Molecular Genetics and Genomics

, Volume 275, Issue 2, pp 114–124

Role of the iron mobilization and oxidative stress regulons in the genomic response of yeast to hydroxyurea

Authors

  • Caroline Dubacq
    • Service de Biochimie et de Génétique Moléculaire,
    • Ecole Normale Supérieure
  • Anne Chevalier
    • Service de Biochimie et de Génétique Moléculaire,
  • Régis Courbeyrette
    • Service de Biochimie et de Génétique Moléculaire,
  • Cyrille Petat
    • Service de Génomique Fonctionnelle du CEA
  • Xavier Gidrol
    • Service de Génomique Fonctionnelle du CEA
    • Service de Biochimie et de Génétique Moléculaire,
    • Department of Biochemistry and Molecular Biology, F. Edward Hébert School of MedicineUSUHS
Original Paper

DOI: 10.1007/s00438-005-0077-5

Cite this article as:
Dubacq, C., Chevalier, A., Courbeyrette, R. et al. Mol Genet Genomics (2006) 275: 114. doi:10.1007/s00438-005-0077-5

Abstract

Hydroxyurea (HU) is a specific inhibitor of ribonucleotide reductase and thus impairs dNTP synthesis and DNA replication. The long-term transcriptional response of yeast cells to hydroxyurea was investigated using DNA microarrays containing all yeast coding sequences. We show that the redox-responsive Yap regulon and the iron-mobilization Aft regulon are activated in yeast cells treated with HU. Yap1 accumulates in the nucleus in response to HU, but HU activation of the Yap regulon was only partially dependent on Yap1 and yap1Δ mutants were not hypersensitive to HU. In contrast, deletion of the AFT1 and AFT2 transcription factor genes blocked the HU activation of a subset of the Aft regulon and the aft1Δ aft2Δ double mutant was hypersensitive to HU in an iron-suppressible manner. These results highlight the importance of the redox and iron mobilization regulons in the cellular response to HU.

Keywords

DNA damage responseHydroxyureaAft1 and Yap1 transcription factorsRibonucleotide reductaseYeast

Abbreviations

HU

Hydroxyurea

RNR

Ribonucleotide reductase

DDR

DNA Damage Response

ESR

Environmental stress response

RT-PCR

Reverse transcriptase polymerase chain reaction

Introduction

Ribonucleotide reductase (RNR) furnishes precursors for DNA synthesis by catalyzing the reduction of ribonucleoside diphosphates into deoxyribonucleoside diphosphates (Eklund et al. 2001). Eukaryotic RNRs are hetero-tetrameric enzymes containing two large subunits and two small subunits. The large subunit (R1) contains both catalytic and allosteric sites, whereas the small subunit (R2) contains an oxo-diferric iron center that is required for the formation of a catalytically essential tyrosyl-free radical (Eklund et al. 2001). Given its essential role in DNA synthesis and repair, small molecule inhibitors of RNR might be effective drugs to treat cancers. Several types of inhibitor are being actively developed. One type involves iron chelators that may inhibit the formation of the oxo-diferric center on the small subunit (Le and Richardson 2002). These chelators are expected to have pleiotropic effects on cells since many enzymes use iron as cofactors, but some studies suggest that RNR activity is preferentially sensitive to their action (Furukawa et al. 1992; Nyholm et al. 1993a). Another type is radical scavengers that quench the tyrosyl radical of R2. Hydroxyurea (HU) is the best-characterized RNR inhibitor of this class (Eklund et al. 2001). Although high concentrations of HU are required for RNR inhibition, its only essential target in the cell appears to be RNR because overexpression or mutation of R2 confers increased cellular resistance to HU (Sneeden and Loeb 2004).

The cellular response to HU has been most extensively studied in Saccharomyces cerevisiae. Inhibition of RNR and dNTP synthesis by HU triggers an S-phase checkpoint response that leads to transcriptional activation of RNR genes, the stabilization of stalled replication forks, and the inhibition of cell-cycle progression (Osborn et al. 2002). The screening of deletion mutants of all non-essential genes has also identified numerous genes that contribute to cellular tolerance to HU in as yet unknown ways (Parsons et al. 2004). SNF1 is an example of a non-essential gene that is required for wild-type levels of resistance to HU (Dubacq et al. 2004). The Snf1 kinase is a member of the AMP-activated kinase family that responds to various stresses, but its exact role in promoting HU tolerance has not been determined. Ctk1, a non-essential kinase that phosphorylates the C-terminal heptapeptide repeats of the large subunit of RNA polymerase II, is also required for cellular resistance to HU (Ostapenko and Solomon 2003). A recent analysis of the transcriptome of yeast cells treated for short periods of time with HU uncovered a role for the Ctk1 kinase in the transcriptional regulation of a subset of HU-induced genes (Ostapenko and Solomon 2003). We were interested in identifying genes that contribute to the cellular resistance to HU during long-term exposure to the drug, as might occur during cancer chemotherapy. We thus carried out transcriptional profiling of yeast cells treated for up to 6 h with high levels of HU.

Materials and methods

Strains and media

The yeast strains used in this work are listed in Table 1. Standard genetic methods were followed and yeast cultures were grown in yeast extract, bactopeptone, 2% dextrose (YPD), or synthetic complete media. Synthetic medium with defined iron concentrations was prepared from YNB without iron (Q Biogene) plus 2% (w/v) dextrose plus all required auxotrophies, and was supplemented with the desired concentration of FeCl3.
Table 1

Strains used in this study

Strain

Genotype

Reference

BY4741

MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0

Giaever et al. (2002)

BY4742

MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0

Giaever et al. (2002)

MCY2649

MATα his3Δ200 leu2-3,112 ura3-52

Vincent et al. (2001)

MCY2916

MATα his3 leu2 lys2 ura3 snf1Δ10

Vyas et al. (2001)

W303-1a

MATa can1-100 ade2-1 his3-11,15 leu2-3,112 trp1-1 ura3-1

Allen et al. (1994)

CM3260

MATα trp1-63 leu2-3,112 gcn4-101 his3-609

Blaiseau et al. (2001)

CM3260Δaft2

MATα trp1-63 leu2-3,112 gcn4-101 his3-609 aft2::KanMX4

Blaiseau et al. (2001)

Y18

MATα trp1-63 leu2-3,112 gcn4-101 his3-609 aft1::TRP1

Blaiseau et al. (2001)

Y18Δaft2

MATα trp1-63 leu2-3,112 gcn4-101 his3-609 aft1::TRP1 aft2::KanMX4

Blaiseau et al. (2001)

The respective backgrounds of these strains are: S288C (MCY2649, MCY2916), W303 (W303-1a), CM3260 (Y18), and BY (BY4741, BY4742)

Quantitative reverse transcriptase-polymerase chain reactions (RT-PCR)

RNA was prepared from cells with an RNeasy extraction kit (Qiagen). RNA was converted to cDNA by reverse transcription with random hexanucleotide primers and 1 μg of total RNA, 1 mM dNTPs, and 1 μl RNasin for 1.5 h at 42°C, followed by a 5 min heat inactivation at 95°C. For each gene, real-time quantitative RT-PCR amplification was performed with an ABI Prism 7,000 according to the manufacturer’s directions.

cDNA microarrays

Strains were grown to mid-log phase in rich medium and submitted to the indicated treatments. Total RNA was isolated using RNeasy extraction kit (Qiagen) and cDNA was synthesized and labeled using the protocol previously described (Fauchon et al. 2002) for the W303 strain with or without HU, or using an indirect protocol with incorporation of aminoallyl-dUTP during cDNA synthesis followed by coupling to CyDye mono-reactive Cy5 or Cy3 (Amersham; see GEO accession numbers GSE1915, GSE1941, GSE1942, and GSE1943 for details). DNA microarrays were prepared by the Service de Génomique Fonctionnelle, CEA/Evry, and data capture and analysis were performed with GenePix and GeneSpring software as described (Fauchon et al. 2002). The cut off for significant spots required that 70% of the pixels be more intense than the background plus twice the standard deviation in at least one channel. Each set of RNA samples was hybridized to two different microarrays performed with an inversion of the fluorochromes to avoid incorporation bias. With these criteria, 4,000–5,000 coding sequences of the approximately 6,200 sequences on the microarrays yielded significant signals depending on the particular experiment and hybridization condition. The complete data set for the arrays was deposited in the GEO database (http://www.ncbi.nlm.nih.gov/projects/geo/; accession numbers GSE1915, GSE1941, GSE1942, and GSE1943).

Fluorescence microscopy

Cells transformed with a YAP1–GFP plasmid (2 μ, LEU2, generously provided by Agnès Delaunay) were cultivated to O.D.600=0.5–0.6 in a synthetic medium lacking leucine, centrifuged and resuspended in PBS prior to observation using a Leica DMRXA fluorescence microscope.

Results

Transcriptional profiling of yeast cells treated with HU

We were interested in the yeast transcriptional response during mid- to long-term exposure to the ribonucleotide reductase inhibitor hydroxyurea. To this end, we carried out competitive hybridizations on microarrays containing all yeast coding sequences of cDNA synthesized from cells grown in the presence or absence of 200 mM hydroxyurea for 1–6 h. At this HU concentration, yeast cells accumulated in early S phase within 2 h and remained in S phase through the 6 h time course (data not shown). Furthermore, cell viability was greater than 90% after 5 h of HU treatment (Dubacq et al. 2004). The expression of 1,735 genes changed at least twofold after 1, 2, 4 or 6 h of growth in the presence of hydroxyurea compared to untreated controls (Fig. 1a). Among those genes, we observed hydroxyurea-inducible genes (855 genes induced at least twofold; 288 at least threefold), hydroxyurea-repressible genes (805 genes repressed at least twofold; 211 at least threefold) and 75 genes whose expression was both induced and repressed by hydroxyurea depending on the duration of the treatment.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig1_HTML.gif
Fig. 1

Changes in genomic expression following long-term exposure to hydroxyurea. a Overview of the expression of 1,735 genes that showed at least twofold variation in levels when MCY2649 cells (S288C genetic background) were incubated with 200 mM hydroxyurea from 1 to 6 h. b Overview of the Environmental Stress Response (ESR) during HU treatment. The expression pattern during HU exposure of the genes that are induced (left panel) or repressed (right panel) in response to environmental stress is shown. c Venn diagrams comparing the ESR (top and right circles) and HU (left) genomic responses for genes showing at least a twofold variation in expression in response to HU. d Venn diagrams showing the overlap of HU-induced genes with ESR-induced genes and HU-repressed genes with ESR-repressed genes for genes showing at least a threefold variation in expression in response to HU. A color scale showing the intensity of the expression ratios is at the left of the figure

The genomic response to hydroxyurea is partly a non-specific stress response

Gasch and collaborators have defined an environmental stress response (ESR) regulon including genes induced or repressed by a variety of environmental stresses (Gasch et al. 2000). We found that about 57% of the ESR-induced genes and 38% of the ESR-repressed genes were amongst the HU inducible or repressible genes, respectively (Fig. 1b, c). To further test the specificity of this response, we limited our analysis to genes that were induced or repressed at least threefold in the presence of HU and compared these to the ESR regulon. 76/288 (26%) of these HU-induced genes were ESR-induced and 42/219 (19%) of these HU-repressed genes were also ESR-repressed (Fig. 1d). Thus, a part of the genomic response to hydroxyurea involves the ESR, as previously shown for exposure to other genotoxic stresses such as methyl methane sulfonate (an alkylating agent) or ionizing radiation (Gasch et al. 2001). However, most genes whose expression is modified during long-term exposure to hydroxyurea do not belong to the ESR suggesting a significant specificity of the HU-induced expression profile. Among the most highly repressed genes following HU addition (at least twofold at three time points or more, or at least threefold at two time points or more), we found genes implicated in cell wall biogenesis, cytokinesis, bud site selection and cell polarity, regulation of mitosis, chromatin structure, and transcriptional regulation (Table 2). Many of these genes are normally transcribed outside of the S phase, so their repression may be due to the S-phase arrest of the HU-treated cells. Repression of histone gene expression was also observed, as expected from previous work (Lycan et al. 1987).
Table 2

Genes repressed during long-term exposure to hydroxyurea

Function

Genes

Cell wall biogenesis

DSE1, DSE2, CWP1, PST1

Cytokinesis

EGT2, CTS1, SCW11, DSE4

Bud site selection, cellpolarity

MSB2, BUD9, AXL2

Regulation of mitosis

AMN1, NIS1

Chromatin structure

HTA2, HTB1, HTA1, SFH1, IOC4, NET1, RAP1, HST3, SGF73

Transcription regulation

SWI5, TUP1, REB1, SET2, SPT6, LEO1

Other functions

Drug response: PDR5; DNA topology: TOF2; RNA helicase: DBP2; Meiotic chromosome segregation: CSM4; rRNA/ribosome processing: UTP10, RRP8, MAK16, NOP15; mRNA degradation: PUF3; Protein modification: PSA1, CDC5, UBP12; Protein synthesis: CBP6, RML2; Protein nucleus import: MLP1; Pseudo-hyphal growth: TEC1, STE12, ASH1; Purine base metabolism: ADE13, MIS1; Conjugation: YNL174W; SPB duplication: KAR1; Thiamin transport: PHO3; translation: TIF4631; vacuole fusion: VTC2; aromatic compound catabolism: PAD1

Unknown function

YOR263C, YGL101W, DSE3, YMR193C-A, YDR428C, YIL130W, YIL158W, YOR315W, YJL051W, YNL057W, YNL058C, YNL087W, YNR066C, YER184C

Wild-type strain MCY2649 was treated with 200 mM HU for 1,2,4 or 6 h and gene expression changes relative to untreated cells were analyzed by competitive hybridization to microarrays containing yeast coding sequences. The genes whose expression was repressed at least twofold at three time points or more, or at least threefold at two time points or more, are listed

HU-induced genes

We were most interested in genes whose expression was increased by HU (Fig. 2), since they could be involved in drug tolerance. Given that the widest genomic response to HU was observed 2 h after the drug addition, we chose to focus on this time point by comparing two genetic backgrounds. The comparison between the W303 and the S288C backgrounds revealed that only part of the genomic changes are common to both strains, and we limited our analysis to the 126 genes induced at least twofold by HU in both genetic contexts. Among those, we found several DNA damage response (DDR) genes (Fig. 2a) that were anticipated from previous work: DUN1, RAD51, RNR2, RNR4, YER004W (Gasch et al. 2001), MIG3 (Dubacq et al. 2004), RNR3 (Elledge and Davis 1990), and DDR48 (Treger and McEntee 1990). Consistent with the analysis of hydroxyurea long-term effects presented earlier, we found that 41 of these HU-induced genes were also induced during the ESR (Fig. 2b). In response to DNA damage provoked by MMS or IR, induction of ESR genes was shown to be largely under the control of the Mec1 DNA checkpoint kinase (Gasch et al. 2001). Since Mec1 is implicated in the S-phase checkpoints (Weinert et al. 1994), the ESR genes activated by HU treatment may also be dependent on Mec1. We thus sought to identify regulons other than the DDR or the ESR that might contribute to HU tolerance.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig2_HTML.gif
Fig. 2

Classification of HU-induced genes. HU-induced genes were partitioned into: a DNA-damage induced genes, b ESR-induced genes, or c genes that were neither induced by DNA damage Nor ESR, but fell into functional categories involving stress response, transport, iron homeostasis, and protein degradation, and d all others. These genes were induced twofold or more in both the S288C (strain MCY2649) and W303 (strain W303-1a) genetic backgrounds following a 2 h HU treatment

Identification of novel regulons induced by HU

After excluding the DDR and ESR genes, we noticed that many of the remaining HU-induced genes were implicated in redox and iron homeostasis, or in various transport functions (Fig. 2c, d). Many genes in these functional categories belong to the Yap and Aft regulons. The Yap1 transcription factor is activated following oxidative stress, or in the presence of some drugs or electrophilic compounds, and controls the expression of genes involved in cellular redox homeostasis and in drug detoxification and transport (Gounalaki and Thireos 1994; Lee et al. 1999; Toone et al. 2001; Azevedo et al. 2003). We defined the Yap regulon as the panel of genes whose expression was affected by deletion of YAP1 during exposure to hydrogen peroxide (Lee et al. 1999; Gasch et al. 2000). Aft1 and Aft2 are partially redundant transcription factors that are implicated in the transcriptional activation of genes involved in iron transport and homeostasis in response to iron limitation (Yamaguchi-Iwai et al. 1995; Blaiseau et al. 2001; Rutherford et al. 2001; Shakoury-Elizeh et al. 2004). We collected data from published papers to define the Aft regulon (Foury and Talibi 2001; Rutherford et al. 2001; Stadler and Schweyen 2002; Shakoury-Elizeh et al. 2004). We found that many of the genes belonging to the Yap and Aft regulons are induced after treating cells with HU (Fig. 3a, b), suggesting that the respective transcription factors could be activated in the presence of HU and participate in HU-induced modifications of gene expression.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig3_HTML.gif
Fig. 3

Overview of changes in the expression of genes belonging to the Yap regulon (a), or the Aft regulon (b) in the presence of hydroxyurea. The expression pattern is shown for the MCY2649 strain (S288C) during 1–6 h of HU exposure (left) or the W303-1a strain (W303) 2 h after addition of 200 mM HU (center). The two rightmost columns show the expression of genes in yap1Δ (a) or aft1Δ (b) mutants relative to the wild type after growth in the presence of 200 mM HU for 2 h. The two columns represent duplicate experiments using isogenic BY4741 and BY4742 strains, respectively. Bold: genes showing twofold or greater increases in mRNA levels for at least one condition of HU exposure. A color scale showing the intensity of the expression ratios is at the left of the figure. Gray blocks indicate data points in which signals were not statistically significant relative to background

Activation of the Yap regulon by HU

Sixty-four genes of the Yap regulon had statistically significant hybridization signals in our experiments. Of these, 42 were induced by HU in at least one of the conditions examined (S288C or W303 genetic context after 1-6 h of HU treatment) (Fig. 3a, bold) and 17 were induced by HU in both strains after a 2 h treatment. We confirmed the increased expression of certain Yap1 targets genes during HU exposure by RT-PCR analyses. The expression changes occurred early (TRX2) or late (PRX1) after HU addition (Fig. 4a).
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig4_HTML.gif
Fig. 4

The Yap1 transcription factor contributes to changes in genomic expression in response to HU. aTRX2 and PRX1 genes showed increased expression in the strain MCY2649 in the presence of 200 mM HU as determined by quantitative RT-PCR analyses. The relative amount of mRNA for the gene of interest was normalized to ACT1 mRNA content and expression in WT without HU was normalized to 1 (arbitrary unit). b Expression of some genes is dependent on Yap1 after 2 h of exposure to 200 mM HU. Gene expression was estimated by competitive hybridization of cDNA from BY4741 yap1Δ or BY4742 yap1Δ strains and cDNA from the isogenic wild-type strain after growth for 2 h in the presence of 200 mM HU. Only genes showing twofold or greater lower expression in yap1Δ versus wild type are shown. Genes earlier identified in the Yap1 regulon are shown bold. c Localization of the Yap1-GFP protein during hydroxyurea treatment. Fluorescent microscopy was performed on wild-type (MCY2649) cells expressing Yap1–GFP after addition of 200 mM HU or 0.3 mM H2O2 to the growth medium. Quantification of the percentage of cells showing mostly nuclear localization of Yap1–GFP is shown below as representative images of the cells

To address the involvement of the Yap1 transcription factor in the HU-induced expression of these genes, a global analysis of gene expression was performed with yap1Δ and wild-type strains during HU exposure (Fig. 3a, yap1Δ). We found that 10/41 of the HU-induced genes belonging to the Yap regulon were repressed twofold or more in the yap1Δ mutant treated with HU (Fig. 4b, bold). Thus, the HU induction of these genes is probably under the direct control of the Yap1 transcription factor. However, many of the HU-induced genes of the Yap regulon were repressed less than twofold following YAP1 deletion (Fig. 3a). Moreover, the yap1Δ null mutant was not hypersensitive to HU (data not shown). We suggest that another functionally redundant transcription factor may participate in the activation of the Yap regulon in response to HU in yap1Δ mutants.

HSP31 is a gene whose induction by HU is dependent on Yap1 (Fig. 4b), although it is not a known target gene of Yap1 in response to oxidative stress. We also identified three genes (CYC1, TRR1, and ROX1) whose expression was repressed in yap1Δ compared to the wild type during HU exposure, but whose expression was not modified in wild-type cells treated with HU. The basal expression of these three genes during normal growth may depend in part on Yap1.

Yap1 is mainly cytosolic in yeast cells under normal growth conditions, but it accumulates in the nucleus and activates the Yap1 regulon in response to oxidative stress, or when cells are treated with electrophilic compounds or with some drugs (Azevedo et al. 2003). We monitored the subcellular localization of a Yap1–GFP fusion protein after treating cells with 200 mM HU in comparison with the same cells treated with 0.3 mM H2O2 (Fig. 4c). As previously reported, Yap1–GFP rapidly and transiently accumulated in the nucleus after treating cells with these low levels of H2O2 (Delaunay et al. 2000). The transient nature of this response is due to the rapid detoxification of H2O2 within cells. Yap1–GFP also accumulated in the nucleus in response to HU, but with slower kinetics relative to H2O2. Nuclear accumulation of Yap1–GFP was observed in 50% of the cells after 1 h of incubation with 200 mM HU and was sustained for the 3 h duration of the experiment in the presence of the drug. The induced nuclear accumulation of Yap1 further confirms a role for Yap1 in the cellular response to HU.

Aft1 and Aft2 are required for tolerance to hydroxyurea

To test the role of the Aft1 and Aft2 transcription factors in the HU-induced expression of the Aft regulon (Fig. 3b), we compared the transcriptomes of aft1Δ, aft1Δ aft2Δ, and wild-type cells in the presence of 200 mM HU for 2 h. The expression of a subset of the Aft regulon was affected in the aft1Δ single mutant or aft1Δ aft2Δ double mutant strains during HU treatment (Figs. 3b, 5a). Combining the results of microarray analysis and RT-PCR, we observed distinct transcriptional regulation of the genes included in the Aft regulon (Figs. 5a, b). On the one hand, the basal expression of SMF3, encoding a vacuolar iron transporter, was independent of Aft1 and Aft2, whereas its HU-induced expression was abolished in an aft1Δ aft2Δ mutant. On the other hand, the basal and HU-induced expression of ENB1, encoding an iron siderophore transporter, was greatly dependent on Aft1. The global transcriptome analyses of the aft1Δ and aft1Δaft2Δ mutant strains suggest that the Aft regulon can be split into two parts: genes whose expression is dependent on Aft1 and Aft2 only in response to stress (HU or iron deprivation), and genes whose basal expression is dependent on Aft1 or Aft2 basal activity. SMF3 seems to be regulated by both Aft1 and Aft2, whereas ENB1 appears to be controlled only by Aft1 (Rutherford et al. 2001; Shakoury-Elizeh et al. 2004), suggesting that the expression of these genes following hydroxyurea treatment might reflect their relative dependence on Aft1 and Aft2.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig5_HTML.gif
Fig. 5

The Aft1 and Aft2 transcription factors contribute to changes in genomic expression in response to HU. a Expression of some genes is dependent on Aft1 or Aft1 and Aft2 during exposure to hydroxyurea. Gene expression was determined by competitive hybridization of cDNA from wild type with the mutant strain Y18 aft1Δ aft2Δ after growth in the presence of 200 mM HU for 2 h. Only genes showing twofold or greater repression in the mutants relative to the wild type are shown. b HU-induced expression of SMF3 expression is dependent on Aft1 and Aft2, but not its basal expression, whereas the basal and HU-induced expression of ENB1 is mostly dependent on Aft1 as determined by quantitative RT-PCR on WT (CM3260) aft1Δ (Y18), aft2Δ (CM3260Δaft2) or aft1Δ aft2Δ (Y18Δaft2) strains, before or after addition of 200 mM HU. The relative amount of mRNA for the gene of interest is normalized versus ACT1 mRNA, and expression in WT without HU was normalized to 1 (arbitrary units). c Levels of SMF3 and ENB1 mRNA relative to ACT1 mRNA were determined by real-time RT-PCR on wild type (MCY2649) or snf1Δ (MCY2916) strains grown in the presence or absence of 200 mM HU. The relative amount of mRNA for the gene of interest is normalized versus ACT1 mRNA and expression in WT without HU was normalized to 1 (arbitrary unit)

To further test the involvement of the Aft1 and Aft2 transcription factors in the cellular response to hydroxyurea, we performed phenotypic analyses of single or double AFT1 and AFT2 deletion mutants. We observed no phenotype for the aft1Δ or the aft2Δ single mutants in a rich YPD medium with or without HU (Fig. 6a), but the aft1Δaft2Δ double mutant was sensitive to HU (Fig. 6a) and could not grow on an iron-deprived synthetic medium (Fig. 6b). In these low-iron conditions, the aft1Δ single mutant was also sensitive to HU, but this sensitivity was suppressed by the addition of excess iron to the medium, as was the HU sensitivity of the aft1Δ aft2Δ double mutant (Fig. 6b). Thus, Aft1 and Aft2 are required for optimal HU tolerance under conditions of iron limitation.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig6_HTML.gif
Fig. 6

AFT1 and AFT2 are required for iron mobilization during the cellular response to hydroxyurea. Strains CM3260 (WT), Y18 (aft1Δ), CM3260Δaft2 (aft2Δ), and Y18Δaft2 (aft1Δaft2Δ) were grown to log phase in YPD and serial one-tenth dilutions were spotted on YPD with or without 100 μM FeCl3 (a) or iron-deprived synthetic medium with 0.1 μM or 100 μM FeCl3 (b), and with the indicated concentrations of hydroxyurea

A limited role for the Snf1 kinase in the transcriptional response to HU

We previously demonstrated a role for the Snf1 kinase in the cellular resistance to HU, but its mechanism of action was unclear (Dubacq et al. 2004). Snf1 has been implicated in transcriptional regulation, so we carried out further transcription profiling to compare the response of snf1Δ and wild-type cells to HU. We identified 95 genes that were repressed at least threefold and 118 that were induced at least threefold in the snf1Δ mutant compared to the wild type during a 1–6 h exposure to HU. We focused on those genes that showed at least twofold changes in expression at multiple time points in the experiment (Fig. 7). Only some of the genes that showed differential expression in the HU-treated snf1Δ mutant compared to the HU-treated wild type also proved to be induced or repressed (twofold change or more) in the wild type upon treatment with HU. Altogether, our results suggest that very few HU-induced genes are under Snf1 control. Snf1 has been implicated in the regulation of both Yap1 and Aft1 (Haurie et al. 2003; Wiatrowski and Carlson 2003); however, we did not observe any defects in the nuclear accumulation of Yap1 (data not shown) or in the induction of the Yap1 regulon in the snf1Δ mutant treated with HU. Furthermore, an snf1Δ mutant had no significant effect on the transcriptional activation of SMF3 or ENB1 in response to HU (Fig. 5c), although the expression of the SIT1 gene, a known member of the Aft regulon, was dependent on the Snf1 kinase (Fig. 7). SIT1 encodes an iron siderophore transporter, but in contrast to the aft1 mutant, HU sensitivity of snf1Δ was not suppressed by the addition of excess iron to the medium (data not shown). The HU sensitivity of the snf1Δ mutant thus cannot be traced to obvious defects in its transcriptional response to HU.
https://static-content.springer.com/image/art%3A10.1007%2Fs00438-005-0077-5/MediaObjects/438_2005_77_Fig7_HTML.gif
Fig. 7

Wild type (MCY2649) and snf1Δ (MCY2916) strains were treated with 200 mM HU for 1, 2, 4 or 6 h and gene expression changes were analyzed by microarray competitive hybridization for each time point. Genes whose expression was repressed or induced in snf1Δ compared to wild type at least twofold at three time points or more are listed. The expression ratios during HU treatment of the snf1Δ mutant compared to the wild type (four leftmost columns), or in the wild type strain treated with HU compared to untreated cells (four central columns), are shown according to the color scale at the bottom of the figure (grey: no data)

Discussion

We used microarrays containing all yeast coding sequences to study the long-term transcriptional response of yeast cells to treatment with the ribonucleotide reductase inhibitor hydroxyurea. In addition to the expected activation of the DNA damage response (DDR) and environmental stress response (ESR) regulons, we observed activation of the Yap1 and Aft regulons. The Yap1 regulon is activated upon treatment of yeast cells with oxidative agents or electrophiles (Azevedo et al. 2003). Two distinct pathways lead to Yap1 nuclear accumulation. H2O2 induces a specific disulfide bond between Cys-303 and Cys-589 that is catalyzed by the hydrogen peroxide sensor Gpx3 (Delaunay et al. 2002). In contrast, thiol-reactive electrophiles, such as alkylating agents and some heavy metals, are believed to directly modify C-terminal cysteines of Yap1 independently of Gpx3 (Azevedo et al. 2003). In both cases, nuclear export of Yap1 is inhibited. Accumulation of Yap1 in the nucleus contributes to activation of the Yap regulon encoding proteins implicated in redox homeostasis, in drug and metal detoxification, and in multidrug efflux from the cell (Gounalaki and Thireos 1994; Toone et al. 2001). Hydroxyurea is not an electrophile and is unlikely to react directly with thiols. The slow kinetics of Yap1 nuclear accumulation in the presence of HU is also consistent with an indirect action of HU leading to Yap1 nuclear accumulation. HU blocks DNA synthesis by inhibiting ribonucleotide reductase through a single electron transfer reaction that quenches a tyrosyl radical contained within the small subunit of the enzyme (Eklund et al. 2001). This reaction would also generate a hydroperoxy radical form of HU that could diffuse away from the reduced RNR and act as an oxidizing agent. However, we found that Yap1–GFP still accumulated in the nucleus in a gpx3Δ mutant in response to HU, as did a Yap1–C598A–GFP mutant in a wild-type strain treated with HU (data not shown). Thus, Yap1 nuclear accumulation in response to HU does not occur through the activation pathway used by H2O2. It remains possible that hydroperoxy radical forms of HU directly or indirectly modify Yap1 cysteines.

The production of hydroperoxy radical forms of HU may constitute an oxidative stress that would be combated by activation of the Yap regulon. Unexpectedly, deletion of YAP1 only partially blocked HU activation of the Yap regulon, and yap1Δ mutants were not hypersensitive to HU. These results suggest that other transcription factors participate in this activation. Candidates include the seven other Yap1-like proteins encoded within the yeast genome, but further experimental work is necessary to test their possible involvement (Fernandes et al. 1997).

In addition to quenching its tyrosyl radical, treatment of mammalian RNR with HU also leads to reduction and loss of the oxo-diiron center from the small subunit (Nyholm et al. 1993b). This diferric iron center is required for regenerating a tyrosyl radical within R2. Cells have poorly studied pathways for assembling this oxo-diiron center, but it is known that two ferrous ions must first bind to apo R2. Active R2 is then formed by reaction with oxygen to give an oxo-diferric iron center and a tyrosyl radical (Eklund et al. 2001). We found that treatment of yeast cells with HU leads to activation of the Aft regulon that is implicated in iron mobilization in response to iron deprivation. Treatment of yeast cells with HU induces the transcription of the RNR2 and RNR4 genes encoding RNR small subunits. Rnr2 binds iron, but Rnr4 lacks three conserved iron-binding residues and does not itself bind iron. Instead, it forms heterodimers with Rnr2 and acts as a cofactor for Rnr2 assembly and activity (Chabes et al. 2000; Sommerhalter et al. 2004). Assembly of iron into newly synthesized Rnr2 subunits, and possibly the reassembly of diferric iron centers in old Rnr2 subunits attacked by HU, could lead to depletion of cytosolic iron and activation of the Aft regulon in order to mobilize iron from the vacuole (Rouault and Tong 2005). It is also possible that a specific chaperone involved in efficient assembly of the Rnr2 diferric iron center is expressed as a part of the Aft regulon. Strikingly, deletion of AFT1 and AFT2, encoding the transcriptional factors required for activation of the Aft regulon, sensitized yeast cells to HU in a manner that could be suppressed by the addition of excess iron to the growth medium. This result illustrates the physiological importance of the Aft regulon in the cellular response to HU.

Among the 126 genes induced in both S288C and W303 genetic contexts by a 2 h treatment with 200 mM HU, we identified 26 genes belonging to known regulons (Crt, Aft, Yap) and 32 additional genes induced during the environmental stress response. Thus, there remain 68 HU-induced genes that have not yet been assigned to known regulons and whose eventual roles in the cellular response to HU remain to be elucidated.

Acknowledgements

We thank P.-L. Blaiseau, A. Dancis, A. Delaunay, Y. Pereira, M. Toledano, and Y. Yamaguchi-Iwai for providing yeast strains, plasmids, and discussions. We are grateful to M. Werner for access to the microarray analysis platform and for general support. C.D. was supported by an Allocation Couplée de l’Ecole Normale Supérieure de Paris. This work was also supported by grants from the Association pour la Recherche sur le Cancer (ARC number 4470).

Copyright information

© Springer-Verlag 2005