OUP user menu

Vmr 1p is a novel vacuolar multidrug resistance ABC transporter in Saccharomyces cerevisiae

Donata Wawrzycka, Iwona Sobczak, Grzegorz Bartosz, Tomasz Bocer, Stanisław Ułaszewski, André Goffeau
DOI: http://dx.doi.org/10.1111/j.1567-1364.2010.00673.x 828-838 First published online: 1 November 2010

Abstract

The Saccharomyces cerevisiae Yhl035p/Vmr1p is an ABC transporter of the MRP subfamily that is conserved in all post Whole Genome Duplication species. The deletion of the YHL035 gene caused growth sensitivity to several amphiphilic drugs such as cycloheximide, 2,4-dichlorophenoxyacetic acid, 2,4-dinitrophenol as well as to cadmium and other toxic metals. Vmr1p-GFP was located in the vacuolar membrane. The ATP-dependent transport of a DNP-S-glutathione conjugate was reduced in a vesicular fraction from the VMR1 deletant. The energy-dependent efflux of rhodamine 6G was increased by VMR1 deletion. Growth sensitivity to cadmium of the VMR1-deleted strain was more pronounced in glycerol/ethanol than in glucose-grown cells. The VMR1 promoter had higher activity when grown in glycerol/ethanol compared with glucose. In glucose, the VMR1 promoter was activated by the deletion of the glucose-dependent repressor ADR1.

Keywords
  • ABC
  • VMR1
  • vacuolar membrane
  • cadmium
  • multidrug resistance
  • ADR1

Introduction

In tumor cell lines, multidrug resistance may result from overexpression of either the 190-kDa multidrug resistance protein (MDR) or the 170-kDa multidrug resistance-associated protein (MRP). Mammalian MRPs (Cole., 1992; Zaman., 1993) are classified within the ABCC subfamily (Dean, 2002). They export a broad range of amphiphilic substrates, which are similar, but not identical to the MDR's substrates (Borst., 2000). These substrates are typically conjugated to S-glutathion, glucuronate or -sulfate. Analyses of eukaryotic genome sequences have unraveled many phylogenetic homologs of the human MRP1 in mammalian, plant, and yeast species (Decottignies & Goffeau, 1997; Taglicht & Michaelis, 1998; Klein., 1999; Sanchez-Fernandez., 2001; Martinoia., 2002). Analyses of the yeast genome have identified six orthologs of the human MRP1 (Decottignies & Goffeau, 1997; Taglicht & Michaelis, 1998). They all contain an N-terminal cytoplasmic extension (NTE) of about 300 amino acid residues that is predicted to contain up to five transmembrane spans (Hipfner., 1997; Bakos., 1998, 2000; Mason & Michaelis, 2002).

So far, five out of the six yeast MRP proteins have been characterized (Paumi., 2009). The Yeast Cadmium Factor, Ycf1p, transports cadmium-GS conjugates into the vacuole (Li., 1996; Wemmie & Moye-Rowley, 1997). The Bilirubin Pigment Transporter, Bpt1p, also contributes to the transport of cadmium-GS conjugates into vacuoles (Petrovic., 2000; Klein., 2002; Sharma., 2002). The Bile Acids Transporter, Ybt1p, formerly called Bat1p, mediates ATP-dependent transport of bile acids across vacuolar membranes (Ortiz., 1997). The two adjacent ORFs, YKR103w and YKR104w, of the genome-sequenced strain are fused in most Saccharomyces strains and named NFT1 (Mason., 2003). The yeast ORF YHL035c is the only one that has not been characterized as yet. Here, we show that Yhl35cp, which forms a phylogenetic cluster with the yeast vacuolar transporter Ybt1p, is located in the vacuolar membrane. It is involved in multiple drug resistance as well as in metal sensitivity. It is preferentially expressed under respiratory conditions. The corresponding gene has been named VMR1 (vacuolar multidrug resistance).

Materials and methods

Strains and plasmids

The single or doubly vmr1Δ- and ycf1Δ-deleted strains DWFY10 (vmr1Δ∷loxPkanMXloxP), DWFY6 (ycf1Δ∷loxPkanMXloxP) and DWFY10,6 (vmr1Δ∷loxP; ycf1Δ∷loxPkanMXloxP) were derived from FY1679-28c (MATa, ura3-52, trp1Δ63, leu2Δ1, his3Δ200, GAL2) (Winston., 1995). The strain AD1-10 was obtained by the deletion of VMR1 in the multideleted strain AD1-9 (Δyor1hisG, Δsnq2hisG, pdr5Δ2hisG, Δpdr10hisG, Δpdr11hisG, Δycf1hisG, pdr3Δ2hisG, Δpdr15hisG, pdr1Δ3hisG) (Rogers., 2001) derived from US50-18c (MATa, ura 3, his-1, PDR1-3) (Balzi., 1987). The strains W303-1A (MATa ade2-1, his3-11,15, leu2-13,112, trp1-1, ura3-1, can1-100) (Thomas & Rothstein, 1989), TYY303 (W303-1A, adr1Δ1LEU2) (Sloan., 1999), Wmsn2/4 (W303-1A, msn2-Δ3HIS3, msn4-Δ1TRP1) (Martinez-Pastor., 1996), BY4741 (Mat a; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0) (Brachmann., 1998), Y00249 (BY4741; YEL009ckanMX4) (Euroscarf collection), SEY6210 (MATα, leu2-3,112, ura3-52, his3−Δ200, trp1−Δ901, lys2-801, suc2-Δ9) and SM10 (SEY6210, yap1Δ∷HIS3) (Wemmie., 1994) were used in β-galactosidase assays.

Start-to-stop disruption of VMR1 and/or YCF1 was performed using a PCR amplified loxP-kanMX-loxP disruption cassette, flanked by short regions homologous to the target gene (Guldener., 1996).

The strain DWFY35GFP (FY1679-28c, VMR1-GFP) containing the chromosomal VMR1-GFP fusion was generated by genomic integration of the PCR cassette, consisting of the GFP-kanMX reporter/marker module flanked by 50 bp upstream and downstream the stop codon of VMR1. It was amplified using the pFA6a-GFPMT-kanMX6 plasmid (GFP, KanMX6) (Wach., 1997).

The pVMR1-GFP plasmid was constructed by in-frame C-terminal fusion of VMR1 and GFP as follows: the centromeric pUG35 plasmid bearing the yeast URA3, the bacterial AmpR, the yeast-enhanced green fluorescent GFP under the control of the MET25 promoter and the CYC1 terminator (H. Hegemann, Heinrich-Heine-Universität, Düsseldorf, Germany) were modified in its multicloning site by K. Flis (Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Warsaw, Poland) and named pKF47 (URA3, MCS, 6xHis-GFP). The VMR1 fragment comprising the 300 bp starting from ATG and 300 bp of the C-terminus of VMR1 without the stop codon was PCR amplified and cloned in the pKF47. The plasmid thus derived was linearized and introduced into the cells. The whole chromosomal VMR1 gene was rescued by homologous recombination and the pVMR1-GFP plasmid was verified by sequencing.

The promoter–reporter plasmid pVMR1-lacZ (CEN, URA3, PVMR1-lacZ) was constructed by in-frame fusion of the VMR1 promoter (from −540 to +15 bp from the ATG) to the lacZ gene in the pSEYC102 plasmid (CEN, URA3, lacZ) (Emr., 1986).

Media, growth conditions and drug sensitivity

Yeast strains were grown at 30 °C either in a rich YPD medium (1% yeast extract, 2% peptone, 2% glucose), a YPGE medium (1% yeast extract, 2% peptone, 2% glycerol, 2% ethanol), a synthetic medium SD (2% glucose, 0.67% yeast nitrogen base without amino acids) or SGE (2% glycerol, 2% ethanol, 0.67% yeast nitrogen base without amino acids) supplemented with appropriate amino acids.

Liquid test: cells grown overnight were used to inoculate 105 cells mL−1 into the fresh medium containing different concentrations of drugs or metal. Growth was measured for 60 h by cell counting.

Plate test: cells were cultured until OD600 nm of 1.0 was reached and sets of 10-fold dilution were spotted onto plates containing various drug concentrations. Growth was monitored for 3–5 days.

Halo test: performed on YPGE plates in the presence of an inhibitor, as described previously (Kolaczkowski., 1998). Each of the 64 inhibitors listed below is followed by the final concentration in mg mL−1 and the source (P, Pfizer; S, Sigma; B, Boehringer; J, Janssen; C, Calbiochem; CG, Ciba-Geigy; A, Aagrunol Groningen; CS, Chem Service; M, Merck; F, Fluka; AO, Acros Organics). The 64 inhibitors were acridine orange, 2.2, J; anisomycin, 50, P; antimycin, 25, S; benzamidine hydrochloride, 250, S; benomyl, 20.8, S; carbazole, 100, M; carbonyl cyanide m-chlorophenyl hydrazole (CCCP), 10, S; carbonyl cyanide p-(trifluoromethoxy) phenyl hydrazone (FCCP), 1.5, S; chloramphenicol, 500, S; chlorbromuron, 25, S; colchicine, 500, S; crystal violet, 1.5, S; cycloheximide, 10, B; daunorubicin hydrochloride, 10, S; deoxycorticosterone, 40, S; 1,4-diaminobutane, 100, J; 2,4-dichlorophenoxyacetic acid, 250, J; 2,4-dinitrophenol, 100, C; diuron, 13, CG; doxorubicin hydrochloride, 20, S; doxycycline, 100, S; erythromycine, 40, J; ethidium bromide, 250, B; ferbam, 5, A; fluconazole, 5, F; glycochenodeoxycholic acid sodium salt, 80, S; 8-hydroxyquinoline, 400, J; hygromycin B, 125, S; ioxynil, 100, CS; ketoconazole, 1, J; lactic acid, 150, M; malachite green, 4, F; menadion, 5, S; metobromuron, 30.3, CG; miconazole, 0.5, S; 4-nitrophenol, 500, M; 4-NQO, 2.5, S; 4-nitrotetrazoliumchloride blue, 4.5, S; nystatin, 100, S; oligomycin, 2, S; 1,10-phenantroline monohydrate, 70, S; phenol red, 100, S; piperidine, 870, M; polidocanol, 400, S; progesteron, 40, S; propanil, 100, CS; quinidine, 30, S; resasurin, 20, F; rhodamine B, 150, M; rhodamine 6G, 100, M; sodium azide, 250, F; tamoxifen, 20, S; taurocholic acid, 400, S; tetradecyl trimethyl ammonium bromide, 45, S; tetracycline, 100, S; thezit, 250, B; thiolutin, 80, B; thiram, 10, CS; trans dehydroandrosteron, 7, S; triton X100, 120, B; 2,3,5-triphenyltetrazolium chloride, 200, S; and vanilic acid, 500, F.

β-Galactosidase reporter assay

The strains were transformed with the promoter–reporter plasmids pVMR1-lacZ or pYCF1-lacZ (Wemmie., 1994) using pSEYC102 as a control. Yeast transformants were grown up to the midlog phase on SD or SGE media supplemented with appropriate inhibitors. The β-galactosidase activity of transformants was determined as described previously (Guarente, 1983). For measurement of the induction of gene expression in the presence of drug, the cells were grown overnight with 5 μM cadmium chloride, 0.2 mM zinc sulfate or 0.2 mM copper sulfate. For the last 2 h of incubation, the drug concentration was increased 10-fold. Measurements were repeated four times each for three independent transformants.

Rhodamine transport in intact cells

Rhodamine 6G fluorescence measurements were performed using the SLM Aminco 48000 S spectrofluorimeter. The uptake of rhodamine 6G was performed as described by van den Hazel. (1999) and the efflux in 10 mM glucose as described by Kolaczkowski. (1996, except that the concentration of rhodamine 6G was 70 μg mL−1. All transport assays were repeated three times.

Fluorescence microscopy

Cells expressing GFP-tagged Vmr1p were grown in a glucose or a glycerol medium lacking methionine to OD600 nm 0.7 and visualized for GFP fluorescence and Nomarski optics using a Zeiss microscope with a × 100 oil objective and photographed using an AxioCam MRc5 camera. To visualize the vacuolar membrane, cells were grown to an OD600 nm of 0.7, incubated with FM4-64 for 60–120 min at 30 °C, washed twice with a fresh SD medium and observed using the CY3 filter.

Protein fractionation and immunoblotting analysis

For immunoblotting, the proteins were extracted using the trichloroacetic acid method as described previously (Riezman., 1983). The total extract was solubilized in Laemmli buffer supplemented with 8 M urea and incubated at 37 °C for 30 min. For the protein fractionation: log-phase cultures (2 × 107 cells mL−1) were harvested by centrifugation. The cells were broken with glass beads in lysis buffer (20 mM Tris-HCl, 2 mM MgCl2, 250 mM sorbitol, pH 7.5) supplemented with protease inhibitors. Cell debris and glass beads were removed by centrifugation at 950 g for 5 min. The resulting whole-cell extract was centrifuged at 15 000 g for 40 min to yield a membrane-enriched pellet (C15/40), which was suspended in MS buffer (2 mM MgCl2; 10 mM imidazole). The resulting supernatant fraction was named S40. Proteins were resolved on 6% SDS-PAGE. The primary antibodies used were mouse anti-GFP (Roche) 1 : 5000 or rabbit anti-PMA1 1 : 15 000 and anti-Pgt1 1 : 2000 (GE Heathcare).

3H-2,4-dinitrophenyl-S-glutathione (3H-DNP-SG) transport

3H-DNP-SG was synthesized from 1-chloro-2,4-dinitrobenzene and 3H-glutathione (Akerboom., 1992). Vesicular fractions were isolated as described previously (Goffeau & Dufour, 1988; Zadzinski., 1996). ATP-dependent transport of 3H-DNP-SG into vesicles was studied using the filtration technique (Akerboom., 1992; Zadzinski., 1996).

Computational analysis

The amino acid sequences of Ycf1p (P39109), Bpt1p (P14772), Ybt1p (P32386), Vmr1p (P38735) Nft1p (P36028) and Yor1p (P53049) MRP proteins were scanned by HMMTOP server (Tusnády & Simon, 1998). The 300 amino acid NTE regions were aligned using clustalw (Thompson., 1994) and subjected to a phylogenetic analysis using the neighbor-joining method (Saitou & Nei, 1987). A phylogram was created by treeview (Page, 1996).

Results

Phylogenetic analysis

The members of the mammalian MRP subfamily possess an N-terminal extension of about 275 amino acid residues that usually comprises five predicted transmembrane spans (Hipfner., 1997; Bakos., 1998). The software hmmtop predicts one to five transmembrane spans in the NTE of the six yeast MRPs. The different NTEs, which are the most variable domains of the yeast MRPs, were aligned and subjected to phylogenetic analysis. Figure 1 shows that three phylogenetic clusters can be distinguished. Cluster 1 contains Vmr1p and Ybt1p. Cluster 2 contains Bpt1p, Ycf1p and Nft1p. Cluster 3 comprises only Yor1p. The phylogenetic relation between Vmr1p and Ybt1p is strong. Five transmembrane spans are predicted in both proteins and 180 out of the 300 NTE amino acid residues are similar. The protein sequence homology between Vmr1p and Ybt1p is confirmed by a comparison of the complete proteins that exhibit 64% identity and 79% similarity (not shown).

1

The computational analysis of the putative NTE region of yeast MRP proteins. The protein sequences comprising the 300 amino terminal residues have been subjected to a multiple sequence alignment by clustal w and the phylogenetic analysis using the neighbor-joining method. treeview was used to obtain the final printout of the resulting data.

Blast analyses (not shown) of the NTE and of the full protein sequence indicate that ‘ancestral’ forms of VMR1 are present in preWGD hemiascomycetous species such as Kluyveromyces lactis, Ashbya gossipi, Candida tropicalis and Debaryomyces hansenii and even in filamentous fungi such as Podospora anserina. Synteny analysis through the Yeast Gene Order Browser (Gordon., 2009) shows that the ancestral block containing the ortholog of VMR1 has been duplicated by WGD yielding in S. cerevisiae VMR1 on chromosome VIII and YBT1 on chromosome XII. Conservation of the duplicated VMR1/YBT1 paralogs (ohnologs) is observed in the post Whole Genome Duplication species such as Candida glabrata and Saccharomyces castelii.

Disruption of the VMR1 gene confers multidrug sensitivity

In the FY1679-28c background, a single disruptant vmr1Δ was screened for sensitivity against a panel of 64 toxic drugs using the inhibitory halo assay on a Petri dish (Kolaczkowski., 1998). After confirmation under liquid growth conditions, it was concluded that the deletion of VMR1 produces sensitivity to hygromycin B, cycloheximide, 2,4-dinitrophenoxyacetic acid, diuron, 1,10-phenantroline, piperidine, polidocanol, 2,3,5-triphenyltetrazolium chloride, 8-hydroxyquinoline, 4-nitroquinoline-N-oxide and 2,4-dinitrophenol (Table 1a). The Vmr1p-dependent multidrug sensitivity was confirmed in multideleted strains as follows: the AD1-9 strain deleted in the plasma membrane ABC transporters YOR1, SNQ2, PDR5, PDR10, PDR11, PDR15 as well as in the vacuolar transporter YCF1 and in the two transcription factors PDR1 and PDR3 (Rogers., 2001) was used as a control. The AD1-10 strain constructed from AD1-9 by the additional deletion of VMR1 was more sensitive to the tested drugs than AD1-9 (Table 1b). This shows that the multiple drug sensitivity produced by the deletion of VMR1 is independent of that of PDR5 or of any other gene deleted in AD1-9 including all major targets of the transcription regulators PDR1 and PDR3. The observation by Kegel. (2006) that the insertion of a strong promoter upstream of VMR1 of K. lactis results in strain resistance to a high level of geneticin confirms the role of Vmr1p in drug resistance.

View this table:
1

Drug sensitivity of the vmr1Δ strains

Drug concentration (μg mL−1)Growth inhibition (% of wild type)*
(A) Single deletion
No drug0
2,4-Dinitrophenol (30)61
2,4-Dichlorophenoxyacetic acid (50)33
Hygromycin B (15)32
Diuron (5)24
4-Nitroquinoline-N-oxide (0.5)49
1,10-Phenantroline (5)50
8-Hydroxyquinoline (7)48
Cycloheximide (0.1)44
Polidocanol (100)40
2,3,5-Triphenyltetrazolium chloride (10)28
Bilirubin (100)13
CCCP (4)0
Ketoconazole (0.05)0
Oligomycin (0.7)0

View this table:

Rhodamine 6G is a substrate of Pdr5p and is a convenient marker for its efflux activity (Kolaczkowski., 1996). Rhodamine 6G preaccumulated in deenergized cells is more actively extruded upon the addition of glucose in vmr1Δ than in the parental strain (Fig. 2). The cellular uptake of rhodamine 6G is not appreciably modified by the deletion of VMR1 in wild-type strains (data not shown), but the supersensitive strain AD1-10 is sensitive to rhodamine 6G and sensitivity is complemented by the introduction of VMR1 on the plasmid (Supporting Information, Fig. S1). The efflux of rhodamine B that is extruded by Yor1p (Decottignies., 1998) is not modified by the deletion of VMR1 (data not shown).

2

Extrusion of rhodamine 6G from the vmr1Δ strain. Efflux of rhodamine 6G from preloaded deenergized yeast cells in the presence of glucose was measured for the wild-type FY1679-28c and vmr1Δ (DWFY10) strains.

Disruption of VMR1 also confers metal sensitivity

As other members of the yeast MRP subfamily such as Ycf1p (Li., 1996) and Bpt1p (Petrovic., 2000; Sharma., 2002) transport the glutathionecadmium complex into the vacuole, the possible contribution of Vmr1p in cadmium resistance was assessed. Figure 3a and b shows that on a solid glucose medium, the ycf1Δ strain is sensitive to cadmium, while the growth of vmr1Δ is not sensitive. In contrast, when grown on ethanol/glycerol, vmr1Δ is much more sensitive to cadmium than the same strain grown on glucose. The cadmium sensitivity of vmr1Δ, but not that of ycf1Δ, is complemented by the wild-type copy of VMR1 (Fig. 3c). In contrast, vmr1Δ is not sensitive to antimonite, another Ycf1p substrate (Ghosh., 1999) (Fig. 3a–c).

3

Cadmium resistance of Vmr1p on a nonfermentable carbon source. (a–b) Yeast strains FY1679-28c (WT), DWFY6 (ycf1Δ), DWFY10 (vmr1Δ) and DWFY10,6 (ycf1Δvmr1Δ) were grown in a liquid medium and 10-fold serial dilutions of cultures were spotted on a medium without (control) and with the addition of cadmium CdCl2 or potassium antimonyl tartrate (SbIII). (c) Growth of mutants transformed with a pVMR1-GFP plasmid carrying a wild copy of the VMR1 gene. (d) Growth of the strains described above cultivated for 60 h in either a liquid glycerol–ethanol medium (open symbols) or with the addition of 5 μM CdCl2 (filled symbols).

The results of the liquid tests confirmed those obtained on plates. Figure 3d shows that after 40 h of growth in ethanol/glycerol in the presence of 5 μM cadmium, the strain ycf1Δ reaches 57% of the cell density observed in the wild type, while the vmr1Δ strain reaches 46% and the ycf1Δvmr1Δ strain reaches only 33% of the wild-type cell density. No effect of cadmium was observed in vmr1Δ grown on glucose. This indicates the prominent role of Vmr1p in cadmium detoxification during growth in ethanol/glycerol, but not in glucose, where Ycf1p is most efficient.

The involvement of the glutathione conjugates in cadmium sensitivity of vmr1Δ is indicated by measurements of the ATP-dependent transport of the DNP-S-glutathione conjugate by a vacuolar-enriched fraction from the ethanol/glycerol-grown vmr1Δ and ycf1Δ strains. Table 2 shows that the accumulation of glutathione conjugates into the vesicles is slightly more decreased by the deletion of VMR1 in ethanol-grown cells (32%) than in glucose-grown cells (27%). The deletion of YCF1 caused over a 90% decrease in the wild-type transport capacity. The combined deletion of VMR1 and YCF1 inhibits DNP-S-glutathione transport almost completely.

View this table:
2

Microsomal uptake of DNP-SG

StrainATP-dependent DNP-SG transport (nmol h−1× mg−1 protein ± SD) (% of wild type)
GlucoseEthanol/glycerol
FY1679-28c (wild type)2.6 ± 0.4 (100)3.8 ± 0.35 (100)
DWFY10 (vmr1Δ)1.9 ± 0.3 (73)2.6 ± 0.35 (68)
DWFY6 (ycf1Δ)0.2 ± 0.01 (7.5)0.4 ± 0.3 (10.5)
DWFY10,6 (vmr1Δycf1Δ)0.04 ± 0.03 (<2)0.07 ± 0.05 (<2)
  • The uptake of [3H] DNP-GS in the presence vs. absence of ATP by microsomal fraction from the wild type and ycf1Δ and vmr1Δ strains cultured on glucose or ethanol/glycerol.

Vmr 1p localizes to the vacuolar membrane

The GFP was fused to the 3′ end of the chromosomal VMR1 copy. The strain DWFY35GFP expressing the Vmr1p–GFP fusion protein exhibited a level of cadmium sensitivity similar to that of the wild-type FY1679-28c strain, indicating normal trafficking of the fusion protein. However, a very slight GFP fluorescence signal was detected, indicating a low expression of the VMR1 gene. To increase the signal, the C terminal-GFP-tagged version of VMR1 was generated on the plasmid. The Vmr1p–GFP fusion expressed from the centromeric plasmid under the MET25 promoter localized exclusively in the vacuolar membrane of log-phase cells grown on glycerol (Figs 4 and 5). When grown on glucose, the log-phase cells exhibited some punctated fluorescence in addition to that of the vacuolar membrane (Fig. 5b). The Vmr1p-GFP signal clearly colocalizes with the vacuolar membrane marker FM4-64 (Fig. 4).

4

Location of GFP-tagged Vmr1p to the vacuolar membrane. GFP fluorescence pattern from the cells expressing the VMR1-GFP driven from the MET25 promoter or carrying the empty vector, grown in a glycerol medium. Nomarski images and FM4-64 labeling pattern of the same cells.

5

Western blot analysis of the GFP-tagged Vmr1p. (a) Cells expressing the VMR1-GFP or carrying the empty vector were grown to the log phase and the total protein extract was prepared from cells. The Vmr1p-GFP expressed from the MET promoter in wild type cells were examined using the anti-GFP antibody; *additional band ∼100 kDa detected with the anti-GFP antibody. Pgk1p was probed as a loading control. (b) Localization pattern of Vmr1p-GFP in glucose- or glycerol-grown cells; the vacuolar localization of Vmr1p-GFP was verified by fluorescence microscopy. (c) Western blot analysis of Vmr1p-GFP-expressing cells. Cells expressing the VMR1-GFP were grown to the log phase in ethanol/glycerol (gly) or glucose (glu) medium, proteins were isolated from cells and fractionated as given in Materials and methods. C15/40 membrane-enriched fraction and S15 (resulting supernatant) fraction was loaded on a gel and probed with anti-GFP. Anti-Pma1 was used as a control of loading and of the membrane fraction purity.

In the stationary phase, the labeling of the vacuolar membrane decreased and some Vmr1p-GFP fluorescence was observed as intracellular punctate structures or as light GFP fluorescence inside the vacuoles (data not shown). Immunoblot analysis of anti-GFP confirmed the particulate nature of Vmr1p as Vmr1p-GFP was detected in the membrane-enriched C15/40 pellet (Fig. 5c). The VMR1–GFP translation product was detected in parental as well as in the Δvmr1 or Δvmr1Δycf1 strains, indicating that native Vmr1p or Ycf1p does not influence the location of the Vmr1p-GFP expressed from the plasmid (data not shown). The Vmr1p-GFP fluorescence decreased after 10 h of expression in glucose, but persisted over 24 h in ethanol/glycerol.

Regulation of VMR1 transcription

Pdr1p, Pdr3p and Yap1p are well-known transcriptional regulators of yeast multidrug resistance (Rogers., 2001). To assess whether the promoter activity of VMR1 is affected by those transcription regulators, a centromeric plasmid in which the 5′-upstream region of VMR1 fused in-frame to the lacZ coding region was introduced into either wild-type, pdr1/pdr3-defective, PDR1-3-expressing, yap1-defective or YAP1-expressing strains. Table 3 shows that according to these assays, neither Pdr1p/Pdr3p nor Yap1p controls the VMR1 promoter.

View this table:
3

Activity of the VMR1 promoter

Strainβ-Galactosidase activity VMR1promoter-lacZ
GlucoseEthanol/glycerol
ControlCdCl2ControlCdCl2
W303–1A (parental)2.8 ± 0.24.0 ± 0.27.5 ± 0.411.0 ± 0.5
Δadr16.6 ± 0.56.0 ± 0.46.2 ± 0.78.0 ± 0.7
Δmsn2/Δmsn49.2 ± 0.39.2 ± 0.67.3 ± 0.57.5 ± 0.4
BY4142 (parental)3.0 ± 0.25.9 ± 0.46.3 ± 0.58.1 ± 0.7
Δyap24.2 ± 0.49.0 ± 0.66.2 ± 0.58.2 ± 0.3
Δgcn45.5 ± 0.47.6 ± 0.46.7 ± 0.48.7 ± 0.5
SEY6210 (parental)4.0 ± 0.49.0 ± 0.511.0 ± 0.713.0 ± 0.9
Δyap14.1 ± 0.39.1 ± 0.411.1 ± 0.713.9 ± 0.7
US50-18c (parental)8.0 ± 0.815.6 ± 1.014.2 ± 0.917.1 ± 0.9
AD1-10 (Δpdr1, Δpdr3)6.0 ± 0.411.0 ± 0.914.0 ± 1.218.0 ± 1.2
FY1679-28c (parental)3.2 ± 0.27.5 ± 0.34.7 ± 0.36.7 ± 0.7
Δycf13.2 ± 0.45.4 ± 0.35.0 ± 0.76.9 ± 0.6
View this table:

Table 3 also shows that the VMR1 promoter tested by the expression of β-galactosidase activity was over twofold more active in cells grown on ethanol/glycerol than on glucose. Moreover, the presence of 5 μM cadmium in either ethanol/glycerol or glucose growth medium increased by twofold the β-galactosidase activity, which was YAP and PDR1/PDR3 independent. Similar measurements made in the presence of other metals show that VMR1 expression was about twofold-induced by zinc sulfate, antimony potassium tartrate, sodium arsenite and sodium arsenate (Table S1), but not by cobalt, cesium, bismuth or copper salts. Under our conditions, the growth of vmr1Δ was not modified by copper either on glucose or on ethanol/glycerol medium. Thus, we could not confirm the report that Vmr1p is involved in copper and iron homeostasis (De Freitas., 2004).

Analysis of the 437 nucleotides sequence upstream of the ATG of VMR1 using the yeast promoter database SCPD (http://rulai.cshl.edu/SCPD/) showed the existence of two MSN2/MSN4-reactive STRE elements in the VMR1 promoter, as well as putative binding sites for the transcription factors Gcn4p and Adr1p. Single putative binding sites for Gcr1p and Afr1p were also detected. Therefore, we tested the effect of the deletion of the MSN2/MSN4 or GCN4 or ADR1 genes on the VMR1 promoter using the β-galactosidase assay. We found that the deletion of ADR1 caused over a twofold increase in the basal VMR1 expression, together with the loss of additional stimulation by ethanol/glycerol (Table 3). As expected, the stimulation of the expression of VMR1 by cadmium and zinc was lost in the deletion mutants of the MSN2/MSN4 and GCN4 genes that are known to control heavy-metal stress (Arndt & Fink, 1986; Martinez-Pastor., 1996; Schmitt & McEntee, 1996). The effect of the deletion of the activator of glucose-repressed genes HAP4 (Forsburg & Guarente, 1989) was also tested. We found that the deletion of Hap4p does not affect the VMR1 expression either on cells grown on glucose or ethanol/glycerol or on cells grown in the presence of cadmium (data not shown).

Discussion

Vmr1p is one of the six members of yeast MRPs (Decottignies & Goffeau, 1997). It exhibits a typical MRP topology including a hydrophobic N-terminal extension of about 270 amino acids comprising five predicted transmembrane spans. Phylogenetic analysis of the NTE from the yeast MRPs shows that Vmr1p and Ybt1p result from the Whole Genome Duplication. The conservation of the two ohnologs in all post Whole Genome Duplication species indicates that these isoforms perform different functions. The conservation of an ‘ancestral’VMR1 in all hemiascomycetous species supports the conclusion that its function is important.

Vmr1p-GFP localizes solely in the vacuolar membrane when cells are grown in ethanol/glycerol, but punctuate structures are observed when grown in glucose. These punctuate structures do not colocalize with the mitochondria or the nucleus and accumulate in some VMR1 mutants (not shown). We estimate that previous identification of Vmr1p in mitochondrial fractions (Reinders., 2006) is likely to be due to contaminants. The dual localization of Vmr1p may be due to mistrafficking from ER to the vacuole and deserves a complete analysis, which is beyond the scope of the present work.

We show that VMR1 deletants exhibit increased sensitivity to at least 11 amphiphilic drugs of unrelated structure (among the 64 drug tested). Using single and multideleted strains devoid of seven traditional plasma membrane multidrug ABC transporters, we confirmed that the Vmr1p is specifically involved in multidrug resistance. The profile of drugs to which the VMR1-deleted strain is sensitive differs from that of any other multidrug pumps reported so far. Moreover, despite the close phylogenetic similarity of Vmr1p and Ybt1p, no difference was found in the vmr1Δ strain for taurocholic acid and glycochenodeoxycholic acid sensitivity (data not shown), which are characteristic substrates of Ybt1p (Ortiz., 1997).

Vmr1p also contributes to cadmium and mercury resistance. No significant resistance to Zn, Cu or Sb salts was observed. Although the cadmium sensitive phenotype of vmr1Δ is observed only on ethanol/glycerol medium, but not on glucose, the cadmium-induced expression of VMR1 was observed in both glucose and glycerol/ethanol media.

The introduction of the vmr1 deletion causes the sensitivity to rhodamine 6G of the multideleted AD1-10 strain on glucose as well as on glycerol media. This indicates a direct role of Vmr1p in drug resistance, which is independent of the seven other full-size ABC transporters deleted in both strains. The single deletion of VMR1 caused an increase in the rhodamine 6G extrusion from the cells. It was shown by Kolaczkowski (1998) that rhodamine 6G is the substrate for Pdr5p, which is the major multidrug efflux pump in yeast (see also Rogers., 2001). We propose that the increased cellular extrusion of rhodamine 6G observed in deleted vmr1 strain results from the activity of Pdr5p (or other pumps), which is activated by the higher concentration of drug in the cytoplasm resulting from the lack of rhodamine 6G transport into the vacuole by Vmr1p.

Measurements of the ATP-dependent DNP-S-glutathione accumulation in vacuolar-enriched fractions from VMR1 and/or YCF1 deleted mutants indicate that similar to the Ycf1p and Bpt1p proteins (Klein., 2002), a conjugated form of cadmium is used as a substrate for Vmr1p-mediated transport. The single deletion of VMR1 decreased accumulation by 32% in ethanol/glycerol-grown cells compared with 27% inhibition in glucose-grown cells. The single deletion of YCF1 is more powerful as it decreased the accumulation of DNP-S-glutathione by over 90%. This inhibition amounts to almost 100% when both VMR1 an YCF1 deletions are combined. These data indicate the existence of some interaction between the vacuolar Vmr1p and Ycf1p. Similar modulation of Ycf1p by Bpt1p has been reported previously (Sharma., 2002).

A novel feature among the yeast MRPs is the induction of Vmr1p on a respiratory substrate, which is reflected by increased cadmium sensitivity as well as by β-galactosidase fusion analyses of ethanol/glycerol grown cells. Our data indicate the predominant role of Ycf1p on glucose and that of Vmr1p on a respiratory substrate. The VMR1 promoter is also activated by cadmium and zinc. The ethanol/glycerol induction and cadmium induction are not additive, indicating a different mechanism of action. Indeed, the deletion of the two general stress-response transcription factors Msn2p and Msn4p had no detectable effect on VMR1 expression on ethanol/glycerol, but led to the lack of cadmium induction.

The screening of several deletions in transcription factors shows that the ethanol/glycerol and the cadmium induction of VMR1 expression are independent of Pdr1/3p and Yap1p. In contrast, ADR1 is necessary for glucose repression of VMR1 and the VMR1 promoter activity is not modified on ethanol/glycerol by the deletion of ADR1. This transcription factor activates the expression of several genes that are regulated by glucose repression and that are required for ethanol/glycerol and fatty acid utilization (Young., 2003). Although there is no predicted binding site for the HAP complex in the VMR1 promoter sequence, we measured the level of expression of VMR1 in a strain lacking HAP4. The HAP complex is responsible for the induction of glucose-repressed genes and controls the expression of cytochrome genes (Forsburg & Guarente, 1989). However, the lack of Hap4p did not influence the VMR1 expression depending either on the carbon source or on the presence of cadmium. Thus, Hap4p is not involved in the regulation of VMR1.

In brief, Vmr1p is a yeast MRP involved in multidrug and metal resistance that is well conserved during evolution. It is located in the vacuolar membrane. It may use glutathione-conjugated substrates. It is preferably expressed in ethanol/glycerol through a Yap1p- and Pdr1/3p-independent pathway. The repression by glucose of the expression of Vmr1p is dependent on the presence of Adr1p in a complex way that remains to be dissected.

Supporting Information

Fig. S1. Deletion of VMR1 in multideleted strain confers metal sensitivity.

Table S1. Induction of the transcriptional activity of the VMR1 promoter by heavy metals.

Acknowledgements

We thank A. Fortuniak for help in the measurements of glutathione-S conjugate transport, C. Walch-Solimena (Dresden) and S. Kohlwein (Graz) for the evaluation of subcellular location data; K. Flis (Warsaw) for the pKF47 plasmid and M. Ghislain (Louvain-la-Neuve) for the Pma1p antibody. This work was supported by grants from the Interuniversity Pôle d'Attractions program of the Belgian Government Office for Scientific, Technical, and Cultural Affairs, and the European Community's Human Potential Program under contract HPRN-CT-2002-00269, plant transporters. The authors are grateful to the Ministry of Science and Higher Education for support of this work (Grant PBZ-MIN-015/PO5/2004).

Footnotes

  • Editor: Monique Bolotin-Fukuhara

References

View Abstract