Clustered amino acid substitutions in the yeast transcription regulator. Pdr3p increase pleiotropic drug resistance and identify a new central regulatory domain.
Ó Springer-Verlag 1997
Mol Gen Genet (1997) 256: 397±405
ORIGINAL PAPER
A. Nourani á D. Papajova á A. Delahodde C. Jacq á J. Subik
Clustered amino acid substitutions in the yeast transcription regulator Pdr3p increase pleiotropic drug resistance and identify a new central regulatory domain Received: 5 February 1997 / Accepted: 20 June 1997
Abstract In the yeast Saccharomyces cerevisiae mutations in the genes encoding the transcription factors Pdr1p and Pdr3p are known to be associated with pleiotropic drug resistance mediated by the overexpression of the eux pumps Pdr5p, Snq2p, and Yor1p. Mutagenesis of PDR3 was used to induce multidrug resistance phenotypes and independent pdr3 mutants were isolated and characterized. DNA sequence analysis revealed seven dierent pdr3 alleles with mutations in the Nterminal region of PDR3. The pdr3 mutants were semidominant and conferred dierent drug resistance patterns on host strains deleted either for PDR3 or for PDR3 and PDR1. Transactivation experiments proved that the mutated forms of Pdr3p induced increased activation of the PDR3, PDR5, and SNQ2 promoters. The amino acid changes encoded by ®ve pdr3 mutant alleles were found to occur in a short protein segment (amino acids 252±280), thus revealing a regulatory domain. This region may play an important role in protein±DNA or protein±protein interactions during activation by Pdr3p. Moreover, this hot spot for gain-of-function mutations overlaps two structural motifs, MI and MII, recently proposed to be conserved in the large family of Zn2Cys6 transcription factors. Key words Saccharomyces cerevisiae á Transcription factor regulation á PDR3 á Multidrug resistance á Mutational analysis
Communicated by C. P. Hollenberg A. Nourani á A. Delahodde á C. Jacq á J. Subik (&) Laboratoire de GeÂneÂtique MoleÂculaire, CNRS URA1302, Ecole Normale SupeÂrieure, 46 rue d'Ulm, F-75230 Paris Cedex 05, France D. Papajova á J. Subik Department of Microbiology and Virology, Comenius University, Mlynska dolina B2, 842 15 Bratislava, Slovak Republic Tel. +4217 796 631; fax +4217 729 064
Introduction PDR1 and PDR3 are two homologous yeast genes which code for transcription regulators of the Zn2Cys6 family. Their target genes so far identi®ed are: (1) the three ATP binding cassette (ABC) transporters PDR5, SNQ2, and YOR1 (Decottignies et al. 1995; Katzmann et al. 1994, 1995; Mahe et al. 1996); and (2) the two hexose transporters HXT9 and HXT11, belonging to the major facilitator superfamily (MFS) (Nourani et al. 1997). Interestingly, while PDR1/PDR3-mediated overexpression of PDR5, SNQ2, and YOR1 leads to a multidrug resistance phenotype, overexpression of HXT9 and HXT11 makes the cell more sensitive to the same set of drugs. This antagonistic eect of the two transcription factors emphasizes their key function in the control of membrane biogenesis. Pdr1p and Pdr3p are members of the Saccharomyces cerevisiae family of Zn2Cys6 transcription factors which contains 52 members, among which Gal4p is the best known representative. General properties of Pdr1p and Pdr3p dier from those of Gal4p in several respects. First, while Gal4p binds to DNA sequences in which the two CGG inverted repeats are separated by a spacer (CGG-N11-CCG), Pdr3p binds to an everted repeat (CCGCGG) (Delahodde et al. 1995; Hellauer et al. 1996). Second, Pdr1p and Pdr3p act together to trigger the drug response. In spite of the fact that the two transcription factors recognize the same DNA sequence, they play dierent and complementary roles in the activation process. The PDR3 promoter contains two pleiotropic drug response elements, which can be recognized by Pdr1p or Pdr3p, suggesting that an autoregulatory loop is important in establishing a two-step regulation of the drug resistance process (Delahodde et al. 1995). On the other hand, Pdr1p/Pdr3p share with Gal4p and most of the Zn2Cys6 family members features which reveal common properties: 1. The DNA-binding domains of Pdr1p/Pdr3p are in the N-terminal parts of the proteins and, at least for
398
Pdr3p, there are two activation domains (amino acids 61±109 and 796±976), a situation which is reminiscent of that found for Gal4p (Delaveau et al. 1994). 2. As for Gal4p, the coactivator/repressor, Ngg1p (also called Ada3p), inhibits transcriptional activation by Pdr1p/Pdr3p. This repression is mediated by direct or indirect interactions between Ngg1p and the C-terminal part of the transcription factor (amino acids 815±967 in the case of Pdr3p) (Martens et al. 1996). 3. It was also recently shown, by sequence comparisons with 62 fungal members of the Zn2Cys6 family, that eight motifs are conserved in evolution (Poch 1997). On the basis of deletion experiments conducted with Gal4p, it was suggested that these conserved motifs might correspond to an inhibitory domain (Stone and Sadowski 1993). Very few mutational analyses have been conducted to examine the putative regulatory properties of these conserved motifs. The present work describes a random mutagenesis of the PDR3 region coding for the N-terminal region (amino acids 1±660) which contains the eight motifs conserved among GAL4 family members. Selection for mutant forms of the transcription factor which confer high levels of resistance to cycloheximide and other unrelated drugs allowed us to identify a speci®c functional inhibitory domain.
Materials and methods Strains, plasmids, and media The S. cerevisiae strains used in this study are: W303-1A (MATa ade2-1 his3-11,15 leu2-3, 118 trp1-1 ura3-1); W303-1A/TD (MATa ade2-1 his3-11,15 leu2-3, 118 trp1-1 ura3-1 pdr3::HIS3); FY 167928C (MATa ura3-52 trp1D63 leu2D1 his3D200); FY1679-28C/TD (MATa ura3-52 trp1D63 leu2D1 his3D200 pdr3::HIS3); FY167928C/TDEC (MATa ura3-52 trp1D63 leu2D1 his3D200 pdr1::TRP1 pdr3::HIS3) (Delaveau et al. 1994). Escherichia coli TG1 was used to propagate all plasmids. The selection-transformation step in the directed mutagenesis procedure was done in the E. coli mutS strain: (supE, thi-1,D(lac-proAB), F¢(proAB+, lac IqZDM15), mutS::Tn10). The PDR3 gene was cloned on centromeric plasmid pFL38-PP3 (ARS1 CEN4 URA3 PDR3) under the control of its promoter (Delaveau et al. 1994). The plasmids used in b-galactosidase assays, PDR3-lacZ and PDR5-lacZ, contained the lacZ gene fused with the corresponding promoters of PDR3 and PDR5 (Delahodde et al. 1995). The SNQ2-lacZ fusion was constructed by inserting a PCR fragment of SNQ2 extending from )1416 bp to +181 bp (relative to the initiator ATG) in the BamHI-HindIII sites of YEp367. Yeast cells were grown on glucose-rich medium or on a minimal medium
Fig. 1 Schematic representation of the mutagenized DNA fragments of PDR3 used in the gap repair approach. DNA fragments (1708-bp HpaI-BstXI and 1793-bp PvuII-HpaI) puri®ed from the mutagenized plasmid pFL38-PP3 were used in two experiments (gap repair 1 and gap repair 2, respectively). The structural motifs composing Pdr3p have been identi®ed either by functional analyses (C-terminal activation domain; Delaveau et al. 1994) or by sequence comparisons within the zinc cluster family of transcriptional regulators. MI, MII, MIV, MV, MVI, stand for the motifs recently identi®ed (Poch 1997), ID1 corresponds to the inhibitory domain identi®ed in Gal4p (Stone and Sadowski 1993), MHR is the region recently described by Schjerling and Holmberg (1996). PDRE indicates the location of two pleiotropic drug response elements recognized by Pdr3p (Delahodde et al. 1995) containing 0.67% yeast nitrogen base without amino acids, 2% agar, and 2% glucose or 2% glycerol plus 2% ethanol. The appropriate nutritional requirements and drugs were added at the indicated concentrations. The drug resistance tests were performed with a drop replicator (Delaveau et al. 1994). Hydroxylamine mutagenesis of plasmid DNA and selection of the mutants Mutations were introduced into PDR3 by hydroxylamine-induced in vitro mutagenesis (Rose et al. 1990). Brie¯y, 10 lg of puri®ed pFL38-PP3 was treated with 1 M hydroxylamine adjusted to pH 7.0 with NaOH. After incubation at 37° C for 24 h the reaction was stopped by adding 10 ll of 5 M NaCl and 50 ll of 1 lg/ml bovine serum albumin. After precipitation, the DNA was resuspended in 100 ll of 10 mM TRIS-Cl pH 7.0, 1 mM disodium EDTA pH 8.0 and precipitated once more with ethanol. The mutagenized plasmid DNA (9314 bp) was cleaved with HpaI plus BstXI or PvuII plus HpaI. The 1708-bp HpaI-BstXI and the 1793bp PvuII-HpaI DNA fragments containing 5¢-terminal parts of the PDR3 gene with and without the promoter, respectively, were isolated. These DNA fragments, together with the corresponding gapped plasmid DNA isolated from the wild-type pFL38-PP3 after digestion with SacII and SpeI (8396 bp, gap repair 1) and SpeI plus BstXI (8315 bp, gap repair 2) (Fig. 1), were used for co-transformation of the strain W303-1A/TD (Delaveau et al. 1994). A total of about 10 000 individual yeast transformants was replica-plated onto minimal glucose media containing 0.6 lg/ml and 1.0 lg/ml cycloheximide. Transformants growing on both plates were picked and used to determine the resistance phenotypes to other unrelated drugs. In vitro directed mutagenesis Directed mutagenesis was performed according to Deng and Nickolo (1992) using a Transformer Site-Directed Mutagenesis kit from Clontech. This method works by simultaneously annealing two oligonucleotide primers to one strand of denatured doublestranded plasmid (pFL38-PP3). One primer introduces the desired mutation (pdr3-8K257E or pdr3-13E236K) and the other primer
399 changes the unique restriction site StuI to AvrII. After DNA elongation, ligation, and digestion by StuI, the DNA was used to transform the mutS E. coli strain de®cient in mismatch repair. All the clones which had a plasmid cleavable bv AvrII and not by StuI contained the expected base changes. Molecular genetic techniques DNA preparation, restriction analysis, and cloning in E. coli were carried out as described in Sambrook et al. (1989). Isolation of plasmid DNA from yeast cells was done by the method described in Rose et al. (1990). Yeast transformation was carried out using the modi®ed lithium acetate protocol of Gietz and Schiestl (1991). The DNA sequence of the pdr3 mutant was determined by the dideoxynucleotide chain-termination technique using double-stranded plasmid DNA puri®ed using Qiagen Plasmid kits and a set of synthetic 19mer oligonucleotide primers (corresponding to the coding sequence) distributed at intervals of about 250±300 bp between the two HpaI restriction sites. b-Galactosidase assays b-Galactosidase activity was determined by using o-nitrophenylbeta-D-galactopyranoside as the substrate (Miller 1972; Rose et al. 1990) on crude extracts of transformants grown to mid-log phase. The activity was normalized with protein concentrations assayed by the method of Bradford (1976). Each value reported is the average of determinations from three or four independent transformants. RNA isolation, radiolabelling, and northern blot analyses Total yeast RNA was isolated using a published procedure (Schmitt et al. 1990) and fractionated by electrophoresis through a 1% agarose gel containing 6% formaldehyde. RNA was transferred to nylon membranes by capillary blotting, and hybridization of the membranes was carried out using standard methods (Sambrook et al. 1989). DNA fragments were radiolabelled using a Nonaprime Labelling kit under conditions recommended by the manufacturer (Appligene). Autoradiographs of northern blots were quanti®ed with a phosphorimager (Fuji).
Results Mutagenesis of the PDR3 gene and selection of mutants showing a drug resistance phenotype The two pdr3 mutants described so far were identi®ed in two unrelated yeast strains, 2D and DRI9-T7 (Subik et al. 1977, 1986). Since their corresponding wild-type strains, D225-5A and DRI9, are not isogenic, the molecular analysis of the pdr3-1 and pdr3-2 mutations would have been dicult to interpret due to strain-speci®c polymorphisms. Therefore, in order to further investigate the molecular mechanisms of the Pdr1p/ Pdr3p-mediated drug resistance phenomenon, we decided to construct a set of isogenic pdr3 mutants by in vitro mutagenesis of a cloned PDR3 gene. Hydroxylamine mutagenesis was used to introduce mutations into the plasmid-borne PDR3 gene. Two mutagenized N-terminal DNA fragments, one overlapping the promoter part of the PDR3 gene (gap repair 1, Fig. 1) and the other included in the PDR3 ORF (gap repair 2, Fig. 1), were co-transformed with a linearized plasmid containing the corresponding gapped wild-type PDR3 gene. Selected transformants were replica-plated onto two cycloheximide-containing plates (0.6 and 1.0 lg/ml). Resistant clones were picked and their crossresistance to chloramphenicol (2 mg/ml), mucidin (0.1 lg/ml), and cycloheximide (1 lg/ml) was determined. Mutants with a multidrug resistance phenotype were selected and their plasmid DNA obtained by transformation of E. coli was retransformed into W3031A/TD to con®rm their drug resistance phenotype. This procedure eliminated all chromosomal mutants in PDR1 or in other drug resistance loci and allowed us to obtain
Table 1 Sequence changes in pdr3 mutants and multiple drug resistance phenotype of W303-1A/TD transformants Mutanta
Originb
Base changec
Amino acid substitutionc
Minimal inhibitory concentration (strain W303/TD)d CYH
SMM
4-NQO
CMP
MUC
OLI
pdr3-4 pdr3-5 pdr3-6 pdr3-7 pdr3-8 pdr3-9 pdr3-10
Random Random Random Random Directed Random Random
C356T A615T A756G G758A A769G T826C G839A
T119I K205N I252M G253E K257E Y276H R280K
1.5 1.5 1.5 1.5 1.5 1.5 1.5
>8 8 >8 >8 >8 >8 >8
1.0 0.5 0.75 0.75 0.75 1.0 1.0
>4 4 >4 >4 >4 >4 >4
0.2 0.1 0.2 0.2 0.2 0.2 0.2
0.75 0.25 0.75 0.75 0.5 0.75 0.75
pdr3-11 pdr3-12 pdr3-13 pdr3-14 Wild-type PDR3
Random Random Directed Directed ±
G706A, A769G T578C, G839A G706A T578C ±
E236K, K257E I193T, R280K E236K I193T ±
1.5 1.5 0.6 0.6 0.6
>8 >8 8 8 8
0.75 0.75 0.5 0.5 0.5
>4 >4 2 2 2
0.2 0.2 0.1 0.1 0.1
0.5 0.75 0.25 0.25 0.25
a
The nomenclature follows on the description of pdr3-1 and pdr3-2 by Subik et al. (1977, 1986) and of pdr3-3 by Katzmann et al. (1994) The mutations were obtained after hydroxylamine mutagenesis (random) or after oligonucleotide-directed mutagenesis (directed) c Nucleotides and amino acids are numbered from the ATG initiation codon d Cells were grown on minimal glucose or glycerol plus ethanol medium. The drug concentrations tested were as follows. Cycloheximide (CYH): 0.6, 1.0, 1.5 and 2 lg/ml; sulfomethuron methyl (SMM): 5.0 and 8.0 lg/ml; 4-nitroquinoline oxide (4-NQO): 0.25, 0.5, 0.75 and 1.0 lg/ml; chloramphenicol (CMP): 2.0 and 4.0 mg/ml; mucidin (MUC): 0.05, 0.1 and 0.2 lg/ml; and oligomycin (OLI): 0.25, 0.5, 0.75 and 1.0 lg/ml. Growth was scored after 5 days, except in the case of SMM, where cells were scored after 3 days b
400
only the mutant alleles of PDR3. From about 104 primary Ura+ transformants, ten independent clones were isolated and further studied. Sequence analysis of pdr3 mutant alleles The pdr3-carrying plasmids corresponding to the ten independent mutants were sequenced in the region corresponding to the mutated DNA fragments introduced by the gap repair approach. The sequence and location of each mutation in the dierent alleles of the PDR3 gene are shown in Table 1 and Fig. 2. As expected, most of the mutations resulted from transitions. One transversion was found in the mutant pdr3-5K205N. Eight out of ten mutants had a single base pair mutation leading to one amino acid substitution. One base change was found twice (corresponding to pdr3-9Y276H) and another base change was found three times (corresponding to pdr310R280K). Two mutants, pdr3-11E236K, K257E and pdr312I193T, R280K, have two base changes. Each of these four mutations was introduced into PDR3 to assess their properties individually. pdr3-8K257E and pdr3-10R280K were found to be responsible for the cycloheximide resistance phenotype (Table 1). Finally, seven dierent single base changes were found to be responsible for the drug resistance phenotype, numbered pdr3-4T119I to pdr3-10R280K (Table 1). Although 333 bp of the pro-
Fig. 2 Localization of gain-of-function mutations relative to the structural motifs of Pdr3p. The mutations characterized in this work are presented by the abbreviated names of the pdr3 alleles (see Table 1). Five mutations in PDR3 were found to be localized within the motifs MI or MII. Pdr1p and Pdr3p sequences (lower lines) are represented in the region of motifs MI and MII (Poch 1997). Identical amino acids are boxed
moter and 1974 bp of the coding region of the PDR3 gene (2307 bp) were mutagenized, all the pdr3 mutations were recovered in a 483-bp DNA fragment encoding the N-terminal moiety of PDR3, upstream of the ®rst BstXI restriction site (Fig. 1). No mutations were found in the promoter region which could have led to overexpression of the gene. Dierential drug resistance patterns of isogenic pdr3 mutants The isogenic set of independent pdr3 mutants was characterized phenotypically. Independent transformants isolated in retransformation experiments done with each individual plasmid were used to determine susceptibility to cycloheximide, sulfomethuron methyl, 4-nitroquinoline oxide, chloramphenicol, and oligomycin. It was observed that the pdr3 alleles conferred dierent drug resistance spectra and variable levels of resistance to individual drugs (Table 1). Most of the mutants
401 Table 2 Multidrug resistance phenotype of pdr3 mutant alleles in strains FY1679-28C/TD and FY1679-28C/TDEC
Mutant
Minimal inhibitory concentration (lg/ml)a FY1679-28C/TD
pdr3-4 pdr3-5 pdr3-6 pdr3-7 pdr3-8 pdr3-9 pdr3-10 Wild-type PDR3
FY1679-28C/TDEC
CYH
MUC
OLI
SMM
CYH
MUC
OLI
SMM
1.5 1.5 1.5 1.5 1.5 2.0 1.5 0.4
0.2 0.1 0.2 0.2 0.2 0.2 0.2 0.1
1.0 0.5 1.0 1.0 0.8 1.0 1.0 0.3
>16 10 >16 >16 >16 >16 >16 10
2.0 1.5 2.0 2.0 2.0 2.0 2.0 0.2
0.4 0.1 0.2 0.4 0.2 0.4 0.4 0.1
1.0 0.8 >1.0 >1.0 >1.0 >1.0 >1.0 0.2
>16 6 >16 >16 >16 >16 >16 6
a
Growth on cycloheximide (CYH) and mucidin (MUC) was scored after 12 days, growth on oligomycin (OLI) and sulfomethuron methyl (SMM) after 3 days
conferred resistance to all six drugs. Only pdr3-5K205N exhibited a resistance spectrum that was limited to cycloheximide and chloramphenicol. Considering the minimal inhibitory concentrations (MIC) of the drugs, it is clear (Table 1) that the various pdr3 mutants diered in their ability to confer resistance to individual drugs. In this respect, the least eective allele was pdr3-5K205N and the most ecient were pdr3-5T119I, pdr3-7G253E, and pdr3-9Y276H. After 5 days, transformants harboring pdr3-7G253E tolerated 0.2 lg/ml of mucidin and after 12 days they also grew at an oligomycin concentration of 1 lg/ml. The same level of oligomycin resistance was observed with pdr3-4T119I. pdr3 mutant alleles are semidominant and confer drug resistance in the absence of the homologous PDR1 gene Wild-type strain FY1679-28C and its derivative with a disrupted pdr3 gene, strain FY1679-28C/TD, were transformed with the centromeric plasmid encoding wild-type PDR3 or its mutant alleles. Six independent transformants for each host strain and each individual plasmid were used to determine drug resistance. All the pdr3 mutants (except pdr3-5K205N) conferred resistance to cycloheximide (1.0 lg/ml) or to mucidin (0.1 lg/ml) independently, in the presence or absence of a chromosomal copy of PDR3 (data not shown). In this respect, pdr3 mutants can be considered as semidominant, a property already observed for the chromosomal mutations both in PDR3 and PDR1 (Balzi et al. 1987; Ruttkay-Nedecky et al. 1992; Subik et al. 1986). The results were dierent when we analyzed the effects of the presence or absence of PDR1 (Table 2). When the host strain was deleted for both PDR1 and PDR3 (FY1679-28C/TDEC, Table 2), most of the transformants were signi®cantly more resistant to cycloheximide and mucidin or oligomycin than were the corresponding transformants in the strain harboring PDR1.
Increased activation of PDR3, PDR5, and SNQ2 promoters by mutant Pdr3p Pdr3p is an interesting transcription factor which can promote the transcription of genes coding for: (1) either ABC transporters, such as Pdr5p, Snq2p or Yor1p (Decottignies et al. 1995; Katzmann et al. 1994, 1995; Mahe et al. 1996), or MFS proteins such as Hxt9p or Hxt11p (Nourani et al. 1997), as well as (2) inducing its own expression through an important autoregulatory mechanism (Delahodde et al. 1995). Therefore we tested the ability of the mutant proteins speci®ed by dierent pdr3 alleles to activate the expression of the PDR3, PDR5, and SNQ2 promoters fused to the lacZ reporter gene. Plasmids carrying a wild-type PDR3-lacZ, PDR5lacZ or SNQ2-lacZ gene fusion were introduced into the FY1679-28C/TD host strain together with the various pdr3 mutant alleles described above and PDR3- PDR5and SNQ2-dependent b-galactosidase activities were determined. All the pdr3 mutants tested led to increased expression of the fusion genes in comparison with the wild-type PDR3 (Table 3). The b-galactosidase activities driven by Table 3 Eects of pdr3 mutations on expression of b-galactosidase driven by various promoters in the strain FY1679-28C/TD pdr3 allele
pdr3-4 pdr3-5 pdr3-7 pdr3-9 pdr3-10 pdr3-11 Wild-type PDR3
b-Galactosidase activity (nmol/min per mg protein) (strain: FY1679-28C/TD)a PDR3-lacZ
PDR5-lacZ
SNQ2-lacZ
9.1 18.6 11.4 17.7 14.6 11.8 4.6
24761 444.5 1307.2 236.1 2877.4 249.2 3630.1 819.3 3094.7 376.3 2263.3 665.4 1654.1 191.3
4.5 4.3 3.9 6.5 5.2 3.1 3.3
2.1 3.1 2.6 4.9 2.1 2.2 1.1
0.4 0.2 0.7 0.8 0.9 0.1 0.7
a The three PDRE-containing promoters PDR3, PDR5 and SNQ2 were fused to the lacZ gene. b-Galactosidase activity was determined as described by Miller (1972). Each value is the average (S.D.) for three or four independent transformants
402
the PDR3 promoter were stimulated 3- to 4-fold and those driven by PDR5 or SNQ2 promoters only 2-fold. The pdr3-9Y276H allele was found to be the most ecient with all three promoters. The expression of b-galactosidase involves transcription and translation. To correlate more precisely the conferred drug resistance with the increased activation function of various pdr3 mutant alleles, northern blot analyses were carried out. As shown in Fig. 3, all the pdr3 mutations increased the activity of the Pdr3p transcription factor. Bearing in mind that several promoters, including those of PDR5, SNQ2, and PDR3, can be activated by Pdr1p and/or Pdr3p, it was interesting to study the eect of the presence or absence of Pdr1p on the properties of the Pdr3p mutants. For this purpose, the degree of overproduction of PDR5/SNQ2/PDR3 transcripts was assessed in the presence of the dierent pdr3 alleles and in two strains diering by the presence or absence of PDR1 (Fig. 3A, B and Fig. 3C, D, respectively). The results are presented as the ratio between the amount of the particular transcript obtained with mutated pdr3 alleles and the corresponding value in the
Fig. 3 Northern blot analyses of total RNA from strains deleted either for PDR3 (FY1679-28C/TD, A) or for PDR1 and PDR3 (FY1679-28C/TDEC, C) and expressing representative pdr3 mutants. Quantitative steady state levels of the dierent RNA species were assessed with phosphorimager analyses and their values relative to the wild-type level are represented (B, D)
presence of the wild-type PDR3 allele. This induction factor was clearly highly dependent on the presence of the PDR1 allele (Fig. 3B, D). For all the mutated pdr3 alleles, the induction factor was much higher in the absence of PDR1 (Fig. 3C, D) than in its presence (Fig. 3A, B). Also, when the dierent pdr3 alleles are considered, it can be seen that, while the induction factors for PDR3 and SNQ2 were dependent on the presence of PDR1, the induction factor for PDR5 was not signi®cantly aected when PDR1 was deleted. These interesting promoter-speci®c eects merit further consideration (see Discussion). We also checked the eects of pdr3 double mutants (pdr3-11E236K, K257E and pdr3-12I193T, R280K) and their respective single mutants (pdr3-8K257E, pdr3-10R280K,
403
pdr3-13E236K, and pdr3-14I193T) on steady-state levels of PDR5, SNQ2, and PDR3 mRNA. The results shown in Fig. 4 demonstrate that only the single mutants pdr3-8K257E and pdr3-10R280K led to an increase in transcript levels of the dierent Pdr3p target genes. This observation is in agreement with the MIC results shown in Table 1 and con®rms that the single mutations pdr3-8K257E and pdr3-10R280K are responsible for the double mutant phenotype (pdr3-11E236K, K257E and pdr3-12I193T, R280K).
Discussion The results presented here unexpectedly showed that random mutagenesis of two-thirds (660 of 976 amino acids) of the sequence coding for the transcription factor Pdr3p led to seven single gain-of-function mutations, ®ve of which are localized in a short fragment of 29 amino acids. The hydroxylamine-based mutagenesis approach used in this work generated ten independent mutants. Interestingly, some amino acid changes such as
Fig. 4 Northern blot analyses of total RNA from strains deleted either for PDR3 (FY1679-28C/TD, A) or for PDR1 and PDR3 (FY1679-28C/TDEC, C) and expressing two random double mutants (pdr3-11E236K, K257E and pdr3-12I193T, R280K) and their corresponding single mutant (pdr3-13E236K, pdr3-8K257E, pdr3-14I193T, and pdr3-10R280K. Quantitative steady state levels of the dierent RNA species were assessed with phosphorimager analyses and their values relative to the wild-type level are represented (B, D)
Y276H or R280K were found several times in totally independent experiments (gap repair 1 or 2, Fig. 1), suggesting that we were actually approaching the saturation limits of the mutagenesis. For the two double mutants, pdr3-11E236K, K257E and pdr3-12I193T, R280K, we engineered the four corresponding single mutations. In each case, only one single point mutation, K257E (pdr3-8K257E) and R280K (pdr3-10R280K), respectively, turned out to be responsible for the drug resistance phenotype (Table 1, Fig. 4). This makes a total of seven point mutations which confer high levels of drug resistance by simply changing one amino acid in the sequence of Pdr3p.
404
Phenotypes of the dierent PDR3 alleles on the expression of PDR5 and SNQ2 All mutants, selected on the basis of their ability to confer a higher level of resistance to cycloheximide, also made the cell more resistant to ®ve other unrelated drugs (Table 1). That the regulation of the pleiotropic drug resistance phenomenon was indeed altered in the pdr3 mutants was con®rmed by analysis of the transcription of the two ABC transporter-encoding genes, PDR5 and SNQ2 (Table 3, Fig. 3). Interestingly, expression of PDR5 was much more aected by the mutants than that of SNQ2 (Fig. 3B). Two non-exclusive reasons might explain these promoter-speci®c eects: 1. Mutants were selected that permit growth on high levels of cycloheximide, a drug which might be more speci®cally recognized by Pdr5p (Balzi et al. 1994; Katzmann et al. 1994). One might thus suspect a relationship between the properties of the mutants and the target drug; this would imply that selection with another drug would have revealed another set of mutations. 2. The two PDR5 and SNQ2 promoters possess different combinations of Pdr1p/Pdr3p-response elements (Delahodde et al. 1995; Katzmann et al. 1996; Mahe et al. 1996) which might induce dierent sensitivities to either Pdr1p or Pdr3p. This is supported by the dierences observed in the transcription of PDR5, SNQ2, and PDR3 as a function of the presence or absence of PDR1 (Fig. 3). These dierences are especially important for the autoregulatory properties of the PDR3 promoter. In the presence of PDR1 (Fig. 3B), the induction factor associated with the pdr3 mutants was considerably reduced, an observation which is in agreement with the reduced ability of the pdr3 mutants to confer resistance to cycloheximide or mucidin when PDR1 was present (Table 2). Competition between Pdr1p and Pdr3p for DNA binding and on Pdr1p/Pdr3p heteromeric associations might explain these eects. These points clearly merit further investigation. The clustered pdr3 mutations and the motifs MI±MII Apart from pdr3-4T119I and pdr3-5K205N, all the pdr3 alleles recovered here are clustered in a small protein fragment. Mutations pdr3-6I252M to pdr3-10R280K are localized in a 29-amino acid fragment totally included in the conserved motifs MI±MII (see Fig. 2) recently described by Poch (1997). A systematic sequence analysis of the GAL4 family members revealed that eight motifs, MI±MVIII, are conserved between these dierent transcription factors. Based on deletion experiments conducted with Gal4p, these motifs have been proposed to act as inhibitory domains (Stone and Sadowski 1993). Our studies support and extend these observations: (1) in Pdr3p, the motifs MI±MII correspond to a functional domain since most of the gainof-function mutations were located in this region; (2) the motifs MI±MII are likely to act as inhibitory
domains since alterations of their wild-type sequences enhance the activity of Pdr3p; and (3) our genetic analysis, which pinpointed only motifs MI and MII, suggests that the other conserved motifs, MIII±MVIII, have dierent functions, an aspect which is reminiscent of the fact that homologous central regions were found to also contain latent activation domains, as observed for Adr1p (Cook et al. 1994). The predominant role of the motifs MI±MII in these Zn2Cys6-type members is also supported by the description of similar mutations in Pdr1p (E. Carvajal et al., in preparation) and in Put3p where a mutation in position 15 of MII led to a non-inducible allele (Marczak and Brandiss 1991). How do the motifs MI±MII control the properties of the transcription factor? The ®ve gain-of-function mutations, pdr3-6I252M to pdr3-10R280K, in motifs MI±MII prove that the integrity of this short region is necessary to keep the transcription factor in a less active form. Several parameters, including protein level, protein modi®cation, and nuclear localization, are known to be involved in the regulation of transcription factors (Struhl 1995) and might explain the phenotypes of the dierent mutants. Hypotheses such as an increased protein stability conferred by the mutations cannot be excluded. However, the special characteristics of motifs MI and MII, mainly composed of a-helices with a 3±4 periodicity of the conservatively maintained hydrophobic residues (Poch 1997), suggest that other hypotheses are worth considering. It was proposed (Poch 1997; Schjerling and Holmberg 1996) that such internal motifs might control the DNA-binding properties of the transcription factor, interact with an inhibitory protein or control a locked conformation of the factor by interacting with another distal domain. Our data cannot discriminate between these non-exclusive hypotheses. Only in the case of an internal locked protein conformation could one suggest that the interacting regions might correspond to motif MI and motif MII or, most probably, to motifs MI± MII and a C-terminal region (after position 660), which was not analyzed in this study. Such intramolecular interactions involving the central and C-terminal parts of the protein have been proposed for Leu3p (Sze et al. 1992) and intramolecular inhibition is well documented in the case of the Ets-1 factor (Donadson et al. 1996). Recent studies on Gal4p (Ding and Johnston 1997) demonstrated that the central region of Zn2Cys6 transcription factors might act as a spacer (between DNA binding and C-terminal activation domains) and/or a surface for intramolecular or Gal4p±Gal4p interactions. In the case of Pdr3p, the mutations described in this work tend to favor the second possibility and Pdr3p± Pdr3p and/or Pdr3p±Pdr1p interactions should be studied by taking advantage of these dierent mutated forms.
405 Acknowledgements Thanks are due to A. Goeau and E. Carvajal for communicating results before publication. This work was supported by grants from CNRS (ACCSV6-9506049), from ARC (1358), and from VEGA (1/4001/97). The research of J. S. was also supported by an International Research Scholars Grant from the Howard Hughes Medical Institute. A. N. was supported by a grant from ARC (France).
References Balzi E, Chen W, Ulaszewski S, Capieaux E, Goeau A (1987) The multidrug resistance gene PDR1 from Saccharomyces cerevisiae. J Biol Chem 262:16871±16879 Balzi E, Wang M, Leterme S, Van Dyck L, Goeau A (1994) PDR5, a novel yeast multidrug resistance conferring transporter controlled by the transcription regulator PDR1. J Biol Chem 269:2206±2214 Cook WJ, Chase D, Audino DC, Denis CL (1994) Dissection of the ADR1 protein reveals multiple, functionally redundant activation domains interspersed with inhibitory regions: evidence for a repressor binding to the ADR1c region. Mol Cell Biol 14:629± 640 Decottignies A, Lambert L, Catty P, Degand H, Epping EA, Moye-Rowley WS, Balzi E, Goeau A (1995) Identi®cation and characterization of SNQ2, a new multidrug ATP binding cassette transporter of the yeast plasma membrane. J Biol Chem 270:18150±18157 Delahodde A, Delaveau T, Jacq C (1995) Positive autoregulation of the yeast transcription factor Pdr3p, which is involved in control of drug resistance. Mol Cell Biol 15:4043±4051 Delaveau T, Delahodde A, Carvajal E, Subik J, Jacq C (1994) PDR3, a new yeast regulatory gene, is homologous to PDR1 and controls the multidrug resistance phenomenon. Mol Gen Genet 244:501±511 Deng WP, Nickolo JA (1992) Site-directed mutagenesis of virtually any plasmid by eliminating a unique site. Anal Biochem 200:81±84 Ding WV, Johnston SA (1997) The DNA binding and activation domains of Gal4p are sucient for conveying its regulatory signals. Mol Cell Biol 17:2538±2549 Donadson LW, Petersen JM, Graves BJ, McIntosh LP (1996) Solution structure of the ETS domain from murine Ets-1: a winged helix-turn-helix DNA binding motif. EMBO J 15:125±134 Hellauer K, Rochon MH, Turcotte B (1996) A novel DNA binding motif for yeast zinc cluster proteins: the Leu3p and Pdr3p transcriptional activators recognize everted repeats. Mol Cell Biol 16:6096±6102 Katzmann DJ, Burnett PE, Golin J, Mahe Y, Moye-Rowley WS (1994) Transcriptional control of the yeast PDR5 gene by the PDR3 gene product. Mol Cell Biol 14:4653±4661 Katzmann DJ, Hallstrom TC, Voet M, Wysock W, Golin J, Volckaert G, Moye-Rowley WS (1995) Expression of an ATPbinding cassette transporter-encoding gene (YOR1) is required for oligomycin resistance in Saccharomyces cerevisiae. Mol Cell Biol 15:6875±6883
Katzmann DJ, Hallstrom TC, Mahe Y, Moye-Rowley WS (1996) Multiple Pdr1p/Pdr3p binding sites are essential for normal expression of the ATP binding cassette transporter protein-encoding gene PDR5. J Biol Chem 271:23049±23054 Mahe Y, Parle-McDermott A, Nourani A, Delahodde A, Lamprecht A, Kuchler K (1996) The ATP-binding cassette multidrug transporter Snq2 of Saccharomyces cerevisiae: a novel target for the transcription factors Pdr1 and Pdr3. Mol Microbiol 20:109± 117 Marczak JE, Brandiss MC (1991) Analysis of constitutive and noninducible mutations of the PUT3 transcriptional activator. Mol Cell Biol 11:2606±2619 Martens JA, Genereaux J, Saleh A, Brandl J (1996) Transcriptional activation by yeast Pdr1p is inhibited by its association with Ngg1p/Ada3p. J Biol Chem 271:15884±15890 Miller JH (1972) Experiments in molecular genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp 352± 355 Nourani A, Wesolowski-Louvel M, Delaveau T, Jacq C, Delahodde A (1997) Multiple-drug-resistance phenomenon in the yeast Saccharomyces cerevisiae: involvement of two hexose transporters. Mol Cell Biol 17:5453±5460 Poch O (1997) Conservation of a putative inhibitory domain in the GAL4 family members. Gene 184:229±235 Rose MD, Winston F, Hieter P (1990) Methods in yeast genetics. A laboratory course manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Ruttkay-Nedecky B, Obernauerova M, Subik J (1992) High level resistance to cycloheximide resulting from an interaction of the mutated pdr3 and cyh genes in yeast. Curr Genet 22:237±239 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Schjerling P, Holmberg S (1996) Comparative amino acid sequence analysis of the C6 zinc cluster family of transcriptional regulators. Nucleic Acids Res 24:4599±4607 Schmitt ME, Brown TA, Trumpower BL (1990) A rapid and simple method for preparation of RNA from Saccharomyces cerevisiae. Nucleic Acids Res 19:3091±3092 Stone G, Sadowski I (1993) GAL4 is regulated by a glucose-responsive functional domain. EMBO J 12:1375±1385 Struhl K (1995) Yeast transcriptional regulatory mechanisms. Annu Rev Genet 29:651±674 Subik J, Kovacova V, Takacsova G (1977) Mucidin resistance in yeast. Isolation, characterization and genetic analysis of nuclear and mitochondrial mucidin-resistant mutants of Saccharomyces cerevisiae. Eur J Biochem 73:275±286 Subik J, Ulaszewski S, Goeau A (1986) Genetic mapping of nuclear mucidin resistance mutations in Saccharomyces cerevisiae. A new pdr locus on chromosome II. Curr Genet 10:665±670 Sze JY, Woonter M, Jaehning JA (1992) In vitro transcriptional activation by a metabolic intermediate: activation by Leu3p depends on alpha-isopropylmalate. Science 258:1143±1145 [erratum (1993) 262:492]
