OUP user menu

RPD3 and ROM2 are required for multidrug resistance in Saccharomyces cerevisiae

Silvia Borecka-Melkusova, Zuzana Kozovska, Imrich Hikkel, Vladimira Dzugasova, Julius Subik
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00352.x 414-424 First published online: 1 May 2008


The PDR5 gene encodes the major multidrug resistance efflux pump in Saccharomyces cerevisiae. In drug-resistant cells, the hyperactive Pdr1p or Pdr3p transcriptional activators are responsible for the PDR5 upregulation. In this work, it is shown that the RPD3 gene encoding the histone deacetylase that functions as a transcriptional corepressor at many promoters and the ROM2 gene coding for the GDP/GTP exchange protein for Rho1p and Rho2p participating in signal transduction pathways are required for PDR5 transcription under cycloheximide-induced and noninduced conditions. Transposon insertion mutations in ROM2, RPD3 and some other genes encoding specific subunits of the large Rpd3L protein complex resulted in enhanced susceptibility of mutant cells to antifungals. In the rpd3Δ and rom2Δ mutants, the level of PDR5 mRNA and the rate of rhodamine 6G efflux were reduced. Unlike rpd3Δ, in rom2Δ mutant cells the drug hypersensitivity and the defect in PDR5 expression were suppressed by PDR1 or PDR3 overexpressed from heterologous promoters and by the hyperactive pdr3-9 mutant allele. The results indicate that Rpd3p histone deacetylase participating in chromatin remodeling and Rom2p participating in the cell integrity pathway are involved in the control of PDR5 expression and modulation of multidrug resistance in yeast.

Key words
  • multidrug resistance
  • PDR3
  • PDR5
  • ROM2
  • RPD3
  • transcription


In Saccharomyces cerevisiae, the pleiotropic drug (multidrug) resistance (PDR) is caused by increased drug efflux from cells due to overexpressed membrane transporters. These efflux pumps belong either to the ATP-binding cassette (ABC) transporter superfamily or major facilitator superfamily (MFS) members (Balzi & Goffeau, 1991; Kolaczkowska & Goffeau, 1999; Jungwirth & Kuchler, 2006). Microarray experiments have indicated that PDR5 encoding the plasma membrane ABC transporter is one of the most extensively induced genes upon activation of the PDR pathway (DeRisi et al., 2000). Pdr5p recognizes hundreds of structurally and functionally unrelated drugs, lipids and metal ions (Miyahara et al., 1996; Kolaczkowski et al., 1998). The expression of PDR5 and many other drug efflux transporter genes is under the control of several transcription factors (Moye-Rowley, 2003; Jungwirth & Kuchler, 2006). The transcription of PDR5 peaks during mitosis and requires serine/threonine protein kinases such as Elm1p and related proteins acting during mitosis (Souid et al., 2006). PDR5 expression is downregulated in the stationary phase of growth in glucose medium (Mamnun et al., 2004) and is dramatically induced in rho0 petite cells (Hallstrom & Moye-Rowley, 2000).

The transcriptional activators Pdr1p (Balzi et al., 1987) and Pdr3p (Subik et al., 1986; Delaveau et al., 1994) belong to the master regulators of PDR5 and the entire PDR gene network (Fardeau et al., 2006). They display some functional redundancy, recognize the same DNA consensus motif (PDRE) in promoters of target genes (Katzmann et al., 1994; Delahodde et al., 1995) and are constitutively bound to the PDR5 promoter (Mamnun et al., 2002; Gao et al., 2004; Fardeau et al., 2006). In spite of this, they respond differently to some extracellular or intracellular signals and differently regulate the expression of specific genes (Broco et al., 1999; Moye-Rowley, 2003; Jungwirth & Kuchler, 2006; MacPherson et al., 2006). Gain-of-function mutations in the PDR1 and PDR3 genes confer resistance to drugs due to altered expression of PDR5 and other drug efflux transporter genes (Carvajal et al., 1997; Nourani et al., 1997; Michalkova-Papajova et al., 2000; Simonics et al., 2000) as well as of permeases and enzymes involved in lipid and cell wall synthesis (DeRisi et al., 2000). On the other hand, the disruption (Delaveau et al., 1994) or specific loss-of-function mutations in PDR3 (Sidorova et al., 2007) enhance the susceptibility of yeast cells to drugs. However, very little is known about the dynamic functioning of the PDR network in general, particularly about the signal transduction induced by drug and transcriptional cascades leading to the PDR phenotype.

Recently, evidence has been provided showing that PDR5 belongs to a group of stress-related genes whose transcription requires the Spt-Ada-Gcn5 acetyltransferase (SAGA) complex. The PDR5 transcription was associated with an enhanced promoter occupancy of coactivator complexes including SAGA, mediator, the chromatin remodeling SWI/SNF complex and the TATA-binding protein, and the cycloheximide induction was found to alter the nucleosome structure at the PDR5 upstream-activating sequence region harboring PDRE (Gao et al., 2004). Histone acetyltransferases and histone deacetylases function in an antagonistic manner to regulate balance of histone acetylation and gene activity. It is widely accepted that histone acetylation correlates with gene expression, while histone deacetylation is associated with gene repression (Struhl, 1998; Roth et al., 2001; Kurdistani & Grunstein, 2003).

Several signal transduction pathways involved in cellular responses to a wide variety of stresses (Hohmann & Mager, 2003) also modulate the susceptibility of yeast cells to antifungal compounds. They involve the cAMP-protein kinase A (Jain et al., 2003; Park et al., 2005), the calcium–calmodulin–calcineurin (Edlind et al., 2002; Sanglard et al., 2003; Fardeau et al., 2006) and the protein kinase C–cell integrity signaling pathways (Reinoso-Martin et al., 2003; Park et al., 2005; Vilella et al., 2005). The crosstalk between the pathways (Hohmann & Mager, 2003; Park et al., 2005; Fardeau et al., 2006) can ensure proper response of yeast cells to different environmental stresses. The expression modulation of protein kinase C target genes requires the mitogen-activated protein kinase module and the Rlm1p and Swi4p/Swi6p transcription factors, which controls expression of several cell wall synthesis and cell cycle genes (Heinisch et al., 1999; Hohmann & Mager, 2003). In S. cerevisiae, the protein kinase C is activated by the small G-protein Rho1p. The GDP/GTP exchange factor of Rho1p is encoded by the ROM2 gene, which can be activated by cell wall damage (Bickle et al., 1998), phosphatidylinositol-4,5-bisphosphate (Audhy & Emr, 2002) or Tor2p (Schmidt et al., 1997).

In this study, using transposon mutagenesis, two novel genes have been identified: RPD3 and ROM2 encoding histone deacetylase and GDP/GTP exchange protein for Rho1p and Rho2p, respectively, which were required for drug resistance and cycloheximide-induced PDR5 transcription in S. cerevisiae. In contrast to ROM2, the disruption of RPD3 also suppressed the drug resistance in gain-of-function pdr3-9 mutant cells. These results point to the role of histone deacetylase and chromatin remodeling in PDR5 expression and development of multidrug resistance in yeast.

Materials and methods

Strains and culture conditions

The genotypes of yeast strains used in this study are listed in Table 1. Cells were grown on glucose-rich (YPD) medium (2% glucose, 1% yeast extract, 2% bactopeptone), glycerol-rich (YPG) medium (as YPD but 2% glycerol was used instead of 2% glucose) or on minimal (YNB) medium containing a 0.67% yeast nitrogen base without amino acids, 2% glucose, 2% galactose or 2% glycerol plus 2% ethanol and appropriate nutritional requirements. The media were solidified with 2% bactoagar. The media containing azole antifungals were buffered with 0.165 M 3-(N-morpholino) propanesulfonic acid, pH 6.6. The Escherichia coli XL1-Blue strain was used as a host for transformation, plasmid amplification and preparation. The bacterial strains were grown at 37 °C in Luria–Bertani medium (1% tryptone, 1% NaCl, 0.5% yeast extract, pH 7.0) supplemented with 50 μg mL−1 ampicillin for selection of transformants. For induction of rho0/rho- mutants, yeast cells were grown for 24 h in the presence of 25 μg mL−1 ethidium bromide in YPD medium.

View this table:

Saccharomyces cerevisiae strains used in this study

FY1679-28C/ECMATa ura3-52 trp1-63 pdr1∷TRP1 leu2-1 his3-200Nourani (1997)
FY1679-28C/EC rpd3∷mTn3 MATa ura3-52 trp1-63 pdr1∷TRP1 leu2-1 his3-200 YNL330c∷Tn3LEU2lacZThis study
FY1679-28C/EC rom2∷mTn3 MATa ura3-52 trp1-63 pdr1∷TRP1 leu2-1 his3-200 YLR371w∷Tn3LEU2lacZThis study
BY4742 MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0EUROSCARF
BY4742 rpd3Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YNL330c∷kanMX4EUROSCARF
BY4742 rom2Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YLR371w∷kanMX4EUROSCARF
BY4742 dep1Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YAL013w∷kanMX4EUROSCARF
BY4742 sds3Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YIL084c∷kanMX4EUROSCARF
BY4742 sin3Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YOL004w∷kanMX4EUROSCARF
BY4742 ume6Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YDR207c∷kanMX4EUROSCARF
BY4742 ash1Δ MATαhis3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 YKL185w∷kanMX4EUROSCARF

Generation of mutants using a transposon insertion yeast genomic library

The LEU2 transposon library was received as a generous gift from M. Snyder's lab (Burns et al., 1994). Escherichia coli XL1-Blue were transformed with the pools of DNA according to the standard bacterial transformation protocol. The mutated yeast DNA from amplified pools containing the 6.6 kb mTn3(lacZ/LEU2) cassette was restricted with NotI and introduced into FY1679-28C/EC using the lithium acetate transformation. Transformants were selected on minimal medium supplemented with uracil and histidine.

Screening and analysis of transposon insertion mutants

Leucine-positive transformants were replica plated onto minimal medium containing 0.3 μg mL−1 cycloheximide. Cycloheximide-sensitive transformants were further analyzed by growth assays, using serial dilutions and spotting onto agar plates containing different concentrations of cycloheximide. Growth was scored after 5 days of incubation at 30 °C.

Identification of disrupted genes that caused the cycloheximide-sensitive phenotype

The sites of mTn3(lacZ/LEU2) insertion in the mutants were determined by rescuing a genomic DNA fragment adjacent to the lacZ sequence and DNA sequence analysis. Cycloheximide-sensitive mutants were transformed with 1 μg of BamHI-digested pRSQ2-URA3 plasmid (http://ygac.med.yale.edu) that carried the E. coli replication origin, the bacterial lacZ gene, the yeast URA3 gene and partial sequences of lacZ at both ends of the digested fragment. Transformants were selected on minimal medium supplemented with histidine. Genomic DNA was extracted from transformants, digested with SalI, circularized with T4 DNA ligase and introduced into E. coli by electroporation. Plasmid DNA was prepared from individual E. coli transformants and digested with BamHI and SalI to verify the presence of 2.85-kb band containing vector sequences and additional bands derived from yeast genomic DNA. The sequence flanking the lacZ was determined by an ABI Prism 3100 DNA sequencer (Applied Biosystems, Foster City, CA) using the primer 5′-GTTTTCCCAGTCACGAC-3′. The position of transposon insertion was determined using the blast service of the Saccharomyces genome database at the NCBI.

Drug susceptibility assays

Drug susceptibility was determined by a spot assay. Five microliters of suspension of three independent clones, grown on solid glycerol-rich (YPG) medium to eliminate rho0 cells, were spotted onto minimal media containing glucose, galactose or glycerol and supplemented with various concentrations of antifungals. Growth at 30 °C was scored after 5 days on media containing glucose and 12 days on media containing glycerol or galactose. The drug concentrations were as follows: cycloheximide: 0.05, 0.1, 0.2, 0.3, 0.4, 0.6, 0.8, 1.0 and 1.5 μg mL−1; fluconazole: 10, 20, 30 and 40 μg mL−1; and rhodamine 6G: 2.5, 5.0, 7.5, 10 and 20 μg mL−1. The susceptibility to antifungals was also assessed using zone-inhibition assays. Approximately 107 stationary-phase cells were plated onto minimal glucose agar. The filter discs (diameter of 6 mm), soaked with an appropriate amount of antifungals, were placed on the plates, which were incubated at 28 °C for 3–5 days before determination of the diameter of the zone of growth inhibition.

Plasmids and recombinant DNA techniques

Plasmid pFL38 (ARS CEN URA3) was used as an empty vector to introduce a plasmid born URA3 gene into host strains. Plasmids pFL38-PDR3, pFL38-pdr3-9 (Nourani et al., 1997) and pYeD1/8-PGAL1-PDR3 (Sidorova et al., 2007) served as a source of the PDR3 gene. The PDR1 gene was expressed from the pFL38-PPGK1-HAPDR1 plasmid (Delahodde et al., 2001). Standard protocols were used for generating recombinant DNAs, the 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 Ward (1990). Yeast transformation was carried out using the modified lithium acetate protocol (Nourani et al., 1997) or by electroporation (Thompson et al., 1998).

Northern blot analysis

Samples of total RNA were prepared from cells of transformants grown in YPD medium to the mid-logarithmic phase by the hot acidic phenol extraction method (Ausubel et al., 1989). RNA concentrations were determined spectrophotometrically. Ten micrograms of 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 et al., 1998). Autoradiographs of Northern blots were quantified with a phosphoimager (Bio-Rad, Hercules, CA).

Rhodamine 6G accumulation and efflux

The accumulation of rhodamine 6G (Sigma-Aldrich, Taukirchen, Germany) was measured by flow cytometry as described (Sidorova et al., 2007). Active efflux of rhodamine 6G was determined as described previously (Kolaczkowski et al., 1996; Nakamura et al., 2001; Mukhopadhyay et al., 2002). Approximately 108 cells from an overnight culture were incubated in 100 mL of YPD medium and grown for 5 h at 30 °C. The cells were pelleted and washed three times with 50 mM HEPES, pH 7.0. Cells were resuspended in 50 mM HEPES containing 2 mM 2-deoxyglucose and 10 μM rhodamine 6G to a concentration of 107 cells mL−1 and shaken for 2 h at 30 °C to exhaust the energy and allow rhodamine 6G accumulation. Cells were then washed three times and resuspended in 50 mM HEPES, pH 7.0, to a concentration of 108 cells mL−1. At specific intervals after the addition of glucose (final concentration, 2 mM) to initiate rhodamine 6G efflux, the cells were centrifuged and 100 μL supernatants were added to Nunc 96-well fluoro-/luminunc plates (Nalge Nunc International, Rochester, NY). Rhodamine 6G fluorescences of the samples were measured at an excitation wavelength of 515 nm and an emission wavelength of 555 nm using a Tecan Saphir II TM spectrofluorimeter (Tecan Austria GmbH, Grödig/Salzburg, Austria). Statistical significance between values in different strains was determined by paired Student's t-test analyses. A value of P<0.05 was considered to be significant.

Real-time PCR

Yeast cells were grown in minimal medium containing glucose to the mid-logarithmic phase. Total RNA was isolated by the hot acidic phenol extraction method (Ausubel et al., 1989). One microgram of RNA was used as a template in each sample with a reaction mix (MBI Fermentas, Vilnius, Lithuania) containing 200 U of Revert Aid H Minus M-MuLV Reverse Transcriptase (MBI Fermentas, Vilnius, Lithuania) in a total volume of 20 μL. The reaction mixture was incubated for 5 min at 37 °C, 60 min at 42 °C and 10 min at 70 °C.

For the quantitative PCR reaction, primers were designed using the primer3 v. 0.4.0 (Rozen & Skaletsky, 2000) and perlprimer v. 1.1.4a (Marshall, 2004) software. Primers and magnesium concentrations were optimized for each gene analyzed to evaluate nonspecific amplifications. Control reactions including RNA instead of cDNA were performed for each gene and under all conditions. The threshold values for Ct determination were set in the lower region of the log-linear part of the amplification curves for all genes. The average Ct value for each sample was calculated from three experiments. The Ct value of the ACT1 gene was used for normalization of variable cDNA levels. The primers used were: ACT1-For GAATTGAGAGTTGCCCCAGA, ACT1-Rev GGCTTGGATGGAAACGTAGA and PDR5-For GTGCTCATAAATGCCCTGCT, PDR5-Rev TGAACGGCCCTGTACTCTTC. The quantitative PCR was prepared using the AmpliQ Real Time universal PCR master mix (Ampliqon, Herleu, Denmark), cDNA equivalent of 30 ng total RNA and 5 pmol of both primers in a final volume of 20 μL. The samples were prepared by adding 10 μL of 2 × RealQ PCR master mix, 5 pmol of the both For and Rev oligonucleotides and H2O to the cDNA to reach a final volume of 20 μL. Amplification was carried out in the 7900 HT Fast Real-Time PCR System (Applied Biosystems, Foster City) using a denaturation step of 15 min at 95 °C and 40 cycles of 95 °C for 30 s and 57 °C for 1 min. A melting curve was obtained after completion of the cycles to verify the presence of a single amplicon. The reporter signals were analyzed using the abi sds 2.2.2 software (Applied Biosystems, Foster City). Target nucleic acids were quantified using relative quantification determining the ration between the amount of target gene and the ACT1 reference gene. The RT-PCR results presented are mean values of three independent experiments.


The screen for novel genes required for development of drug resistance

To identify novel genes required for Pdr3p-mediated multidrug resistance in yeast, the S. cerevisiae mutant strain FY1679-28C/EC disrupted in the PDR1 gene was chosen for transposon mutagenesis. About 4500 transformants were first selected on minimal plates. From them, only Leu+ transformants reproducibly displaying sensitivity to 0.2 μg mL−1 cycloheximide were selected. The disruption of the PDR3 gene in selected transposon insertion mutants was ruled out due to observed maintenance of hypersensitivity to cycloheximide of their cells after transformation with the PDR3 gene carried on a centromeric plasmid pFL38-PDR3. In this study, the physiological consequence of two mutations conferring hypersensitivity to cycloheximide is examined.

The identification of genes disrupted by transposon insertion revealed that RPD3 (YNL330c) and ROM2 (YLR371w) are involved in the modulation of multidrug resistance in selected yeast mutants. The RPD3 gene (1302 bp) encodes histone deacetylase involved in transcriptional regulation (Rundlett et al., 1996; Ekwall, 2005), transcriptional silencing (Vannier et al., 1996) and in mitotic exit (Ye et al., 2005). The ROM2 gene (4071 bp) encodes a GDP/GTP exchange protein for Rho1p and Rho2p participating in the regulation of 1,3-β-glucan synthase, protein kinase C and several signal transduction pathways (Drgonova et al., 1996; Vay et al., 2004; Park et al., 2005; Vilella et al., 2005). The transposon insertion point was found at 177 and 2661 bp in the coding regions of RPD3 and ROM2, respectively.

Transposon insertion mutations in RPD3 and ROM2 resulted in enhanced susceptibility to different antifungals

In two transposon insertion mutants studied, the cycloheximide hypersensitivity was found to be accompanied by increased sensitivities to fluconazole, rhodamine 6G and other azole antifungals (Tables 2 and 3). The susceptibility to all drugs of the pdr1Δrpd3∷mTn3 mutant cells was significantly higher than that of the pdr1Δrom2∷mTn3 mutant cells. In addition, on glucose-rich medium, the pdr1Δrpd3∷mTn3 mutant cells were also heat shock (6 min at 55 °C) sensitive and displayed sensitivity to 12 μg mL−1 calcofluor white and 0.5 M LiCl. The pdr1Δrom2∷mTn3 mutant cells were also sensitive to calcofluor white and in addition displayed a higher sensitivity to zymolyase 20T (2.5 μg mL−1) treatment (data not shown). The pleiotropic phenotypes of mutant cells indicate that the enhanced susceptibility of mutant strains to drugs resulted from the suppression of their stress-response mechanisms.

View this table:

Minimal inhibitory concentrations of drugs in transposon insertion mutant strains

Minimal inhibitory concentrations (μg mL−1)
StrainGenotypeCycloheximideFluconazoleRhodamine 6G
FY1679-28C/EC pdr1Δ0.30207.5
BY4742 PDR10.402020
  • NT, not tested.

  • * Minimal medium containing glucose.

  • Glycerol rich medium.

View this table:

Susceptibilities of transposon insertion mutant strains to different azole antifungals determined by agar diffusion assay

Diameter of growth inhibition zone (mm)
FY1679-28C/EC pdr1Δ312137373953
  • Minimal medium containing glucose was buffered to pH 6.6 with 0.165 M MOPS. Drug amounts per disk: 25 μg of fluconazole and 30 μg of other azole antimycotics.

The susceptibility of yeast cells to cycloheximide was also determined in the rpd3 and rom2 null mutants purchased from the collection of EUROSCARF. These mutant strains were generated in the genetic background of the BY4742 strain containing intact chromosomal PDR1 and PDR3 genes. As shown in Table 2, the drug hypersensitivity was also observed in the rpd3∷kanMX4 and rom2∷kanMX4 mutant cells in which a part of the gene coding sequence was replaced with kanMX4. Moreover, the sensitivity to 0.1 μg mL−1 cycloheximide was also observed in isogenic mutant strains disrupted either in the DEP1, SIN3 or SDS3 genes encoding components of the histone deacetylase complex. SDS3 encodes protein that is a subunit unique to the Rpd3L protein complex (Vannier et al., 2001; Carrozza et al., 2005). Surprisingly, deletion of the UME6 and ASH1 genes encoding two sequence-specific DNA-binding proteins of Rpd3L (Carrozza et al., 2005) did not affect the sensitivity of mutant cells to cycloheximide. These results indicate that independent of the presence of the master multidrug resistance regulator Pdr1p, the large Rpd3L histone deacetylase protein complex targeting deacetylation to the promoter as well as Rom2p are essential for a proper response of yeast cells to stress induced by antifungal agents.

Saccharomyces cerevisiae cells that have lost their mitochondrial genome strongly induce a Pdr3p-dependent transcription of multidrug resistance genes and display lower susceptibility to drugs as the wild-type cells (Hallstrom & Moye-Rowley, 2000). Such retrograde signaling also appears to be intact in the pdr1Δrpd3∷mTn3 and pdr1Δrom2∷mTn3 mutant strains. The rho0/rho cells induced by ethidium bromide mutagenesis in transposon insertion mutants were found to be more resistant to cycloheximide than their rho+ parental strains (Fig. 1). Nevertheless, the degree of resistance to cycloheximide of rho0/rho cells in the rpd3∷mTn3 mutant strain was significantly lower than that observed in the wild type and rom2∷mTn3 mutant strains, indicating that the expression of drug efflux pumps in the rpd3 mutant lacking functional mitochondria is severely diminished. Essentially the same results were also observed with the rpd3Δ, sin3Δ and rom2Δ mutants derived from the BY4742 strain.


Pdr3p-mediated retrograde regulation of drug resistance in the rpd3Δ and rom2Δ mutant cells.

Mutations in RPD3 and ROM2 diminish the PDR5 mRNA level and rhodamine 6G efflux

The enhanced susceptibilities of the rpd3Δ and rom2Δ mutant strains to cycloheximide and other drugs correlated with a reduced expression of the PDR5 gene encoding the main multidrug resistance efflux pump. The levels of PDR5 mRNA in the rpd3, sin3 and rom2 mutants determined by Northern blot analysis were significantly lower than in the corresponding wild-type strains. Although the PDR5 mRNA levels in mutant strains were still responsive to cycloheximide, they were substantially reduced after 45 min of exposure to drug (Fig. 2). In the BY4742 genetic background, cycloheximide increased the level of PDR5 mRNA in the wild-type strain, rpd3Δ, sin3Δ and rom2Δ mutants 2.01, 1.73, 1.61 and 1.22 times, respectively. In the absence of functional PDR1 gene, the level of PDR5 expression in response to cycloheximide was lower. Induction levels for the wild-type strain, rpd3Δ and rom2Δ mutants were 1.61, 1.48 and 1.52, respectively.


Northern blot analysis of PDR5 mRNA in the wild type, rpd3Δ, sin3Δ and rom2Δ mutant strains grown in glucose-rich medium in the absence and in the presence of cycloheximide. (a) Saccharomyces cerevisiae FY1679-28C/EC; cycloheximide (0.025 μg mL−1). (b) Saccharomyces cerevisiae BY4742; cycloheximide (0.050 μg mL−1).

Rhodamine 6G is an acknowledged substrate of the Pdr5p efflux pump (Kolaczkowski et al., 1996). Compared with parental strains, its accumulation determined by fluorescence cytometry was only slightly enhanced both in the rpd3Δ and the rom2Δ mutants derived from the FY1679-28C/EC strain as well as in the rom2Δ mutant derived from the BY4742 strain (Table 4). The rate of the energy-dependent rhodamine 6G efflux was determined by measuring the drug concentrations in supernatants of starved cells after incubation in drug-free medium. The addition of glucose induced the rhodamine 6G efflux that displayed a relatively longer lag in the rom2Δ mutant cells (Fig. 3). Compared with the wild-type strain BY4742, the drug efflux rates (130.3, 71.1 and 84.2 pmoles mL−1 10−8 cells for the wild type, rpd3Δ and rom2Δ mutant strains, respectively) calculated from the slope of the efflux curves as well as the total amounts of drug expelled 6 min after addition of glucose were significantly lower (P values ranged from 0.001 to 0.039) in the rpd3Δ and rom2Δ mutant cells, corroborating their reduced level of an PDR5 expression. The lack of exact correlation among the sensitivity to rhodamine 6G on glycerol medium, its accumulated amounts in YPD medium and its glucose-induced efflux rates probably reflects the differences in the experimental conditions used in the corresponding methods. Nevertheless, the results clearly show that RPD3 and ROM2 are required for proper expression of the PDR5 gene, determined in the absence or in the presence of cycloheximide, resulting in a decreased rate of drug efflux from mutant cells.

View this table:

Rhodamine 6G accumulation in the wild type strains FY1679-28C/EC, BY4742 and corresponding rpd3Δ and rom2Δ mutants determined by fluorescence cytometry

The geometric mean fluorescence
StrainGenotype− Rhodamine 6G+ Rhodamine 6G
FY1679-28C/EC pdr1Δ13.06 ± 1.0732.88 ± 5.80
28C/EC pdr1Δrpd3∷mTn318.42 ± 1.6869.68 ± 10.91
pdr1Δrom2∷mTn321.24 ± 1.4283.88 ± 1.57
BY4742 PDR113.89 ± 2.0728.55 ± 0.34
rpd3∷kanMX415.93 ± 0.5830.80 ± 3.13
rom2∷kanMX420.73 ± 0.5866.50 ± 1.45
  • The results are mean values of three independent experiments.


Energy-dependent rhodamine 6G efflux from cells of the wild-type strain BY4742 and its rpd3Δ and rom2Δ mutants. Glucose (final concentration 2 mM) was added at time zero. The results are the mean±SD for the four independent experiments.

Mutations in RPD3 suppress multidrug resistance and PDR5 transcription even in the presence of overexpressed or activated Pdr3p

To find out whether a diminished PDR5 transcription and suppressed multidrug resistance in mutant strains are caused by the defect in the transcription of the PDR1 and PDR3 genes or in the activation of Pdr3p, we determined the susceptibility to drugs of their transformants expressing a plasmid-borne PDR3 gene from its own promoter, expressing PDR1 and PDR3 from heterologous promoters or containing a plasmid born gain-of-function pdr3-9 mutant allele. In the BY4742 genetic background, in contrast to a single-copy plasmid-borne PDR3 gene, the expression of PDR1 from the PPGK1 promoter as well as the activation by galactose of the PGAL1-PDR3 fusion gene bypassed the defect caused by the rom2Δ mutation and resulted in an increased level of resistance to cycloheximide (Table 5). The highest level of drug resistance, similar to that displayed in wild-type strains, was observed in rom2Δ transformants carrying the gain-of-function pdr3-9 allele. On the other hand, rpd3Δ transformants harboring the PPGK1-PDR1 or PGAL1-PDR3 fusion genes failed to display cycloheximide resistance. A small but significant increase in drug resistance was only observed in the presence of the pdr3-9 allele. These results indicate that the Rpd3p histone deacetylase is required for proper PDR5 transcription and establishment of multidrug resistance in yeast independently on promoters and cellular concentrations of Pdr1p/Pdr3p, even in the presence of the activated form of the Pdr3p transcriptional activator.

View this table:

Minimal inhibitory concentrations of cycloheximide in transformants of the rpd3Δ and rom2Δ mutant strains

Minimal inhibitory concentrations (μg mL−1)
FY1679-28C/EC pdr1ΔPDR30.
28C/EC pdr1Δrpd3∷mTn30.
BY4742 PDR1PDR30.

Real-time quantitative reverse-transcription PCR (qRT-PCR) was used to measure mRNA levels for the PDR5 gene in the wild type, rpd3Δ and rom2Δ transformants containing the gain-of-function pdr3-9 allele on the centromere plasmid. Transformants were grown to the mid-log phase, total RNA was isolated and the relative transcript levels of each gene were determined. As shown in Table 6, the level of overexpressed PDR5 driven by the hyperactive pdr3-9 allele was significantly decreased in the rpd3Δ mutant cells while in the rom2Δ mutant it was comparable to that in the wild-type strain. After cycloheximide treatment, a further increase in the PDR5 transcript level was-observed in all pdr3-9-transformed strains but its level in the rpd3Δ and rom2Δ mutants was significantly reduced compared with wild-type cells. These results indicate that RPD3 is required not only for basal and drug-induced expression of PDR5 but also for the PDR5 expression driven by the hyperactive mutant form of the Pdr3p transcriptional activator. Apparently, the absence of Rom2p did not influence the pdr3-9-driven PDR5 expression but reduced its further induction by cycloheximide.

View this table:

Relative PDR5 transcript abundance in transformants of the wild type, rpd3Δ and rom2Δ mutant cells containing a plasmid-borne gain-of-function pdr3-9 allele

PDR5 expression compared to calibrator
Strain− CYH+ CYH
BY47421.00 ± 0.081.92 ± 0.12
BY4742 rpd3Δ0.66 ± 0.091.09 ± 0.05
BY4742 rom2Δ1.05 ± 0.191.35 ± 0.22
  • Total RNA was isolated from cells grown in minimal medium containing glucose and after real-time quantitative RT-PCR the relative amounts of PDR5 normalized to ACT1 transcript levels were determined. The PDR5 expression in wild-type cells untreated by cycloheximide were used as calibrator.


In this study, using a genetic screen for the selection of S. cerevisiae drug-sensitive mutants resulting from the transposon disruption of chromosomal genes, two genes were identified: RPD3 and ROM2, involved in the control of multidrug resistance. Usually, multidrug resistance results from overexpression of the PDR5 gene due to gain-of-function mutations in either Pdr1p or Pdr3p transcriptional activators. Independent of the presence of Pdr1p, the isolated rpd3Δ and rom2Δ mutant strains displayed an enhanced susceptibility to antifungals, a reduced level of PDR5 mRNA under both induced and noninduced conditions and a decreased rate of rhodamine 6G efflux from cells. The ability of cycloheximide or dysfunctional mitochondria to enhance the PDR5 mRNA levels was apparently preserved in both mutant strains. Unlike rpd3Δ, in rom2Δ mutant cells, the drug hypersensitivity was suppressed by overexpressed PDR1 and PDR3 as well as by the gain-of-function pdr3-9 mutant allele.

One possible mechanism linking disruption of ROM2 to impaired protein kinase C signaling and downregulation of PDR5 may involve Swi4p/Swi6p. Rom2p is a GDP/GTP exchange protein for Rho1p and Rho2p participating in the regulation of 1,3-β-glucan synthase (Drgonova et al., 1996) and protein kinase C-mediated cell integrity signaling (Hohmann & Mager, 2003). Cell integrity signaling results in dual phosphorylation and activation of Slt2p serine/threonine mitogen-activated protein kinase that exerts its effects via two transcription factors: Rlm1p and the Swi4p/Swi6p cell cycle box factor complex. Rlm1p, Swi4p and Swi6p are phosphorylated in an Slt2p – dependent manner. Moreover, Swi6p is also the regulatory subunit of the Mbp1p/Swi6p MCB transcription factor complex. These transcription factors activate many genes involved in cell wall synthesis, cytoskeletal organization and cell cycle progression (Madden et al., 1997; Heinisch et al., 1999; Vay et al., 2004; Park et al., 2005; Vilella et al., 2005). In S. cerevisiae, mitotic entrance, also displaying PDR5 mRNA accumulation (Souid et al., 2006), is coupled with morphogenesis (Sakchaisri et al., 2004). In fact, the PDR5 promoter contains Swi4p and Mbp1p cell cycle response elements, and the interaction of Swi4p (Workman et al., 2006) and Mbp1p (Chua et al., 2006) with PDR5 has already been documented. Disruption of the SLT2 gene resulted in drug hypersensitivity both in S. cerevisiae (Reinoso-Martin et al., 2003) and in pathogenic Candida species (Edlind et al., 2005), corroborating the significance of this pathway in the control of multidrug resistance. Whether Rlm1p and Swi4p/Swi6p regulate PDR5 expression and drug resistance directly or indirectly remains to be determined.

Chromatin structure exerts vital control over gene expression, DNA replication, recombination and repair (Workman & Kingston, 1998; Khorasanizadeh, 2004). RPD3, the second gene identified in the authors' screen, encodes histone deacetylase participating in chromatin remodeling (Rundlett et al., 1996; Ekwall, 2005). Acetylation and deacetylation of lysine residues of histone are, respectively, linked to transcriptional activation or repression (Workman & Kingston, 1998; Roth et al., 2001; Kurdistani & Grunstein 2003). In S. cerevisiae, two distinct complexes of Rpd3, large Rpd3L (1.2 MDa) and small Rpd3S (0.6 MDa), have been identified. They are sharing Rpd3p, Sin3p and Ume1p but certain subunits are specific only for one protein complex. Rpd3S contains two unique subunits, Rco1p and Eaf3p, while Sds3p is a subunit unique to Rpd3L (Carrozza et al., 2005). As shown in Table 2, the sds3Δ mutant cells were hypersensitive to drugs and displayed a diminished level of PDR5 mRNA like the rpd3Δ mutant. This suggests that the large Rpd3L complex is responsible for regulating PDR5. This complex containing the DNA-binding proteins Ume6p and Ash1p was found to be responsible for the targeted deacetylation of the promoters of genes (Robyr et al., 2002; Carrozza et al., 2005). However, deleting either the UME6 or the ASH1 gene did not affect the susceptibility of mutant cells to cycloheximide (Table 2). Recently, the Ume6p- and Ash1p-independent activation of DNA damage-inducible RNR3 and HUG1 genes by the large Rpd3L protein complex has been demonstrated (Sharma et al., 2007).

While histone deacetylation is generally associated with the gene repression (Struhl, 1998; Roth et al., 2001; Kurdistani & Grunstein, 2003), its role in transcription activation has been appreciated only recently (Bernstein et al., 2000). In fact, histone deacetylases have been found to be required for the activation of osmotic stress genes (Proft & Struhl, 2002; De Nadal et al., 2004), galactose-induced genes (Wang et al., 2002), DNA damage-induced RNR3 and HUG1 genes (Sharma et al., 2007) as well as the anaerobic DAN/TIR genes encoding the cell wall mannoproteins (Sertil et al., 2007). Thus, Rpd3p can function as a positive regulator of transcription in addition to its usual role as a repressor.

Exactly how Rpd3p affects the changes in the PDR5 mRNA level is not clear. Whether Rpd3p activates PDR5 and other multidrug resistance genes by counteracting repressor function or facilitating activator binding and nucleosome depletion is currently unknown. The PDR5 downregulation caused by deletion of RPD3 may also occur indirectly by changes in expression of other genes (Robyr et al., 2002). The changes in PDR5 mRNA stability also cannot be ruled out.

The positive roles of Rom2p and Rpd3p in the regulation of multidrug resistance in yeast pave the way for isolating novel inhibitors interfering with the functions of these proteins and enhancing the sensitivity of pathogenic yeast to antifungals. In fact, specific histone deacetylase inhibitors were found to reduce the azole-induced upregulation of the CDR1, ERG1 and ERG11 genes in Candida albicans and enhance its sensitivity to azole antimycotics (Smith & Edlind, 2002). Discovery of uracil-based histone deacetylase inhibitors able to reduce the acquired resistance to antimycotics and trailing growth in C. albicans (Mai et al., 2007) as well as its pathogenicity and virulence (Simonetti et al., 2007) may prove to be useful in the combined treatment of infection caused by this pathogen.


The authors are grateful to Drs C. Jacq and K. Kuchler for plasmid born fusion genes, and Drs M. Snyder and A. Kumar for transposon library used in this study. The authors thank Drs J. Sedlak and D. Cholujova for advice with FACS measurements and Dr D. Drahovska for help in RT PCR experiments. The authors also thank Z. Kratochvilova for help in some experiments. This work was supported by grants from the Slovak Research and Developmental Agency (APVV-20-00604, LPP-0022-06), Slovak Grant Agency of Science (VEGA 1/3250/06) and Comenius University (UK/151/2007).


View Abstract