Factors governing the transcriptome changes and chronological lifespan of fission yeast during phosphate starvation

Starvation of Schizosaccharomyces pombe for inorganic phosphate elicits adaptive transcriptome changes in which mRNAs driving ribosome biogenesis, tRNA biogenesis, and translation are globally downregulated, while those for autophagy and phosphate mobilization are upregulated. Here, we interrogated three components of the starvation response: upregulated autophagy; the role of transcription factor Pho7 (an activator of the PHO regulon); and upregulated expression of ecl3, one of three paralogous genes (ecl1, ecl2, and ecl3) collectively implicated in cell survival during other nutrient stresses. Ablation of autophagy factor Atg1 resulted in early demise of phosphate-starved fission yeast, as did ablation of Pho7. Transcriptome profiling of phosphate-starved pho7Δ cells highlighted Pho7 as an activator of genes involved in phosphate acquisition and mobilization, not limited to the original three-gene PHO regulon, and additional starvation-induced genes (including ecl3) not connected to phosphate dynamics. Pho7-dependent gene induction during phosphate starvation tracked with the presence of Pho7 DNA-binding elements in the gene promoter regions. Fewer ribosome protein genes were downregulated in phosphate-starved pho7Δ cells versus WT, which might contribute to their shortened lifespan. An ecl3Δ mutant elicited no gene expression changes in phosphate-replete cells and had no impact on survival during phosphate starvation. By contrast, pan-ecl deletion (ecl123Δ) curtailed lifespan during chronic phosphate starvation. Phosphate-starved ecl123Δ cells experienced a more widespread downregulation of mRNAs encoding aminoacyl tRNA synthetases vis-à-vis WT or pho7Δ cells. Collectively, these results enhance our understanding of fission yeast phosphate homeostasis and survival during nutrient deprivation.

Starvation of Schizosaccharomyces pombe for inorganic phosphate elicits adaptive transcriptome changes in which mRNAs driving ribosome biogenesis, tRNA biogenesis, and translation are globally downregulated, while those for autophagy and phosphate mobilization are upregulated.Here, we interrogated three components of the starvation response: upregulated autophagy; the role of transcription factor Pho7 (an activator of the PHO regulon); and upregulated expression of ecl3, one of three paralogous genes (ecl1, ecl2, and ecl3) collectively implicated in cell survival during other nutrient stresses.Ablation of autophagy factor Atg1 resulted in early demise of phosphate-starved fission yeast, as did ablation of Pho7.Transcriptome profiling of phosphate-starved pho7Δ cells highlighted Pho7 as an activator of genes involved in phosphate acquisition and mobilization, not limited to the original three-gene PHO regulon, and additional starvationinduced genes (including ecl3) not connected to phosphate dynamics.Pho7-dependent gene induction during phosphate starvation tracked with the presence of Pho7 DNA-binding elements in the gene promoter regions.Fewer ribosome protein genes were downregulated in phosphate-starved pho7Δ cells versus WT, which might contribute to their shortened lifespan.An ecl3Δ mutant elicited no gene expression changes in phosphate-replete cells and had no impact on survival during phosphate starvation.By contrast, pan-ecl deletion (ecl123Δ) curtailed lifespan during chronic phosphate starvation.Phosphate-starved ecl123Δ cells experienced a more widespread downregulation of mRNAs encoding aminoacyl tRNA synthetases vis-à-vis WT or pho7Δ cells.Collectively, these results enhance our understanding of fission yeast phosphate homeostasis and survival during nutrient deprivation.
Inorganic phosphate is an essential nutrient acquired by cells from their environment.Cells from all domains of life respond to acute phosphate starvation by inducing the transcription of phosphate acquisition genes.In the fission yeast Schizosaccharomyces pombe, phosphate acquisition (PHO) genes pho1 (cell surface acid phosphatase), pho84 (phosphate transporter), and tgp1 (glycerophosphodiester transporter) are repressed under phosphate-replete conditions by upstream long noncoding RNA (lncRNA)-mediated transcriptional interference (1) and derepressed during acute phosphate starvation over 4 h (2).Induction of the PHO regulon during acute phosphate starvation depends on the transcription factor Pho7 that binds to target DNA sequences in the PHO mRNA promoters (3)(4)(5)(6).
In a recent study, we characterized the adaptations of fission yeast to chronic phosphate starvation, during which cells enter a state of quiescence, initially fully reversible upon replenishing phosphate after 2 days of starvation (7).Time-resolved analyses of transcriptome changes revealed coherent perturbations of gene expression, whereby the mRNAs encoding the cellular machineries for ribosome biogenesis, tRNA biogenesis, and protein translation are globally downregulated, whereas those for autophagy and phosphate mobilization are upregulated (Fig. S1).At the proteome level, phosphate starvation resulted in depletion of ribosome assembly factors, 60S and 40S proteins, tRNA-modifying enzymes, and translation factors.The phosphate starvation-induced downregulation of ribosome biogenesis genes, ribosomal protein genes, translation factor genes, and tRNA biogenesis genes was abolished or severely curtailed in the presence of cycloheximide (7).Also, the upregulation of the PHO genes pho84, pho1, and tgp1 was squelched by cycloheximide treatment due to a failure to shut off the production of upstream interfering lncRNAs.We surmised that phosphate starvation-induced transcriptional shutoff of the translation machinery mRNAs and PHO-regulatory lncRNAs requires new synthesis of one or more repressive proteins.
WT fission yeast cells progressively lose viability in the interval between 2 days (100% viable), 14 days (44% viable), and 28 days (4% viable) of phosphate starvation (7).We hypothesized that prolonged phosphate deficiency sets up bifurcated pathways of cell quiescence or cell death and that the balance between these outcomes might be altered by mutations in genes that affect these pathways.An initial clue to pathway choice was that maf1 mRNA and Maf1 protein were upregulated during a 48-h period of phosphate starvation.Maf1 is an endogenous negative regulator of RNA polymerase III (Pol3) transcription.We proceeded to show that maf1Δ fission yeast cells undergo accelerated demise between 1 and 2 days of phosphate starvation.Starved maf1Δ cells upregulate genes for phosphate mobilization and autophagy and downregulate the machinery for production and processing of rRNA and tRNA and for protein synthesis akin to starved WT cells.The 24 to 48 h temporal window during which phosphate-starved maf1Δ cells begin to expire is associated with overproduction of tRNA and the accumulation of polyadenylated tRNAs, introncontaining pre-tRNAs, and unspliced tRNA fragments (7).We proposed that Maf1 prolongs chronological lifespan during phosphate starvation by repressing Pol3 and preventing a death pathway associated with aberrant tRNA metabolism.
In the present study, we focus on three different arms of the transcriptional response to chronic phosphate starvation: (i) upregulated autophagy; (ii) Pho7-dependent transcription; and (iii) upregulated expression of ecl3, one of three paralogous genes (ecl1, ecl2, and ecl3) that are collectively implicated in cell survival during nutrient stress.

Limited transcriptome changes upon transition of fission yeast from YES medium to phosphate-replete ePMGT medium
Our phosphate starvation protocol entailed transfer of logphase cultures of S. pombe cells grown at 30 C from YES medium (yeast extract with supplements) to a defined medium ePMGT (enhanced Pombe Minimal Glutamate with Thiamine) containing either 15.5 mM phosphate (ePMGT+PO 4 ) or lacking phosphate (ePMGT-PO 4 ).Fission yeast continued to grow normally after transfer from YES to ePMGT+PO 4 , that is, the doubling time of 147 min in ePMGT+PO 4 was virtually identical to the doubling time of 143 min in YES (7), and there was no increase in cell surface acid phosphatase activity (a sensitive indicator of phosphate starvation response).Upon transfer from YES to ePMGT-PO 4 (starvation medium), cells rapidly accumulated surface acid phosphatase (Pho1) and underwent two to three rounds of cell division before entering a state of G0 quiescence, as gauged by flow cytometry.Previously, we interrogated the transcriptional adaptations to phosphate starvation by performing RNA-seq on poly(A) + RNA isolated from WT fission yeast cells prior to (time 0) and 4, 8, 12, 24, 36, and 48 h after transfer from YES to ePMGT-PO 4 .This analysis revealed that the mRNAs encoding the cellular machineries for ribosome biogenesis, tRNA biogenesis, and protein translation are globally downregulated, whereas those for autophagy and phosphate mobilization are upregulated (7).Transfer from YES to ePMGT-PO 4 also resulted in upregulation at 4 h of a subset of core environmental stress response genes that had been identified by Bähler and colleagues (8) as induced in response to oxidation, heavy metal exposure, heat shock, osmotic stress, and DNA damage, thereby implying that fission yeast cells experience acute phosphate starvation as a stress.
Although we attributed the observed transcriptome changes versus the YES time 0 dataset solely to phosphate limitation, we proceeded here to query whether and how the change from complex to defined phosphate-replete medium might impact gene expression, by performing RNA-seq on poly(A) + RNA isolated from WT fission yeast cells prior to (time 0) and 2, 4, 8 h after transfer from YES to ePMGT+PO 4 .Three biological replicates were sequenced for each time point.As before, we imposed the criterion that genes be differentially expressed ±2fold at two or more of the time points to be deemed genuinely responsive to the switch in media.We thereby identified 72 annotated protein-coding transcripts that were upregulated by between 2-fold and 90-fold in ePMGT+PO 4 (Table S1).Constituents of the fission yeast iron homeostasis regulon (n = 9) and noniron transmembrane transporters (n = 21) were the predominant class of genes upregulated during the change to ePMGT+PO 4 .Forty seven of the 72 genes up in ePMGT+PO 4 were also upregulated after transfer from YES to ePMGT-PO 4 .We attribute the increased expression of iron regulon genes to the fact that the concentration of iron in YES (3.2 μM) is 4-fold higher than in ePMGT (0.74 μM).
The key point here is that growth in ePMGT+PO 4 did not increase the expression of the key genes highlighted previously that were induced at equivalent times during phosphate starvation in ePMGT-PO 4 , that is, 25 autophagy genes, 23 Pol2 transcription factors, the Pol3 repressor Maf1, and 40 phosphate-mobilizing proteins/enzymes (7).Two transcription factors (Cbf12, Toe1) that were noted previously to be upregulated in ePMGT-PO 4 (7) were upregulated to a similar extent in ePMGT+PO 4 .Two phosphate-mobilizing genes-SPBPB2B2.06cand SPAC1039.02-encodingpredicted extracellular 5 0 -nucleotidase enzymes (which hydrolyze extracellular AMP to adenosine and phosphate) that were upregulated by 21,000-fold and 140-fold, respectively, after 8 h in ePMGT-PO 4 (7) were increased to a lesser extent after 8 h in ePMGT+PO 4 (by 13-fold and 4-fold, respectively) from which we conclude that they are bona fide phosphate-regulated genes.
Only five coding genes were downregulated by 2-to 4-fold after transfer from YES to ePMGT+PO 4 .Here, the salient point is that growth in ePMGT+PO 4 did not result in downregulation of the several hundred ribosome biogenesis, tRNA biogenesis, and protein translation genes seen during an equivalent interval of phosphate starvation (7).

Autophagy prolongs the lifespan of phosphate-starved fission yeast
Autophagy is an inducible response of eukaryal cells to changes in their environment, such as nutrient deprivation, whereby cytoplasmic components are enclosed within membranous autophagosomes that undergo lysosome/vacuole fusion, leading to degradation of the autophagosome contents (9).Autophagy thereby allows for salvage of the building blocks of preexisting macromolecules to restore nutrients.Autophagy is initiated and executed by a cascade of ubiquitinlike conjugation and kinase events performed by a conserved set of "Atg" proteins, named after their budding yeast Fission yeast adaptive responses to phosphate starvation orthologs (9), in concert with additional autophagy factors, none of which is essential for vegetative fission yeast growth under nutrient-rich conditions (10).A key finding from our RNA-seq experiments was that 25 of the fission yeast genes in the autophagy pathway were upregulated at the transcriptional level in response to phosphate starvation (Fig. S1).The greatest fold increases were noted for mRNAs specifying Atg12 (a ubiquitin-like protein modifier conjugated to Atg5) and Atg1 (a serine/threonine protein kinase), which were up by 10-fold and 5-fold, respectively.Proteomics analysis showed that the levels of 12 of the autophagy proteins were increased by 2-to 12-fold within 24 h of phosphate starvation, including Atg1, which increased 4-fold (7).
A pertinent question is whether the transcriptional induction and increased production of autophagy pathway proteins aid in prolonging the chronological lifespan of phosphatestarved fission yeast once they become quiescent.To address this point, we deleted the atg1 gene and compared the survival of WT and atg1Δ cells that were subjected to phosphate starvation then allowed to recover growth on phosphatereplete medium.Aliquots of cells were collected prior to (time 0) and 2, 4, and 7 days after transfer from YES to ePMGT-PO 4 and counted with a hemacytometer.Serial dilutions were plated on YES agar medium and incubated at 30 C. Viable colony counts were normalized to the time 0 control and percent survival was plotted as a function of starvation time (Fig. 1).WT cells retained full viability after 2 days of starvation before gradual demise ensued (66% survival after 7 days of starvation).The instructive finding was that deletion of Atg1 resulted in accelerated death of phosphate-starved cells, with 50% survival after 2 days of phosphate deprivation and less than 6% survival after 7 days (Fig. 1).

Interrogating the role of transcription factor Pho7
The 738-amino acid Pho7 protein is a member of the zinc binuclear cluster (Zn 2 -Cys 6 ) family of fungal DNA-binding transcription regulators (11).Dennis Wykoff and colleagues identified Pho7 as a key player in phosphate homeostasis by screening a fission yeast gene KO collection for strains defective in the elaboration of Pho1 cell surface acid phosphatase activity in response to phosphate starvation (3).They showed that pho7 deletion interdicted the phosphate starvation-induced increase in mRNAs encoding Pho1 acid phosphatase, the Pho84 phosphate transporter, and the Tgp1 glycerophosphodiester transporter (2, 3).These three genes comprise a PHO regulon; their expression is repressed under phosphate-replete conditions by lncRNA-mediated transcriptional interference (1) and derepressed during phosphate starvation when lncRNA synthesis abates (7).Microarray analyses of gene expression in pho7 + and pho7Δ strains in phosphate-replete and phosphate-starved cells affirmed the role of Pho7 in the acute phosphate starvation response, while highlighting an additional function of Pho7 in driving expression of multiple stress-response genes independent of phosphate status (2).Their microarray analysis identified 149 protein-coding genes that were upregulated at least 2-fold at 4 h of phosphate starvation, 113 of which were upregulated by at least 2-fold at 4 h in our RNA-seq analysis of phosphate starvation.Carter-O'Connell et al. (2) noted that one-third of the set of 22 genes induced after 2 h of phosphate starvation were pho7-dependent.In keeping with the greater sensitivity of RNA-seq, we detected 710 mRNAs that were upregulated at 4 h and at least one other time point (7).Thus, we were interested in applying RNA-seq to a pho7Δ strain to more fully delineate a "Pho7 regulon" and its impact on adaptation to phosphate starvation.

Transcriptional profiling of pho7Δ cells under phosphatereplete conditions
We performed RNA-seq on poly(A) + RNA isolated from WT and pho7Δ cells during logarithmic growth in YES medium at 30 C. Three biological replicates were sequenced for each strain.A cut-off of ±2-fold change in normalized transcript read level in pho7Δ versus WT and a p value of ≤0.05 were the criteria applied to derive a list of differentially expressed genes (Table S2).We thereby identified 361 annotated protein-coding RNAs that were downregulated in phosphate-replete pho7Δ cells (by between 2-and 86-fold), including those of known PHO regulon genes pho1, pho84, and tgp1 (each down by 3-fold) and pho842 and pho843 encoding inorganic phosphate transmembrane transporters (down by 24-fold and 2-fold, respectively).The other pho7Δ downregulated mRNAs (15 of which were decreased by more than 16-fold) had no obvious connections to phosphate homeostasis.mRNAs encoding several DNA-binding transcription Fission yeast adaptive responses to phosphate starvation factors were downregulated in pho7Δ cells: Atf31 (down 24fold); Cuf2 (9-fold), Prz1 (7-fold), and Mbx1 (5-fold) (Fig. 2, time 0).
We were surprised to find that 385 protein-coding RNAs were upregulated in phosphate-replete pho7Δ cells (by between 2-and 72-fold).Of the genes identified presently as upregulated in phosphate-replete pho7Δ cells by RNA-seq, only four-SPAC1039.02 (up 17-fold), urg2 (up 16-fold), urg1 (up 14-fold), and SPAC521.03(up 8-fold)-had been flagged as upregulated in the earlier microarray analyses of phosphatereplete pho7Δ cells by Carter-O'Connell et al. (2).It is noteworthy that our RNA-seq analysis showed that these same four genes were upregulated during the shift of WT cells from YES to ePMGT+PO 4 as follows: SPAC1039.02(up 8-fold), urg2 (up 20-fold), urg1 (up 32-fold), and SPAC521.03(up 4-fold).Indeed, a total of 30 of the mRNAs that were upregulated in phosphate-replete pho7Δ cells were also upregulated during the shift of WT cells from YES to ePMGT+PO 4 (Table S1).This suggests that absence of Pho7 might simulate a stress-like state in cells growing in rich YES medium.This is consistent with the fact that the pho7Δ strain displays severe ts and cs growth defects on YES agar at 37 C and 18 C, respectively (4).

pho7Δ cells mount a global transcriptional response to phosphate starvation
We performed RNA-seq on poly(A) + RNA isolated from pho7Δ cells harvested 4, 8, 12, 24, 36, and 48 h after transfer from YES to ePMGT-PO 4 medium.Three biological replicates were sequenced for each time point.A cut-off of ±2-fold change in normalized transcript read level compared to the WT time 0 control read level (i.e., in YES medium prior to starvation) and a p value of ≤0.05 were the criteria applied to derive an initial list of differentially expressed genes as a Fission yeast adaptive responses to phosphate starvation function of starvation time.A secondary criterion was that genes be differentially expressed at two or more time points to be deemed genuinely responsive to phosphate starvation.We mined the RNA-seq data to see whether loss of Pho7 interdicted any of the adaptive and (presumably) lifespanprolonging arms of the transcriptional reprogramming observed in starved WT cells (7).We found that pho7Δ cells manifested a forme fruste of the WT gene expression pattern (shown as a heat map in Fig. S1) in which genes involved in phosphate mobilization and autophagy were upregulated and the protein synthetic machinery was downregulated in response to phosphate starvation, albeit with fewer ribosome subunit genes affected in pho7Δ cells versus WT (heat map in Fig. 2).To summarize: (i) mRNAs encoding 34 proteins involved in phosphate dynamics were upregulated; (ii) mRNAs encoding 17 proteins driving autophagy were upregulated; (iii) mRNAs for 29 proteins of the 60S ribosome and 15 proteins of the 40S ribosome were transiently downregulated at 4 to 12 h (versus 129 ribosomal protein mRNAs that were downregulated in WT cells;  S3 and Fig. 2).
We showed previously that phosphate starvation of WT cells elicits endonucleolytic cleavages of 18S and 28S rRNAs to generate specific fragments that accumulated by 8 h and persisted for 48 h of starvation (7; data reprised in Fig. 3).The apparent site-specificity of the rRNA cleavages and the stability of the cleavage products suggested that the starvation-induced repression of ribosomal protein mRNAs and accompanying depletion of ribosome proteins exposes regions of the rRNAs to cellular endonucleases.The bioanalyzer profiles of the total RNA samples from three biological replicates of pho7Δ cells indicated that the large ribosomal RNAs underwent Figure 3. Phosphate starvation of pho7Δ cells triggers endonucleolytic cleavage of 18S and 28S rRNA.Total RNAs (5 μg) from pho7Δ cells harvested prior to (0 h) and 4, 8, 12, 24, 36, and 48 h after transfer to phosphate-free medium were resolved by urea-PAGE in parallel with 5 0 32 P-labeled denatured DNA size markers (pBR322 MspI digest ladder).The gel contents were electro-transferred to a nylon membrane, which was then serially hybridized to 32 P-labeled DNA oligonucleotide probes complementary to the 5 0 and 3 0 ends of 18S rRNA (top panels) and 28S rRNA (bottom panels) as described previously (7).Annealed probes were detected by autoradiography.The positions and sizes (nt) of the DNA markers are indicated on the right.The 18S and 28S rRNAs are depicted alongside the gels as thin lines with dots indicating the 5 0 ends.The annealed 5 0 end and 3 0 end oligonucleotide probes are shown as thick lines with arrows indicating the 3 0 end of the probe.The pho7Δ Northern blots are displayed adjacent to analogous Northern blots for 18S and 28S rRNAs from phosphate-starved WT cells reported previously (7).
Fission yeast adaptive responses to phosphate starvation fragmentation during the starvation time course.In the new experiments presented in Figure 3, we tracked the fate of 18S and 28S rRNA during phosphate starvation of pho7Δ cells via Northern blotting with 32 P-labeled oligonucleotide probes complementary to the very 5 0 and 3 0 ends of the mature 1842nt 18S rRNA and 3485-nt 28S rRNA.We observed a starvation time-dependent accumulation of stable rRNA fragments cleaved at discrete internal sites at distance from the 5 0 and 3 0 ends corresponding to their size.The major cleavage products detected in starved pho7Δ cells with the 5 0 probes were the same as those detected in WT cells (Fig. 3).Whereas the larger (>400 nt) rRNA fragments detected with the 3 0 probes were the same as in WT cells, 3 0 fragments smaller than 400 nt were less abundant in pho7Δ cells vis-à-vis WT (Fig. 3), likely because fewer ribosomal protein mRNAs were downregulated by phosphate starvation in pho7Δ cells.

Pho7-dependent gene induction during long-term phosphate starvation
Focusing on the genes that were upregulated during phosphate starvation of WT cells, we set a 4-fold decrement in mRNA induction at two or more times during phosphate starvation of pho7Δ cells as the criterion for designating a starvation-induced gene as Pho7-dependent.The expression profiles of 19 such genes are shown in Figure 4.The dependencies on Pho7 were high for several genes involved in phosphate mobilization.To wit: (i) SPBPB2B2.06cencoding an extracellular 5 0 -nucleotidase is the most highly upregulated mRNA in phosphate-starved WT cells in terms of fold induction, by a factor of 28,500 at 12 h, but is only up by 2200fold in pho7Δ cells; (ii) pho1 mRNA increased by 315-fold at 12 h in WT cells compared with 19-fold in pho7Δ cells; (iii) tgp1 mRNA was upregulated by 55-fold at 36 h in WT cells versus 9-fold in pho7Δ cells; and (iv) pho842 expression at 12 h in WT cells was 32-fold higher than in pho7Δ cells.Of the three canonical PHO regulon genes, pho84 (upregulated by 39fold in WT cells) was the least dependent on Pho7 for its expression during phosphate starvation, which was reduced by only a factor of 2 in pho7Δ versus WT (Fig. 4).

Pho7-dependent induced genes with Pho7-binding sites in their promoters comprise a Pho7 regulon
The RNA-seq data per se do not discriminate whether the contributions of Pho7 to the expression of specific genes are direct or indirect.Indirect effects apply if deleting pho7 affects the transcription of genes that regulate the expression of other genes identified as Pho7-dependent, whereas direct regulation by Pho7 implies that Pho7 binds to one or more cis-acting DNA sequence elements in the promoter of the Pho7dependent gene.The Pho7 DNA-binding domain (DBD) comprises the polypeptide segment from aa 279 to 368 that includes the Zn 2 Cys 6 module.DNase footprinting and electrophoretic mobility shift assays located the Pho7 DBD recognition sites in the promoters of the pho1 and tgp1 genes to a 12-nt motif 5 0 -TCG(G/C)(A/T)xxTTxAA ( 4).The pho1 promoter contains two Pho7 12-mer recognition elements (sites 1 and 2) in direct repeat orientation and separated by a 20-nt spacer.The Pho7-DBD binds independently and noncooperatively to these two sites in the pho1 promoter.By contrast, the tgp1 promoter contains a single Pho7 binding site.Crystal structures of Pho7 DBD in complex with its target site in the tgp1 promoter (5 0 -TCGGACATTCAAAT) or site 2 in the pho1 promoter (5 0 -TCGGAAATTAAAAA) highlighted two distinctive features of the Pho7 DBD: (i) it binds DNA as a monomer, unlike most other fungal zinc-cluster factors that bind as homodimers and (ii) it makes extensive interactions with its asymmetric target sequence over a 14-bp footprint that entails direct and/or water-mediated hydrogen bonding to individual nucleobases and backbone phosphates within, and remote from, the CGG triplet typically recognized by other Zn 2 Cys 6 DBDs (5,6).Comparison of the two Pho7 DBD-DNA structures revealed shared determinants of target site specificity as well as variations in the protein-DNA interface that accommodate different promoter DNA sequences.Nucleobase mutations in either of the pho1 sites or in the tgp1 site that eliminate Pho7 DBD binding in vitro result in loss of pho1 or tgp1 promoter activity in vivo, to the same extent as does deleting the pho7 gene (4).
Using epitope-tagged Pho7 (Pho7-TAP), Carter O'Connell et al. (2) performed chromatin immunoprecipitation-seq analyses that identified 1676 peaks of Pho7 occupancy in the genome of phosphate-replete fission yeast cells.They found that the occupancy of 367 sites increased during phosphate starvation.Whereas it was not possible at the time to discern a Pho7 DNA binding motif, a revisited in silico analysis of the separate phosphate-starved and phosphate-replete Pho7-TAP chromatin immunoprecipitation-seq datasets identified a consensus dodecamer, 5 0 -TCG(G/C)AxTTTxAA, that agreed with the Pho7 DBD sites in the pho1 and tgp1 promoters.The prevalence of genomic Pho7 sites in vivo, and the suggestion that their occupancy is dynamic with phosphate status, accords with the present RNA-seq data on the number of genes for which expression is altered when Pho7 is absent.

Fission yeast adaptive responses to phosphate starvation
To query which Pho7-dependent protein-coding genes shown in Figure 4, in addition to pho1, tgp1, and pho84, are likely to be directly regulated by Pho7, we inspected the genomic sequences upstream of the mapped transcription start sites (12) for Pho7 binding motifs containing the 5 0 -TCG(G/C) element or its inverse 5 0 -(C/G)CGA that is directly recognized by the Pho7 DBD.The transcription start sites are preceded by a TATA-box, upstream of which is at least one, and as many as five, candidate Pho7 DNA-binding sites, in either forward or inverted orientation (Fig. S2).We propose that this gene set comprises a phosphate starvationresponsive Pho7 regulon.

Pho7 contributes to the chronological lifespan of phosphatestarved fission yeast
We maintained quiescent WT and pho7Δ cells at 30 C in phosphate starvation ePMGT-PO 4 medium and gauged their survival upon restoration of growth on phosphate-rich YES medium.pho7Δ cells displayed reduced viability after 4 and 7 days of starvation (22% and 14% survival, respectively) (Fig. 5).We surmise that Pho7 promotes the expression of gene(s) that extend the chronological lifespan of phosphatestarved cells.
ecl3, alone among the three-member ecl gene family, is upregulated by phosphate starvation The fission yeast polypeptides Ecl1 (80 aa), Ecl2 (84 aa), and Ecl3 (89 aa) comprise a family of protein paralogs containing an N-terminal tetracysteine zinc-binding motif-CxxCG….LYCSxEC (13, 14) (Fig. 6A).The name Ecl (extender of chronological lifespan) derives from initial findings that the ecl1, ecl2, or ecl3 genes on a high-copy plasmid extended the viability of WT cells after entry into stationary phase (13,15).Subsequent studies by the Aiba lab showed that the expression of the three ecl genes is responsive to distinct signals, as follows: (i) ecl1 mRNA increased by 100-fold in sulfate-starved cells in a manner that depends on transcription factor Zip1 ( 16); (ii) ecl2 expression was increased only 3-fold by sulfate starvation, whereas ecl3 expression was unchanged ( 16); (iii) ecl1 mRNA increased by 100-to 200-fold upon leucine starvation or magnesium starvation of leucine-auxotrophic fission yeast in a manner that depends on transcription factor Fil1 (17,18); (iv) ecl2 expression was increased only 2-fold by leucine or magnesium starvation, whereas ecl3 expression was unaffected (17,18); (v) ecl2 mRNA was rapidly upregulated by 4-fold during heat shock, whereas ecl1 mRNA was unaffected and ecl3 mRNA declined by 5-fold (19).
ecl3 is expressed at a low basal level during vegetative growth and little was known about physiological cues that induce its expression or the transcription factors involved.Our RNA-seq analysis of phosphate-deprived S. pombe (7) showed that the ecl3 transcript was upregulated by 170-fold at 4 h, 300-fold at 8 h, and 400-to 440-fold at 24 to 48 h of phosphate starvation (Figs. 4 and 6B).ecl1 expression was unaffected (<2fold change) during the 48-h period of phosphate starvation, whereas ecl2 mRNA decreased by half at 36 and 48 h (Fig. 6B).ecl3 expression did not change during the switch from YES to ePMGT+PO 4 .Thus, ecl3 transcription is specifically and selectively responsive to phosphate status and not to the several other nutrients that affect ecl1 (e.g., sulfate, leucine, magnesium).Simultaneous with our report (7), Ohtsuka et al. (20) also found that ecl3 + is induced by phosphate starvation.
We find here that full induction of ecl3 during phosphate starvation relies on transcription factor Pho7, absent which ecl3 RNA was more modestly induced and with a temporal delay (26-fold at 4 h, 45-fold at 8 h, 68-fold at 24 h, 107-fold at 36 h, and 180-fold at 48 h of phosphate starvation) (Figs. 4 and  6B) compared to WT cells.The ecl3 gene is located on chromosome II, adjacent to and in opposite orientation to the prt2 lncRNA gene of the prt2-pho84-prt-pho1 gene cluster of the PHO regulon.ecl3 RNA is coregulated with the Pho7dependent pho84 and pho1 transcripts in a variety of fission yeast mutant strains under phosphate-replete growth conditions (21-26) (Fig. S3).Whereas expression of ecl2 remained unaffected during phosphate starvation of pho7Δ cells, the ecl1 transcript was upregulated at 8 to 24 h of starvation by 8-to 11-fold (Fig. 6B).

The ecl gene family collectively prolongs the lifespan of phosphate-starved fission yeast
The Aiba lab concluded, via characterization of single ecl deletion mutants and a triple-deletion mutant, that the three Ecl paralogs play functionally redundant roles in prolonging the chronological lifespan of cells during stationary phase, particularly in the absence of sulfate, leucine, or zinc (14,16,17,27).Here, after a preliminary experiment showed that an ecl3Δ strain we constructed did not differ significantly from WT with respect to survival during 4 weeks of phosphate starvation, we proceeded to construct an ecl123Δ triple mutant.The instructive finding was that simultaneous absence of the three Ecl paralogs shortened the lifespan of phosphatestarved fission yeast, leading to reduced viability after 4 and Fission yeast adaptive responses to phosphate starvation 7 days of starvation (27% and 13% survival, respectively) (Fig. 6C).

RNA-seq analysis of phosphate-replete ecl3Δ and ecl123Δ cells
A role for the Ecl family in gene expression was suggested based on a microarray analysis of logarithmically growing fission yeast cells in which the ecl1, ecl2, or ecl3 gene was provided on a high-copy plasmid (28).Ohtsuka et al. (28) identified 65 coding genes that were upregulated by more than 1.7-fold when any of the three ecl genes was present in high copy, including the PHO regulon genes pho1 (up 7-to 12-fold), pho84 (up 3-to 4-fold), and tgp1 (up 4-fold).The microarray study identified nine coding genes that were downregulated by more than 1.7-fold when any of the three ecl genes was present in high copy.Because the effects of forced overexpression do not necessarily signify a gene regulatory phenotype under physiological conditions, we undertook to gauge the impact of Ecl3 ablation, either singly or simultaneously with Ecl1 and Ecl2, on the fission yeast transcriptome.We performed RNAseq on poly(A) + RNA isolated from WT, ecl3Δ, and ecl123Δ cells during logarithmic growth in YES medium at 30 C. Three biological replicates were sequenced for each strain.We found that no transcripts (other than ecl3 itself) were dysregulated by ±2-fold in ecl3Δ cells.Thus, Ecl3 per se does not impact gene expression during growth in rich medium.
In ecl123Δ cells, we detected 32 coding RNAs that were upregulated by at least 2-fold of which four were increased by between 4-and 6-fold.A set of 56 mRNAs was downregulated by at least 2-fold in ecl123Δ cells of which 11 were reduced by between 4-and 49-fold (Table S4).None of the PHO mRNAs were downregulated in ecl123Δ cells.We reasoned that if a gene is truly subject to regulation by the Ecl proteins, we would expect the gene to be downregulated by pan-Ecl deletion and upregulated by Ecl overexpression.By comparing our RNA-seq data for ecl123Δ cells with the earlier microarray study of Ecl plasmid-bearing cells (28), we highlighted 9 genes that meet this criterion-mfm1, mam2, ste11, spk1, map1, mfm2, ste4, rgs1, and mam1-each of which is involved in mating and pheromone signal transduction.By contrast, there was no overlap between the genes identified by RNA-seq as upregulated in ecl123Δ cells and those identified by microarray as downregulated in Ecl plasmid-bearing cells.These results implicate Ecl proteins as a positive influence on the expression of mating-related genes; whether this reflects a direct effect of Ecl proteins on transcription is not clear.

RNA-seq analysis of phosphate-starved ecl123Δ cells
To gauge whether the Ecl family is involved in expression of the PHO regulon during phosphate starvation, we first assayed cell surface-associated Pho1 acid phosphatase activity of WT and ecl123Δ cells as a function of time after transfer from YES to ePMGT(-PO 4 ) medium.Pho1 activity, normalized to cell density, increased linearly during the first 12 h of phosphate starvation and continued to rise between 12 and 48 h, albeit with a shallower slope.The induction profiles were virtually identical in WT and ecl123Δ cells (Fig. S4).
To probe how loss of Ecl proteins affects the global transcriptional responses to chronic phosphate starvation, we performed RNA-seq on poly(A) + RNA from ecl123Δ cells harvested prior to (time 0) and 4, 8, 12, 24, 36, and 48 h after transfer from YES to ePMGT-PO 4 .(The Bioanalyzer profiles of the total RNA samples from ecl123Δ cells indicated that the large rRNAs underwent fragmentation during the starvation time course akin to that observed for WT and pho7Δ cells [not shown]) Three biological replicates were sequenced for each time point.Here again, a cut-off of ±2-fold change in normalized transcript read level compared to the time 0 control read level and a p value of ≤0.05 were the criteria applied to derive an initial list of differentially expressed genes as a function of starvation time.Genes differentially expressed at two or more time points were deemed responsive to phosphate starvation.The findings are summarized as follows.mRNAs encoding 22 proteins driving autophagy were upregulated during phosphate starvation of ecl123Δ cells (Table S5).mRNAs encoding 22 enzymes/proteins involved in phosphate dynamics were upregulated (Table S5).mRNAs for 21 Pol2 transcription factors were upregulated (Table S5) versus 26 in WT cells (7) (overlap = 13).By contrast, mRNAs for 47 proteins of the 60S ribosome, 35 proteins of the 40S ribosome, 116 ribosome assembly factors, eight subunits of the Pol1 transcription machinery, 29 tRNA biogenesis factors, and 45 translation factors were downregulated (Table S5).mRNAs for ten constituents of the Pol3 transcription machinery were downregulated, while the mRNAs encoding TFIIIB subunits Brf1 and Bdp1, and the Pol3 repressor Maf1 were upregulated (Table S5).Thus, the major transcriptome changes in phosphate-starved ecl123Δ cells were similar to those seen previously in WT cells (7).
A noteworthy finding, not commented on previously (7), was that mRNAs encoding 11 aminoacyl tRNA synthetase (aaRS) enzymes were downregulated during phosphate starvation of WT cells; proteomics analysis affirmed that the levels of 11 aaRS proteins were reduced by 2-to 3-fold after 24 h of phosphate starvation (7).Whereas we found here that ten aaRS mRNAs were downregulated in pho7Δ cells, phosphate starvation elicited a greater effect on aaRS gene expression in ecl123Δ cells, whereby 25 aaRS mRNAs were downregulated by 2-to 8-fold (Table S5).Dampening tRNA charging is of a piece with the general trend toward attenuation/alteration of protein synthesis as a key component of fission yeast adaptation to sustained phosphate starvation.

Discussion
We reported previously that chronic phosphate starvation of fission yeast elicits G0 quiescence associated with coherent changes in the transcriptome and proteome (7) that we presume are adaptive for survival.Activation of the autophagy pathway is typical of the response to nutritional stress that allows for recycling of macromolecular components.Beyond activating the autophagy pathway (29), phosphate starvation of fission yeast triggers increased expression of a large ensemble of autophagy genes (7).The factors governing transcription of the autophagy genes during phosphate starvation have not been defined, but there are preliminary indications that at least two separate regulatory schemes apply, based on the effects of inhibiting new protein synthesis during the shift to phosphatefree medium.To wit: (i) the starvation-induced increase in mRNAs encoding 11 autophagy factors (fun14, atg43, atg1802, atg8, atg7, atg10, atg15, aut12, nbr1, nrf1, and scs22) was abolished in the presence of cycloheximide, whereas (ii) the induction of mRNAs encoding nine other autophagy factors (atg1, atg4, atg1801, atg20, atg2402, atg101, atg13, atg6, and isp6) was not affected by cycloheximide (7).Here, we found that genetic ablation of key autophagy factor Atg1 (a serine/ threonine protein kinase) results in early demise of phosphatestarved fission yeast.Though not surprising, this observation is in keeping with the idea that autophagy-driven catabolism dampens metabolic demand and provides life-sustaining metabolites.
Fission yeast respond to acute phosphate starvation by alleviating the lncRNA-mediated repression of a three-gene PHO regulon that encodes a cell surface acid phosphatase (Pho1) and transmembrane transporters of inorganic phosphate (Pho84) and glycerophosphocholine (Tgp1) (1).Synthesis of the three PHO mRNAs is driven by the Zn 2 -Cys 6 family transcription activator Pho7 (2)(3)(4)(5)(6).The present transcriptomic analysis of phosphate-replete and phosphatestarved pho7Δ cells affirms and extends earlier microarray studies conducted by Wykoff and colleagues (2).We see that Pho7 contributes in several ways to the phosphate starvation response: (i) as a transcriptional activator of genes involved in phosphate acquisition and phosphate mobilization, not limited to the original three-gene PHO regulon and (ii) as an activator of additional starvation-induced genes with no clear connection to phosphate dynamics.We draw a connection between Pho7-dependence of gene induction during phosphate starvation and the presence of candidate Pho7 DNA-binding sites (4)(5)(6) in the promoter regions of the Pho7dependent genes, the implication being that Pho7 is directly driving their transcription.Pho7 is distinguished from most other fungal Zn 2 -Cys 6 transcription factors in that it binds to an asymmetric DNA target site and does so as a monomer (4)(5)(6).The 738-aa Pho7 protein is predicted to be almost entirely disordered, except for its DBD (aa 279-339).Outstanding issues regarding Pho7 function include: (i) how the protein segments flanking the DBD direct the Pol2 transcription machinery to Pho7-responsive genes; (ii) whether distinct disordered regions within Pho7 specialize in regulating specific subsets of responsive genes; and (iii) whether Pho7 collaborates with one or more other DNA-binding transcription factors (e.g., via heterodimerization) to activate its target genes.
The RNA-seq results also implicate Pho7 as a subtle contributor to the global downregulation of the protein synthesis machinery during chronic phosphate starvation.mRNAs for 44 ribosomal proteins and 85 ribosome assembly factors were downregulated during phosphate starvation of pho7Δ cells, which represents an attenuated response compared to the 129 ribosomal protein mRNAs and 115 ribosome assembly factor mRNAs that were downregulated in WT cells (7).Depletion of ribosomal protein mRNAs during phosphate starvation of pho7Δ cells correlated with the stable fragmentation of 18S and 28S rRNAs, as noted previously for starved WT cells (7).

Fission yeast adaptive responses to phosphate starvation
We found that the absence of Pho7 resulted in shortened cellular survival during phosphate starvation.Although the magnitude of the pho7Δ effect on lifespan was similar to that of atg1Δ, we suspect that pho7Δ is not acting via defective autophagy, insofar as the autophagy genes are induced in phosphate-starved pho7Δ cells.It is conceivable that dampening of the ribosomal protein and ribosome assembly factor changes in pho7Δ cells might account for the effect on survival.Interrogating the functional consequences of ribosomal protein depletion and rRNA incision (e.g., via analysis of the abundance and compositions of polysomes and free ribosomes in phosphate-starved cells) will be essential to understanding the nexus between protein synthesis and chronological lifespan.
The ecl3 gene, which is strongly induced during phosphate starvation (7), is located adjacent to and in opposite orientation to the prt2 lncRNA gene of the prt2-pho84-prt-pho1 gene cluster of the PHO regulon.The annotated ecl3 mRNA transcription start site is 715 nt upstream of the prt2 lncRNA transcription start site.We consistently noted in prior studies (21-24) that ecl3 RNA was coregulated with the pho84 and pho1 transcripts in a variety of fission yeast mutant strain backgrounds that either derepressed or hyperrepressed the PHO genes under phosphate-replete growth conditions (Fig. S3).Here, we find via RNA-seq that ecl3 induction during phosphate starvation was squelched in pho7Δ cells.Similar findings were reported recently by Ohtsuka et al. (20), who measured ecl3 mRNA by RT-qPCR during a 6-h interval of phosphate starvation of WT and pho7Δ cells.The PHO genes are turned on during phosphate starvation when synthesis of the upstream interfering lncRNAs is shut off (7), thereby allowing access of Pho7 to the PHO mRNA promoters.The pho84 promoter resides within the prt2 gene, which includes at least four Pho7 DNA-binding sites, as gauged by EMSA assays with overlapping genomic DNA fragments spanning 1100-bp upstream of the pho84 transcription start site (30).One of the high-affinity Pho7 binding sites was mapped via DNase foot printing to a 12-mer sequence ( -977 TCGGTCTTTGAA -988 ) that adheres to the consensus Pho7 recognition element (4,30).A parsimonious model is that shut-off of the prt2 lncRNA promoter allows access of Pho7 to one or more of its binding sites in the prt2 DNA and that this binding then activates the ecl3 mRNA promoter.We reported that the upregulation of pho84, pho1, and tgp1 during acute phosphate starvation was quashed by cycloheximide treatment, owing to a failure to shut off the production of the upstream interfering lncRNAs (7).Though not explicitly mentioned previously, it is noteworthy in the present context that induction of the ecl3 transcript during acute phosphate starvation was also interdicted by cycloheximide (7), consistent with a scenario in which prt2 lncRNA synthesis interferes with expression of both pho84 and ecl3.
Our characterization of an ecl3Δ mutant reveals that Ecl3 per se plays no apparent role in fission yeast gene expression and has no impact on cell survival during phosphate starvation.Aiba and colleagues have implicated the Ecl1, Ecl2, and Ecl3 paralogs collectively in cell survival during various nutritional stresses via short lifespan phenotypes associated with an ecl123Δ triple deletion.Here, we present evidence that pan-ecl deletion curtails lifespan during chronic phosphate starvation.Transcriptome profiling of phosphate-starved ecl123Δ cells uncovered no obvious "smoking gun" to account for their accelerated demise.Rather, the gene expression changes elicited by phosphate starvation of ecl123Δ cells were much the same as those observed for WT cells.In particular, we see that phosphate-starved ecl123Δ cells mount an induction of autophagy genes à la WT cells.This observation contrasts with the effects of pan-ecl deletion on the fission yeast response to sulfate starvation, whereby the 3-to 4-fold sulfate starvation-induced increase in mRNAs encoding autophagy proteins Atg1, Atg3, Atg4, Atg8, Atg13, and Atg20 was effaced or attenuated in an ecl123Δ strain background (31).A potentially interesting difference was that phosphatestarved ecl123Δ cells experienced a more widespread downregulation of mRNAs encoding aminoacyl tRNA synthetases vis-à-vis WT or pho7Δ cells.How this phenotype might connect to the Ecl proteins is unclear.

Phosphate starvation
Fission yeast strains were grown at 30 C in YES medium to A 600 of 0.5 to 0.8.The cells were harvested by centrifugation, washed with ePMGT medium without phosphate, and resuspended at A 600 of 0.3 in ePMGT(-PO 4 ).The phosphate-free cultures were incubated at 30 C and diluted with ePMGT(-PO 4 ) to not exceed A 600 of 0.8.At the times specified: (i) a volume of culture containing 10 A 600 units of cells was rapidly cooled by introducing it into ice and cells were harvested by centrifugation at 4 º C. The cells were transferred to 1.5-ml screw-cap tubes and then frozen on dry ice for subsequent RNA analyses.

RNA-seq analysis
Total RNA was extracted via the hot phenol method from 10 A 600 units of pho7Δ cells or ecl123Δ cells harvested prior to Fission yeast adaptive responses to phosphate starvation (time 0) and at 4, 8, 12, 24, 36, and 48 h after transfer to ePMGT(-PO 4 ) medium or from WT cells harvested prior to (time 0) and at 2, 4, and 8 h after transfer to ePMGT(+PO 4 ), or from ecl3Δ cells that were grown in liquid YES medium at 30 C. The integrity and quantity of total RNA were gauged with an Agilent Technologies 2100 Bioanalyzer and TapeStation.The Illumina TruSeq stranded mRNA sample preparation kit was used to purify poly(A) + RNA from 500 ng of total RNA and to carry out the subsequent steps of poly(A) + RNA fragmentation, strand-specific complementary DNA synthesis, indexing, and amplification.Indexed libraries were normalized and pooled for paired-end sequencing performed by using an Illumina NovaSeq 6000 system.FASTQ files bearing pairedend reads of length 51 bases (total paired reads of 14.9 million to 37.7 million per biological replicate) were mapped to the S. pombe genome using HISAT2-2.1.0with default parameters (32).Mapped reads comprised 89% to 99% of the total reads per replicate.The resulting SAM files were converted to BAM files using Samtools (33).Count files for individual replicates were generated with HTSeq-0.10.0 (34) using exon annotations from Pombase (GFF annotations, genome-version ASM294v2; source "ensembl").Reads per kilobase per million mapped reads analysis and pairwise correlations (Pearson coefficients of 0.963-0.989)were performed as described previously (35).Differential gene expression and fold change analysis was performed in DESeq2 (36).Cut-off for further evaluation was set for genes that were up or down by ≥2-fold at two or more starvation time points compared to the WT time 0 control, with a p value (Benjamini-Hochberg adjusted) of ≤0.05.Genes were further filtered on the following criteria: (i) ≥2-fold up and the average normalized read count at a given time point was ≥100 and (ii) ≥2-fold down and the average normalized read count for WT time 0 was ≥100.This served to cull poly(A) + transcripts that were present at very low levels in all samples.After deriving a list of differentially expressed genes for each time point, all genes that did not have a 2-fold change in expression at two consecutive time points were culled.

S. pombe atg1Δ strain
Genomic DNA segments (619-nt upstream of the atg1 translation start codon and 618-nt downstream of the stop codon, respectively) were PCR-amplified and cloned upstream and downstream of the hygMX antibiotic resistance cassette in a bacterial plasmid.The linear atg1Δ::hygMX gene disruption cassette was excised from the plasmid and transfected into diploid S. pombe cells.Hygromycin-resistant transformants were selected and analyzed by Southern blotting to confirm correct integration at the atg1 locus.Confirmed heterozygous diploids were sporulated and hygromycin-resistant atg1Δ haploid progeny were selected.

S. pombe eclΔ strains
PCR amplification and standard cloning methods were used to construct plasmids in which an antibiotic resistance cassette is flanked by 532-to 610-bp DNA segments corresponding to genomic sequences upstream and downstream of the ecl1, ecl2, and ecl3 ORFs.The resulting ecl1Δ::hygMX, ecl2Δ::natMX and ecl3Δ::kanMX disruption cassettes were excised from the plasmids and transfected into diploid S. pombe cells.Transformants resistant to hygromycin, nourseothricin, and G418 were selected and analyzed by Southern blotting to confirm correct integration at one of the ecl loci.Antibiotic-resistant ecl1Δ, ecl2Δ, and ecl3Δ haploids were isolated after sporulation of the heterozygous diploids.Serial pairwise matings of the eclΔ strains and random spore analysis with selection for two or three drug-resistance makers inserted in lieu of the ecl ORFs yielded an ecl123Δ triple-deletion haploid strain.Strain genotypes are shown in Table S6.

Figure 1 .
Figure 1.Deletion of atg1 shortens survival during phosphate starvation.Viable colony counts of phosphate-starved WT and atg1Δ cultures were normalized to the time 0 control (100%).Percent survival is plotted as a function of starvation time.The 2-day survival data are the average of four independent experiments ± SD.The 4-day survival data are the average of five experiments ± SD.The 7-day WT survival data are the average of eight experiments ± SD.The 7-day atg1Δ survival data are the average of three experiments ± SD.

Figure 2 .
Figure 2. Transcriptome changes in pho7Δ cells during phosphate starvation.poly(A) + RNA-seq data from pho7Δ cells grown in phosphate-replete conditions (0 h) or after 4, 8, 12, 24, 36, and 48 h of phosphate starvation was compared to that of WT cells grown in phosphate-replete conditions.The log2 fold changes of individual mRNAs, output in DESeq2, within the indicated functional classes of genes are colored according to the scale shown on the top left.Red signifies upregulation and blue signifies downregulation.Gray shading denotes less than a 2-fold change (and/or p value >0.05) in transcript level at the times specified.Each row represents a gene (specified at right), and each column represents the indicated sampling time (h).Pho7-dependent genes are denoted by blue asterisks.

Figure 4 .
Figure 4. Pho7-dependent gene induction during long-term phosphate starvation.The fold increases in the indicated mRNA levels in phosphatestarved WT () versus pho7Δ ( ○ ) cells, normalized to the time 0 control are plotted as a function of starvation time.Fold increase is plotted on a linear scale on the y-axis; the values are derived from the log2 fold increases determined from RNA-seq experiments and output in DESeq2.The log2 fold change SE values ranged from 0.052 to 0.67 (corresponding to linear fold change SE values of 1-1.6) and are not visible on the graphs because they are smaller than the data symbols.SE, standard error.

Figure 5 .
Figure 5. Deletion of pho7 shortens survival during phosphate starvation.Viable colony counts of phosphate-starved WT and pho7Δ cultures were normalized to the time 0 control (100%).Percent survival is plotted as a function of starvation time.Each datum in the graph is the average of three independent experiments ± SD.

Figure 6 .
Figure 6.Expression and function of the ecl gene family during phosphate starvation.A, alignment of the primary structures of the fission yeast Ecl1, Ecl2, and Ecl3 polypeptides.Positions of amino acid side chain identity/similarity are denoted by dots above the alignment.The putative zinc-binding cysteines are shaded in gray.B, the log2 changes in ecl1, ecl2, and ecl3 mRNA levels in phosphate-starved WT (left panel) and pho7Δ (right panel) cells versus time 0 controls are plotted as a function of starvation time.C, pan-ecl deletion shortens survival during phosphate starvation.Viable colony counts of phosphate-starved WT and ecl123Δ cultures were normalized to the time 0 control (100%).Percent survival is plotted as a function of starvation time.The 2day survival data are the average of five independent experiments ± SD.The 4-day survival data are the average of four independent experiments ± SD.The 7-day survival data are the average of three independent experiments ± SD.