Identification of Saccharomyces cerevisiae Isoleucyl-tRNA Synthetase as a Target of the G1-specific Inhibitor Reveromycin A*

To dissect the action mechanism of reveromycin A (RM-A), a G1-specific inhibitor, aSaccharomyces cerevisiae dominant mutant specifically resistant to RM-A, was isolated from a strain in which the genes implicated in nonspecific multidrug resistance had been deleted. The mutant gene (YRR2–1) responsible for the resistance was identified as an allele of the ILS1 gene encoding tRNAIle synthetase (IleRS). The activity of IleRS, but not several other aminoacyl-tRNA synthetases examined in wild type cell extract, was highly sensitive to RM-A (IC50 = 8 ng/ml). The IleRS activity of the YRR2–1 mutant was 4-fold more resistant to the inhibitor compared with that of wild type. The mutation IleRSN660D, near the KMSKS consensus sequence commonly found in the class I aminoacyl transferases, was found to be responsible for RM-A resistance. Moreover, overexpression of theILS1 gene from a high-copy plasmid conferred RM-A resistance. These results indicated that IleRS is a target of RM-Ain vivo. A defect of the GCN2 gene led to decreased RM-A resistance. IleRS inhibition by RM-A led to transcriptional activation of the ILS1 gene viathe Gcn2-Gcn4 general amino acid control pathway, and this autoregulation seemed to contribute to RM-A resistance.

Reveromycin A (RM-A) 1 (see Fig. 1), which was discovered in the culture medium of Actinomycetes due to its inhibitory activity on the epidermal growth factor-dependent responses of mouse epidermal cells, blocks the cell division cycle of mammalian cells in G 1 phase (1). 2 However, the mechanism by which RM-A elicits the cell-cycle-specific growth arrest in mammalian cells is unknown. Yeast provides an ideal system to investigate the diverse biological activity of drugs because of the observations that various drugs that are effective to both yeast and higher eukaryotic cells often elicit their effects by an identical mechanism, sharing structurally and functionally conserved drug target molecules. Because RM-A exhibits an antifungal activity, we attempted to dissect its mode of action by identifying its target molecule by the power of yeast genetics (1,2).
Our previous genetic approaches in Saccharomyces cerevisiae have demonstrated that the ATP-binding cassette superfamily anionic drug transporter Yrs1/Yor1 and the Zn(II) 2 Cys 6type transcription factor Yrr1 are important for RM-A resistance by enhancing the Yor1-mediated multidrug resistance mechanism (2)(3)(4). However, the molecular mechanism by which RM-A inhibits the growth of yeast cells has still remained unknown. On the basis of these findings, the purpose of this study is to identify as an yet unknown RM-A target molecule in yeast by isolating and characterizing mutants resistant to RM-A in a genetic background in which several genes implicated in multidrug resistance have been deleted. We revealed that a dominant mutation (YRR2-1) in the ILS1 gene encoding isoleucyl-tRNA synthetase (IleRS) is responsible for the RM-A resistance. The IleRS activity, but not those of several other aminoacyl-tRNA synthetases examined in the wild type (WT) cell extract, was highly sensitive to RM-A, whereas the IleRS activity in the YRR2-1 strain was more resistant to the inhibitor, indicating that IleRS is a major target for RM-A in yeast. The transcription of ILS1 gene is activated by the Gcn2-Gcn4 general amino acid control pathway (for review, see Ref. 5). We also demonstrated that a defect of the GCN2 gene led to an increased RM-A sensitivity, indicating that the elevation of ILS1 expression by the Gcn2-Gcn4-mediated autoregulatory mechanism also contributes to RM-A resistance.
Drug Sensitivity Test-The effect of RM-A or other drugs on the growth of yeast strains was compared by spotting cell suspension on solid medium. A 3-fold serial dilution of yeast cells suspended in water (ϳ2 ϫ 10 7 cells/ml) was applied on YPD solid medium containing various concentrations of drugs using an applicator. The plates were incubated at 28°C for 2-3 days.
[ 35 S]Methionine Incorporation-Strains were grown to log phase in * This work was supported in part by a grant-in aid for scientific research from Ministry of Education, Science, Culture, Arts, and Sports of Japan and Special Coordination Funds for Promoting Sciences and Technology from the Sciences and Technology Agency of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence and reprint requests should be addressed. Morphological Observations and Fluorescence-activated Cell Sorter Analysis-Samples for fluorescence-activated cell sorting were prepared and analyzed as described previously (6). Cells were stained with 4Ј,6Ј-diamidino-2-phenylindole to visualize nuclei after incubation with RM-A for various periods of time. The cells were fixed with formaldehyde to a final concentration of 3.7% and viewed by differential interference contrast microscopy.
Cloning of the YRR2-1 Mutant Gene-Total genomic DNA of the YRR2-1 strain was partially digested with the restriction endonuclease Sau3AI and the DNA fragments of 6 -10 kb were recovered by sucrose gradient centrifugation. A YRR2-1 genomic DNA library was constructed in the pRS315 vector (LEU2 marker) and used to transform the MLC30 strain to obtain about 60,000 Leu ϩ colonies. The Leu ϩ transformants were replica-plated onto a YPD plate containing 0.5 g/ml RM-A. After 3 days of incubation at 28°C, RM-A-resistant transformants were isolated. Plasmid-born resistance was confirmed with 2 of these transformants by plasmid rescue and retransformation into the MLC30 strain (designated pYM2-1 and pYM2-2). Restriction enzyme mapping and Southern blot analyses showed that the two plasmids contained a 4-kb fragment in common (data not shown).
In Vitro Enzyme Assay-Exponentially growing cells of WT or YRR2-1 strain were harvested by centrifugation for 10 min at 5000 ϫ g, washed twice with 100 mM Tris-HCl buffer (pH 8.0), 10 mM MgCl 2 , and 1 mM dithiothreitol, and resuspended in the same buffer (1 ml/g wet cells) containing 1 mM phenylmethylsulfonyl fluoride. The suspension was kept on ice for 10 min, and cells were disrupted by repeated vortexing for 30 s (ten times) with 1 g/ml glass beads (425-600-m diameter, Sigma). The suspension was centrifuged for 15 min at 10,000 ϫ g. The supernatant was analyzed for aminoacyl-tRNA synthetase activity by aminoacylation of tRNA as previously described (7,8). The protein concentration of the extract was measured by protein assay (BioRad). Aminoacyl-tRNA synthetase activities were assayed in a reaction mixture containing cell extract (5 g of protein), 20 mM imidazole-HCl (pH 7.5), 75 mM KCl, 5 mM MgCl 2 , 0.5 mM dithiothreitol, 4 mg/ml bovine serum albumin, 30 M amino acids, 3.5 mg/ml total yeast tRNA (Roche Molecular Biochemicals), 3 mM ATP, and 1 Ci of either H]methionine (80 Ci/mmol), respectively, in a total volume of 50 l. All radioisotopes were purchased from Amersham Biosciences. Each sample was pretreated with RM-A at the indicated concentration for 10 min and then amino acids, yeast tRNA, and ATP were added and further incubated for 20 min at 25°C. Charged tRNA was precipitated with cold 5% trichloroacetic acid , washed three times with 5% trichloroacetic acid, and counted in a liquid scintillation counter.
Expression of ILS1 in Escherichia coli and Assay for Charging Activity-The expression plasmids for E. coli, pET-21b-ILS1, pET21-b-ILS1(G646A) and pET21-b-ILS1(N660D) were constructed by using a T7 promoter vector, pET-21b. Cells of BL21(DE3) containing the plasmids were grown in LB broth at 28°C and recombinant Ils1 proteins were expressed by the addition of 0.2 mM of isopropyl-thio-␤-galactopyranoside. The cells were harvested and disrupted by sonication in 50 mM Tris-HCl buffer (pH 8.0), 1 mM EGTA, 25% sucrose, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM dithiothreitol. The supernatant was obtained by centrifugation, and isoleucyl-tRNA synthetase activity was assayed using the crude protein extracts at 10 g/ml. Extracts of cells containing vector alone showed no aminoacylation activity toward the yeast tRNA.

RESULTS
Physiological Effect of RM-A on Yeast-The RM-A treatment of mammalian cells blocks both protein synthesis and cell growth, accumulating G 1 cells (1). 2 We first investigated the effect of RM-A on protein synthesis in yeast by assaying the incorporation of L-[ 35 S]methionine into proteins in the presence of the drug. The cells were treated with RM-A (0.6 g/ml) for various periods of time at a concentration that inhibits cell growth. The immediate cessation of protein synthesis by the drug treatment suggested that RM-A is a potent inhibitor of protein synthesis (data not shown). We next examined the effect of RM-A on cell cycle progression by fluorescence-activated cell sorting analysis and budding indexes. The results of fluorescence-activated cell sorting analysis showed that the cells with 1C DNA content accumulated when the cells were cultivated with RM-A ( Fig. 2A). Consistent with this data, the population of the unbudded cells after 6 h of RM-A treatment increased from 20 to 57% (Fig. 2B). These results indicated that RM-A treatment of yeast cells blocks the cell cycle in G 1 .

Isolation and Genetic Analysis of Dominant RM-A-Resistant
Mutants-If the interaction between the drug and its cellular target is impaired by mutation, it would lead to a dominant phenotype specifically resistant to the drug. Based on this assumption, to dissect the mode of RM-A action in yeast, we performed a screen for the mutants that exhibited dominant RM-A resistance. Our previous studies in S. cerevisiae have demonstrated that the YRS1/YOR1 gene encoding an ATPbinding cassette superfamily drug transporter protein and the YRR1 gene encoding a Zn(II) 2 Cys 6 -type transcription factor are important for RM-A resistance, and that a dominant YRR1 mutation caused the elevation of the YOR1 transcription leading to RM-A resistance (2)(3)(4). Two homologous genes, PDR1 and PDR3, encode major transcription factors that regulate the expression of various genes that function in the multidrug resistance (for review, see Ref. 9). Some mutations in YRR1, PDR1, and PDR3 genes confer dominant drug resistance through the overexpression of the drug transporters (3, 9 -11).
To eliminate the resistance that arises by mutations in the previously characterized genes implicated in multidrug resistance, we constructed a parental strain in which YRR1, PDR1, and PDR3 genes, together with the YOR1 gene, had been deleted. The growth of the ⌬yrr1 ⌬pdr1 ⌬pdr3 ⌬yor1 quadruplex deletion mutant (MLC30) was hypersensitive to RM-A (Fig. 3A). To isolate spontaneous RM-A-resistant mutants, exponentially growing cells were spread on YPD agar plates containing 0.7 g/ml RM-A at a cell density of about 1 ϫ 10 7 cells per plate. Colonies formed after 3 days of incubation at 28°C were picked up, each strain was crossed with WT haploid strain, and the resulting diploids were tested for RM-A resistance. Of the 12 mutants isolated, six exhibited a dominant phenotype for the resistance. By tetrad analysis of the spores derived from these six diploid strains, five strains exhibited a 2:2 segregation pattern for RM-A resistance, indicating a single mutation. Complementation analysis was performed by crossing among the haploid strains. The results of tetrad analysis indicated that all five mutants belong to a single complementation group designated YRR2 (for yeast reveromycin A resistance 2). The resistance conferred by the YRR2 mutations was highly specific to RM-A and exhibited no significant crossresistance to various other drugs examined, which included cycloheximide and tautomycin ( Fig. 3B and data not shown).
Cloning and Identification of the YRR2-1 Mutant Gene-To investigate the role of the YRR2 gene product in RM-A drug action, we cloned and characterized the YRR2 gene. For gene cloning, we first tested if the YRR2 dominant mutants may exhibit a recessive phenotype useful for cloning from a WT genomic library (e.g. high-or low-temperature sensitivity). However, because no appropriate phenotype was found, we decided to isolate the YRR2 mutant gene (YRR2-1) from a mutant genomic library constructed with YCp50 vector, and this library was used to transform the MLC30 strain. The transformants were selected for the ability to grow on a plate containing 0.7 g/ml RM-A. Two plasmids designated pYM2-1 and pYM2-2 were recovered. Partial sequencing of an overlapping fragment of the inserts revealed that they contained AUT7 and ILS1 genes and the YBL077W open reading frame in common. Subcloning of the fragments indicated that the ILS1 gene, which encodes IleRS belonging to class I aminoacyl-tRNA synthetases, is responsible for the resistance (12). Linkage of the YRR2-1 mutation to the ILS1 locus was verified by tetrad analysis of the spores derived from the diploid produced by crossing WT strain and an integrative transformant with a LEU2 marker at the ILS1 locus (data not shown). The wild type YRR2 allele was screened from a YCp50-based genomic DNA library of the W303-1A strain by the colony hybridization method. By comparing the amino acid sequences of the mutant IleRS with that of WT, two mutations, G646A and N660D, were revealed near the KMSKS consensus sequence commonly found in the class I aminoacyl transferases (13). Of these mu-tations, N660D was found to be responsible for the RM-A resistance by site-directed mutagenesis (Fig. 4A). The Asn-660 residue is conserved between yeast and human IleRS (14). Disruption of the ILS1 gene led to a lethal phenotype as previously revealed by the whole genome approach of yeast (15) (data not shown). Consistent with the notion that IleRS is a target of RM-A action, overexpression of ILS1 (WT) from a 2-based high-copy number plasmid conferred strong RM-A resistance (Fig. 4B).
IleRS Is Sensitive to RM-A in Vitro-Because IleRS was suggested to be an in vivo target for RM-A, we next examined if IleRS is sensitive to RM-A in vitro. Cell-free extract prepared from WT strain was assayed for the activity to catalyze the transfer of 3 H-labeled isoleucine onto tRNA in the presence of various concentrations of RM-A (0, 10, 100, and 1,000 ng/ml). For comparison, the activities of several other aminoacyl-tRNA synthetases were assayed using radiolabeled leucine, valine, methionine, glutamic acid, and phenylalanine as substrate. Aminoacyl synthetases for methionine, leucine, and valine belong to type I aminoacyl synthetases. As shown in Fig. 5A, the IleRS activity decreased to about 20% of control in the presence of 10 ng/ml RM-A, a concentration almost two orders of magnitude below that for the growth inhibition. In contrast, the aminoacyl transferases for leucine, valine, methionine, glutamic acid, and phenylalanine were virtually insensitive to RM-A. These results indicated that IleRS, among various other aminoacyl-tRNA synthetases, is extremely sensitive to RM-A. The considerable difference in the inhibitory concentrations in vivo (growth inhibition) and in vitro (IleRS inhibition) is likely to reflect the poor permeability of the ionizing drug RM-A across the membrane (1).
The YRR2-1 Mutation Causes IleRS with Increased RM-A Resistance-Next, we compared RM-A sensitivity of IleRS from WT and YRR2-1 strains in vitro (Fig. 5B). In the absence of the inhibitor, the specific activities of IleRS from the two strains were comparable (data not shown). However, IleRS from the

FIG. 4. Effect of ILS1 mutations (A) and overexpression of the ILS1 gene from 2 plasmid (B) on RM-A resistance.
A, The RM-A resistance of G646A and N660D mutant alleles was compared with that of WT and YRR2-1 alleles. The growth of the MLC-30 strain transformed with the low-copy plasmid pRS315 containing WT, YRR2-1, G646A, or N660D allele of the ILS1 gene was compared on a plate containing 0.6 g/ml RM-A. B, the YEp24 plasmid with or without the ILS1 gene was introduced into the W303-1A strain and tested for RM-A resistance. The train W303, instead of the RM-A-hypersensitive MLC30 strain, was used as recipient strain because an appropriate selective marker for the plasmid was not available with the MLC30 strain. mutant was more resistant to RM-A than that from the WT strain. The IC 50 values for IleRS from the WT and YRR2-1 strains were 8 and 32 ng/ml, respectively, showing that the mutant enzyme was 4-fold more resistant to the inhibitor. It was confirmed by the in vitro assay that IleRS N660D , but not IleRS G646A , mutation was responsible for the RM-A resistance (Fig. 5B).
Defect of GCN2 Leads to RM-A Sensitivity-Previously, we isolated and characterized the yeast mutants that exhibited increased RM-A sensitivity (2). This study identified the genetic complementation groups yrs2, -3, and -4 in addition to the previously characterized gene yrs1 that encodes a multidrug resistance-associated protein-type ATP-binding cassette transporter responsible for RM-A resistance (2). 3 Of these RM-Asensitive mutants, the yrs2 mutation was identified as a mutant allele of the GCN2 gene encoding a protein kinase that positively regulates GCN4 expression. The GCN4 gene encodes a transcriptional activator for various amino acid biosynthetic genes implicated in general amino acid control in response to starvation for single amino acids (for review, see Ref. 5). In fact, the ⌬gcn2 strain exhibited increased RM-A sensitivity (Fig.  6A). Moreover, in accordance with the idea that ILS1 is a downstream component of GCN2, a gene contained in a lowcopy suppressor of the RM-A sensitivity of the yrs2/gcn2 mutation was identified as the ILS1 gene (data not shown). Furthermore, Northern blot analysis of ILS1 mRNA revealed that the ILS1 expression level was elevated during RM-A treatment in a partially GCN2-dependent manner (Fig. 6B). However, the ILS1 gene expression in the ⌬gcn2 strain was still inducible by RM-A, predicting an additional ILS1 activation mechanism(s) in response to the RM-A stress (Fig. 6B). These results indicated that RM-A treatment causes the transcriptional activation of the ILS1 gene through the GCN2-GCN4 pathway and that the increased IleRS activity by the feedback mechanism contributes to the resistance to RM-A significantly. DISCUSSION The genetic and biochemical approaches on the yeast RM-A resistance have revealed that IleRS is a target of RM-A drug action. The physiological effects of RM-A on yeast, such as the inhibition of growth and protein synthesis, can be explained by the essential role of IleRS for growth and protein synthesis. Of several aminoacyl-tRNA synthetases examined in vitro, only IleRS was highly sensitive to RM-A. Furthermore, both the dominant YRR2-1 mutation and overexpression of ILS1 (WT) from a high-copy plasmid conferred resistance specifically to RM-A, and only the ILS1 mutant alleles were found among the dominant RM-A-resistant mutations. Based on these results, we concluded that IleRS is a major in vivo target of RM-A action in yeast. The molecular basis for the extreme specificity of RM-A to IleRS still remains to be clarified. Under starvation conditions, when nutrients are limited, S. cerevisiae cells accumulate in G 1 (16,17) (for review, see Ref. 18). When nutrients are limiting, haploid yeast cells do not proceed to START in late G 1 , but instead exit the mitotic cell cycle in early G 1 and enter a stationary or G 0 phase. The cell-cycle-specific growth inhibition may be the result of the cessation of protein synthesis and thereby leads to physiological changes characteristic of starved cells entering stationary phase (G 0 ). In fact, we found by reciprocal shift experiments that the RM-A restriction point was in early G 1 before the ␣-factor arrest point (data not shown). A similar mode of growth inhibition was reported with rapamycin, which ultimately blocks translational initiation (19). As suggested for rapamycin action, the cell-cycle-specific growth inhibition by RM-A may be elicited by the decreased translation of an unstable protein(s) essential for G 1 progression, such as G 1 cyclins, which are unstable and limiting in G 1 (19). The entry into the stationary phase enables cells to maintain viability for long periods when nutrients are not available.
Our genetic approach to the mode of RM-A action in yeast revealed that the GCN2 gene plays a role in RM-A resistance. The expression of the ILS1 gene is under the control of the GCN2-GCN4 pathway implicated in general amino acid control that responds to starvation for single amino acids in a manner dependent on Gcn2 (for review, see Ref. 5). A simple interpretation for the involvement of Gcn2 in RM-A resistance is that RM-A treatment leads to a decreased level of charged tRNA Ile , which in turn activates Gcn2 kinase (probably via the Gcn2p HisRS-like domain sensing uncharged tRNAs), and thereby induces the ILS1 expression via the GCN2-GCN4 pathway. In fact, the transcription of the ILS1 gene was activated in response to RM-A stress (Fig. 6, A and B). Gcn2 phosphorylates the ␣ subunit of the translation initiation factor-2 (eIF2-␣) leading to the inhibition of protein synthesis and thereby mediates translational control of GCN4 (20). The feedback regulation mediated by the GCN2-GCN4 pathway is reminiscent of the autoregulatory model described for the yeast GCD5/KRS1 gene, which encodes lysyl-tRNA synthetase. In this model, the gcd5-1 mutation causes a reduced charging of tRNA Lys , and this starvation signal activates the mechanism that compensates for the defect of lysyl-tRNA synthetase by increasing GCN4 expression at the translational level, which in turn stimulates transcription of GCD5/KRS1 and leads to increased lysyl-tRNA synthetase activity (8). The autoregulation of ILS1 expression by a similar mechanism would lead to increased RM-A resistance through the elevation of the cellular abundance of the drug target molecule. It was previously demonstrated that the temperature-sensitive Ils1-1 allele causes derepression of the general control system in amino acid-complete medium at a non-permissive temperature (21). The gcn2 mutants are sensitive to analogs of various amino acids, indicating that similar autoregulation is also inducible in response to depletion of other amino acids caused by the analogs (22).
In this study, the dominant mutants specifically resistant to RM-A were the key to the identification of the drug target molecule. Dominant multidrug resistance is known to arise in the genes that encode Pdr1, Pdr3, and Yrr1 transcription factors in the presence of drugs (3, 9 -11). In this regard, the use of the genetic background in which various genes implicated in nonspecific multidrug resistance had been eliminated was effective, as indicated by the result that all five mutations that exhibited a dominant resistance phenotype fell into a single genetic complementation group YRR2/ILS1. Analogous approaches would be useful in dissecting the mode of drug action with other drugs of biological or pharmacological interest. For gene cloning, we constructed and screened a genomic library of the YRR2-1 mutant because the appropriate recessive phenotype for screening from the WT genomic library was not available with the mutation. However, if a recessive phenotype, such as low-or high-temperature sensitivity, that facilitates gene cloning using a WT genomic library is linked to the dominant mutation, this approach would become more useful for the identification of a drug target molecule.