Expression of FLR1 Transporter Requires Phospholipase C and Is Repressed by Mediator*

In budding yeast, phosphoinositide-specific phospholipase C (Plc1p encoded by PLC1 gene) is important for function of kinetochores. Deletion of PLC1 results in benomyl sensitivity, alterations in chromatin structure of centromeres, mitotic delay, and a higher frequency of chromosome loss. Here we intended to utilize benomyl sensitivity as a phenotype that would allow us to identify genes that are important for kinetochore function and are downstream of Plc1p. However, our screen identified SIN4, encoding a component of the Mediator complex of RNA polymerase II. Deletion of SIN4 gene (sin4Δ) does not suppress benomyl sensitivity of plc1Δ cells by improving the function of kinetochores. Instead, benomyl sensitivity of plc1Δ cells is caused by a defect in expression of FLR1, and the suppression of benomyl sensitivity in plc1Δ sin4Δ cells occurs by derepression of FLR1 transcription. FLR1 encodes a plasma membrane transporter that mediates resistance to benomyl. Several other mutations in the Mediator complex also result in significant derepression of FLR1 and greatly increased resistance to benomyl. Thus, benomyl sensitivity is not a phenotype exclusively associated with mitotic spindle defect. These results demonstrate that in addition to promoter-specific transcription factors that are components of the pleiotropic drug resistance network, expression of the membrane transporters can be regulated by Plc1p, a component of a signal transduction pathway, and by Mediator, a general transcription factor. The results thus suggest another layer of complexity in regulation of pleiotropic drug resistance.

In budding yeast, phosphoinositide-specific phospholipase C (Plc1p encoded by PLC1 gene) is important for function of kinetochores. Deletion of PLC1 results in benomyl sensitivity, alterations in chromatin structure of centromeres, mitotic delay, and a higher frequency of chromosome loss. Here we intended to utilize benomyl sensitivity as a phenotype that would allow us to identify genes that are important for kinetochore function and are downstream of Plc1p. However, our screen identified SIN4, encoding a component of the Mediator complex of RNA polymerase II. Deletion of SIN4 gene (sin4⌬) does not suppress benomyl sensitivity of plc1⌬ cells by improving the function of kinetochores. Instead, benomyl sensitivity of plc1⌬ cells is caused by a defect in expression of FLR1, and the suppression of benomyl sensitivity in plc1⌬ sin4⌬ cells occurs by derepression of FLR1 transcription. FLR1 encodes a plasma membrane transporter that mediates resistance to benomyl. Several other mutations in the Mediator complex also result in significant derepression of FLR1 and greatly increased resistance to benomyl. Thus, benomyl sensitivity is not a phenotype exclusively associated with mitotic spindle defect. These results demonstrate that in addition to promoter-specific transcription factors that are components of the pleiotropic drug resistance network, expression of the membrane transporters can be regulated by Plc1p, a component of a signal transduction pathway, and by Mediator, a general transcription factor. The results thus suggest another layer of complexity in regulation of pleiotropic drug resistance.
In yeast cells, PLC (Plc1p encoded by PLC1) and three inositol polyphosphate (InsPs) kinases (Ipk2p/Arg82p, Ipk1p, and Kcs1p) constitute a nuclear signaling pathway that affects transcriptional control (4) and export of mRNA from the nucleus (5). Recently, InsPs produced by a Plc1p-dependent pathway have been shown to regulate the activity of chromatin remodeling complexes in vivo and in vitro (6,7). The induction of the phosphate-responsive PHO5 gene, chromatin remodeling of its promoter, and recruitment of Swi/Snf and Ino80 chromatin remodeling complexes are impaired in the ipk2/arg82 mutant strain (7). In vitro, nucleosome mobilization by the yeast Swi/Snf complex is stimulated by IP 4 and IP 5 , whereas IP 6 inhibits nucleosome mobilization by yeast Isw2 and Ino80 complexes and Drosophila NURF complex (6). A possible mechanism by which InsPs affect chromatin remodeling may involve effects on protein conformation of the chromatin remodeling complexes (7). Alternatively, IP 4 or IP 5 might affect the interaction between chromatin remodeling complexes and chromatin, as has been shown for phosphatidylinositol 4,5-bisphosphate and the Swi/Snf complex (8).
In Saccharomyces cerevisiae, the gene coding for phosphatidylinositol-specific PLC (PLC1) is not essential; however, its deletion results in a number of phenotypes (9). We found that Plc1p associates with kinetochores and appears to regulate binding of microtubules to kinetochores (10,11). Kinetochores are specialized protein complexes that assemble at centromeric DNA and bind to spindle microtubules. This attachment is essential for proper chromosome segregation and cell cycle progression. We have found that cells with deletion of the PLC1 gene (plc1⌬) display a higher frequency of chromosome loss, nocodazole sensitivity, and mitotic delay. In addition, chromatin extracts from plc1⌬ cells exhibit reduced microtubule binding to minichromosomes. PLC1 displays strong genetic interactions with components of the inner kinetochore, and plc1⌬ cells display alterations in core centromeric chromatin structure. Chromatin immunoprecipitation experiments indicate that Plc1p localizes to the centromeric loci independently of microtubules (11). These results are consistent with the view that Plc1p affects kinetochore function, possibly by modulating centromeric chromatin structure.
In this study, we planned to utilize benomyl sensitivity of plc1⌬ cells as a phenotype that would allow us to identify genes that function downstream of PLC1 in a pathway important for kinetochore activity, spindle function, and chromosome transmission. However, we found that mutations in the Mediator complex suppress benomyl sensitivity of plc1⌬ cells by a mechanism that is independent of the kinetochore and spindle. Benomyl sensitivity of plc1⌬ cells is caused by a defect in FLR1 expression. FLR1 encodes a multidrug membrane transporter that belongs to a major facilitator superfamily (12)(13)(14). FLR1 overexpression confers resistance to benomyl and several other drugs (15)(16)(17). Mutations in the Mediator complex result in derepression of FLR1 expression and increased resistance to benomyl. The results thus demonstrate that mutations that affect transcriptional regulation may result in an altered expression pattern of multidrug permeases and dramatic changes in cellular drug resistance.

EXPERIMENTAL PROCEDURES
Strains and Media-All yeast strains used in this study are isogenic to W303 and are listed in Table 1. Standard genetic techniques were used to manipulate yeast strains (18). Cells were grown in rich medium (YPD; 1% yeast extract, 2% Bacto-peptone, 2% glucose) or under selection in synthetic complete medium (SC) containing 2% glucose and, when appropriate, lacking specific nutrients to select for a plasmid or strain with a particular genotype. Meiosis was induced in diploid cells by incubation in 1% potassium acetate.
Screen for Suppressors of Benomyl Sensitivity of plc1⌬ Cells-To isolate pbr mutants (plc1⌬ benomyl-resistant), we spread plc1⌬ cells onto YPD plates, and we replica-plated ϳ50,000 resulting colonies onto YPD plates containing 10 g/ml benomyl. After retesting on the same plates, we isolated 13 plc1⌬ strains with second site spontaneous mutations that suppress benomyl sensitivity (plc1⌬Ben R ). Because plc1⌬/plc1⌬ homozygous diploids do not sporulate, we backcrossed the mutants three times to the wild-type strain W303-1a to purify the genetic background and to eliminate mutants in which the benomyl resistance is caused by mutations in multiple genes. To assess the dominance/recessivity, the resulting 10 mutants were mated with the parental plc1⌬ strain. Only three of the resulting diploids displayed benomyl resistance; therefore, the corresponding mutations were considered dominant. However, seven diploids were benomyl-sensitive, and thus the mutations were considered recessive. The complementation grouping was accomplished by crossing the individual mutants and determining whether the resulting plc1⌬/plc1⌬ diploids were sensitive or resistant to benomyl. When the diploid strain behaved like the original haploid strains (Ben R ), then the mutation was concluded to be in the same gene, and the original haploid strains belonged to the same complementation group. When the diploid strain, unlike the original haploid strains, was sensitive to benomyl, then the mutation was concluded to be in two different genes, and the corresponding haploids belonged to different complementation groups. The seven plc1⌬Ben R mutants belong to three complementation groups, designated pbr1 (four mutants), pbr2 (two mutants), and pbr3 (one mutant). Because some tub2 alleles (TUB2 encodes ␤-tubulin) result in recessive benomyl resistance, it was possible that one of the complementation groups represents mutations in the TUB2 gene. To test this possibility, we performed linkage analysis. Representatives from the three complementation groups were mated with a strain with a marked TUB2 gene (haploid segregant of strain CUY409; Ref. 19). In this strain, the HIS3 marker was inserted just downstream of the TUB2 gene. When the resulting diploids were sporulated and dissected, the resistance to benomyl segregated randomly with respect to TUB2-HIS3, demonstrating that none of the complementation groups represented mutations in the TUB2 gene. Cloning of the wild-type gene responsible for benomyl resistance of the plc1⌬ mutant was accomplished by screening one representative from the pbr1 complementation group (pbr1-1) with a yeast genomic library on a plasmid carrying CEN and LEU2 (American Type Culture Collection, Manassas, VA). About 5,000 transformants were allowed to grow on SC-Leu and were subsequently replica-plated onto SC-Leu containing 10 g/ml benomyl. Plasmids were recovered from the transformants that grew well on SC-Leu but failed to grow on SC-Leu containing 10 g/ml benomyl and were reintroduced back into the pbr1-1 mutant to confirm the phenotype. The inserts in six plasmids isolated in this way were identified by sequencing. The only gene that was not truncated and was present on all six inserts was SIN4. To determine whether SIN4 is allelic to the pbr1-1 mutation, the pbr1-1 mutant was crossed with strain DY1702 (sin4::TRP1; Ref. 20). The resulting diploid was benomyl-resistant, and upon transformation with the SIN4 plasmid (sin4 homozygous diploid does not sporulate), sporulation, and dissection, all plc1⌬ haploids cured of the SIN4 plasmid were also benomyl-resistant. Thus, no recombination between the pbr1-1 locus and sin4::TRP1 was detected, and we conclude that pbr1-1 is allelic to SIN4. S1 Nuclease Analysis-Oligonucleotides complementary to the genes assayed by S1 nuclease analysis are as follows:  Total RNA was isolated from cultures grown in YPD medium to A 600 nm ϳ1.0 by the hot phenol method as described previously (21). S1 probes were end-labeled in a 25-l reaction mixture (5 pmol of oligonucleotide, 125 Ci of [␥-32 P]ATP, 6,000 Ci/mmol (PerkinElmer Life Sciences), 1 ϫ T4 polynucleotide kinase buffer and 20 U T4 polynucleotide kinase (New England Biolabs)) at 37°C for 1 h. The reaction mixture was diluted with 25 l of water; T4 polynucleotide kinase was inactivated at 65°C for 20 min, and the labeled oligonucleotides were purified using MicroSpin G-25 columns (Amersham Biosciences). The labeled oligonucleotides (0.5 pmol) were hybridized with 20 -40 g of total RNA in a 50-l reaction mixture (0.3 M NaCl, 1 mM EDTA, 40 mM HEPES, pH 7.0, and 0.1% Triton X-100) for 12 h at 55°C and treated with S1 nuclease (PerkinElmer Life Sciences) as described previously (21). The samples were analyzed on 20% denaturing polyacrylamide gels, and quantification was performed using a PhosphorImager (PerkinElmer Life Sciences).
Minichromosome Stability Assay-Mitotic minichromosome stability was measured as a fraction of cells that retained the plasmid after growth in nonselective medium (22,23). Briefly, wild-type, plc1⌬, sin4⌬, and plc1⌬sin4⌬ cells were transformed with pRS413 plasmid (CEN, HIS3). For each transformant, five single colonies were inoculated separately into medium nonselective for pRS413 plasmid (YPD medium or YPD medium containing benomyl at 2 g/ml) and grown for about 24 h at 28°C. At the end of the growth, the cultures were still in the exponential phase, as determined by counting the cells with a hemocytometer and by measuring A 600 nm . For each culture, the frequency of His ϩ cells was determined at the time of inoculation (F o ) and at the end of nonselective growth (F end ) by plating appropriately diluted cultures on YPD plates and subsequent replica plating onto synthetic complete medium lacking histidine (SCϪHis). The rate of plasmid loss per generation was determined as described (22,23), according to the following equation: where F D is the fraction of cells in each generation that retain the minichromosome, and G is the number of generations. The fraction of cells that lose the minichromosome per generation is 1 Ϫ F D . The number of cell doublings was calculated by counting the total cell number at the beginning and at the end of nonselective growth.
Minichromosome-Microtubule Binding Assay-Preparation of yeast lysates containing minichromosomes and the minichromosome-microtubule binding assay were done as described previously (24,25). Briefly, yeast cultures were grown to an A 600 nm of ϳ0.6 after maintaining the cells in exponential growth for several generations. Nocodazole was added to the cultures to 15 g/ml for 6 h, and nocodazole was also present during preparation of spheroplasts (24). Cells were spheroplasted by glusulase and osmotically lysed in EBB buffer (10 mM Tris-Cl, pH 7.4, 10 mM MgCl 2 , 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 0.1 mM dithiothreitol). Minichromosomes were eluted from nuclei by adding 0.3 M NaCl. After 5 min of incubation, the extracts were 3-fold diluted by adding EBB buffer and subjected to two subsequent centrifugations, each at 15,000 ϫ g for 20 min. The clear supernatant was removed, supplemented with 10 M Taxol, and used for the microtubule-binding assay. Purified tubulin was polymerized in vitro into microtubules (MTs) of an average length of 2 m as directed by manufacturer (ICN Biochemicals, Inc), and stabilized by addition of the MTs stabilizing drug Taxol (10 M final concentration). Different amounts of stabilized MTs were added to 500-l aliquots of the cleared extracts to initiate the microtubule-binding assay. After 15 min of incubation at room temperature, the reaction mixtures were centrifuged at 15,000 ϫ g for 8 min. The supernatant and pellet fractions were separated, and the amount of minichromosomes in each fraction was determined by Southern blot using an Amp r probe. The calibration of the band intensities was performed by loading different amounts of the Amp r DNA fragment on the gels, and the band intensities of the scanned images were quantified using UN-SCAN-IT software (Silk Scientific).
Fluorescence Microscopy-Cells in 1 ml of media were fixed for 1 h by the addition of formaldehyde to 3.7% concentration. Cells were centrifuged, washed two times with phosphate-buffered saline, resuspended in 1 ml of 70% ethanol, and kept at 23°C for 30 min. After centrifugation, washing, and sonication for 10 s, cells were stained with 0.5 g/ml 4Ј,6-diamidino-2-phenylindole and observed using Nikon Eclipse 800 microscope equipped with SPOT RT CCD camera, UV filter set, and ϫ100/1.4 n.a. oil immersion objective.

RESULTS
Isolation of pbr Mutants-Hypersensitivity to the microtubule-destabilizing drugs benomyl and nocodazole is a phenotype shared by mutants with defects in kinetochore, mitotic spindle, and mitotic checkpoint (26 -30). Because deletion of the PLC1 gene, encoding phospholipase C, causes hypersensitivity to nocodazole and benomyl (Ref. 10; plc1⌬ cells cannot grow in the presence of 5 g/ml nocodazole or 10 g/ml benomyl), we used this phenotype to identify genes functioning downstream of PLC1. We speculated that chromosomal suppressors of the benomyl sensitivity of the plc1⌬ strain may arise because of the mutations in genes functioning in a pathway that either regulates or affects the function of the mitotic spindle. We isolated seven recessive mutations that include three complementation groups as follows: pbr1 (four mutants), pbr2 (two mutants), and pbr3 (one mutant). In this study, we described the identification of SIN4, a gene that corresponds to the pbr1 complementation group (see "Experimental Procedures"). Subsequent experiments demonstrated that pbr2 is allelic to MED2 (see "Results"; mutations in several components of the Mediator cause increased expression of FLR1 and resistance to benomyl). Because the plc1⌬pbr1-1 mutant was phenotypically identical to the plc1⌬ sin4⌬ strain, we used the latter strain in further experiments aimed at elucidating the mechanism of suppression. SIN4 encodes a component of the Mediator complex of RNA polymerase II. However, mutations in SIN4 exhibit pleiotropic phenotypes reminiscent of mutations of histones, suggesting a role for Sin4p in the regulation of chromatin structure (20,31,32).
Characterization of plc1⌬ sin4⌬ Mutant-Deletion of PLC1 results in alterations in the structure of centromeric chromatin (11), higher frequency of minichromosome loss, mitotic delay, and reduced binding of minichromosomes to microtubules (10). Because Sin4p was implicated in regulation of chromatin structure (20,32), it was possible that suppression of benomyl sensitivity in plc1⌬ cells by sin4⌬ mutation (Fig.  1A) occurs at the level of chromatin structure and activity of the kinetochore. To test this possibility, we performed three assays aimed at assessing the function of the kinetochore in wild-type, plc1⌬, sin4⌬, and plc1⌬sin4⌬ strains.
Because the mitotic stability of minichromosomes depends on the function of kinetochores, we determined the stability of the pRS413 minichromosome in wild-type, plc1⌬, sin4⌬, and plc1⌬sin4⌬ strains in the absence of benomyl and in the presence of a low concentration of benomyl (Fig. 1B). The stability of the pRS413 minichromosome was measured as a fraction of cells that retained the minichromosome after growth in nonselective medium (22). As we reported previously, plc1⌬ cells displayed about 5-fold higher minichromosome loss rate than wild-type cells (10). If sin4⌬ suppresses benomyl sensitivity of plc1⌬ cells by improving the function of the kinetochore, then we would expect that plc1⌬sin4⌬ cells would have a lower rate of minichromo-some loss than plc1⌬ cells. However, sin4⌬ mutation did not suppress the minichromosome loss rate in plc1⌬ cells (Fig. 1B). Benomyl (2 g/ml) increased the minichromosome loss rate less than 2-fold in the wild-type strain but almost 5-fold in plc1⌬ cells. Most interestingly, the minichromosome loss rate was almost unaffected in plc1⌬sin4⌬ cells by the presence of benomyl. This result suggests that sin4⌬ mutation does not improve performance of kinetochores in plc1⌬ cells in the absence of benomyl but, somehow, protects plc1⌬ cells from the effects of benomyl.
To gain more insight into the role of SIN4 in the regulation of kinetochore activity, we used an assay developed by Kingsbury and Koshland (24,25) to measure the ability of minichromosomes formed in vivo to bind microtubules in vitro. Wild-type, plc1⌬, sin4⌬, and plc1⌬sin4⌬ strains were transformed with centromeric plasmid (minichromosome) pRS414, which was then assayed in clarified extracts of these cultures for binding to Taxol-stabilized microtubules. To improve the sensitivity of the assay and to exclude the possibility that any difference in microtubule binding activity between the strains is merely the result of different cell cycle profiles, the assay was performed with lysates prepared from nocodazole-arrested cells (Fig. 1C). The G 2 /M transition is the phase when the kinetochore is required for progression of the cell cycle through mitosis. It is also when the kinetochore is in its most active state (24,25). Kinetochores of cells arrested with nocodazole at the G 2 /M transition exhibit the highest microtubule binding activity (10). Our data for the wild-type strain correspond to these previous results. The saturating concentration of microtubules pelleted about 48 and 50% of the plasmid in the wild-type and sin4⌬ strains, respectively, but only 27% in the plc1⌬ strain. Again, sin4⌬ mutation did not improve the activity of kinetochores in plc1⌬ cells, and microtubules pelleted about 28% of minichromosomes from plc1⌬sin4⌬ lysates (Fig. 1C).
We have shown previously that Plc1p is important for high fidelity chromosome segregation and activity of kinetochores. Consequently, plc1⌬ cells experience delay at the G 2 /M stage of the cell cycle because the kinetochore-based mitotic checkpoint control system detects defects in kinetochore-microtubule interaction in plc1⌬ cells and mediates G 2 /M delay. To assess whether this G 2 /M delay is eliminated in plc1⌬sin4⌬ cells, we examined cell and nuclear morphology of wildtype and plc1⌬ cells during exponential growth at 30°C ( Table 2). The frequency of large budded cells (diameter of the bud is at least 75% of the diameter of the mother cell) with a single nucleus within 50% of the mother cell proximal to the neck (33, 34) was 11 and 13% for the wildtype and sin4⌬ strains, respectively. The frequency of large budded cells for plc1⌬ strain was 30% and for plc1⌬sin4⌬ strain was 29%.
Thus, the above results suggest that the sin4⌬ mutation does not suppress the mitotic phenotypes of plc1⌬ cells by increasing the kinetochore activity or by improving some other aspect of spindle function. However, sin4⌬ suppresses benomyl sensitivity of plc1⌬ cells (Fig. 1A) and significantly enhances the stability of minichromosomes in plc1⌬ cells in the presence of benomyl (Fig. 1B).
plc1⌬ Cells Fail to Induce FLR1 Expression-Another possibility we considered that could explain sensitivity of plc1⌬ cells to benomyl and its suppression by sin4⌬ mutation is cellular resistance to benomyl mediated by Flr1p. Flr1p is a plasma membrane transporter (35) that belongs to the major facilitator superfamily (13,14). Disruption of FLR1 gene results in benomyl sensitivity, and FLR1 overexpression results in increased resistance to benomyl (15). Expression of FLR1 is induced in the presence of benomyl and is mediated by at least three transcriptional regulators, Yap1p, Pdr3p, and Yrr1p (15,16,17,35,36,37).
To test the possibility that sensitivity of plc1⌬ cells and resistance of plc1⌬sin4⌬ cells to benomyl are caused by altered regulation of FLR1 expression, wild-type, plc1⌬, sin4⌬, and plc1⌬sin4⌬ cells were grown in YPD medium, and expression of FLR1 was induced by benomyl addition. Compared with wild-type cells, expression of FLR1 is dramatically FIGURE 1. Benomyl sensitivity of plc1⌬ cells is suppressed by sin4⌬ mutation independently of kinetochore function. A, benomyl sensitivity of plc1⌬ cells is suppressed by sin4⌬ mutation. Cells of the indicated strains were streaked on three YPD plates containing 10 g/ml benomyl and were incubated for 3 days at 30°C. A typical plate is shown. B, mitotic stability of minichromosomes in WT, plc1⌬, sin4⌬, and plc1⌬ sin4⌬ cells. The minichromosome stability was measured for at least five independent transformants as a fraction of cells that retained the minichromosome after growth in nonselective YPD medium (containing or not containing benomyl) as described under "Experimental Procedures." The results are reported as the percentage of minichromosome loss per generation at 30°C (standard deviations are indicated). C, minichromosome binding assay in wild type, plc1⌬, sin4⌬, and plc1⌬ sin4⌬ cells. Cleared lysates were prepared from the indicated strains transformed with centromeric plasmid pRS414. Cells were arrested at the G 2 /M stage of cell cycle by incubating for 6 h in the presence of 15 g/ml nocodazole. Under these conditions, 90% of the cells of all four strains arrested as large budded cells. Different amounts of bovine microtubules were added to the lysates and incubated for 15 min at 20°C. Microtubules were pelleted, and the percentage of minichromosomes that co-sedimented with microtubules was determined. The results represents the means of three independent experiments, which agreed to within 15%. reduced in plc1⌬ cells, although sin4⌬ and plc1⌬sin4⌬ cells express significant amounts of FLR1 even without induction with benomyl, and expression of FLR1 in these two strains appears to be more persistent (Fig. 2). These results thus provide a mechanism for benomyl sensitivity of plc1⌬ cells and its suppression by sin4⌬ mutation. IP 4 and/or IP 5 Are Required for FLR1 Induction and Benomyl Resistance-The Plc1p-dependent pathway produces multiple InsPs with diverse functions. IP 3 , produced by Plc1p, is converted into IP 4 and IP 5 by Ipk2p (4,5). IP 4 and IP 5 were implicated in chromatin remodeling (6,7). IP 6 , produced from IP 5 by Ipk1p, regulates mRNA export from the nucleus (5). Kcs1p produces inositol pyrophosphates, PP-IP 4 and PP-IP 5 , that are involved in homologous DNA recombination (38) and regulation of telomere length (39,40). In addition, inositol pyrophosphates are able to phosphorylate proteins in vivo by a nonenzymatic mechanism (41). Because plc1⌬ cells are completely devoid of all InsPs (4, 5), we wanted to identify the specific inositol polyphosphate that is required for expression of FLR1 and benomyl resistance. We compared the ability of wild-type, plc1⌬, ipk2⌬, ipk1⌬, and kcs1⌬ strains to grow in the presence of benomyl (Fig.  3A). Although the ipk1⌬ strain was almost as resistant to benomyl as the wild-type strain, the kcs1⌬ strain was noticeably less resistant, and the ipk2⌬ strain was as sensitive to benomyl as the plc1⌬ strain. To determine whether benomyl resistance correlates with the ability to activate the FLR1 gene, we treated wild-type, plc1⌬, ipk2⌬, ipk1⌬, and kcs1⌬ cells with benomyl and determined FLR1 expression (Fig. 3B). Similarly to plc1⌬ cells, ipk2⌬ cells fail to induce FLR1, although ipk1⌬ and kcs1⌬ cells express intermediate levels of FLR1. Thus, the ability to induce expression of FLR1 in the presence of benomyl requires synthesis of IP 4 and/or IP 5 and correlates with the ability to grow in the presence of benomyl (Fig. 3). Lack of IP 6 in ipk1⌬ cells and PP-IP 4 and PP-IP 5 in kcs1⌬ cells results in a somewhat decreased level of FLR1 expression. However, it appears that the level of FLR1 expression in ipk1⌬ cells is sufficient to provide wild-type levels of benomyl resistance.  -1a), plc1⌬ (HL1-1), sin4⌬ (DY1702), and plc1⌬sin4⌬ (ND96) growing in YPD at 30°C were stained with 4Ј,6-diamidino-2-phenylindole, and nuclear and bud morphologies were scored. At least 300 separated cells were scored for each sample in each experiment. The experiment was performed three times, and standard deviations are indicated. UB means unbudded cells; SB, small budded cells (diameter of the bud less than 75% of the diameter of the mother cell); LB, large budded cells with single nuclei (cells with nuclei within 50% of the mother cell proximal to the neck and with diameter of the bud at least 75% of the diameter of the mother cell); LB, large budded cells with two separate nuclei (one nucleus in the mother cell and one nucleus in the bud); DB, cells with two buds (occasionally cells with more than two buds were observed and were also scored as DB). FLR1 expression in plc1⌬ cells. A, the indicated strains were grown in YPD medium at 30°C to A 600 nm ϭ 1.0. Benomyl was subsequently added to 5 g/ml, and the samples were collected at 0, 1, 3, and 9 h. Total RNA was prepared, and S1 nuclease protection assays were performed using 20 g (ACT1) or 40 g (FLR1) of the total RNA. A typical experiment is shown. B, the RNA levels in three independent experiments were quantitated by PhosphorImager, normalized by dividing by the value for ACT1 in the corresponding strain, and expressed as a fold change of the WT value at time 0 h. The results are shown as means Ϯ S.D.  (20,42,43), and yeast cells lacking the MED1 gene, which encodes a subunit of the Rgr1 module, display defects in both repression and induction of the GAL gene expression (44).

Mutations in Several Components of Mediator Cause Increased Expression of FLR1 and Resistance to
To determine whether in addition to sin4⌬, mutations in other components of the Mediator suppress benomyl sensitivity of plc1⌬ cells, we constructed corresponding double mutants and tested their ability to grow in the presence of benomyl (Fig. 4). In addition to sin4⌬, med1⌬, med2⌬, med1⌬plc1⌬, and med2⌬plc1⌬ displayed greatly increased levels of resistance to benomyl, although srb8⌬, srb10⌬, srb11⌬, and tup1⌬ are only somewhat more resistant to benomyl than the wild-type cells (Fig. 4).
To test whether one of these mutations is allelic to our pbr2 or pbr3 mutations, we crossed pbr2-1 and pbr3-1 mutants (see "Experimental Procedures") to med1⌬, med2⌬, sin4⌬, srb8⌬, srb10⌬, and srb11⌬ strains. The resulting diploids were sporulated and dissected, and the benomyl sensitivity of both the diploids and haploid segregants was determined. Except for the benomyl-resistant diploid carrying pbr2-1 and med2⌬ mutations, all other diploids were benomyl-sensitive. Upon dissection, the diploid carrying pbr2-1 and med2⌬ mutations consistently yielded four benomyl-resistant haploids. Thus, no recombination between pbr2 locus and med2⌬ was detected, and we conclude that pbr2 is allelic to MED2.
Because the sin4⌬ mutation caused derepression of FLR1 transcription even in the absence of benomyl (Fig. 2), we determined FLR1 expression in a series of strains with mutations in other components of the Mediator. As expected, med2⌬, med1⌬, sin4⌬, and also srb11⌬ and srb10⌬ display increased expression of FLR1 (Fig. 5). Because the multidrug membrane transporters involved in drug resistance, such as Pdr5p, Snq2p, Ycf1p, and Yor1p display overlapping substrate specific-ity and are regulated by the same transcriptional factors (13,14,36), we wanted to determine whether their expression is also affected by mutations in components of the Mediator (Fig. 5). It appears that mutations in the Mediator complex (med2⌬, med1⌬, sin4⌬, srb11⌬, and srb10⌬) cause most significant derepression of FLR1. Expression of the other transporters is affected by Mediator mutations to a much smaller extent (Fig. 5). Benomyl induces expression of FLR1 significantly more than expression of the other transporters (Fig. 6). In addition, FLR1 is the only transporter that shows in the presence of benomyl significantly higher expression in wild-type and plc1⌬ sin4⌬ strain than in plc1⌬ strain. Thus, it appears that the expression of FLR1 and not YOR1, YCF1, PDR5, and SNQ2 correlates with the resistance to benomyl. This is in agreement with the finding that flr1⌬ cells are sensitive to benomyl, and overexpression of FLR1 confers resistance to benomyl (15).
In addition to derepression of FLR1, several Mediator mutations also caused mild derepression of YOR1, YCF1, PDR5, and SNQ2 (Fig.  5). To determine whether this derepression caused a multidrug resistance phenotype, we spotted corresponding Mediator mutants on plates containing model substrates 4-nitroquinoline 1-oxide (4-NQO), cycloheximide, and oligomycin (Fig. 7). Snq2p is responsible for detoxification of 4-NQO, although cycloheximide is a substrate for Pdr5p. Pdr5p and Snq2p transport many structurally and functionally unrelated drugs with significant overlap in substrate specificity. Oligomycin is a substrate for Yor1p, and Ycf1p transports glutathione conjugates into the vacuole (for review see Ref. 13). It appears that Mediator mutations do not cause resistance to any of the tested compounds (Fig. 7). Although the med1⌬ strain does not display increased sensitivity to any of the tested compounds, the sin4⌬ strain is slightly more sensitive to cycloheximide than the wild-type strain, and the med2⌬ mutant has increased sensitivity to both 4-NQO and cycloheximide. Double mutants plc1⌬ sin4⌬, plc1⌬ med1⌬, and plc1⌬med2⌬ are more sensitive to the tested FIGURE 3. InsP 4 and/or InsP 5 are required for FLR1 induction and benomyl resistance. A, cells of each strain were grown in three independent cultures to log phase at 30°C, and 10-fold serial dilutions were spotted onto three sets of YPD plates and three sets of YPD plates containing 10 g/ml benomyl. Cells were grown at the same temperature for 3 days (0 g/ml benomyl) or 5 days (10 g/ml benomyl). The following strains were used: wild-type (WT; W303-1a), plc1⌬ (HL1-1), ipk2⌬ (A0003), ipk1⌬ (A0004), and kcs1⌬ (LSY507). B, the same strains were grown in YPD medium at 30°C to A 600 nm ϭ 1.0. Benomyl was subsequently added to 5 g/ml, and samples were collected at 0 and 3 h. Total RNA was prepared, and S1 nuclease protection assays were performed using 20 g (ACT1) or 40 g (FLR1) of the total RNA. The experiment was repeated three times, and the results agreed within 15%. Results of a typical experiment are shown. compounds than plc1⌬ or single Mediator mutations. Thus, despite the fact that med1⌬, med2⌬, or sin4⌬ mutations dramatically elevate resistance to benomyl, they do not confer a multidrug resistance phenotype.

DISCUSSION
We have shown previously that in budding yeast, Plc1p and InsPs produced by the Plc1p-dependent pathway affect function of kinetochores, probably by modulating centromeric chromatin structure (10,11). Because plc1⌬ cells are sensitive to benomyl and nocodazole, we speculated that this phenotype is a consequence of the partially compromised kinetochore function. In this study, we intended to utilize benomyl sensitivity as a phenotype that would allow us to identify genes that are important for the function of the kinetochore or spindle and that are downstream of Plc1p and InsPs. However, our screen identified SIN4, encoding a component of the Mediator complex. Because Sin4p was implicated in the regulation of chromatin structure (20), we speculated that the suppression of benomyl sensitivity of plc1⌬ cells by sin4⌬ mutation is because of the altered chromatin structure of the centromeres in plc1⌬sin4⌬ cells that results in an improved kinetochore function. However, our characterization of the plc1⌬sin4⌬ mutant showed that sin4⌬ mutation does not suppress benomyl sensitivity of plc1⌬ cells by improving the function of kinetochores ( Fig. 1 and Table 2). In the absence of benomyl, the minichromosome loss rate, binding of FIGURE 5. Expression of multidrug membrane transporters in strains with mutations of different components of the Mediator. The indicated strains were grown in YPD medium at 30°C to A 600 nm ϭ 1.0. Total RNA was prepared, and S1 nuclease protection assays were performed using 20 g (ACT1) or 40 g (all other probes) of the total RNA. The RNA levels were quantitated by Phos-phorImager, normalized by dividing by the value for ACT1 in the corresponding strain, and expressed as a fold change of the WT value. The same set of strains as in Fig. 4 was used. The experiment was repeated three times, and the results agreed within 15%. Results of a typical experiment are shown.  Benomyl induction of FLR1, YOR1, YCF1, PDR5, and SNQ2. WT, plc1⌬, sin4⌬, and plc1⌬sin4⌬ strains were grown in YPD medium at 30°C to A 600 nm ϭ 1.0. Benomyl was subsequently added to 5 g/ml, and samples were collected at 0 and 3 h. Total RNA was prepared, and S1 nuclease protection assays were performed using 20 g (ACT1) or 40 g (all other probes) of the total RNA. The RNA levels were quantitated by Phospho-rImager, normalized by dividing by the value for ACT1 in the corresponding strain, and expressed as a fold change of the WT value at time 0 h. The experiment was repeated three times, and the results agreed within 15%. Results of a typical experiment are shown. FIGURE 7. Mediator mutants do not display multiple drug resistance. Cells were grown to log phase at 30°C, and 10-fold serial dilutions were spotted onto YPD plates containing 4-nitroquinoline (0.1 g/ml), cycloheximide (0.05 g/ml), and oligomycin (50 g/ml). Three plates containing each chemical were incubated at 30°C for 3 days, and typical plates are shown. minichromosomes to microtubules in vitro (Fig. 1), and mitotic delay ( Table 2) are indistinguishable in plc1⌬ and plc1⌬sin4⌬ strains. The minichromosome loss rate, however, is significantly lower in plc1⌬sin4⌬ cells than in plc1⌬ cells in the presence of benomyl. This result thus suggested that sin4⌬ improves kinetochore function of plc1⌬ cells only in the presence of benomyl. Subsequent experiments demonstrated that sin4⌬ mutation causes derepression of the membrane transporter FLR1 that mediates resistance to benomyl (15).
Benomyl is an antimitotic drug that destabilizes microtubules and inhibits microtubule-mediated processes, including nuclear division, migration, and fusion (30). Mutations in components of the mitotic spindle result in benomyl sensitivity or resistance (28,29). Benomyl was also used with great success as a tool that allowed identification of genes of the mitotic checkpoint (26,27). Sensitivity to benomyl is generally considered a phenotype indicative of a mitotic spindle defect (for review see Refs. 30 and 45). However, this study demonstrates that mutations that affect expression of FLR1 result in benomyl sensitivity or resistance. Thus, caution is warranted when interpreting benomyl sensitivity as an indication of a spindle defect. Specifically, mutations that affect chromatin or components of general transcription machinery and result in benomyl sensitivity should be evaluated for FLR1 expression before concluding that benomyl sensitivity is a consequence of a mitotic spindle defect in the particular mutant.
Our finding of increased FLR1 expression in med1⌬, med2⌬, and sin4⌬ mutants suggests that at the FLR1 promoter the Mediator functions as a co-repressor. This is not entirely surprising, because the Mediator plays a role not only in transcriptional activation but also in repression of gene transcription (20,42,43). Mediator displays three distinct submodules: a head, middle, and tail region (46,47). The RNA polymerase II contacts the head and middle regions, although the elongated tail region that represents the largest part of the Mediator is responsible for interaction with several different activators, including Gal4p (48,49), SBF (50), and Swi5p (51). These activators directly interact with components of the Mediator tail region Sin4p, Gal11p, Med2p, and Med3p (48,49,52), thereby recruiting the Mediator and RNA polymerase II to regulated promoters. The repressive role of the Mediator requires function of the Srb8 -11 module that is believed to inhibit association of RNA polymerase II with the Mediator and formation of the preinitiation complex (53,54). The co-repressor complex Tup1p/Ssn6p recruits Mediator to repressed promoters by physically interacting with Mediator components Med3p (55), Srb7p (56), Srb10p (57), and Srb11p (58). However, Tup1p/Ssn6p is probably not involved in recruitment of the Mediator to the FLR1 promotor, because FLR1 expression and benomyl resistance in the tup1⌬ mutant are increased much less than in med1⌬, med2⌬, and sin4⌬ mutants (Figs. 2 and 5). Currently, we do not know the identity of the repressor or co-repressor that recruits Mediator to the FLR1 promoter.
The fact that plc1⌬ cells fail to induce FLR1 expression when challenged with benomyl is also not unprecedented. Cells with PLC1 deletion are osmosensitive because of the inability to induce expression of GPD1 (encoding glycerol-3-phosphate dehydrogenase; see Ref. 59). The possible explanation for the two transcriptional defects in plc1⌬ cells is that InsPs produced by the Plc1p-dependent pathway regulate the activity of chromatin remodeling complexes at the GPD1 and FLR1 promoters. InsPs are required for induction of the phosphate-responsive PHO5 gene, recruitment of Swi/Snf and Ino80 chromatin remodeling complexes, and chromatin remodeling of its promoter (7). However, swi2⌬ cells are not sensitive to benomyl (data not shown). It is possible that the failure to induce FLR1 expression in plc1⌬ cells is caused by a defect in recruitment or deregulation of another chromatin remodeling complex at the FLR1 promoter. Chromatin remodeling mediated by this chromatin remodeling complex may be required for efficient binding of transcriptional activators or co-activators to the FLR1 promoter. Expression of FLR1 is regulated by three transcriptional activators, Yap1p, Pdr3p, and Yrr1p (15)(16)(17)(35)(36)(37). On the basis of analogy with the PHO5 promoter, where InsPs and the Swi/Snf complex are required for efficient recruitment of Pho4p activator (7), we speculate that chromatin remodeling facilitates recruitment of Yap1p, Pdr3p, and/or Yrr1p to the FLR1 promoter. In support of this possibility, we note that overexpression of Yap1p also suppresses benomyl sensitivity of plc1⌬ cells. 4 However, it is also possible that InsPs affect function or recruitment of another transcriptional complex, such as the Mediator or SAGA, which may be required for assembly of the preinitiation complex at the FLR1 promoter.
Multidrug resistance of cancer cells results in failure of chemotherapy and remains a major problem in cancer treatment. In yeast cells, cellular resistance to diverse drugs is mediated by the PDR network that includes several transcription factors that regulate expression of the multidrug membrane transporters (for review see Refs. 13 and 14). In this study we have demonstrated that in addition to these promoterspecific transcription factors, expression of the transporters can be regulated by component(s) of signal transduction pathway(s) (Plc1p) and by general transcription factor(s) (Mediator). The results thus suggest another layer of complexity in regulation of the PDR network.