3.0 kb, respectively, after insertion of the URA3 flipper (lanes 6 and 7) or the URA3 gene (lanes 8 and 9). In all 22 analysed ura-auxotrophic derivatives of strains SFLUM3A and SFLUM3B, the URA3 flipper was excised by intrachromosomal recombination between the flanking FRT sites, as demonstrated for strains SFLUM5A and SFLUM5B (lanes 10 and 11) by the generation of a 1.7 kb EcoRV fragment from the 3.4 kb fragment. Therefore, as in the construction of the cdr4 mutants (see above), interchromosomal recombination between FRT sites on homologous chromosomes which would have resulted in the generation of a 1.9 kb EcoRV fragment had not occurred at a detectable frequency. Excision of the URA3 flipper from the MDR1 locus in strains SFLUM1 and SFLUM3 was even more efficient than from the CDR4 locus, occurring at a frequency of about 40%. These results demonstrate that the URA3-flipping method is generally applicable for sequential gene disruption in C. albicans and does not depend on the transcriptional state of the target gene.
As mentioned above, MDR1 is not detectably expressed in vitro in strain CAI4 and other fluconazole-susceptible C. albicans strains. Correspondingly, disruption of both copies of the gene in strains SFLUM3, SFLUM4 and SFLUM5 did not result in hypersensitivity against fluconazole or any of the other drugs tested (data not shown).
The URA3-flipping method developed in this study for sequential gene disruption in C. albicans has several advantages over the commonly used URA3-blasting strategy. The efficiency of marker regeneration by induced FLP-mediated excision (between 8% and 40% in the experiments described here) is several orders of magnitude higher than loss of the URA3 gene by spontaneous recombination between the hisG repeats, which occurs at a frequency of about 10−5–10−6 (Fonzi and Irwin, 1993). This eliminates the need for selection of ura3-negative clones on plates containing FOA, an extremely expensive compound that, in addition, may have adverse effects on cells still containing some orotidine-5′-phosphate decarboxylase enzyme immediately after plating. On plates containing low amounts of uridine, ura3-negative clones obtained after FLP-mediated excision of the URA3 flipper can easily be differentiated from the prototrophic parent and picked after 2 days of growth, in contrast to FOA selection, which usually requires 5–6 days of growth.
A major advantage of the URA3-flipping method is that homologous mitotic recombination that occurs at about the same frequency as recombination between the hisG repeats (although this depends in each case on the distance of the integrated URA3 gene from the centromere), can be neglected as a source of ura3-negative clones because of its roughly 105-fold lower frequency. Accordingly, none of the 56 ura3-negative clones tested in our study from eight independent parent strains was generated by interchromosomal homologous recombination. Generation of ura3-negative clones by mitotic recombination is a problem in the URA3-blasting protocol after the second round of marker deletion when the same disruption cassette is used for the first and second transformations because, in the absence of restriction site polymorphisms, such mitotic recombinants that are not truly isogenic to the parent strain cannot be differentiated from specific mutants. This problem was pointed out in the original study by Fonzi and Irwin (1993), and it is usually avoided by using the URA3-prototrophic homozygous mutants for comparison with the wild-type strain, especially in virulence studies, because auxotrophic mutants are avirulent (Kirsch and Whitney, 1991; Cole et al., 1995). However, for the disruption of additional genes, the URA3 marker must also be deleted from the second copy of the target gene and mitotic recombinants may be obtained.
An analogous situation might also lead to mitotic recombinants by FLP-mediated recombination during the second round of marker deletion. Such mitotic recombinants would not be generated by interchromosomal recombination of any homologous sequences located between the centromere and the URA3 marker but by recombination of FRT sites on homologous chromosomes. To test this possibility, we used two different insertions of the URA3 flipper for disruption of the two alleles of the target genes. In the 44 ura3-negative clones analysed from four independent parent strains, interchromosomal recombination between FRT sites on homologous chromosomes was never observed. This is probably because FLP-mediated recombination is more efficient between FRT sites located in the vicinity of each other. We have indeed observed that the efficiency of FLP-mediated recombination decreases with increasing distance of the FRT sites in the C. albicans genome (P. Staib and J. Morschhäuser, unpublished). These results suggest that, using the URA3-flipping method for the construction of homozygous C. albicans mutants, it is not necessary to use two different insertions for disruption of the two alleles of a target gene. Although many thousands of ura3-negative clones would have to be analysed to exclude the possibility of interchromosomal recombination, the risk of obtaining such false ‘specific’ mutants is at least minimized compared with the URA3-blasting method.
FLP-mediated, site-specific recombination has already been used by other researchers to remove selection markers from the genome of bacterial and yeast transformants (Cregg and Madden, 1989; Cherepanov and Wackernagel, 1995; Schweizer, 1998). In these cases, deletion of the marker was performed by introducing the FLP gene on a plasmid into transformants containing a marker flanked by the FLP recognition sequence and subsequent curing of the plasmid. In our URA3-flipping method we have avoided these additional transformation and curing steps by integrating an inducible FLP gene within a cassette containing the marker. After insertion into the target gene, the induced FLP recombinase not only removes the selection marker but also its own gene, allowing the subsequent use of the URA3 flipper for additional transformations. The efficiency of induction of the SAP2P–FLP fusion and, concomitantly, the maximum percentage of ura3-negative clones that can be obtained depends on the growth conditions (e.g. age and size of the colony used for inoculation). However, one usually needs only a single ura3-negative clone, and, for convenience, we routinely isolated ura3-negative derivatives after overnight growth in YCB–BSA medium. The differences in the efficiency with which the URA3 flipper was excised from the different loci suggest that it is also influenced to some degree by the site of insertion.
Because the SAP2 promoter is induced by host signals during infection (Staib et al., 1999), loss of the URA3 flipper would lead to avirulent auxotrophic cells. For this reason, we also constructed homozygous mutants containing the URA3 gene alone in the second allele of the target gene instead of the URA3 flipper. These mutants are stable prototrophs under all conditions and would therefore be the strains of choice for in vivo experiments. Nevertheless, removal of the URA3 marker from the second disrupted allele is necessary for additional disruptions of other genes and for reintroduction of a complete gene in complementation experiments, which are mandatory to demonstrate that phenotypic effects are indeed caused by disruption of the target gene and are not unspecific effects caused by the transformation procedure.
An additional advantage of efficient marker removal by site-specific recombination is that this strategy is also applicable for other markers for which no negative selection scheme is available. A dominant selection marker for C. albicans transformation conferring resistance to mycophenolic acid has now been developed (G. Köhler, personal communication, Staib et al., 1999). Use of such a dominant selection marker in combination with the FLP recombinase would allow the generation of homozygous mutants from C. albicans wild-type strains. This is important for the analysis of strains with characteristics absent in the widely used host strain CAI4, for example clinical strains that are resistant to antifungal drugs. The new strategy for sequential gene disruption in C. albicans therefore allows completely novel approaches for the molecular analysis of this major human fungal pathogen.
Strains and growth conditions
Candida albicans strains used in this study are listed in Table 1. Strains were maintained on minimal agar plates containing 6.7 g of yeast nitrogen base without amino acids (YNB; BIO 101), 2 g of glucose and 0.77 g of complete supplement medium without uracil (CSM-URA; BIO101) per litre. Other media used were YPD (20 g of peptone, 10 g of yeast extract, 20 g of glucose per litre) and YCB–BSA (23.4 g of yeast carbon base, 4 g of bovine serum albumin per litre, pH 4.0). Strains were routinely grown at 30°C. For growth of ura3-auxotrophic strains 100 μg ml−1 uridine was added to the media; for differentiation between URA3-positive and ura3-negative colonies 10 μg ml−1 uridine was added to minimal agar plates. Escherichia coli strain DH5α was used as a host for cloning experiments.
Table 1. . C. albicans strains used in this study (Gillum et al. (1984)). a.URA3-FLIP denotes the FRT-SAP2P–FLP-URA3-FRT cassette (URA3 flipper).b. All mutant strains described in this study were constructed two times independently (for example SFLUC1A and B, see text).
Construction of plasmids
Construction of pSFUC2.
Plasmid pSFL26 has been described previously (Staib et al., 1999). It contains, on a BamHI–PstI fragment, a genetically engineered FLP gene under control of the promoter of the C. albicans SAP2 gene (SAP2P), fused to the transcriptional termination sequence of the ACT1 gene (ACT1T) and the URA3 selection marker. For construction of the URA3 flipper and insertion into the CDR4 gene, the CDR4 5′ portion including upstream sequences from nucleotide positions −377 to 304 was PCR-amplified using the primers CDR7 (5′-ATATAGGATCCGAAGTTCCTATTCTCTAGAAAGTATAGGA ACTTCCTCGAGCTGCGTTGAATTCAGGTGAGTTCGGG-3′) and CDR8 (5′-TCTGTGGTACCATTATTAGTGCCACTC TCC-3′). Primer CDR7 contains the 34 bp FRT sequence (− strand, bold letters), flanked by BamHI and XhoI restriction sites (underlined), in addition to CDR4 sequences. In primer CDR8 a KpnI site (underlined) was introduced into the CDR4 sequence. The PCR product was digested with KpnI/BamHI, and the 5′CDR4–FRT fragment was ligated together with the BamHI–PstI fragment from pSFL26 into the KpnI/PstI-digested pBluescript to obtain plasmid pSFUC1. Subsequently, a part of the CDR4 coding region from nucleotide positions 2770–3683 was PCR amplified with the primers CDR5 (5′-ATATACTGCAGGAAGTTCCTATACTTTCTAGAGAATAGGAACTTCAGATCTATTGGGTATGTTCAAC-3′) and CDR6 (5′-GTGTGAAGAGCTCACAAGGAAC-3′). Primer CDR5 contains the 34 bp FRT sequence (+ strand, bold letters), flanked by PstI and BglII restriction sites (underlined), in addition to CDR4 sequences. In primer CDR6 a SacI site (underlined) was introduced into the CDR4 sequence. The PCR product was digested with PstI/SacI and the FRT–3′CDR4 fragment was ligated into the PstI/SacI-digested pSFUC1, resulting in pSFUC2 (see Fig. 1). Recombination proficiency of the FRT sites in pSFUC2 was confirmed with the FLP test strain E. coli 294-FLP (Buchholz et al., 1996).
Construction of pSFUC4, pCDR4M, pCDR4K and pCDR4N
Plasmid pCDR1 contains CDR4 sequences from nucleotide positions 559–2664 cloned in the SmaI site of pUC18 (Franz et al., 1998b). A derivative, pCDR1A, was obtained by cloning this fragment between the KpnI/BamHI sites in pBluescript and subsequent removal of the BamHI site by digestion with BamHI/SacII, blunting of the ends and religation. The primers CDRX1 (5′-CAAGTGATCTCGAGACGTGCTTGATGCGC-3′) and CDRX2 (5′-GGGTTTGTGGATCCTACAC CAAATATG-3′) were then used to delete the region between nucleotide positions 1444 and 2036 by divergent PCR and add XhoI and BamHI restriction sites (underlined) into which the XhoI–BglII fragment with the URA3 flipper from pSFUC2 or an XhoI–BamHI fragment with the URA3 gene (Morschhäuser et al., 1998) was inserted, yielding plasmids pSFUC4 and pCDR4M respectively. For complementation analysis, an XbaI–HindIII fragment containing the complete CDR4 gene from positions −207 to 4908 downstream of the transcription termination sequences was ligated together with a SalI–XbaI fragment containing the URA3 gene into pBluescript, resulting in plasmid pCDR4K. A BamHI site at position 311 of the CDR4 gene was used to linearize pCDR4K for integration into one of the disrupted CDR4 alleles.
Plasmid pCDR4N was constructed by ligating a 2.5 kb fragment from the 3′CDR4 region (positions 2765–5231) together with a 0.8 kb fragment from the coding region of the ACT1 gene in pBluescript. The CDR4–ACT1 fragment from pCDR4N was used as a probe for simultaneous detection of CDR4 and ACT1 mRNAs by Northern hybridization.
Construction of pSFUM1, pSFUM2 and pMDR1M
To facilitate insertion of the URA3 flipper cassette into target genes, plasmid pSFU1 was constructed by cloning the XhoI–BglII fragment with the URA3 flipper from pSFUC2 into the XhoI/BamHI-digested vector pBluescript. In this plasmid the URA3 flipper is flanked by several unique restriction sites (KpnI, ApaI and XhoI on one side, NotI, SacII and SacI on the other) and can conveniently be inserted into any cloned target gene in a single step by divergent PCR. Plasmid pMDR1 contains the MDR1 coding region in pUC18 (Franz et al., 1998a). Deletions in the MDR1 gene were constructed by divergent PCR with the primer pairs MDR5 (5′-TTGAACCGCGG AATGGACCAAAACTAGGACC-3′) and MDR6 (5′-GGCTAA AAGCTCGAGAGCCATCACCGG-3′) or MDR3 (5′-CCGGCA ATATTATTTACCGCGGCAGTGGGG-3′) and MDR7 (5′-TT TCGCTCGAGTTAAACATTTCACCCTCGTTGAATTGGG-3′) while adding SacII and XhoI restriction sites (underlined) into which an XhoI–SacII fragment with the URA3 flipper from pSFU1 or an XhoI–SacII fragment with the URA3 gene were inserted, resulting in plasmids pSFUM1, pSFUM2 and pMDR1M (see Fig. 5).
Candida albicans transformation
Candida albicans strain CAI4 and derivatives were transformed by electroporation (Köhler et al., 1997) using the gel-purified KpnI–SacI fragments from plasmids pSFUC2, pSFUC4 and pCDR4M, the SacI–SphI fragments from plasmids pSFUM1, pSFUM2 and pMDR1M, and the BamHI-linearized plasmid pCDR4K. URA3 transformants were selected on minimal agar plates without uridin.
Isolation of chromosomal DNA and Southern hybridization
Chromosomal DNA from C. albicans strains was isolated as described previously (Millon et al., 1994). Southern hybridization with enhanced chemiluminescence (ECL)-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 CDR4–ACT1 fragment from plasmid pCDR4N was labelled with a random primer DNA labelling kit (Boehringer), and Northern hybridization was performed under stringent conditions using standard protocols (Ausubel et al., 1989).
Drug susceptibility testing
To test for differences in drug susceptibility of cdr4 or mdr1 mutants and wild-type strains, overnight cultures grown in YPD medium were diluted 1:100 in fresh YPD medium or in YPD medium containing the drugs listed below and incubated at 30°C, 250 r.p.m. Growth was monitored for 10 h by determining the optical density of the cultures at 600 nm at 2 h intervals. The following drug concentrations, which resulted in a reduced growth of the wild-type strains SC5314 and CAI4 compared with drug-free medium, were used: fluconazole (20 μg ml−1), ketoconazole (1 μg ml−1), terbinafine (2 μg ml−1), amorolfine (0.05 μg ml−1), cycloheximide (100 μg ml−1), brefeldin (10 μg ml−1), fluphenazine (20 μg ml−1), 4-nitrochinoline-N-oxide (0.05 μg ml−1), nocodazole (10 μg ml−1), cerulenin (1 μg ml−1), crystal violet (0.05 μg ml−1).
Nucleotide sequence accession numbers
Accession numbers for the CDR4 and MDR1 genes are AF044921 (CDR4) and X53823 (MDR1). Nucleotide positions referred to in the text are with respect to the start codons of the genes.
This study was supported by the Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie (BMBF grant O1 K18906-0), the Deutsche Forschungsgemeinschaft (DFG grant MO 846/1-1) and the Fonds der Chemischen Industrie. Peter Staib is the recipient of a grant from the Studienstiftung des deutschen Volkes. We thank Bill Fonzi for the gift of strains SC5314 and CAI4, and Frank Buchholz for providing the FLP test strain E. coli 294-FLP. Joachim Reidl and Gerwald Köhler are acknowledged for critical reading of the manuscript.