OUP user menu

Loss-of-function pdr3 mutations convert the Pdr3p transcription activator to a protein suppressing multidrug resistance in Saccharomyces cerevisiae

Michaela Sidorova, Eva Drobna, Vladimira Dzugasova, Imrich Hikkel, Julius Subik
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00174.x 254-264 First published online: 1 March 2007

Abstract

The PDR1 and PDR3 genes encode the main transcription activators involved in the control of multidrug resistance in Saccharomyces cerevisiae. To identify the amino acids essential for Pdr3p function, the loss-of-function pdr3 mutants were isolated and characterized. Two plasmid-borne pdr3 alleles, pdr3-E902Ter and pdr3-D853Y, which failed to complement drug hypersensitivity in the Δpdr1Δpdr3 mutant strain, were isolated. The E902Ter mutation resulted in a truncated protein lacking the C-terminal activation domain. The D853Y mutation allowed the expression of entire Pdr3p, but its transactivation function was lost. When overexpressed from the PGAL1 promoter, the two mutant alleles increased the sensitivity of wild-type cells to cycloheximide and fluconazole and suppressed drug resistance in gain-of-function pdr1 and pdr3 mutant strains. The drug-sensitizing effect of overexpressed loss-of-function pdr3 mutant alleles correlated with their ability to suppress PDR5 transcription and rhodamine 6G accumulation in transformants of the wild-type and Δpdr1 mutant strains. These results demonstrate that amino acid residue Asp853 is essential for Pdr3p function, and indicate that specific loss-of-function pdr3 mutations can convert the Pdr3p transcription activator to a multicopy suppressor of multidrug resistance.

Keywords
  • loss-of-function mutation
  • multidrug resistance
  • PDR3
  • repressor
  • Saccharomyces cerevisiae
  • transcription factor

Introduction

Multidrug resistance is a defense mechanism used by cells to cope with rapidly changing growth and stress conditions, allowing them to survive in the presence of toxic compounds. It is based on drug efflux by membrane-bound transporters that are overproduced due to gain-of-function mutations in transcription factors regulating their expression. This phenomenon, known also as pleiotropic drug resistance (PDR), has been extensively studied in the baker's yeast Saccharomyces cerevisiae (Balzi & Goffeau, 1991; Kolaczkowska & Goffeau, 1999; Moye-Rowley, 2005; Jungwirth & Kuchler, 2006). Both ATP-binding cassette transporters, including Pdr5p, Snq2p and Yor1p, and major facilitator superfamily transporters, including Atr1p, Sge1p, Flr1p and Tpo1p, are involved in multidrug resistance (Kolaczkowska & Goffeau, 1999; Sa-Correia & Tenreiro, 2002; Teixeira & Sa-Correia, 2002; Jungwirth & Kuchler, 2006). Microarray experiments have indicated that PDR5 is one of the most extensively induced genes upon activation of the PDR pathway (DeRisi, 2000). Its transcription is dramatically induced in ρ0 petite cells (Hallstrom & Moye-Rowley, 2000a) and downregulated in the stationary phase of growth in glucose medium (Mamnun, 2004). Pdr5p recognizes hundreds of structurally and functionally unrelated drugs, lipids and metal ions (Kolaczkowski, 1998; Miyahara, 1996).

The expression of drug efflux transporter genes is under the control of several transcription factors, either zinc cluster proteins such as Pdr1p, Pdr3p and Yrr1p (Kolaczkowska & Goffeau, 1999), or bZip protein family members such as Yap1p (Rodrigues-Pousada, 2004). The transcriptional activators Pdr1p (Balzi, 1987) and Pdr3p (Delaveau, 1994) belong to the family of master regulators of the PDR gene network. They recognize everted repeat CCGCGG motifs as pleiotropic drug resistance response elements (PDREs) (Katzmann, 1994) in promoters of target genes, including the PDR3 gene, pointing to the autoregulation of PDR3 by its own gene product (Delahodde, 1995).

Zinc cluster proteins contribute differently to the regulation of specific multidrug resistance genes and respond differently to extracellular or intracellular signals. For example, induced expression of the FLR1 gene involved in multiple drug resistance is dependent on Pdr3p but not on Pdr1p (Broco, 1999). Conversely, deletion of PDR1 greatly diminishes expression of PDR5, whereas removal of PDR3 has only a marginal effect (Mahe, 1996). Deletion of PDR1 and PDR3 results in increased or decreased expression of PDR15 (Wolfger, 1997; Decottignies, 1998). Transcriptional activation by Pdr1p was found to be inhibited by its association with Ngg1p to form a coactivator–repressor complex in yeast (Martens, 1996). On the other hand, the Hsp70 protein encoded by the PDR13 gene is known as a positive regulator of Pdr1p (Hallstrom & Moye-Rowley, 2000a). Mutations that influence the functional state of mitochondria lead to activation of PDR3 but not of PDR1 (Hallstrom & Moye-Rowley, 2000b). Growth conditions also modulate PDR3 expression (Mamnun, 2004).

Pdr1p and Pdr3p have recently been shown to form homodimers and heterodimers (Mamnun, 2002), indicating that different combinations of homodimers and heterodimers may regulate the expression of different genes. Pdr1p and Pdr3p can also form heterodimers with Rdr1p, acting as a repressor of PDR5 (Hellauer, 2002). On the other hand, another zinc cluster protein, Stb5p, acting through PDRE, was found predominantly as a heterodimer with Pdr1p (Akache, 2004).

Previously, mutational analysis of PDR3 has demonstrated that sets of mutations in the central regulatory region (Nourani, 1997) and in the C-terminal activation domain (Simonics, 2000) significantly increase the activity of Pdr3p, which enhances the expression of PDR3, PDR5 and SNQ2, leading to multidrug resistance. In this article, we show that other mutations in the PDR3 gene can result in a loss of transactivation function of Pdr3p and conversion of this transcription activator to a multicopy suppressor of drug resistance.

Materials and methods

Strains and culture conditions

The S. cerevisiae strains used in this study were as follows: ZK11-1 [MATa ura3-52 trp1-63 leu2-1 his3-200 pdr1::TRP1 pdr3::HIS3, containing two plasmids pYE-PGAL1-PDR3 (2 μm URA3) and pYCp-PPDR5-pma1(D378N) (ARS1 CEN4 LEU2)] (Kozovska & Subik, 2003), JS3-7D (MATαade1 lys2 ura3 pdr3-1 cyh) (Delaveau, 1994), FY1679-28C (MATa ura3-52 trp1-63 leu2-1 his3-200), FY1679-28C/EC (MATa ura3-52 trp1-63 leu2-1 his3-200 pdr1::TRP1), FY1679-28C/TDEC (MATa ura3-52 trp1-63 leu2-1 his3-200 pdr1::TRP1 pdr3::HIS3) (Delaveau, 1994; Carvajal, 1997) and W303-1A (MATa ade2-1 ura3-1 trp1-1 leu2-1 his3-11 can1-100) (Nourani, 1997; Simonics, 2000). Cells were grown on glucose-rich (YPD) medium (2% glucose, 1% yeast extract, 2% Bacto peptone) or on minimal (YNB) medium containing 0.67% yeast nitrogen base without amino acids, 2% glucose or 2% galactose and appropriate nutritional requirements. The media were solidified with 2% Bacto agar. The Escherichia coli XL1 Blue strain was used as host for transformation, plasmid amplification and preparation. The bacterial strains were grown at 37°C in Luria–Bertani medium (LB) (1% tryptone, 1% NaCl, 0.6% yeast extract, pH 7.5) supplemented with 100 μg mL−1 ampicillin for selection of transformants.

Selection of pdr3 mutants

Spontaneous or hydroxylamine-induced (Nourani, 1997) loss-of-function pdr3 mutants in the genetic background of the S. cerevisiae strain ZK11-1 were selected on minimal medium containing galactose. The Ura+Gal+ colonies were picked up, purified on the same medium, and used to determine susceptibilities to cycloheximide and fluconazole on minimal media containing glucose and galactose.

Drug susceptibility assay

Drug susceptibility was determined by spot assay. Five microliters of suspension of three independent clones grown overnight at 30°C were spotted onto minimal medium supplemented with various concentrations of antifungals. Growth was scored after 5 days of incubation at 30°C.

Recombinant DNA techniques and plasmids

Standard protocols were used for generating recombinant DNAs, restriction enzyme analysis, gel electrophoresis and hybridization, as described by Sambrook (1989). Plasmid DNA from E. coli was prepared by the alkaline lysis method. Plasmid DNA from yeast cells was extracted according to the method of Alister & Ward (1990). Yeast transformation was carried out using the modified lithium acetate protocol (Nourani, 1997) or by electroporation (Thompson, 1997). Directed in vitro mutagenesis was carried out using the GeneEditor™in vitro Site-Directed Mutagenesis System from Promega (Promega GmbH, Mannheim, Germany). The following primers were used: (1) 5′-GAA CTC TGG ATA GGA ATC GTA C-3′ to introduce the pdr3-E902Ter mutation into the pFL38-PDR3 (Nourani, 1997) and pYeD1/8-PGAL1-PDR3 plasmids (Kozovska & Subik, 2003); (2) 5′-TTG TAT AGA TAC AAA TTG GCC-3′ to introduce the D853Y mutation into the pFL38-PDR3, pYeD1/8-PGAL1-PDR3 and pYeD1/8-PGAL1-PDR3-GFP plasmids; (3) 5′-GCC ACT GGT AGT ATT CCA GAC-3′ to introduce the N859S mutation into the pFL38-PDR3, pYeD1/8-PGAL1-PDR3 and pYeD1/8-PGAL1-PDR3-GFP plasmids; and (4) 5′-TTT GTA TAG ATA CAA ATT GGC CAC TGG TAG TAT TCC AGA C-3′ to introduce both the D853Y and N859S mutations into the same three plasmids. Plasmid pYES2 (2 μm URA3) was used as an empty vector to introduce a plasmid-borne URA3 gene into host strains. Plasmid pBFG1-HAPDR3-GFP (Delahodde, 2001) served as a source of the PDR3 DNA sequence fused with green fluorescent protein (GFP). The plasmids used in β-galactosidase assays, PPDR3-lacZ and PPDR5-lacZ, contain the lacZ gene fused with the corresponding promoters of PDR3 and PDR5 (Delahodde, 1995; Nourani, 1997).

DNA sequence analysis

The DNA sequence of the pdr3 mutants was determined with the ABI Prism 3100 DNA sequencer (Applied Biosystems, Foster City, CA) using double-stranded plasmid DNA purified using Qiagen Plasmid kits (Qiagen, Hilden, Germany) and a set of synthetic 19-mer oligonucleotide primers corresponding to the PDR3 coding sequence distributed at intervals of about 300–350 bp. Sequence data were analyzed using programs based on the blast algorithm at the NCBI.

Rhodamine 6G accumulation

The accumulation of rhodamine 6G (Sigma-Aldrich, Taukirchen, Germany) was measured by flow cytometry in a Calibur fluorescence-activated cell cytometer (Becton Dickinson, Heidelberg, Germany). Rhodamine 6G was added to exponentially growing yeast culture to give a final concentration of 5 μM. After 1 h of incubation at 30°C, the culture was diluted in ice-cold phosphate-buffered saline at pH 7.0 and analyzed by fluorescence-activated cell sorting (FACS). In total, 10 000 cells were scanned with a 488-nm laser and an FL-1 filter. Cultures without rhodamine 6G were also analyzed, and served as unstained controls. Data were analyzed with the cellquest (Becton Dickinson) program. The geometric mean fluorescence was used for calculation (Tsai, 2004).

RNA isolation, radiolabeling and Northern blot analysis

Samples of total RNA were prepared from cells of transformants grown in YNB minimal medium to mid-logarithmic phase by the hot acidic phenol extraction method (Ausubel, 1989). RNA concentrations were determined spectrophotometrically. Twenty-microgram aliquots of total RNA samples were separated per lane on a 1.2% agarose gel and blotted onto a nylon membrane using the LKB2016 VacuGene vacuum blotting system (Pharmacia LKB, Uppsala, Sweden). Hybridization was performed with an α-[32P]dCTP-labeled DNA probe specific for PDR5 (Ogawa, 1998). Autoradiographs of Northern blots were quantified with a phosphoimager (Bio-Rad, Hercules, CA).

β-Galactosidase assays

The transactivational effect of each pdr3 allele on the PDR5 target promoter was assessed by measurement of β-galactosidase activity in crude extracts of transformants grown to mid-log phase, as described by Miller (1972). The activity was normalized with protein concentrations assayed by the method of Bradford (1976). Each reported value represents the average of determination (±SD) from four independent transformants.

Results

Isolation of loss-of-function pdr3 mutants

Recently, we developed a genetic screen suitable for the identification of multidrug resistance reversal agents acting at the level of gene expression as well as for positive selection of loss-of-function pdr3 mutants (Kozovska & Subik, 2003; Kozovska, 2004). The screen is based on the restoration of growth on medium containing galactose of the hypersensitive S. cerevisiaeΔpdr1Δpdr3 mutant test strain harboring two plasmid-borne fusion genes PGAL1-PDR3 and PPDR5-pma1. Galactose-induced expression of the pma1-D378N dominant lethal allele of the gene encoding the yeast plasma membrane ATPase causes growth arrest of the test strain. Any reversal agent or mutation preventing pma1 expression is expected to restore yeast growth on medium containing galactose. Using the test strain ZK11-l, several spontaneous Gal+ mutants were isolated. After their growth on minimal medium containing leucine, the Ura+Leu transformants lacking plasmid containing the PPDR5-pma1 fusion gene were selected and used for plasmid pYeD-PGAL1-PDR3 extraction and transformation of E. coli. Then, amplified and extracted plasmid DNA containing the PGAL1-PDR3 fusion gene was used to transform both the S. cerevisiae FY1679-28C/TDEC Δpdr1Δpdr3 mutant strain and its transformants bearing the PPDR5-pma1 fusion gene on centromeric plasmids. The resulting Ura+ transformants were isolated on minimal medium containing glucose, and analyzed for their ability to grow on medium containing galactose and for cycloheximide sensitivity (inability to tolerate cycloheximide at a concentration of 0.1 μg mL−1). Only plasmid DNA conferring the Gal+CyhS phenotype and apparently containing loss-of-function pdr3 mutations was selected and used for further studies.

Truncated Pdr3p acts as a multicopy suppressor of drug resistance

Nine loss-of-function pdr3 mutant alleles expressed from the regulatable PGAL1 promoter on the pYeD1/8-PGAL1-PDR3 plasmid were studied in more detail. They were introduced by transformation not only to the Δpdr1Δpdr3 mutant strain but also to the wild-type strain FY1679-28C, its Δpdr1 derivative FY1679-28C/EC, and another wild-type strain, W303-1A, with a different genetic background, and used in studies with gain-of-function pdr3 mutants (Nourani, 1997; Simonics, 2000). The susceptibilities to drugs of corresponding transformants bearing three selected pdr3 alleles (pdr3-a, pdr3-b and pdr3-c) and growing on minimal medium containing galactose are summarized in Table 1, and are shown in Fig. 1 for the wild-type strain W303-1A. As can be seen, the pdr3 mutant alleles did not complement the cycloheximide sensitivity of the Δpdr1Δpdr3 transformants on medium containing galactose. However, all the transformants of the Δpdr1 mutant and the two wild-type strains were more sensitive to cycloheximide than those containing an empty pYES2 vector, regardless of the presence of intact PDR3 or both the PDR1 and PDR3 chromosomal genes. The drug-sensitizing effect of pdr3 alleles was not restricted to cycloheximide only, but was also demonstrated in the presence of fluconazole (Table 1, Fig. 1). Essentially, the same results were observed with all nine loss-of-function pdr3 alleles. These results indicate that pdr3 mutant alleles overexpressed from the PGAL1 promoter may suppress the expression of major multidrug resistance transporter genes, rendering yeast cells hypersensitive to antifungals.

View this table:
1

Minimal inhibitory concentrations of cycloheximide and fluconazole in S. cerevisiae transformants containing different alleles of the PDR3 gene expressed from the PGAL1 or its own promoter and growing on minimal medium containing galactose

PlasmidMinimal inhibitory concentrations (μg mL−1)
CycloheximideFluconazole
FY1679-28CW303-1AFY1679-28CW303-1A
Δpdr1Δpdr3Δpdr1PDR3PDR1 PDR3PDR1 PDR3Δpdr1Δpdr3Δpdr1PDR3PDR1 PDR3PDR1 PDR3
Empty vector pYES20.050.40.50.55404040
pYeD1/8-PGAL1-PDR30.5>0.8>0.8>0.880>80>80>80
pYeD1/8-PGAL1-pdr3-50.050.050.10.1552020
pYeD1/8-PGAL1-pdr3-70.050.050.10.1552020
pYeD1/8-PGAL1-pdr3-160.050.050.10.1552020
pYeD1/8-PGAL1-PDR3-GFP0.30.80.8>0.880>80>80>80
pYeD1/8-PGAL1-pdr3-D853Y-GFP0.10.10.1NT555NT
pYeD1/8-PGAL1-pdr3-N859S-GFP0.30.80.8NT80>80>80NT
pYeD1/8-PGAL1-pdr3-D853Y-N859S-GFP0.10.10.10.155510
pYeD1/8-PGAL1-pdr3-D853Y0.10.10.1NT555NT
pFL38-pdr3-D853Y0.050.30.4NT54040NT
  • NT, not tested.

1

Susceptibilities to cycloheximide and fluconazole of Saccharomyces cerevisiae W303-1A transformants carrying loss-of-function pdr3 alleles spotted onto minimal medium containing galactose and different concentrations of drugs.

In order to determine the molecular nature of loss-of-function pdr3 alleles, they were subjected to DNA sequence analysis. To our surprise, all nine mutants were found to contain the same point mutation, resulting in the appearance of a premature chain-terminating TAG codon at the position of the 902nd amino acid residue (GAG codon), corresponding to glutamic acid in Pdr3p. When this E902Ter mutation was transferred by site-directed mutagenesis into the PDR3 gene carried on centomeric plasmid pFL38-PDR3 and introduced into the wild-type strain FY1679-28C or its isogenic Δpdr1 mutant strain, the ability to enhance the susceptibility of transformants to cycloheximide of this loss-of-function pdr3 mutation expressed from its own promoter was significantly reduced as compared with the galactose-induced and overexpressed pdr3 allele (Fig. 2). The results indicate that the pdr3-E902Ter mutant allele encoding truncated Pdr3p lacking transactivation function can act as a multicopy suppressor of drug resistance in S. cerevisiae.

2

Susceptibilities to cycloheximide on minimal medium containing galactose of transformants of the wild-type and Δpdr1 mutant strains carrying the pdr3-E902Ter allele expressed from its own promoter and the PGAL1 promoter, respectively.

Isolation of loss-of-function pdr3 mutant expressing a full-length form of Pdr3p

To isolate loss-of-function pdr3 mutants possessing a full-length form of Pdr3p, we modified the genetic screen using the PDR3-GFP fusion gene expressed under the control of the GAL1 promoter. The corresponding plasmid pYeD1/8-PGAL1-PDR3-GFP was constructed by exchange of the EcoRI–EcoRI DNA fragment in pYeD1/8-PGAL1-PDR3 for a correctly oriented EcoRI–EcoRI fragment derived from the pBFG1-HAPDR3-GFP plasmid containing the in-frame C-terminal activation domain of PDR3 fused to GFP (Delahodde, 2001). Expression of functional and complete Pdr3p-GFP fusion protein in transformants of the FY1679-28C/TDEC Δpdr1Δpdr3 mutant strain was indicated by green fluorescence of cells after UV illumination and by their resistance to cycloheximide (0.3 μg mL−1) on medium containing galactose compared to cycloheximide-sensitive growth inhibited by cycloheximide at a concentration 0.1 μg mL−1 on medium containing glucose. To increase the frequency of pdr3 mutants in screening tests, in vitro mutagenesis by hydroxylamine of the PDR3 gene was used. The mutagenized PvuII–HpaI fragment of PDR3 was introduced into the FY1679-28C/TDEC Δpdr1Δpdr3 mutant strain harboring plasmid-borne PPDR5-pma1-D378N fusion gene by cotransformation with the BstXI linearized and truncated pYeD1/8-PGAL1-PDR3-GFP plasmid (Kozovska & Subik, 2003).

Direct selection of loss-of-function pdr3 mutants was ensured by the appearance of large colonies grown on the background of cells that were not able to grow, due to expression of the PPDR5 promoter-driven suicide pma1-D378N allele. Seven hundred and thirty-six Gal+ colonies were isolated from screening of more than 35 000 Ura+ transformants in three independent experiments. Among these, 75 colonies were found to be sensitive to cycloheximide (0.1 μg mL−1), and only one contained cells displaying green fluorescence of nuclei in UV light, indicating complete translation of entire Pdr3p fused to GFP (Fig. 3a). Retransformation experiments showed that this Gal+CyhSGfp+ phenotype was conferred by the pYeD1/8-PGAL1-PDR3-GFP plasmid containing the loss-of-function pdr3 mutation. When the FY1679-28C wild-type and corresponding FY1679-28C/ECΔpdr1 mutant strains instead of the Δpdr1Δpdr3 mutant strain were transformed by this plasmid, the resulting transformants displayed cycloheximide and fluconazole sensitivity on medium containing galactose, indicating that the overexpressed loss-of-function pdr3 mutant allele fused with GFP acts as a suppressor of drug resistance in S. cerevisiae (Fig. 3b).

3

Green fluorescence of cells (a) and susceptibilities to cycloheximide (b) on minimal medium containing galactose of transformants carrying the loss-of-function pdr3 mutation on the pYeD1/8-PGAL1-PDR3-GFP plasmid in the genetic background of the wild-type, Δpdr1 and Δpdr1Δpdr3 mutant strains.

Identification of amino acids essential for transactivation function of Pdr3p

The molecular nature of the loss-of-function mutation in the PDR3-GFP fusion gene conferring the Gal+CyhSGfp+ phenotype in the Δpdr1Δpdr3 mutant strain harboring the plasmid-borne PPDR5-pma1-D378N fusion gene was revealed by DNA sequence analysis. Two point mutations in PDR3, resulting in D853Y and N859S amino acid substitutions in Pdr3p, were identified. To our surprise, the N859S mutation already occurred in both the pBFG1-HAPDR3GFP and pYeD1/8-PGAL1-PDR3-GFP plasmids. In order to determine whether the D853Y mutation or its combination with N859S is responsible for the drug-sensitive phenotype of yeast transformants and whether this phenotype is dependent on overexpression of the fusion gene or on the presence of GFP, individual mutations and their combinations were introduced by site-directed mutagenesis into PDR3 borne on three different plasmids, such as pYeD1/8-PGAL1-PDR3-GFP, pYeD1/8-PGAL1-PDR3 and pFL38-PDR3. Mutations in the resulting plasmids were checked by DNA sequence analysis and then introduced by transformation into the FY1679-28C/TDEC Δpdr1Δpdr3 and FY1679-28C/EC Δpdr1 mutant strains as well as into the FY1679-28C wild-type strain. The drug susceptibilities of the resulting transformants were determined on media containing glucose or galactose and different concentrations of cycloheximide and fluconazole. It was found that the D853Y mutation, independently of GFP, is responsible and sufficient for loss of the transactivation function of mutated Pdr3p, resulting in the drug sensitivity of the corresponding Δpdr1Δpdr3 transformants (Table 1). Along with the loss of transactivation function, the pdr3-D853Y allele overexpressed from the PGAL1 promoter on medium containing galactose was able to increase the drug susceptibility of transformants regardless of the presence of wild-type chromosomal PDR3 or both the PDR1 and PDR3 genes. Very weak suppression of cycloheximide and fluconazole resistance was also observed when the pdr3-D853Y allele on centromeric plasmid pFL38 was expressed from its own promoter (Table 1), indicating that the drug-sensitizing activity of the pdr3-D853Y allele is dependent on the level of its expression. Taken together, overexpression of loss-of-function pdr3 alleles, either truncated or bearing the D853Y mutation, is able to enhance the susceptibilities to antifungals in both the wild-type and pdr1 mutant strains.

The pdr3 mutant alleles suppress drug resistance in gain-of-function pdr1 and pdr3 mutants

Gain-of-function mutations in either the PDR1 or PDR3 genes increase the resistance of yeast mutants to several antifungals (Carvajal, 1997; Nourani, 1997; Simonics, 2000). In order to determine whether isolated loss-of-function pdr3 mutant alleles are able to suppress drug resistance in the pdr1-3, pdr1-6, pdr1-8 and pdr3-1 mutants, the susceptibilities to cycloheximide in the corresponding mutant strains EC61, EC62, EC63 and JS3-7D transformed with an empty pYES2 vector and the pYeD1/8-PGAL1-PDR3 plasmid bearing either the wild-type, pdr3-D853Y, pdr3-E902Ter or pdr3-D853Y-GFP alleles were determined. The minimal inhibitory concentration (MIC) of cycloheximide in pdr1 mutant strains reached 2 μg mL−1 (Carvajal, 1997; Simonics, 2000). However, overexpression of the pdr3-D853Y allele in transformants grown on medium containing galactose resulted in suppression of cycloheximide resistance in all the pdr1 mutants tested, and their transformed cells were sensitive to a cycloheximide concentration of 1.4 μg mL−1. Transformants bearing the pYeD1/8-PGAL1-pdr3-D853Y-GFP plasmid displayed even higher cycloheximide sensitivity (Fig. 4). A significant reduction of cycloheximide resistance caused by galactose-induced expression of the pdr3-D853Y and pdr3-E902Ter alleles was also observed in the JS3-7D mutant strain displaying a high level of cycloheximide resistance (MIC >6 μg mL−1), due to a cyh mutation together with the pdr3-1 mutation (Ruttkay-Nedecky, 1992; Delaveau, 1994). In contrast to pdr3-D853Y, overexpression of the pdr3-E902Ter allele in pdr1 mutant strains resulted in a weak reduction of drug resistance that was observed only with pdr1-6 (data not shown).

4

The pdr3-D853Y mutant allele expressed from the PGAL1 promoter is able to sensitize to cycloheximide the drug-resistant pdr1 and pdr3 mutants grown on minimal medium containing galactose.

Diminished PDR5 expression and enhanced rhodamine 6G accumulation in transformants bearing loss-of-function pdr3 mutations

The ability of loss-of-function pdr3 mutant alleles to enhance the susceptibility of yeast to antifungals correlated with the suppression of PDR5 expression and increased accumulation of rhodamine 6G in galactose-grown transformants containing the pYeD1/8-PGAL1-PDR3 plasmid bearing either the E902Ter or D853Y mutation. Northern blot analysis of total RNA extracted from transformants of the wild-type and Δpdr1 mutant strains grown in minimal medium containing galactose revealed strong suppression of PDR5 signals in cells overexpressing the pdr3 alleles in comparison with those containing an empty vector pYES2 only (Fig. 5). Similar results were also observed when the β-galactosidase activities of wild-type cells transformed with plasmids containing loss-of-function pdr3 alleles and the PPDR5-lacZ fusion gene were determined and compared with those of cells containing an empty vector pYES2 (Table 2). These results indicate that Pdr3p mutant forms interfere with the function of Pdr1p and Pdr3p wild-type forms, resulting in the suppression of PDR5 expression and increased drug accumulation in cells.

5

Northern blot analysis of total RNA of transformants of the wild-type and Δpdr1 mutant strains grown in minimal medium containing galactose using 20 μg of isolated RNAs separated onto each lane and blotted onto a nylon membrane. The ribosomal RNAs served as controls for equal loading. 1, Δpdr1PDR3+pYES2; 2, Δpdr1PDR3+pdr3-D853Y; 3, Δpdr1PDR3+pdr3-E902Ter; 4, wild type+pYES2; 5, wild type+pdr3-D853Y; 6, wild type+pdr3-E902Ter.

View this table:
2

Effect of loss-of-function pdr3 mutations on expression of β-galactosidase driven by the PDR5 promoter in transformants of the wild type strain

Plasmidβ-Galactosidase activity (nmol min−1 mg protein−1)
Empty vector pYES22606.33 ± 590.20
pYeD1/8-PGAL1-pdr3-E902Ter282.85 ± 60.83
pYeD1/8-PGAL1-pdr3-D853Y298.58 ± 111.85

In fact, the accumulation of rhodamine 6G, an acknowledged substrate of the Pdr5p efflux pump (Kolaczkowski, 1996), determined by fluorescence cytometry in the wild-type and Δpdr1 mutant cells grown on minimal medium containing galactose, was higher in transformants containing overexpressed pdr3 alleles than in the same strains containing an empty vector pYES2. The geometric mean fluorescence values were significantly higher in transformants of the Δpdr1 mutant than in those of the wild-type strain. At the same time, transformants containing an overexpressed pdr3-D853Y allele accumulated more rhodamine 6G than those bearing the pdr3-E902Ter allele (Fig. 6).

6

Rhodamine 6G accumulation in transformants of the wild-type strain FY1679-28C and the Δpdr1 mutant strain FY1679-28C/EC grown on minimal medium containing galactose as determined by FACS analysis. The geometric mean fluorescence is given in parentheses. Curves on the left (pYES2) are for unstained strains containing an empty pYES2 vector, and curves in with bold are for transformants after rhodamine 6G uptake.

Discussion

In this article, we describe the isolation and characterization of loss-of-function mutations in the multidrug resistance transcription factor encoded by the PDR3 gene in S. cerevisiae. The pdr3 mutant alleles were selected using the screening system recently developed in our laboratory for identification of multidrug resistance reversal agents, genetic suppressors of multidrug resistance, and loss-of-function mutations in transcription factors controlling multidrug resistance in yeast (Kozovska & Subik, 2003; Kozovska, 2004). Two specific mutations in the PDR3 gene resulting in loss of the transactivation function of Pdr3p were identified. The first mutation, generating a stop codon at the position of the 902nd amino acid, led to a truncated protein lacking 75 amino acids from the C-terminal activation domain of the Pdr3p transcription factor. The second mutation selected in the genetic screen for completely translated Pdr3p was identified as the D853Y substitution resulting in loss of the transactivation function of the encoded transcription factor. This amino acid substitution extended by four amino acid residues the predicted α-helical secondary structure of the amino acid sequence consisting of potential 15 amino acid residues in the C-terminal activation domain of Pdr3p. The α-helical structure of the homologous subdomain was also found in ScPdr1p and CgPdr1p, potentially consisting of 12 and 16 amino acid residues, respectively (Fig. 7). These results indicate that specific amino acid residues, such as D853, in the activation domain and apparently the size of the corresponding α-helical subdomain are critical for the transactivation function of Pdr3p. Substitution of such amino acids may inactivate the C-terminal activation domain of Pdr3p or prevent its interaction with other components of the transcription machinery due to its enhanced intramolecular interactions with central regulatory domain (Nourani, 1997).

7

Alignment and secondary structure prediction of amino acid sequences encoded by the PDR3 gene, its loss-of-function pdr3-D853Y allele and PDR1 homologs in Saccharomyces cerevisiae (Sc) and Candida glabrata (Cg) (Vermitsky & Edlind, 2004; Tsai, 2006). Alignment was generated by clustal w (http://clustalw.genome.ad.jp). The secondary structure analysis was carried out using the predator program (http://npsa-pbil.ibcp.fr). Underlined ScPdr3p residues represent amino acids conserved in ScPdr1p, CgPdr1p or both. Blue boxes correspond to α-helix (h), red to extended strand (e) and the rest of the structure to random coil (c). The arrow indicates the D853Y amino acid substitution.

Along with loss of the transactivation function of encoded transcription factors, the isolated loss-of-function pdr3 alleles overexpressed from the PGAL1 promoter were able to sensitize both the wild-type and Δpdr1 mutant strains to cycloheximide and fluconazole. The suppression of drug resistance was observed even in multidrug-resistant mutant strains containing gain-of-function pdr1 or pdr3 mutations. The suppression of multidrug resistance is apparently allele-specific, as in pdr1 mutant strains it was clearly demonstrated only in the presence of the pdr3-D853Y mutant allele and its GFP derivative.

The drug-sensitizing effect of loss-of-function pdr3 mutant alleles was accompanied by a decreased level of PDR5 expression, as revealed by Northern blot analysis, determination of β-galactosidase activities of the PPDR5-lacZ fusion gene, and increased uptake of rhodamine 6G. As Pdr3p reaches the yeast transcription machinery using a different pathway from that used by Pdr1p (Delahodde, 2001), reduced expression of PDR5 in wild-type strains harboring pdr3 mutant alleles rules out the possibility that a mutated Pdr3p prevents the transport of wild-type Pdr3p into the nucleus. Rather, mutant forms of Pdr3p may inhibit the activities of Pdr3p homodimers or Pdr1p/Pdr3p heterodimers (Mamnun, 2002), which are expected to be formed under physiologic conditions in the presence of overproduced mutant forms of Pdr3p, or activate the Rdr1p repressor of PDR5, forming heterodimers with Pdr3p (Hellauer, 2002). However, as suppression of the PDR5 transcription was found to be dependent on a high level of pdr3-E902Ter and pdr3-D853Y expression from the PGAL1 promoter, this effect could be also explained by the competition of loss-of-function Pdr3p with the wild-type Pdr1p and Pdr3p transcription factors for PDREs in promoters of their target genes (DeRisi, 2000), including PDR5 and the chromosomal copy of autoregulated PDR3 (Delahodde, 1995).

The ability of overexpressed loss-of-function pdr3 mutant alleles to sensitize wild-type cells to drugs and suppress multidrug resistance in S. cerevisiae mutant strains may have certain practical consequences. It could also be applied to homologous multidrug resistance transcription factors, opening the way for the development of novel strategies to combat fungal pathogens and their offspring that are resistant to currently used antifungals.

Acknowledgements

This work was supported in part by grants from the Science and Technology Assistance Agency (APVT-20-0502 and APVV-20-0604), the Slovak Grant Agency of Science (VEGA 1/3250/06) and the Slovak Ministry of Education.

Footnotes

  • Editor: André Goffeau

References

View Abstract