The Fission Yeast GATA Factor, Gaf1, Modulates Sexual Development via Direct Down-Regulation of ste11+ Expression in Response to Nitrogen Starvation

Gaf1 is the first GATA family zinc-finger transcription factor identified in Schizosaccharomyces pombe. Here, we report that Gaf1 functions as a negatively acting transcription factor of ste11+, delaying the entrance of cells exposed to transient nitrogen starvation into the meiotic cycle. gaf1Δ strains exhibited accelerated G1-arrest upon nitrogen starvation. Moreover, gaf1Δ mutation caused increased mating and sporulation frequency under both nitrogen-starved and unstarved conditions, while overexpression of gaf1+ led to a significant impairment of sporulation. By microarray analysis, we found that approximately 63% (116 genes) of the 183 genes up-regulated in unstarved gaf1Δ cells were nitrogen starvation-responsive genes, and furthermore that 25 genes among the genes up-regulated by gaf1Δ mutation are Ste11 targets (e.g., gpa1 +, ste4 +, spk1 +, ste11 +, and mei2 +). The phenotype caused by gaf1Δ mutation was masked by ste11Δ mutation, indicating that ste11+ is epistatic to gaf1+ with respect to sporulation efficiency, and accordingly that gaf1+ functions upstream of ste11+ in the signaling pathway governing sexual development. gaf1Δ strains showed accelerated ste11+ expression under nitrogen starvation and increased ste11+ expression even under normal conditions. Electrophoretic mobility shift assay analysis demonstrated that Gaf1 specifically binds to the canonical GATA motif (5′-HGATAR-3′) spanning from −371 to −366 in ste11+ promoter. Consequently, Gaf1 provides the prime example for negative regulation of ste11+ transcription through direct binding to a cis-acting motif of its promoter.


Introduction
The fission yeast Schizosaccharomyces pombe reproduces asexually by mitosis under favorable conditions.When haploid cells are starved of nutrients, particularly of nitrogen, they arrest the cell cycle at G 1 and undergo sexual differentiation [1].Cells of opposite mating types, h + and h 2 , fuse to form a diploid zygote, which undergoes meiosis to give four haploid ascospores that remain dormant until they encounter favorable growth conditions [2].The transition from the mitotic cell cycle into meiosis is tightly regulated by a network of positive and negative factors that are controlled at various levels of gene expression, from transcription initiation [3][4][5][6] to protein modification [7][8][9][10][11].
One important regulatory component of S. pombe sexual development is Ste11.Ste11 positively regulates transcription of the mating type genes, matP and matM, and the mei2 + gene, which is essential for commitment to meiosis, by binding to an upstream cis-acting element under conditions of nitrogen starvation [12].ste11D mutants are completely defective in mating and sporulation, while ectopic expression of ste11 + leads to sexual differentiation irrespective of nutritional conditions.The activity of Ste11 is regulated by two antagonistic protein kinases, Pat1 and Spk1, via the pheromone signaling pathway [13] and by the TOR protein kinase, Tor2, which is activated in the presence of nitrogen and represses sexual differentiation by directly interfering with the function of Ste11 and Mei2 [11].Expression of ste11 + is regulated by at least three different signal transduction pathways: mating pheromone signaling (RAS/MAPK pathway), cyclic AMP (cAMP)-dependent protein kinase A (PKA), and stress-activated protein kinase (SAPK) in conjunction with MAPKKKs (Wis4 and Win1), MAPKK (Wis1), and MAPK (Sty1/Spc1/Phh1) [14,15].So far, only positive regulatory factors of ste11 + expression, such as Atf1 [6,16], Pcr1 [5], Rst2 [4], Prr1 [17], and Ste11 itself [4] have been reported.Furthermore, only Rst2 and Ste11 are transcription factors that directly activate ste11 + expression.No transcription factor that directly represses the expression of ste11 + has been identified.
Here, we explore the role of Gaf1, the first GATA transcription factor identified in S. pombe [18,19], in the expression of ste11 + .GATA family transcription factors have a wide range of functions, from terminal differentiation in vertebrates [20][21][22] to nitrogen metabolism, siderophore biosynthesis, photoinduction, and mating type switching in fungi [23].In S. pombe, Ams2 is a cell cycleregulated GATA factor that is required for centromere function [24] and Fep1/Gaf2 occupies a central role in iron homeostasis by coordinating the reductive and non-reductive iron transport systems [25,26].GATA factors recognize a six base-pair consensus sequence, 59-HGATAR-39 (where H can be A/C/T and R can be A/G), contained in the promoters of their target genes.Although the C-terminal fragment of Gaf1 (Gaf1 565-855 ) has been shown to bind specifically to the GATA motif DAL7 UAS, a canonical GATA motif of Saccharomyces cerevisiae [18], little is known about the function of Gaf1 in S. pombe.We present evidence that Gaf1 downregulates the transcription of ste11 + via direct binding to its promoter and consequently delay the shift of nitrogen-starved cells from the vegetative cycle to the meiotic cycle.

Materials and Methods
S. pombe strains, media, and general procedures S. pombe strains used in this study are listed in Table 1.Cells were maintained on complete medium (YES) containing 0.5% yeast extract, 3% glucose, 2% Bacto agar, adenine (225 mg ml 21 ), leucine (225 mg ml 21 ), and uracil (225 mg ml 21 ).Edinburgh minimal medium (EMM2) [27][28][29] was used as a minimal selective medium.EMM-N (EMM2 without NH 4 Cl) was used for starvation of nitrogen source, and EMM-G (EMM2 containing 0.5% instead of 2% glucose) for glucose-restriction experiments.All the minimal media were supplemented with required auxotrophic nutrients (adenine, leucine, and uracil) at the concentrations of 225 mg ml 21 each, which led to the presence of a starved amount of organic nitrogen source in EMM-N.Thiamine was added to the medium at a final concentration of 20 mM to repress expression from the thiamine-repressible nmt42 + (no message in thiamine) promoter.Transformation was performed by the lithium acetate procedure [27].Standard techniques were used for genetic manipulation and analysis [29].

Construction of plasmids
To construct pREP-Gaf1 with a full-length open reading frame (ORF) of gaf1 + downstream of the nmt42 + promoter, a 2.6-kb fragment was amplified by polymerase chain reaction (PCR) with the following primers: P1 (59AACCCGGGCCATGGATC-TAAAGTTTTCC39) and P2 (59AACCCGGGCATAACGCTA-TACCAATC39) in which underlines designate SmaI sites.The resulting PCR product was cloned into pGEM-T (Promega) to yield pGEM-Gaf1.After confirming the absence of PCR artifact by sequence analysis, the gaf1 + ORF was excised by digestion with SmaI and ligated to SmaI-digested middle-copy expression vector pREP42 [30] to yield pREP-Gaf1.For construction of pGEX4T3-Gaf1 encoding the glutathione S-transferase (GST)-tagged Gaf1 (GST-Gaf1), the 2.6-kb gaf1 + ORF was excised from pGEM-Gaf1 by SmaI-digestion and cloned in frame into the SmaI site of pGEX4T3 (Amersham).

Gene disruption
Construction of gaf1D strains was performed by direct chromosomal integration as described previously [28].The 2.6kb genomic regions corresponding to the entire gaf1 + ORF (855 amino acids) of the wild-type strains, 972, ED665, JY4, and ED668, carrying different auxotrophic markers or mating type were replaced with the 1.5-kb PCR-amplified gaf1::kanMX deletion cassette derived from pFA6a-kanMX4 [32].Stable transformants were selected by resistance to G418, and the disruptions were confirmed by PCR with appropriate primers (data not shown) yielding the gaf1D strains, KL230, KL210, KL211, and KL216, respectively.ste11D strain JZ396 [kindly donated by Dr. Yamamoto [12]] was transformed with the gaf1::kanMX deletion cassette to yield the gaf1D ste11D double-deletion mutant (KL213).To construct a gaf1D strain carrying the hphMX marker, the gaf1::kanMX allele of KL210 strain was replaced with the 1.6-kbgaf1::hphMX deletion cassette amplified from pFA6a-hphMX6 [33].Stable transformants were selected by resistance to hygromycin B and sensitivity to G418 (indicating loss of kanMX), and the disruptions were confirmed by PCR (data not shown) to yield strain KL240.

Growth tests
To analyze growth by plate assays, S. pombe cells grown to midexponential phase in EMM2 for 18 h were washed and resuspended in EMM2-N (for nitrogen starvation) or EMM-G (for glucose restriction) to a concentration of 5610 6 cells ml 21 .The cell suspensions were serially diluted in 5-fold steps, and then a 5 ml aliquot of each dilution was spotted onto EMM, EMM-N, and EMM-G plates.The plates were incubated at 30uC and photographed after 3 d.

Preparation of total RNA
For preparation of total RNA, wild-type (ED668) and gaf1D (KL216) strains were grown to mid-log phase in EMM2 for 18 h at 30uC.Cells were harvested from the mid-log phase cultures and used as unstarved cell preparations (+N).To prepare nitrogenstarved cell samples (2N), cells harvested from the mid-log phase cultures in EMM2 were washed with distilled water and shifted to EMM-N.Nitrogen-starved cells used for microarray analysis were harvested 4-h cultivation in EMM-N.For Northern blot analysis, aliquots of the nitrogen-starved culture were removed at intervals, i.e., at 0, 3, 6, 9, and 18 h after shift.The cells were washed twice with distilled water and frozen immediately at 270uC for total RNA preparation.Total RNA samples were extracted from approximately 2610 8 cells using a bead beater as described previously [12].

Microarray analysis
Thirty-microgram total RNA of each sample was further purified using RNeasy (QIAGEN) columns and submitted for microarray analysis by the SeouLin Bioscience (Korea).Probes were generated and hybridized to the GeneChip Yeast Genome 2.0 Array (Affymetrix), and the data were analyzed using GeneSpring GX software (Agilent).

Flow cytometric analysis
For flow cytometry, cells grown to mid-log phase (5610 6 cells ml 21 ) were harvested and fixed in 70% ethanol containing 50 mM sodium citrate overnight at 4uC.After brief centrifugation, cell pellets were washed twice with 1 ml of 50 mM sodium citrate buffer (pH 7) and treated with RNase A (10 mg ml 21 ) at 37uC for 2 h.The cells were stained with propidium iodide (16 mg ml 21 ) and analyzed using a BD FACScalibur Flow Cytometer as described previously [28].Data were analyzed using WinMDI software, version 2.8.

Mating and sporulation assay
In order to monitor the efficiency of sporulation, S. pombe cells grown to mid-log phase were prepared by two successive transfers of young colonies on EMM2 agar plates.The cells were then collected and washed three-times with distilled water.Suspensions of homothallic haploid cells (h 90 ) or mixtures of mating pairs (h + and h 2 ) were spotted in 10 ml aliquots (1.0610 9 cells) onto EMM2 and EMM-N agar plates.Cultures were grown at 30uC, and the cells were observed at intervals by DIC microscopy to determine sporulation frequencies.At least 400 cells from three independent experiments were evaluated, and mating and sporulation frequencies (F M ) were calculated using the following equation [34]: F M (%) = (2Z+2A+0.5S)6100/(H+2Z+2A+0.5S), where Z stands for the number of zygotes, A for the number of asci, S for the number of free spores, and H for the number of cells that failed to mate.When necessary, sporulation was visualized by iodine vapor staining.

b-Galactosidase reporter assay
S. pombe cells containing b-galactosidase reporter plasmids were pre-grown to mid-log phase in liquid EMM2 medium for 18 h and shifted to EMM2 or EMM-N medium.Cells were harvested at intervals by centrifugation, washed, and resuspended at a concentration of 5610 6 cells ml 21 .After the cells were permeabilized with 0.1% sodium lauroylsarcosinate, b-galactosidase activity was determined by measuring hydrolysis of the chromogenic substrate, o-nitrophenyl-b-D-galactoside, as described previously [35].

Electrophoretic mobility shift assay
To prepare GST and GST-Gaf1 fusion proteins for electrophoretic mobility shift assay (EMSA), cells of Escherichia coli BL21 strains carrying pGEX4T3 or pGEX4T3-Gaf1 were cultured in LB medium with 50 mg ml 21 of ampicillin at 30uC to A 600 of 0.5.Isopropyl b-D-thiogalactopyranoside was added to a final concentration of 1 mM and the cells were grown for 4 h at 25uC.Harvested cells were resuspended in Buffer A (20 mM Tris, pH 7.6; 137 mM NaCl) containing 0.1% Tween 20, 1 mM phenylmethylsulfonyl fluoride, and 1 mg ml 21 lysozyme.Resuspended cells were lysed by sonication, and the lysate was cleared by centrifugation at 24,0006g for 20 min at 4uC.The protein extracts were purified through a Glutathione Sepharose 4B column (Amersham).
To prepare EMSA probes and cold competitors, the upstream 0.60-kb SphI-NdeI (P A ) and downstream 0.25-kb NdeI-EcoRV (P B ) segments of the partial ste11 + promoter region [4] were amplified by PCR from the genomic DNA of S. pombe 972 using the following primers: P17 (59GCATGCCATCTCCAGGGA39) and P18 (59ACATATGATGCGAAAGCATT39) for P A ; and P19 (59CATATGTTACTTTAACCCCT39) and P20 (59GGA-TATCCTTTTAATATATGCT39) for P B .Specific doublestranded oligonucleotide probes for wild-type (P W ) and mutant (P M ) GATA motifs corresponding to nucleotides 2385 to 2352 of the ste11 + promoter were prepared by annealing complementary pairs of the following single-stranded oligonucleotides: P21 (59CATTTTGCCTTGCGCTATCTCCCTCAACGAAAAG39) and P22 (59CTTTTCGTTGAGGGAGATAGCGCAAGG-CAAAATG39) for P W ; and P10 and P9 for P M .The duplex DNAs were end-labeled with [c-32 P] ATP by T 4 -polynucleotide kinase and purified using a G-50 or G-25 column (Amersham).All binding reactions were carried out in 20 ml binding buffer (10 mM Tris pH 7.5, 2 mM MgCl 2 , 50 mM NaCl, 1 mM DTT, 5% glycerol) containing 2 mg of poly (dI?dC) as a non-specific competitor, 0.1-1 mg of recombinant GST-Gaf1 protein, and 5 rmol 32 P-end-labeled double-stranded probe, at room temperature for 20 min.For competition experiments, DNA binding reactions were allowed to reach equilibrium and a 50-or 100-fold excess of unlabeled specific competitor DNA was added to the binding reaction mixture.For detection of a specific DNA-protein complex, samples were loaded onto a 6% non-denaturation polyacrylamide gel in 0.56Tris-glycine buffer and electrophoresed at 10 V cm 21 at room temperature.The gels were run for 2 h, dried on a gel dryer and autoradiographed at 270uC using Fuji Xray film.Putative binding sites for transcription factors were searched by using TRANSFAC program (http://gene-regulation. com/pub/databases.html).

Deletion of gaf1 + causes accelerated G 1 -arrest under nitrogen-starved conditions
Previously, we suggested that gaf1 + might function as a transcriptional activator, although its biological relevance was unclear as a deletion mutant showed no significant defects [18].However, recent genome analysis revealed that the gaf1 + sequence previously reported (Accession No. AAC35593) contains only a partial ORF corresponding to the C-terminal 290-amino acid segment of Gaf1, and that the complete gaf1 + sequence comprises a 2,565-bp ORF encoding an 855-amino acid protein (Accession No. Q10280) [36].
In the present study, we deleted the entire gaf1 + ORF from the genome of S. pombe (Table 1) and evaluated the response of gaf1D cells to nitrogen starvation and glucose restriction by plate assay.Cells of the gaf1D strain did not show any growth defect on EMM2 or EMM-G plate, indicating that the gaf1 + gene is dispensable for mitotic growth under normal and glucose-restricted conditions (Figure 1A).However, gaf1D cells showed significantly reduced growth on EMM-N which lacked the inorganic nitrogen source, NH 4 Cl, compared to wild-type cells that did grow to a limited extent by utilizing the supplementary auxotrophic nutrients, adenine, leucine, and uracil, as organic nitrogen sources.
To confirm that the sensitivity of the gaf1D mutant to nitrogen starvation is due to the loss of Gaf1 activity rather than acquisition of abnormal activity, we constructed a system in which the production of Gaf1 could be shut off artificially using the thiaminerepressible nmt42 + promoter.In the absence of thiamine, the gaf1D/pREP-Gaf1 cells carrying ectopic copies of gaf1 + under the control of the nmt42 + promoter exhibited growth similar to the wild-type strains (WT/pREP42 and WT/pREP-Gaf1) on EMM-N plates (Figure 1B).In the presence of thiamine, gaf1D/pREP-Gaf1 cells exhibited as poor growth on EMM-N medium as gaf1D/pREP42 cells.These results suggest that the gaf1 + gene is dispensable for mitotic growth under normal conditions, but apparently plays a significant role in sustaining growth, though to a limited extent, under nitrogen-starved conditions.
In S. pombe, as key nutrients become limited, cells exit the mitotic cycle and enter either G 0 stationary phase or a program of sexual differentiation [37].During early nitrogen starvation, S. pombe cells undergo several rounds of rapid cell division and then arrest at the G 1 phase [8,13].In this study, the gaf1D strain accumulated G 1arrested cells after 4 h of nitrogen starvation, but the wild-type strain did not accumulate any detectable amount of G 1 -arrested cells until 6 h after the nitrogen shift (Figure 2A).Correspondingly, the homothallic haploid strain h 90 gaf1D/pREP-Gaf1 began to accumulate G 1 -arrested cells after 2 h of nitrogen starvation in the presence of thiamine, however, it did not show any signs of G 1arrest even after 8 h of nitrogen starvation in the absence of thiamine (Figure 2B).Therefore, the deletion of gaf1 + causes accelerated entrance into G 1 under nitrogen-starved conditions.Accordingly, the function of gaf1 + might be to delay the shift of nitrogen-starved cells from the vegetative cycle to the meiotic cycle, helping to sustain the vegetative cycle upon transient nitrogen starvation.

Sporulation efficiency is enhanced by gaf1D mutation and reduced by gaf1 + overexpression
To evaluate the function of gaf1 + in sporulation under nitrogenstarved conditions, we spotted suspensions of pre-grown homothallic haploid cells (h 90 ) and mixtures of mating pairs (h + and h 2 ) with or without gaf1D on EMM2 (+N) and EMM-N (2N) agar plates and estimated their mating and sporulation efficiencies (F M ) after 3-d cultivation.The homothallic haploid gaf1D strain showed a significantly higher F M value (24%) than the wild-type strain (7%) on EMM2 plates (Table 2).In addition, the F M value of the h + gaf1D6h 2 gaf1D mating mixture (63%) was approximately 5fold higher than the h + 6h 2 mating pair of wild-type (13%) on EMM2 plates.The homothallic gaf1D strain (85%) and h + gaf1D6h 2 gaf1D mating mixture (80%) showed approximately 10-15% higher F M values than the homothallic wild-type strain (69%) and h + 6h 2 mating mixture of wild-type strains (71%) on EMM-N plates.These results indicate that the gaf1D mutation elevates mating and sporulation efficiency by making cells more sensitive to nitrogen starvation.
We also evaluated the role of gaf1 + in sporulation using an overexpression system of gaf1 + under the control of the nmt42 + promoter in homothallic strains.In liquid EMM-N medium, no sporulation was observed among h 90 gaf1D/pREP-Gaf1 and h 90 WT/pREP-Gaf1 cells even after 18-h exposure to nitrogenstarved conditions, while a considerable portion of the h 90 gaf1D/ pREP42 and h 90 WT/pREP42 cells showed mating behavior after 3-and 18-h exposure, respectively (Figure 3A).On EMM-N plates, the h 90 WT/pREP-Gaf1 and h 90 gaf1D/pREP-Gaf1 strains exhibited negligible levels of iodine staining and F M value (,0.01%) in the absence of thiamine, but moderate levels of iodine staining and F M values (23-25%) in the presence of thiamine (Figure 3B).On the contrary, the spots of h 90 WT/ pREP42 and h 90 gaf1D/pREP42 strains were stained dark brown with iodine vapor and exhibited high F M values (69-85%) after 3-d cultivation regardless of the presence of thiamine, indicating luxuriant sporulation under nitrogen-starved conditions.Together, these results demonstrate that overexpression of gaf1 + leads to a significant reduction of sporulation efficiency.
The overlapping and non-overlapping portions of the three groups (Group 2G, 2N, and 2N/2G) are designated as Subgroup A to G (Figure 4), which enables more detailed clustering of the genes of interest.Approximately 63% (116 genes) of the members of Group 2G were up-regulated by both gaf1D mutation and nitrogen starvation (Subgroup B+G, Table S5 and  S10), while most of the remainders (61 genes) were up-regulated solely by gaf1D mutation (Subgroup A, Table S4).Therefore, it is suggested that Gaf1 down-regulates the basal expression levels of the nitrogen starvation-responsive genes included in Subgroup B+G.Although majorities of the members in Group 2N and 2N/ 2G were shown to be up-regulated by nitrogen starvation in both gaf1D and wild-type cells (1,043 genes, Subgroup D+G, Table S7 and S10), a considerable portions of them were up-regulated by nitrogen starvation in either gaf1D (375 genes, Subgroup E+F, Table S8 and S9) or wild-type cells (262 genes, Subgroup B+C, Table S5 and S6).
Furthermore, by comparing our data with the result of microarray analysis using the wild-type, Ste11-overexpressing, and ste11D strains [40], we found that 25 genes among those induced by gaf1D mutation are included in the list of the 61 Ste11 target genes (Table 3).Among the genes exhibiting increased expression in gaf1D cells, 10 genes including mei2 + , spk1 + , and ste11 + were up-regulated even under unstarved conditions (Subgroup G), while the remainders (15 genes) including ran1 + , tht1 + and ste6 + were induced only under nitrogen-starved conditions (Subgroup D+E).This result supports the speculation that ste11 + is possibly a direct target of Gaf1 that negatively regulates its expression at the transcriptional level.
ste11 + is epistatic to gaf1 + Epistasis analysis was performed to determine whether Gaf1 functions in the same pathway as Ste11, a key transcription factor for sexual development [12].Pre-grown cells of the homothallic   S1), nitrogenstarved wild-type cells (Group 2N, Table S2), and nitrogen-starved gaf1D cells (Group 2N/2G, Table S3).The overlapping and nonoverlapping portions of the three groups (Group 2G, 2N, and 2N/2G) are designated as Subgroup A to G. The lists of the genes included in Subgroup A-G are provided in Tables S4, S5, S6, S7, S8 S9, and S10, respectively, in the Supporting Information.doi:10.1371/journal.pone.0042409.g004haploid strains, h 90 gaf1D, h 90 ste11D, and h 90 gaf1D ste11D, were spotted on both EMM2 and EMM-N plates and the sporulation efficiencies were monitored after 3 d.The F M value of h 90 gaf1D reached approximately 85% on EMM-N and 24% on EMM2, but h 90 ste11D and h 90 gaf1D ste11D cells, and the h + ste11D6h 2 gaf1D mating mixture exhibited negligible levels of sporulation efficiencies even on EMM-N (Table 2).Therefore, the phenotype of the gaf1D mutation, i.e., accelerated initiation of sporulation and elevated sporulation efficiency, is masked by the ste11D mutation, causing impaired sporulation.Thus, ste11 + is epistatic to gaf1 + with respect to sporulation efficiency, and furthermore, gaf1 + functions upstream of ste11 + in the signaling pathway governing sexual development.

Expression of ste11 + is accelerated and increased in gaf1D cells
To test whether gaf1 + is responsible for the transcriptional regulation of ste11 + , we analyzed the expression of ste11 + under both nitrogen-starved and normal conditions in gaf1D and wildtype strains by Northern blot analysis.Under normal conditions, gaf1D cells showed increased level of ste11 + compared to wild-type cells (Figure 5A).When the wild-type cells were exposed to nitrogen starvation, the level of gaf1 + transcript increased considerably during the first 6 h of nitrogen starvation, followed by a subsequent decline.Unexpectedly, the levels of ste11 + transcript in both the wild-type and gaf1D cells harvested from the nitrogen-starved cultures at the time point of 0 h were considerably higher than those in the corresponding cells from the unstarved culture, which might be due to the very short exposure of the cells to EMM-N medium followed by washing with distilled water.This result is similar to that observed in the previous study conducted by other research group [41].Therefore, we adopted the data from the unstarved cells of corresponding strains cultivated in EMM2 for 18 h, rather than those from the nitrogen-starved cells sampled at the time point of 0 h, as references.As shown in Figure 5A, the amount of ste11 + transcript increased at a much slower rate than that of gaf1 + transcript, and it did not reach its highest level until at least 9 h after the transfer.The level of ste11 + transcript in gaf1D cells, however, increased steeply up to the plateau within 3 h and was maintained during Table 3. List of the Ste11 target genes up-regulated by gaf1D mutation.Relative expression of the genes was measured by microarray assay in the nitrogen-starved (2N) and unstarved (+N) cells of wild-type (ED668) and gaf1D (KL216) strains.Unstarved (+N) cells were prepared from the mid-log phase cultures in EMM2.Nitrogen-starved (2N) cells were prepared by shifting the mid-log phase cells to EMM-N for 4 h as described in Materials and Methods.b Genes up-regulated by gaf1D mutation under both nitrogen-starved and unstarved conditions (Subgroup G, p,0.05).
In accordance with the result of Northern blotting, the expression of Ste11 (p) -lacZ was higher in gaf1D cells than in wild-type cells under unstarved conditions (Figure 5B).When cells were subjected to nitrogen starvation, the expression of Ste11 (p) -lacZ was induced more rapidly in gaf1D cells than in wild-type cells.These results suggest that deletion of gaf1 + causes accelerated induction of ste11 + expression under nitrogen-starved conditions and increased ste11 + expression even under unstarved conditions.
Gaf1 protein binds to the promoter region of ste11 + We performed EMSA to address whether Gaf1 can directly bind to the promoter region of ste11 + .The upstream probe (P A ), encompassing the region from 2835 to 2227 of the ste11 + promoter, was amplified by PCR and end-labeled with T4 polynucleotide kinase (Figure 6A).A DNA-protein complex was observed in the reaction mixture containing the 32 P-labeled P A probe and the recombinant GST-Gaf1 (Figure 6B, lanes 2 and 5).The DNA-protein complex was specifically reduced in the presence of 50-and 100-fold molar excess of cold P A competitor probe (Figure 6B, lanes 3 and 4), but not in the presence of 50-or 100-fold molar excess of cold P B competitor (Figure 6B, lanes 6 and 7).This suggests that the Gaf1 protein specifically binds to a cis-element contained in the upstream region (from 2828 to 2227) of the ste11 + promoter.
We also performed competition experiments using cold oligonucleotide probes spanning base pairs 2385 to 2351: P W containing a canonical GATA motif and P M containing mutations in the GATA motif (Figure 6A).The DNA-protein complex between the 32 P-labeled P A probe and GST-Gaf1 protein was diminished by the addition of 100-fold molar excess cold P W (Figure 6C, lanes 4 and 6), but not by a similar amount of cold P M (Figure 6C, lanes 4 and 5).Accordingly, the amount of P W -GST-Gaf1 complex increased in proportion to the amount of GST-Gaf1 (Figure 6D, lanes 3-6) and was diminished to the basal level by addition of 100-fold molar excess of cold P W (Figure 6C, lane 6).The P M probe did not produce any detectable amount of DNAprotein complex with GST-Gaf1 (Figure 6D, lane 10) or exhibit competition against the P W probe to GST-Gaf1 even at a 100-fold molar excess (Figure 6D, lane 11).These results reflect that the canonical GATA motif from 2371 to 2366 in the promoter of ste11 + (Figure 6A) is the target sequence of the Gaf1 protein.It is thus suggested that the GATA motif in ste11 + promoter functions as a cis element to delay and attenuate the induction of ste11 + expression via the interaction with Gaf1.

Discussion
The S. pombe protein Ste11, which activates a number of genes required for mating and meiosis, is a pivotal regulator of sexual differentiation induced by nutrient starvation or environmental stress [12].In the present study, we identified Gaf1, an S. pombe GATA factor, as a negative regulator of ste11 + expression.
Deletion of gaf1 + caused no growth defects under normal conditions.However, under nitrogen-starvation, it led to reduced mitotic growth (Figure 1), accelerated entrance into G 1 (Figure 2), and elevated mating and sporulation efficiency (Table 2).Overexpression of gaf1 + resulted in a remarkable reduction of sporulation efficiency under nitrogen-starved conditions (Figure 3).It seems likely that Gaf1 functions as a modulator of the mitosis-meiosis transition, delaying the entrance of growing cells into the meiotic cycle during the initial stages of nitrogen starvation and signaling the optimal time for promoting sexual development.The delay in G 1 -arrest and subsequent sporulation may provide a safety mechanism allowing cells to revert to vegetative growth when nutrient availability again becomes favorable [6,16,42,43].In accordance with the present result, overexpression of tor2 + encoding the TOR protein kinase Tor2 strongly represses mating, meiosis and sporulation efficiency, whereas its inactivation has the opposite effect, leading to cell differentiation regardless of nutritional conditions [11].In S. cerevisiae, it has been shown that Tor kinase and GATA transcription factor are involved in nitrogen catabolite repression (NCR), a regulatory event in which transcription of certain genes is down-regulated by a good nitrogen source such as glutamine but up-regulated by a poor nitrogen source such as proline [44].Therefore, the involvement of TOR kinase and GATA transcription factor in nitrogen signaling may be a widely conserved phenomenon among various organisms.
Microarray analysis of the global gene expression profiles in gaf1D and wild-type cells under nitrogen-starved and unstarved conditions enabled us to search for a cluster of genes controlled by the action of Gaf1.Approximately 63% of the genes induced by the deletion of gaf1 + under normal conditions (Subgroup B+G, Figure 4) overlap with those induced by nitrogen starvation in wild-type cells.In addition, many of the Subgroup B+G genes are identified to be induced during mating and sporulation [38].Thus, it is likely that Gaf1 plays an important role not only in nitrogen signaling pathway but also in mating response.In accordance with the present result, recent microarray analysis using temperaturesensitive tor2 mutants revealed that a total of 151 of 194 genes induced by the loss of Tor2 function are included in the list of roughly 1,000 genes found to be induced by nitrogen starvation in S. pombe [45].This result pointed to an important role that Tor2 plays in nitrogen starvation and mating response.Interestingly, we also found that 13 genes among those induced by the deletion of gaf1 + overlap with those induced by the loss of Tor2 function, suggesting that the two genes might be involve in the same signaling pathway activated by nitrogen starvation.Thus, it would be interesting to determine the relationship between gaf1 + and tor2 + genes.
Comparison of our microarray data with the genome-wide view of Ste11 target genes reported previously [40] enabled us to speculate that ste11 + is a strong candidate for a direct transcriptional target of Gaf1 (Table 3).Especially, it is noteworthy that the genes involved in Ras/MAPK signaling pathway stimulated by pheromone such as gpa1 + , ste4 + , spk1 + , ste11 + , and mei2 + are transcriptionally induced in response to the loss of gaf1 + function under normal conditions.In addition, this finding is in agreement with the observation that deletion of gaf1 + causes accelerated entrance of cells into meiotic cell cycle (Figure 2) and elevated mating and sporulation efficiency on exposure to nitrogen starvation (Table 2).
In accordant with the result of microarray analysis, cells of the gaf1D ste11D strain were completely defective in mating and sporulation (Table 2), indicating that ste11 + is epistatic to gaf1 + .Deletion of gaf1 + not only increases the expression of ste11 + in unstarved cells but also accelerates the induction of ste11 + transcription in nitrogen-starved cells (Figure 5).In addition, the result of EMSA provides compelling evidence that Gaf1 binds to the canonical GATA motif (59-HGATAR-39) spanning from  5).Thus, it becomes evident that Gaf1 functions as a negative regulator of ste11 + transcription, via direct interaction with the GATA motif in the ste11 + promoter.The expression of ste11 + is regulated directly by two transcription factors, Rst2 and Ste11, that bind to the upstream activating sequence (UASst; 59-CCCCTC-39) and the T-rich box (TR box; 59-TTCTTTGTTY-39) in the ste11 + promoter, respectively [4].No other proteins that either activate or repress the transcription of ste11 + through direct binding to its promoter had been identified.Our research indicates that Gaf1 provides the prime example for negative regulation of ste11 + transcription through direct binding to a cis-acting motif of its promoter.
Figure 7 shows a simplified view of the proposed role of Gaf1 in the nitrogen-signaling pathways governing the expression of ste11 + and consequent sexual differentiation in S. pombe together with the cAMP-dependent PKA and stress-activated MAPK pathways determined previously.We found in the present study that nitrogen starvation causes induction of gaf1 + expression, and Gaf1, in turn, represses the expression of ste11 + via direct interaction with its promoter.Starvation of carbon or nitrogen source leads to decrease in the level of cAMP and subsequent drop of PKA activity, which consequently induces expression of ste11 + through the activation of Rst2 that can bind to the promoter region of ste11 + [4,7,46].Nitrogen starvation also causes activation of the stress-activated MAPK pathway including Wis4, Win1, Wis1, and Sty1 [14,15], leading to activation of Atf1 by phosphorylation [16,47,48].Activated Atf1, in turn, forms a complex with another cAMP response element-binding protein Pcr1 to yield an Atf1-Pcr1 heterodimeric transcription factor, which is also required for expression of ste11 + [5,6,49].It has not yet been determined whether the Atf1-Pcr1 complex directly regulates the expression of ste11 + or not.However, it has been reported that the Atf1-Pcr1 heterodimer directly activates the expression of cgs2 + encoding a phosphodiesterase that has a major role in regulating the single cAMP-dependent PKA pathway [3].In addition, phosphodiesterase is most likely stimulated by PKA activity to create a feedback mechanism [50].Thus there exists a direct connection between the MAPK and PKA pathways mediated by the action of Atf1-Pcr1 complex.It remains to be determined whether the expression of gaf1 + is subject to the control of either the cAMP-dependent PKA or the stress-activated MAPK pathway.Nitrogen starvation causes induction of gaf1 + expression, and Gaf1, in turn, represses the expression of ste11 + via direct interaction with its promoter.It has been previously known that nitrogen starvation leads to the activation of Rst2 via the cAMP-dependent PKA pathway [14,15] as well as Atf1-Pcr1 via the Sty1 MAPK pathway [3,5,6,[14][15][16][47][48][49], consequently resulting in induction of ste11 + expression.In addition, phosphodiesterase is most likely stimulated by PKA activity to create a feedback mechanism [50]

Figure 3 .
Figure 3. Overexpression of gaf1 + results in reduction of sporulation efficiency.Cells and colonies of the homothallic (h 90 ) strains, WT(JY4)/ pREP42, gaf1D(KL211)/pREP42, WT(JY4)/pREP-Gaf1, and gaf1D(KL211)/pREP-Gaf1, were analyzed for sporulation by DIC microscopy and iodine staining, respectively.(A) Cells were grown to mid-log phase in EMM2 for 18 h and shifted to EMM-N medium.Samples were taken from the cultures at intervals, and the morphological characteristics of cells were observed by DIC microscopy.Bar, 10 mm.(B) Cells were grown to mid-log phase in EMM2 for 18 h and spotted onto EMM-N plates with (+B 1 ) or without (2B 1 ) 20 mM thiamine.Sporulation was monitored by iodine vapor staining of colonies after 3-d incubation at 30uC.The F M value presented under each panel was determined by observing the cells under a DIC microscope.The values represent the average of at least three independent assays carried out in triplicate.WT denotes the wild-type (gaf1 + ).doi:10.1371/journal.pone.0042409.g003

Figure 5 .
Figure5.Deletion of gaf1 + results in accelerated induction of ste11 + expression under nitrogen-starved conditions.(A) Northern blot analysis of ste11 + and gaf1 + mRNA from wild-type and gaf1D cells exposed to nitrogen starvation.Cells of wild-type (972) and gaf1D (KL230) strains pre-grown in EMM2 (+N) were shifted to EMM-N (2N) and cultured with constant shaking.At indicated time points, cells were harvested and washed twice with distilled water, and total RNAs were extracted from the cells.RNA blots were hybridized with 32 P-labeled PCR-amplified gaf1 + and ste11 + probes.For internal control, all blots were stripped and subsequently rehybridized with32 P-labeled actinspecific probe (act1 + ).(B) b-Galactosidase reporter assay for analysis of ste11 + expression in wild-type and gaf1D cells subjected to nitrogen starvation.Cells of wild-type (ED665) and gaf1D (KL240) strains carrying pJLC-Ste11 (p) -LacZ were cultivated to mid-log phase in EMM2 (+N) and shifted to EMM-N (2N).At indicated time points, cells were harvested and washed twice with distilled water, and the level of ste11 + expression was estimated by measuring the activity of b-galactosidase in each sample.Values are the mean 6 standard error of three independent experiments carried out in triplicate, n = 9. *, p,0.01; **, p,0.05 (two-tailed Student's t-test, versus wild-type).WT denotes the wild-type (gaf1 + ).doi:10.1371/journal.pone.0042409.g005

Figure 6 .
Figure 6.Gaf1 protein specifically binds to the GATA binding motif of ste11 + promoter in vitro.(A) Schematic diagram of the promoter region of ste11 + .The open-and closed-ovals represent binding sites for Rst2 (UASst) and Ste11 (TR1 and TR2), respectively.The major transcription start site (position number +1) is designated by a hinged arrow.The regions contained in the upstream probe P A (2834,2225) and the downstream probe P B (2231,+10) are shown as bars.The dark square represents the GATA motif spanning from 2385 to 2352 for which wild-type (P W ) and mutant (P M ) double-stranded oligonucleotide probes were designed.Restriction sites: Sp, SphI; N, NdeI; RV, EcoRV.(B) Search for Gaf1-binding region in ste11 + promoter by EMSA.GST-Gaf1 protein (1 mg) was incubated with labeled P A probe containing the 0.6-kb upstream region of ste11 + promoter in the absence (lanes 2 and 5) or presence of excess cold competitor, P A (lane 3, 50-fold; lane 4, 100-fold) or P B (lane 6, 50-fold; lane 7, 100-fold).(C) Analysis of Gaf1 binding to the GATA motif of ste11 + promoter by EMSA.GST-Gaf1 protein (lane 3, 0.1 mg; lanes 4-6, 1 mg) was incubated with labeled P A probe in the absence (lanes 3 and 4) or presence of 100-fold excess of cold competitor P M (lanes 5) or P W (lanes 6).(D) Mutational dissection of the Gaf1-binding GATA motif of ste11 + promoter by EMSA.Varying amounts of GST-Gaf1 protein (lane 3, 0.1 mg; lane 4, 0.2 mg; lane 5, 0.5 mg; lanes 6, 7, 11, and 12, 1 mg) were incubated with labeled P W oligonucleotide probe in the absence (lanes 3 and 4) or presence of 100-fold excess of cold competitor P M (lanes 5) or P W (lanes 6).Mock reaction mixtures without GST-Gaf1 or with GST protein were used as negative controls in (B)-(D).The closed and open arrow heads in (B)-(D) represent shifted bands and free probes, respectively.doi:10.1371/journal.pone.0042409.g006

Figure 7 .
Figure 7. Schematic diagram showing the proposed function of Gaf1 in the nitrogen-signaling pathways in S. pombe.Nitrogen starvation causes induction of gaf1 + expression, and Gaf1, in turn, represses the expression of ste11 + via direct interaction with its promoter.It has been previously known that nitrogen starvation leads to the activation of Rst2 via the cAMP-dependent PKA pathway [14,15] as well as Atf1-Pcr1 via the Sty1 MAPK pathway [3,5,6,14-16,47-49], consequently resulting in induction of ste11 + expression.In addition, phosphodiesterase is most likely stimulated by PKA activity to create a feedback mechanism [50].The pathway addressed in this study is shown in thick lines, and other paths previously determined are shown in thin lines.Activation and inhibition are indicated by arrows and crossing bars, respectively.Dotted lines indicate pathways remained to be fully determined.doi:10.1371/journal.pone.0042409.g007 Figure 7. Schematic diagram showing the proposed function of Gaf1 in the nitrogen-signaling pathways in S. pombe.Nitrogen starvation causes induction of gaf1 + expression, and Gaf1, in turn, represses the expression of ste11 + via direct interaction with its promoter.It has been previously known that nitrogen starvation leads to the activation of Rst2 via the cAMP-dependent PKA pathway [14,15] as well as Atf1-Pcr1 via the Sty1 MAPK pathway [3,5,6,14-16,47-49], consequently resulting in induction of ste11 + expression.In addition, phosphodiesterase is most likely stimulated by PKA activity to create a feedback mechanism [50].The pathway addressed in this study is shown in thick lines, and other paths previously determined are shown in thin lines.Activation and inhibition are indicated by arrows and crossing bars, respectively.Dotted lines indicate pathways remained to be fully determined.doi:10.1371/journal.pone.0042409.g007

Table 1 .
S. pombe strains used in this study.

Table 2 .
Mating and sporulation frequency of S. pombe strains.