Cellular responses to long-term phosphate starvation of fission yeast: Maf1 determines fate choice between quiescence and death associated with aberrant tRNA biogenesis

Abstract

Inorganic phosphate is an essential nutrient acquired by cells from their environment. Here, we characterize the adaptative responses of fission yeast to chronic phosphate starvation, during which cells enter a state of quiescence, initially fully reversible upon replenishing phosphate after 2 days but resulting in gradual loss of viability during 4 weeks of starvation. Time-resolved analyses of changes in mRNA levels revealed a coherent transcriptional program in which phosphate dynamics and autophagy were upregulated, while the machineries for rRNA synthesis and ribosome assembly, and for tRNA synthesis and maturation, were downregulated in tandem with global repression of genes encoding ribosomal proteins and translation factors. Consistent with the transcriptome changes, proteome analysis highlighted global depletion of 102 ribosomal proteins. Concomitant with this ribosomal protein deficit, 28S and 18S rRNAs became vulnerable to site-specific cleavages that generated temporally stable rRNA fragments. The finding that Maf1, a repressor of RNA polymerase III transcription, was upregulated during phosphate starvation prompted a hypothesis that its activity might prolong lifespan of the quiescent cells by limiting production of tRNAs. Indeed, we found that deletion of maf1 results in precocious death of phosphate-starved cells via a distinctive starvation-induced pathway associated with tRNA overproduction and dysfunctional tRNA biogenesis.

Figure 1.

Fission yeast response to phosphate starvation. (A) S. pombe cells were grown to mid-log phase in YES medium, harvested, washed with water, and then transferred to ePMGT medium containing different concentrations of phosphate as indicated on the x-axis. The cells were incubated at 30°C for 18 h, then harvested, washed with water, and assayed for cell-surface Pho1 acid phosphatase activity by conversion of p-nitrophenylphosphate to p-nitrophenol. Acid phosphatase activity (y-axis) is plotted as a function of phosphate concentration in the medium. The data were fit to a dose-response model (EC₅₀ shift) in Prism with a goodness of fit correlation coefficient of 0.99 and the dose of half maximal response of 0.048 ± 0.006 mM. (B) Mid-log cultures of wild-type or pho7Δ cells grown in YES medium were harvested, washed with water, transferred to ePMGT medium with no phosphate, and incubated for 48 h at 30 °C. Cells harvested prior to (0 h) and at sequential times after transfer to phosphate-free medium were assayed for Pho1 acid phosphatase activity (left y-axis). The growth of replicate cultures of wild-type cells (n = 3) as a function of time after transfer to phosphate-free medium was monitored by A₆₀₀ (right y-axis). (C) S. pombe cells sampled prior to (0 h) and at sequential times after transfer to phosphate-free medium were visualized by brightfield microscopy (630× magnification, lower panels) or processed for flow cytometry (upper panels). Histogram plots of DNA content measured by flow cytometry for each time are shown, where x- and y-axes denote fluorescence intensity and cell number, respectively. The DNA content as estimated by fluorescence intensity is indicated above representative peaks. (D and E) Phosphate-starved quiescent fission yeast return to growth in phosphate-rich medium. Cells were starved for phosphate for 2 days as described in panel B and then monitored over time prior to (time 0) and after transferring aliquots of the starved cells to phosphate-rich liquid medium (YES) to evaluate resumption of growth. (D) A plot of A₆₀₀ (right y-axis) and cell surface acid phosphatase activity (left y-axis) as a function of time after restoration of phosphate. (E) Cells sampled prior to (0 h) and at sequential times after transfer to phosphate-replete medium were visualized by brightfield microscopy (630× magnification, lower panels) or processed for flow cytometry (upper panels).

Figure 2.

Phosphate starvation triggers upregulation of genes involved in phosphate metabolism. The list of upregulated genes is stratified roughly according to the fold increase in transcript level at the indicated times post-starvation, expressed as the log₂ of the fold change compared to the pre-starvation ‘time 0’ control. The known or imputed functions of the gene products are listed in the right-most column. The set of genes upregulated by at least 16-fold (log₂ of ≥4) is highlighted in gold shading. The pho1, tgp1 and pho84 genes that comprise the original PHO regulon are highlighted in bold font.

Figure 3.

New protein synthesis is required for de-repression of the three-gene PHO regulon during acute phosphate starvation. Strand-specific RNA-seq read densities (counts/base/million) are plotted as a function of position across the (A) nc-tgp1–tgp1, (B) prt2–pho84 and (C) prt–pho1 loci, respectively, from cells in phosphate-replete conditions (0 h), or at the specified time intervals after transfer to phosphate starvation medium, either in the absence (–) or presence (+) of 100 μg/ml cycloheximide (CHX). The top and bottom RNA-seq time-course for the specified loci depict the same data at different read density scales (indicated on the y-axes), with the bottom series highlighting the read densities corresponding to the PHO-regulatory lncRNAs. The x-axis scales are denoted by the 1 kb bars. The RNA-seq read densities were determined from cumulative counts of three RNA-seq replicates. The individual lncRNAs, lncRNA–mRNA read-through (RT) transcripts, or mRNAs are labeled and shown to scale as black arrows in the direction of their synthesis. The broken arrow indicates an extended nc-tgp1–tgp1 RT lncRNA.

Figure 4.

Phosphate starvation triggers endonucleolytic cleavage of 18S and 28S ribosomal RNA. (A) Total RNA (5 μg) from cells harvested prior to (0 h) and 12, 24, 36, and 48 h after transfer to phosphate-free medium was resolved by urea-PAGE in parallel with denatured DNA size markers (pBR322 MspI digest ladder). Nucleic acids were visualized by staining the gel with Sybr Green II dye. The positions and sizes (nt) of the DNA markers are indicated on the right. (B–I) Northern analysis. Equal aliquots of total RNA from cells harvested prior to (0 h) and 4, 8, 12, 24, 36 and 48 h after transfer to phosphate-free medium was resolved by urea-PAGE in parallel with 5′ ³²P-labeled denatured DNA size markers. The gel contents were electro-transferred to membranes, which were then hybridized to ³²P-labeled DNA oligonucleotide probes complementary to the 5′ and 3′ ends of 18S rRNA (panels B and C), the 5′ and 3′ ends of 28S rRNA (panels D and E), 5S rRNA (panel F), 5.8S rRNA (panel G), tRNA-Trp (panel H), or tRNA-Val-AAC 5′ exon (panel I). Annealed probe and labeled DNA size markers were detected by scanning the Northern blots with a Typhoon FLA 7000 phosphorimager and capturing the data in ImageQuant. The positions and sizes (nt) of the DNA markers are indicated on the right in each panel. The signal intensities of full-sized 5S rRNA, 5.8S rRNA, tRNA-Trp and tRNA-Val-AAC were quantified in ImageQuant and the values, normalized to the respective RNA levels at time 0 (defined as 1), are shown below the lanes in panels F–I.

Figure 5.

Deletion of maf1 acutely curtails the lifespan of phosphate-starved fission yeast cells. (A) Chronological lifespan of wild-type fission yeast during 4 weeks of phosphate starvation. Viable colony counts were normalized to the time 0 control (100%). Percent survival is plotted as a function of starvation time. (B) Survival of wild-type and maf1Δ cells during 7 days of phosphate starvation. (C) Wild-type and maf1Δ cells were assayed for Pho1 acid phosphatase activity prior to (time 0) and at the indicated time after transfer to phosphate starvation medium. (D) maf1Δ cells sampled prior to (0 h) and at sequential times after transfer to phosphate-free medium were visualized by brightfield microscopy (top panels) and analyzed by flow cytometry (bottom panels). (E–G) Cells were starved for phosphate for 4 days and then monitored over time after transferring aliquots of the starved cells to phosphate-rich liquid YES medium to evaluate resumption of growth. (E) Wild-type and maf1Δ cells sampled 4 days after phosphate starvation and at sequential times after transfer to phosphate-replete medium were processed for flow cytometry. (F) A plot of A₆₀₀ as a function of time after restoration of phosphate. Each datum in the graph is the average absorbance of two independent phosphate-starved cultures with error bars indicating the range. (G) Wild-type and maf1Δ cells sampled 4 days post-starvation and at 12 h after transfer to phosphate-replete medium were visualized by brightfield microscopy (630× magnification).

Figure 6.

Accumulation of polyadenylated tRNAs in phosphate-starved maf1Δ cells. Log₂ fold increases of poly(A)⁺ RNAs derived from the indicated tRNA and 5S rRNA genes in maf1Δ cells (versus the time 0 values of wild-type cells) are plotted as a function of the duration of phosphate starvation.

Figure 7.

Polyadenylated tRNAs in phosphate-starved maf1Δ cells have unprocessed 3′ extensions. Strand-specific RNA-seq read densities (top panels; counts/base/million), or the poly(A) sites (bottom panels; raw counts), from maf1Δ cells after 48 h of phosphate starvation are plotted as a function of position across the indicated chromosomal tRNA loci. The chromosomal regions corresponding to the mature tRNAs are depicted above the RNA-seq reads. tRNA exons are denoted by thick black arrows in the direction of synthesis and the PRO.02 intron is denoted by a thin bar connecting the exons. The common x-axis scale is indicated on the bottom left. Each poly(A) site is represented by a vertical bar at single-nucleotide resolution. The RNA-seq read densities and poly(A) sites were determined from cumulative counts of three RNA-seq replicates.

Figure 8.

Phosphate-starved maf1Δ cells overproduce tRNA. (A) Total RNA (5 μg) from wild-type and maf1Δ cells harvested prior to (0 h) and 12, 24, 36, and 48 h after transfer to phosphate-free medium was resolved by urea-PAGE (on the same gel) in parallel with denatured DNA size markers. Nucleic acids were visualized by staining the gel with Sybr Green II dye. The positions and sizes (nt) of the DNA markers are indicated on the right. The wild-type RNA gel is the same as shown in Figure 4A and is reprised here to highlight the overproduction of tRNAs and accumulation of tRNA fragments (tRFs) in phosphate-starved maf1Δ cells. (B–F) Northern analysis of equal aliquots of total RNA from maf1Δ cells harvested prior to (0 h) and after 4, 8, 12, 24, 36, and 48 h of phosphate starvation was performed as described in Figure 4. Membranes hybridized to ³²P-labeled DNA oligonucleotide probes complementary to 7SL RNA (panel B), 5S rRNA (panel C), 5.8S rRNA (panel D), tRNA-Trp (panel E) or tRNA-Phe (panel F). The positions and sizes (nt) of the DNA markers are indicated on the right in each panel. The signal intensities of the ³²P probe annealed to full-sized 7SL rRNA, 5.8S rRNA, tRNA-Trp, and tRNA-Phe, normalized to the respective RNA levels at time 0 (defined as 1), are shown below the lanes in panels B–F.

Figure 9.

Phosphate-starved maf1Δ cells accumulate intron-containing pre-tRNAs and unspliced tRNAs fragments. Northern blots of RNA isolated from maf1Δ cells prior to (0 h) and at the times specified after transfer to phosphate-free medium were hybridized to ³²P-labeled DNA oligonucleotide probes complementary to the 5′ and 3′ exons of tRNA-Lys-CTT (panels A and F), tRNA-Tyr (panels B and G), tRNA-Val-AAC (panels C and H), and tRNA-Leu-CAA (panels D and I) or to the introns of tRNA-Leu-CAA (panel E) and tRNA-Val-AAC (panel J). The positions and sizes (nt) of the DNA markers are indicated on the right in each panel. The signal intensities of the ³²P exon probes annealed to full-sized tRNAs, normalized to the respective tRNA levels at time 0 (defined as 1) are shown below the lanes in panels A–D and F–I. The signal intensities of the ³²P intron probes annealed to intron-containing pre-tRNAs, normalized to the respective pre-tRNA levels at time 0, are shown below the lanes in panels E and J.

Figure 10.

Phosphate-starved maf1Δ cells accumulate 5′-extended and 3′-extended pre-tRNA-Val.08 and unspliced 5′-extended and 3′-extended pre-tRNA-Val.08 exons. Northern blots of RNA isolated from maf1Δ cells prior to (0 h) and at the times specified after transfer to phosphate-free medium were hybridized to ³²P-labeled DNA oligonucleotide probes complementary to the 5′ leader (left panel) or 3′ trailer (right panel) of tRNA-Val.08. The positions and sizes (nt) of the DNA markers are indicated. The terminally extended tRNA species migrating between the 90-nt and 110-nt DNA markers are taken to correspond to 5′- or 3′-extended intron-containing pre-tRNA-Val and 5′- or 3′-extended spliced tRNA-Val, respectively. The species migrating between the 34-nt and 67-nt markers are 5′- or 3′-extended unspliced tRNA exon fragments. The signal intensities of the three species were quantified and summed for each time point, then normalized to the time 0 values (defined as 1) and expressed as fold increase below each lane.