HIRA, a Conserved Histone Chaperone, Plays an Essential Role in Low-dose Stress Response via Transcriptional Stimulation in Fission Yeast*

Background: HIRA is a conserved histone chaperone required for regulation of chromatin structure. Results: Genes that encode HIRA proteins are responsible for cross-tolerance. Specifically, stress-responsive gene expression was most profoundly compromised in HIRA disruptants. Conclusion: HIRA is involved in cross-tolerance via regulation of stress-responsive gene expression. Significance: This study provides evidence that fission yeast HIRA functions in stress response. Cells that have been pre-exposed to mild stress (priming stress) acquire transient resistance to subsequent severe stress even under different combinations of stresses. This phenomenon is called cross-tolerance. Although it has been reported that cross-tolerance occurs in many organisms, the molecular basis is not clear yet. Here, we identified slm9+ as a responsible gene for the cross-tolerance in the fission yeast Schizosaccharomyces pombe. Slm9 is a homolog of mammalian HIRA histone chaperone. HIRA forms a conserved complex and gene disruption of other HIRA complex components, Hip1, Hip3, and Hip4, also yielded a cross-tolerance-defective phenotype, indicating that the fission yeast HIRA is involved in the cross-tolerance as a complex. We also revealed that Slm9 was recruited to the stress-responsive gene loci upon stress treatment in an Atf1-dependent manner. The expression of stress-responsive genes under stress conditions was compromised in HIRA disruptants. Consistent with this, Pol II recruitment and nucleosome eviction at these gene loci were impaired in slm9Δ cells. Furthermore, we found that the priming stress enhanced the expression of stress-responsive genes in wild-type cells that were exposed to the severe stress. These observations suggest that HIRA functions in stress response through transcriptional regulation.

Cells are equipped with stress response mechanisms at various levels in order to survive and proliferate under ever-changing environmental stresses. Cross-tolerance is one of such stress response mechanisms. Cells that have been pre-exposed to mild stress (priming stress) are known to acquire transient resistance to subsequent severe stress. If the two stresses are of the same type, the phenomenon is called acquired tolerance. It is also known that this increased survival happens even under combinations of different types of stresses, such as heat stress and oxidative stress. This phenomenon is called cross-tolerance. It has been reported that acquired tolerance and crosstolerance occur in a wide variety of species, including bacteria, plants, yeasts, and mammals (1)(2)(3)(4)(5)(6)(7)(8).
Hormesis is a widely accepted term that more comprehensively describes cross-tolerance (9,10). This phenomenon represents a biphasic dose response to toxins and stressors, with beneficial effects at low doses and harmful ones at high doses. Recent studies have provided new insights into hormesis as an application in anti-aging research (11,12). Thus, an understanding of the response to low-dose stress is important. However, generally it is difficult to detect the response to low-dose stress because the low-dose stress does not induce a significant phenotype. Considering that the response to priming stress is important for survival under subsequent severe stress, the analysis of cross-tolerance is expected to lead to further understanding of the response mechanism to low-dose stress.
In the fission yeast Schizosaccharomyces pombe, it is well known that a wide range of stresses lead to the activation of stress-activated mitogen-activated protein kinase (MAPK) Spc1/Sty1. The inactivation of this kinase causes hypersensitivities to various stresses (13)(14)(15)(16). There are common stress-responsive genes called core environmental stress response (CESR) 3 genes whose expression is induced more than 2-fold under at least four of five types of stress conditions examined (17). CESR genes were regulated predominantly by Spc1 via the transcription factor Atf1. It has been proposed that the crosstolerance depends on nascent protein synthesis (7) and requires the induction of CESR genes (17). However, the molecular mechanism of the cross-tolerance remains unclear.
Chromatin structure should be highly regulated in many cellular processes, such as DNA replication, repair, or transcription. Accumulating evidence has shown that histone chaper-ones are one of the key proteins involved in those processes (18). Histone chaperones are known to associate with histones and facilitate the assembly and disassembly of nucleosomes. HIRA/HIR is one of the major histone chaperones that are conserved in many eukaryotic organisms (19). Whereas higher eukaryotes have a single HIRA protein (19 -22), the fission yeast possesses two HIRA proteins (Slm9 and Hip1) (23,24), same as the budding yeast Saccharomyces cerevisiae (Hir1 and Hir2) (25). Fission yeast HIRA proteins stably associate with two other proteins, Hip3 and Hip4, and form a tetrameric complex (HIRA complex) (26,27). Recently, Cabin1 and UBN1 were identified as the human counterparts of Hip3 and Hip4, respectively (28 -30), suggesting that the HIRA complex is evolutionarily conserved. HIRA is the histone chaperone for histone H3-H4 and is involved in the replication-independent nucleosome deposition pathway, whereas another histone chaperone CAF-1 is coupled to DNA replication (31)(32)(33)(34).
HIRA has been shown to function in transcription as well. HIRA proteins were first identified in the budding yeast as a negative regulator of histone gene expression (25,35). It has been reported that the budding yeast HIR complex interacts with nucleosomes and prevents the remodeling activity of the SWI/SNF complex (36). The ectopic expression of HIRA in human cells also represses the transcription of histones (37). In the fission yeast, HIRA is required for the suppression of Tf2 long terminal repeat retrotransposons, normally repressed genes, or cryptic antisense transcripts (38). Consistent with its repressive role in transcription, HIRA also functions in heterochromatin assembly and silencing. In human cells, the formation of senescence-associated heterochromatin foci depends on HIRA (39). Loss of the fission yeast HIRA complex components results in silencing defects at the centromere and mating type loci (27). A recent study has also demonstrated that a complex formed by the histone chaperone Asf1 and HIRA spreads across silenced regions via its association with the chromodomain protein Swi6 to facilitate histone deacetylation and heterochromatin spreading in the fission yeast (40). On the other hand, HIRA can also act as a positive regulator of transcription. The N-terminal and C-terminal halves of chicken HIRA regulate different sets of cell-cycle-related genes positively and negatively, respectively (41). Mutations in the budding yeast HIR genes display strong synthetic defects or lethality when combined with mutations in the genes encoding the transcription elongation factor FACT components (42). In higher eukaryotes, HIRA is involved in the incorporation of H3.3 variant histones into transcriptionally active genes (33,43,44). However, it is not clear whether HIRA is involved in transcriptional activation in the fission yeast.
In this study, we found that the fission yeast slm9 ϩ is responsible for the cross-tolerance. The disruption of each component gene of the HIRA complex led to defects in the cross-tolerance. In wild-type cells, Slm9 was located at several stress-responsive gene loci under the stress condition, and this localization is dependent on Atf1. HIRA disruption caused impaired stressresponsive gene expression, stress-dependent RNA polymerase II (Pol II) recruitment, and nucleosome eviction. Moreover, it was suggested that the priming stress facilitates stress-responsive gene expression in wild-type cells under the severe stress.
Together, these results highlight the novel function of the fission yeast HIRA in stress response.

EXPERIMENTAL PROCEDURES
Yeast Strains and General Techniques-S. pombe strains used in this study are listed in Table 1. Growth media and basic techniques for the fission yeast have been described previously (45,46).
Stress Experiments-For the cross-tolerance and acquired tolerance experiments, cells were grown in duplicate to the logarithmic phase in YES medium at 32°C. Two cultures each were subjected and not subjected to the priming stress, respectively, for 1 h and centrifuged gently (780 ϫ g for 1 min) to remove the medium. Subsequently, both cultures were resuspended in YES medium, and severe stress was applied for 1 h. The stress conditions are described below. (P) and (S) indicate priming stress and severe stress, respectively. For oxidative stress, H 2 O 2 was added to make a final concentration of 0.1 mM (P) or 25 mM (S). For heat stress, cells were cultured in a 40°C (P) or 46°C (S) water bath. For osmotic stress, YES medium containing 2.4 M KCl was added to make a final concentration of 0.6 M (P) or 2.4 M (S). After the above stress treatment, the cells were immediately collected by gentle centrifugation (400 ϫ g for 2 min) and diluted with YES medium. Five hundred cells per plate were plated onto YES plates, and the number of colonies was counted after incubation for 4 days at 32°C. Viability was calculated as the percentage of the number of colonies for 500 cells.
For the colony-spotting assay, the cells were grown to the logarithmic phase in YES medium at 32°C. Then, the cells were serially diluted from 5 ϫ 10 6 to 5 ϫ 10 3 cells/ml (10-fold dilution), and 5 l of each was spotted onto YES plates. Incubation was carried out for 3-4 days at 32°C. For heat stress, cells that were spotted onto YES plates were incubated for 3-4 days at 37°C or for 1 h at 47°C followed by incubation for 3-4 days at 32°C. For the other stresses, YES plates containing the following compounds were used: 2 mM H 2 O 2 , 50 M menadione, 1.2 M KCl, 2 M sorbitol, and 0.1 M CaCl 2 .
Chromatin Immunoprecipitation-The cells were grown in duplicate to the logarithmic phase in YES medium at 32°C. One aliquot was used as the unstressed control, and the other aliquot was exposed to 40°C for 15 min. Subsequently, the cells (2.5 ϳ 4ϫ10 8 cells) were cross-linked by adding 1% formaldehyde for 30 min at 25°C, and the cross-linking was stopped by treating with 125 mM glycine on ice for 5 min. The cell pellets were washed twice with ice-cold water and twice with lysis buffer 1 (50 mM HEPES-KOH (pH 7.5), 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate). The cell pellets were resuspended in lysis buffer 1 containing 50 mM NaF, 0.1 mM Na 3 VO 4 , 1 mM PMSF, and 1ϫ Complete (Roche Applied Science) and broken with zirconia beads using a Multi-Beads Shocker (Yasui Kikai) at 4°C. The lysates were sonicated with Sonifier 250 (Branson) to yield chromatin fragments having an average size of 500 bp. The sonicated lysates were spun at 17,800 ϫ g for 15 min at 4°C. The supernatant was immunoprecipitated with mouse anti-Myc antibody (sc-40 Santa Cruz) or mouse anti-RNA polymerase II CTD antibody (05-623 Millipore) for 2 h at 4°C, and this was followed by the addition of magnetic beads (Dynal). After incubation for 1.5 h at 4°C, the beads were washed once with lysis buffer 1 containing 50 mM NaF, 0.1 mM Na 3 VO 4 , 1 mM PMSF, and 1ϫComplete (Roche Applied Science); once with lysis buffer 1 containing 500 mM NaCl and 1 mM PMSF; once with lysis buffer 2 (10 mM Tris-HCl (pH 8.0), 1 mM EDTA, 0.25 M LiCl, 0.5% Nonidet P-40, 0.5% sodium deoxycholate) and twice with TE (10 mM Tris-HCl (pH 8.0), 1 mM EDTA). The beads were resuspended in TE containing 10 g/ml RNase and incubated for 15 min at 37°C. The samples were adjusted to 0.25% SDS and 250 g/ml proteinase K and incubated at 37°C overnight. This was followed by another incubation for 6 h at 65°C. The eluted DNA was subjected to phenol/chloroform extraction and precipitated with ethanol. The purified DNA was analyzed by real-time PCR using the StepOnePlus TM Real-Time PCR System and Power SYBR Green PCR master mix (Applied Biosystems). The nucleotide sequences of the primer sets are listed in supplemental Table S1.
RT-PCR-Total RNA was isolated as previously described (47) and treated with 0.625 units of RNase-free DNase I (TaKaRa)/g of RNA to digest genomic DNA. cDNA samples were synthesized using AMV Reverse Transcriptase (Life Sciences Advanced Technologies, Inc.) and Random Primer (nonadeoxyribonucleotide mixture). The primer sequences for PCR are available on request. Real-time PCR was performed using the StepOnePlus TM Real-Time PCR System and Power SYBR Green PCR master mix (Applied Biosystems). The nucleotide sequences of the primer sets for real-time PCR are listed in supplemental Table S1.
Microarray Analysis-The cells were grown in quadruplicate at 32°C to the logarithmic phase, and an aliquot was collected as the unstressed control. The other three aliquots were exposed to 40°C for 1 h, 25 mM H 2 O 2 for 1 h, or 40°C for 1 h followed by 25 mM H 2 O 2 for 1 h, respectively. Total RNA was purified as described for RT-PCR. All the 12 RNA samples were analyzed with GeneChip Yeast Genome 2.0 Array (Affymetrix) according to the manufacturer's instructions. After using the RMA algorithm to obtain the summarized probeset-level expression data, the array data were transferred to GeneSpring 7.3 (Agilent Technologies) microarray analysis software for gene ontology (GO) analysis. Standard hypergeometric distribution was used to calculate the p values for individual GO terms. Significant enrichment of GO was selected using a p value of Ͻ0.05. To avoid the detection of false positives, the Benjamini-Yekutieli correction method was applied to obtain the corrected p values. The microarray data are available at Gene Expression Omnibus (GEO) under accession number GSE35281.
Preparation of Mononucleosomal DNA-The cells were grown in duplicate to the early logarithmic phase in YES medium at 32°C. One aliquot was used as the unstressed control, and the other aliquot was exposed to 40°C for 15 min. Subsequently, mononucleosomal DNA was obtained as described previously (48) with some modifications. Cell wall was digested with 1 mg/ml Zymolyase 100T (Seikagaku Corp.) for 40 min at 35°C. Micrococcal nuclease digestion was performed with a final concentration of 133 units/ml micrococcal nuclease. Mononucleosomal DNA fragments were purified from the agarose gel using QIAquick Gel Extraction Kit (Qiagen).
Nucleosome-scanning Analysis-The nucleosome-scanning analysis was performed as described previously (49,50). Genomic DNA was obtained from the same protocol as for the preparation of mononucleosomal DNA, without the cross-linking, micrococcal nuclease treatment, and gel purification step. Five nanograms of purified mononucleosomal and genomic DNA were analyzed by real-time PCR using the StepOne-Plus TM Real-Time PCR System and Power SYBR Green PCR master mix (Applied Biosystems). Ten overlapping primer pairs were set downstream of nucleosome-depleted region (50,51). The nucleotide sequences of the primer sets were designed by reference to the previous report (50) as listed in supplemental Table S1.

Cross-tolerance in Fission
Yeast-We first confirmed crosstolerance in the fission yeast. Wild-type (JK317) cells were treated or not treated with mild (priming) stress and subsequently subjected to severe stress using combinations of heat, oxidative, and osmotic stresses, and cell viability was compared ( Fig. 1). In contrast to the case of the budding yeast (7), the combination of mild oxidative stress and severe heat stress

Fission Yeast HIRA in Stress Response
induced cross-tolerance. Other combinations examined also induced cross-tolerance. Overall, we conclude that in the fission yeast, various combinations of stresses generally induce cross-tolerance.
Identification of Responsible Gene of Cross-tolerance-defective Mutant-To identify the factor involved in cross-tolerance, we performed a genetic screen for cross-tolerance-defective mutants and isolated several mutants. 4 Among them, 7-4 mutant, which showed clearly the cross-tolerance-defective phenotype, was chosen for further analysis (supplemental Fig.  S1A). 7-4 mutant was backcrossed three times with the parental wild-type strain. Tetrad analysis revealed that the cross-tolerance-defective phenotype of 7-4 mutant was caused by a single mutation. In addition to the cross-tolerance-defective phenotype, we found that 7-4 mutant showed strong sensitivity to heat shock (37°C) and 0.1 M CaCl 2 treatment (supplemental Fig. S1B). The responsible gene of 7-4 mutant was cloned by complementation of heat and CaCl 2 sensitivities with S. pombe genomic library (pTN-L1), and subsequent sequencing revealed its identity as slm9 ϩ . We verified that a single nucleotide deletion occurred at nucleotide 1451 in the ORF of slm9 ϩ in 7-4 mutant (supplemental Fig. S1C). Slm9 is the homolog of mammalian HIRA in the fission yeast. HIRA is a histone chap-erone that is involved in the replication-independent nucleosome deposition pathway and transcriptional control.
HIRA Complex Is Involved in Cross-tolerance-The fission yeast has two HIRA/HIR proteins, Slm9 and Hip1 (23,24). To verify the function of the fission yeast HIRA in cross-tolerance, we constructed slm9⌬ and hip1⌬ strains and examined the cross-tolerance phenotype. Both slm9⌬ and hip1⌬ cells showed the cross-tolerance-defective phenotype under the combination of mild heat stress and severe oxidative stress ( Fig. 2A). As Slm9 and Hip1 form a complex with two other proteins, Hip3 and Hip4 (26,27), we also examined these HIRA complex subunit disruptants. We found that both hip3⌬ and hip4⌬ cells showed the cross-tolerance-defective phenotype, similar to slm9⌬ and hip1⌬ cells (Fig. 2B). These results suggest that the fission yeast HIRA functions as a complex to confer cross-tolerance.
To examine the possibility that HIRA disruptants are specifically sensitive to the combination of mild heat stress and severe oxidative stress, the cells were treated with other stress combinations. slm9⌬ and hip1⌬ cells also showed the cross-tolerance-defective phenotype when treated with combinations of mild oxidative stress and severe heat stress, and mild osmotic stress and severe oxidative stress (Fig. 2C). These experiments demonstrated that the fission yeast HIRA is involved in crosstolerance regardless of the stress combination. Moreover, the 4 Y. Tarumoto, J. Kanoh, and F. Ishikawa, submitted for publication.  JULY 6, 2012 • VOLUME 287 • NUMBER 28

JOURNAL OF BIOLOGICAL CHEMISTRY 23443
cells were also treated with combinations of same types of stresses, namely, mild oxidative stress and severe oxidative stress (acquired tolerance-inducible stress). Acquired tolerance was defective in slm9⌬ and hip1⌬ cells as well as cross-tolerance (Fig. 2D). Hereafter, we used the combinations of mild heat stress and severe oxidative stress as the cross-toleranceinducible stress.
To determine whether the function of HIRA in stress response is a general feature of histone chaperones, we analyzed other histone chaperone mutants, pcf1⌬ and nap1⌬. Pcf1 is a large subunit of histone chaperone CAF-1 that loads histone H3-H4 onto DNA and is involved in the replication-dependent nucleosome deposition pathway (34,52). Nap1 is involved in the transfer of histone H2A-H2B from the cytoplasm to the nucleus and the deposition of histones onto DNA (53). Crosstolerance occurred in both pcf1⌬ and nap1⌬ cells, similar to the case of wild-type cells (Fig. 2E). These results raise an interesting possibility that among histone chaperones, HIRA is specifically involved in cross-tolerance.
HIRA Functions Particularly under Low-dose Stress-As we examined the cross-tolerance phenotype under various combinations of heat, oxidative, and osmotic stresses, we next investigated the viability of cells lacking each subunit of the HIRA complex under the single stress condition (Fig. 3). The HIRA subunit disruptants (slm9⌬, hip1⌬, hip3⌬, and hip4⌬) were not so sensitive to osmotic stress (2 M sorbitol and 1.2 M KCl) and oxidative stress caused by 2 mM H 2 O 2 , whereas spc1⌬ mutant, which is known to be sensitive to a wide variety of stresses, showed strong sensitivity. On the other hand, the HIRA subunit disruptants showed severe sensitivity to heat shock (37°C, 3 days) and oxidative stress caused by 50 M menadione. Although the HIRA subunit disruptants showed varied sensitivities to distinct forms of stress, the cross (acquired)-tolerance-defective phenotype of slm9⌬ and hip1⌬ cells was observed under different stress combinations (Fig. 2, A, C, and  D). These results further suggest the stress type-independent function of HIRA in the cross-tolerance. It is known that menadione generates intracellular reactive oxygen species and exerts weak oxidative stress on the cells (54). It should be noted that the HIRA subunit disruptants showed higher sensitivity to menadione than H 2 O 2 . Moreover, they were more sensitive to weak and chronic heat shock (37°C, 3 days) than strong and acute heat shock (47°C, 1 h). These observations suggest that HIRA responds to low-dose stress specifically, consistent with its response to the priming stress in the cross-tolerance.
HIRA Is Localized at Stress-responsive Gene Loci upon Stress Treatment-As the fission yeast HIRA has a function in the cross-tolerance, we examined the protein levels of Slm9 and Hip1 under the stress condition. Considering the protein levels of loading control (Cdc2), it appeared that Hip1 was more abundant than Slm9. The protein levels of both Slm9 and Hip1 did not change significantly during the course of the priming stress treatment (Fig. 4A). We also examined Slm9 localization using a strain whose chromosomal copy of slm9 ϩ was tagged with the GFP sequence. In cells expressing Slm9-GFP, fluorescent signals were observed in the nuclei, as reported previously (23). This nuclear localization was not altered significantly under the stress condition (supplemental Fig. S2A). In addition, the chromatin fractionation assay was performed to determine HIRA localization biochemically (supplemental Fig. S2B). Slm9 and Hip1 were both enriched in the chromatin-bound fractions, and the distributions of Slm9 and Hip1 among different fractions did not change notably under the stress condition (supplemental Fig. S2C).
To further investigate whether HIRA is localized at specific chromatic loci upon stress treatment, the chromatin immunoprecipitation (ChIP) assay was performed. Slm9 was enriched  . HIRA is recruited to stress-responsive gene loci upon stress treatment. A, slm9-12myc (MC3793) and hip1-12myc (MC3795) cells were exposed to 40°C for the indicated times. Whole cell extracts were prepared and analyzed by Western blotting with anti-Myc and anti-Cdc2 (control) antibodies. B, slm9-12myc (MC3793) and slm9-12myc atf1⌬ (MC4219) cells were exposed or not exposed to 40°C for 15 min. ChIP assay using anti-Myc antibody was performed. Purified DNA was analyzed by real-time PCR using primer sets for the promoter (prom) and coding (ORF) regions of stress-responsive genes (ctt1 ϩ , gpx1 ϩ , and hsp9 ϩ ), non-stress-responsive gene (pol1 ϩ ), and heterochromatic locus (outer repeats of centromere, dh). Values shown were normalized to cdc2 ϩ promoter. Results are the means of three independent experiments, and error bars represent S.E..

Fission Yeast HIRA in Stress Response
at both the promoters and ORFs of CESR genes (ctt1 ϩ , gpx1 ϩ , and hsp9 ϩ ) in a stress-dependent manner. However, such physical association of Slm9 was not observed at non-stress responsive loci (pol1 ϩ ORF and dh) (Fig. 4B). Thus, whereas the protein levels and the chromatin association of HIRA, revealed by Western blotting and chromatin fractionation assay, did not show distinct alteration, HIRA localization at chromatin revealed by the ChIP assay changes under the stress condition.
CESR genes were primarily regulated by Spc1 through its downstream b-ZIP transcription factor Atf1 (17). We hypothesized that specific localization of Slm9 at CESR gene loci is determined by Atf1. Indeed, Slm9 recruitment to CESR gene loci is almost totally dependent on Atf1 (Fig. 4B). This result suggests that Atf1 determines the stress-dependent HIRA recruitment to chromatin.
HIRA Is Required for Stress-responsive Gene Transcription-Histone chaperones have been surmised to play important roles in transcriptional regulation (18). As the fission yeast HIRA was localized at the stress-responsive gene loci upon stress treatment, we hypothesized that the fission yeast HIRA complex may play a role in the stress response through the transcriptional control of stress-responsive genes. RT-PCR was performed to examine the expression of several CESR genes under the stress conditions in wild-type, slm9⌬, and hip1⌬ cells. slm9 and hip1 disruption decreased the expression of many CESR genes (ctt1 ϩ , gpx1 ϩ , gpd1 ϩ , and tps1 ϩ ) under the priming stress compared with wild-type cells (Fig. 5A). There were some exceptions, such as hsp9 ϩ and hsp16 ϩ , that showed increased basal expression in the mutants (Fig. 5A). This basal up-regulation of some CESR genes is consistent with previous reports (26,38). In general, all the genes examined showed smaller differences in transcriptional levels between the non-stress condition and the priming stress condition in the mutant cells compared with the wild-type cells. In contrast, slm9 and hip1 disruption increased the expression of those genes under the severe stress condition (supplemental Fig. S3). Taken together, the results suggest that HIRA is required for the proper expression of stress-responsive genes. In addition to the gene expression, ChIP assay was carried out to examine the transcriptional kinetics at several gene loci in the wild-type and slm9⌬ cells under the stress condition. As expected, Pol II was recruited to both promoters and ORFs of CESR genes (ctt1 ϩ , gpx1 ϩ , and hsp9 ϩ ) in the wild-type cells subjected to stress treatment, whereas this recruitment was impaired in slm9⌬ cells. On the other hand, basal stress-independent Pol II recruitment to non-stress responsive loci (pol1 ϩ ORF and dh) was unaffected in slm9⌬ cells (Fig. 5B). This result is consistent with the expression of CESR genes (Fig. 5A). Therefore, our results indicate that HIRA plays an important role in Pol II recruitment and progression, and the transcriptional activation of stress-responsive genes under the low-dose stress conditions.
HIRA Particularly Regulates Stress-responsive Genes under Stress Conditions-To determine whether HIRA specifically regulates stress-responsive genes or not, microarray analysis was carried out on wild-type, slm9⌬, and hip1⌬ cells under four conditions: control, priming stress alone, priming stress followed by severe stress, and severe stress alone. Signal concor-dance between two arrays was evaluated using Pearson's correlation coefficient (r 2 ). A strong correlation (r 2 Ͼ 0.998) was noted between slm9⌬ and hip1⌬ samples under all conditions (supplemental Fig. S4), consistent with the previously reported strong correlation of gene expression between slm9⌬ and hip1⌬ cells under the normal conditions (38).
We identified genes that exhibited 2-fold or greater change in the stress-treated samples compared with the control and categorized them by GO classification. The GO terms that were enriched in induced and repressed genes under all conditions are listed in supplemental Tables S2 and S3, respectively. Among these results, we focused on the priming stress condition because the response to the priming stress would be important for survival under the subsequent severe stress. The principal GO terms that are most significantly associated with the priming-stress-induced genes are selected from supplemental Table S2 and are shown in Table 2. GO analysis identified "Cellular response to stress," "Meiosis," and "M phase" as the major enriched biological functions in both wild-type cells and mutants. Whereas the p values of "Meiosis" and "M phase" were similar between the wild-type cells and the mutants, the number of genes enriched into "Cellular response to stress" was much larger in the wild-type cells than in the mutants. In addition, GO terms, including "Cellular response to oxidative stress" and "Oxidoreductase activity," were only found in the wild-type cells. These results are consistent with the reduced CESR gene expression in the mutants under the priming stress FIGURE 5. HIRA is required for stress-responsive gene expression and Pol II recruitment. A, wild-type (JK317), slm9⌬ (MC3749), and hip1⌬ (MC3725) cells were exposed or not exposed to 40°C for 1 h. Total RNA was analyzed by RT-PCR using primer sets for stress-responsive genes (ctt1 ϩ , gpx1 ϩ , gpd1 ϩ , tps1 ϩ , hsp9 ϩ , and hsp16 ϩ ). act1 ϩ is shown as the loading control. B, wild-type (JK317) and slm9⌬ (MC3749) cells were exposed or not exposed to 40°C for 15 min. ChIP assay using anti-Pol II antibodies was performed. Purified DNA was analyzed in the same way as described in Fig. 4B. Results are the means of three independent experiments, and error bars represent S.E.

Fission Yeast HIRA in Stress Response
condition (Fig. 5A). On the contrary, the number of genes enriched into "Cellular response to stress" was much larger in the mutant cells than in the wild-type cells under the severe stress condition, which was again consistent with the results of RT-PCR (supplemental Fig. S3 and Table S2) (see "Discussion").
To further confirm the difference in expression between the wild-type cells and the mutants under the priming stress condition, the fold change of gene expression was plotted under the priming stress condition compared with the control condition. The fold change in expression of all genes was smaller in the mutants than in the wild-type cells (Fig. 6A). Similarly, the fold change in expression of CESR genes and genes whose expression was increased more than 2-fold in the wild-type cells decreased significantly in the mutants (Fig. 6, B and C). More-over, among the priming stress-induced genes of the wild-type cells, we selected genes that showed 2-fold or higher change in the wild-type cells compared with each mutant and performed GO analysis of those genes. GO analysis identified "Cellular response to stress" as the most significant term (Table 3). Thus, although HIRA may be required for global gene expression, it particularly plays an important role in regulating the stressresponsive genes.
HIRA Regulates Stress-responsive Gene Expression via Nucleosome Eviction-To explore the mechanism by which HIRA regulates stress-responsive gene transcription, the nucleosome-scanning analysis of ctt1 ϩ region was performed. In wildtype cells, the position of ϩ1 to ϩ3 (relative to the transcription start site (TSS)) nucleosomes, as reflected by micrococcal nuclease sensitivity, were detected as peaks, and these peaks were diminished upon heat stress (Fig. 7), as previously reported in H 2 O 2 -treated cells (50). Although positioned nucleosomes were also observed in slm9⌬ cells, decreased nucleosome peaks upon stress treatment were not observed in slm9⌬ cells (Fig. 7). Considering that regulatory regions of ctt1 ϩ gene such as TATA box and Atf1 binding site are located in close proximity to the TSS (55) and overlapped with ϩ1 nucleosome position (Fig. 7), it seems that HIRA is required for recruitment of Pol II to these regulatory regions. Furthermore, the position of nucleosomes downstream of the TSS was also altered in slm9⌬ cells (Fig. 7), consistent with the result of the ChIP assay detecting impaired Pol II recruitment to ORFs of CESR genes (Fig. 5B). Taken together, these results suggest that HIRA is required for nucleosome eviction to regulate Pol II accessibility and/or progression, and expression of stress-responsive genes.
Priming Stress Facilitates Expression of Stress-responsive Genes under Subsequent Severe Stress-As HIRA plays a role in the transcriptional control of stress-responsive genes under the stress conditions, we characterized the expression of several CESR genes (ctt1 ϩ , gpx1 ϩ , and hsp9 ϩ ) in the wild-type cells during cross-tolerance by quantitative RT-PCR. We found that the severe stress alone induced only less than a 3-fold increase in CESR gene expression (Fig. 8A). In contrast, when the cells were treated with the priming stress prior to the severe stress, the expression of those genes increased Fold change in expression of all genes (A), CESR genes (B), and genes whose expression was increased more than 2-fold in wild-type cells (C) is shown. Significant difference in expression between wild-type and mutant cells was determined by Student's t test (*, p Ͻ 0.05; **, p Ͻ 0.01).  (Fig. 8A). Furthermore, the fold change relative to the control condition of CESR gene expression in the wildtype cells under the stress conditions was plotted using microarray data. Similar to the results of quantitative RT-PCR (Fig. 8A), the fold change of CESR gene expression under the severe stress condition showed a dramatic increase when the cells were exposed to the priming stress (Fig. 8B). These findings indicate that the priming stress enhanced stress-responsive gene expression, and as a result, the cells acquired resistance to impending stress.

DISCUSSION
We have demonstrated that the fission yeast HIRA complex is involved in cross-tolerance. We also found that in the crosstolerance, the expression levels of stress-responsive genes under the severe stress were augmented when the cells were exposed to the priming stress. Although the fission yeast HIRA has been shown to be implicated in gene silencing and the heterochromatin assembly (26,38,40,56), our results showed that the fission yeast HIRA is required for the transcriptional activation of stress-responsive genes under the low-dose stress conditions. Therefore, HIRA would regulate transcription both positively and negatively.
HIRA disruption decreased stress-responsive gene expression under the priming stress condition (Fig. 5A, Tables 2 and  3). As the expression of stress-responsive genes under the severe stress was enhanced by the priming stress (Fig. 8), the defect of the cross-tolerance in the HIRA disruptants may have come from the impaired expression of those genes during the priming stress. Thus, cells may stimulate stress-responsive gene expression under the low-dose stress conditions to deal with stress and to prepare for future stress, consistent with a previous report of the budding yeast (7). In contrast, severe stress alone increased stress-responsive gene expression in slm9⌬ and hip1⌬ cells (supplemental Fig. S3 and Table S2). Considering that the severe stress alone did not lead to a marked increase of stress-responsive gene expression in the wild-type cells (Fig. 8), we speculate that the induction of those gene expression may be inhibited or occur at only a low level in the wild-type cells under the severe stress conditions. One possible explanation is that the induction of stress-responsive genes for survival would be too late after the high-dose stress, so cells may cease the transcription of those genes under the highdose stress conditions in order not to consume extra, yet futile, energy. However, such regulation may be compromised in HIRA disruptants, and this may lead to the up-regulation of stress-responsive genes. It will be important to study the dosedependent stress response in detail to test if this hypothesis is plausible or not. . HIRA is required for nucleosome eviction at stress-responsive gene locus. Wild-type (JK317) and slm9⌬ (MC3749) cells were exposed or not exposed to 40°C for 15 min. Mononucleosomal DNA was isolated from the cells, and nucleosome-scanning analysis was performed. Real-time PCR was carried out using 10 overlapping primer sets along ctt1 ϩ gene. Nucleosomal DNA enrichment is defined by the ratio of the amplified products with mononucleosomal DNA to genomic DNA. Results are the means of three independent experiments, and error bars represent S.E. Inferred locations of nucleosomes (light gray ovals) with respect to ORF (white rectangle) and TSS (black arrow) are shown. TATA box (Ϫ30 to Ϫ23) and Atf1 binding site (Ϫ57 to Ϫ50) are also represented as black rectangle and dark gray rectangle, respectively (55).

Fission Yeast HIRA in Stress Response
The viability of HIRA disruptants under the severe stress condition was comparable to that of the wild-type cells ( Fig.  2A). Furthermore, HIRA disruptants showed high sensitivity to rather low-dose stress (Fig. 3). Thus, HIRA seems to control the transcription of stress-responsive genes specifically under lowdose stress. The accessibility of the transcription machinery is regulated by chromatin assembly and disassembly (57). Slm9 is recruited to the stress-responsive gene loci in the wild-type cells (Fig. 4B), and stress-responsive gene expression and stress-dependent Pol II enrichment are impaired in HIRA disruptants (Fig. 5). Moreover, nucleosome eviction at ctt1 ϩ gene region upon stress treatment is hampered by slm9 disruption (Fig. 7). Taken together, HIRA likely regulates the transcription of stress-responsive genes through the eviction of histones. In addition, the position of nucleosomes was slightly different between wild-type and slm9⌬ cells (Fig. 7). Although the significance of this displacement is unclear, it may contribute to the accessibility of proteins involved in transcription. Thus, HIRA may also have a role in the regulation of nucleosome position-ing. One recent study has shown that histone H3 acetyltransferase Gcn5 facilitates Pol II progression along stress-responsive genes in fission yeast (50). Therefore, both HIRA and Gcn5 should stimulate the eviction of histones at these loci to allow Pol II to progress, and stress-responsive gene expression is induced by stress-activated MAPK Spc1 and its downstream transcription factor Atf1 to respond to stress.
One possible factor that enhances the recruitment of HIRA and Gcn5 may be the 19S ATPase subunits of the proteasome. Mass spectrometric analysis of the fission yeast Slm9 has led to the identification of the 19S ATPase complex as the interacting proteins (27). The 19S ATPase complex is recruited to chromatin in a histone H2B ubiquitylation-dependent manner and plays a nonproteolytic role in Pol II transcriptional elongation in the budding yeast (58 -60). Moreover, in the fission yeast, histone H2B ubiquitylation is required for transcriptional elongation, and HIRA mutations are synthetically lethal with htb1-K119R, the mutation in the conserved ubiquitin acceptor site of histone H2B, indicating the role of histone H2B ubiquitylation in chromatin assembly during transcription (61). Indeed, the overexpression of ubiquitin-conjugating enzyme gene confers enhanced stress tolerance in plants (62) and RAD6, which encodes ubiquitin-conjugating enzyme in the budding yeast, is involved in the heat shock induction of bleomycin resistance (cross-tolerance) (63). Transcriptional regulation through HIRA recruitment may be required for not only stress response but also other biological responses in general. For instance, a recent study has revealed that Hip3, a component of the HIRA complex, is engaged in the repression of meiosis-specific gene SPCC663.14c expression under the vegetative state (56). Considering that Slm9 recruitment to stress-responsive gene loci depends on Atf1 (Fig. 4B), other factors such as transcription factor, which function together with general factors including HIRA, may determine the specificity of the response. In this way, HIRA may stimulate or repress transcription via mediation of nucleosome states, depending on the situation. Thus, it will be important to reveal the relationship among all these factors as they may cooperatively regulate stress-responsive gene expression.