The ecl family gene ecl3⁺ is induced by phosphate starvation and contributes to sexual differentiation in fission yeast

There are three ecl family genes in Schizosaccharomyces pombe, namely, ecl1⁺, ecl2⁺ and ecl3⁺, which encode small proteins composed of less than 100 amino acids that are predicted to bind Zn²⁺ through conserved cysteine residues (Ohtsuka and Aiba, 2017; Shimasaki et al., 2017). These genes act on the extension of chronological lifespan (CLS), which is defined as the survival period of the nondividing cell population, measured by the viability after entry into the stationary phase in yeast (Naito et al., 2014; Masumura et al., 2019; Lee et al., 2021; Romila et al., 2021; Ohtsuka et al., 2021b, 2022c; Legon and Rallis, 2022). These genes are conserved in fungi, including Saccharomyces cerevisiae (Azuma et al., 2012; Ohtsuka and Aiba, 2017; Zhang et al., 2021). In S. pombe, ecl family genes are induced by various signals, including starvation of the nitrogen source, sulfur source, amino acids and Mg²⁺, and then induce CLS extension, sexual differentiation and autophagy (Miwa et al., 2011; Shimasaki et al., 2020; Ohtsuka et al., 2021a,c, 2022b). The Ecl family proteins also inhibit the activity of target of rapamycin complex 1 (TORC1) and contribute to the regulation of the cell cycle under starvation (Ohtsuka et al., 2022a). Under sulfur starvation, ecl1⁺ is induced by the transcription factor Zip1 (Ohtsuka et al., 2017). Under starvation of amino acids or Mg²⁺, the transcription factor Fil1 induces ecl1⁺ through the general amino acid control system (Ohtsuka et al., 2019, 2021a). Oxidative stress induces ecl1⁺ through the transcription factor Atf1 (Shimasaki et al., 2014). Heat stress induces ecl2⁺ through the heat shock transcription factor Hsf1 (Ohtsuka et al., 2011). However, although several conditions that induce ecl1⁺ and ecl2⁺ have been identified, any environmental conditions inducing ecl3⁺ have not been clarified.

In S. pombe, intracellular phosphate levels are maintained by the phosphate signal transduction pathway (PHO pathway), which regulates gene expression in response to phosphate starvation (Henry et al., 2011; Estill et al., 2015; Zhou et al., 2016). The PHO pathway is regulated by three phosphate uptake genes that are primarily controlled by the phosphate homeostasis regulon (Garg et al., 2022): pho1⁺, which encodes a cell-surface acid phosphatase, a secreted glycoprotein required for extracellular phosphate uptake (Schweingruber et al., 1992); pho84⁺, which encodes an inorganic phosphate transporter, one of five phosphate transporters in S. pombe (Sawada et al., 2021); and tgp1⁺, which encodes a glycerophosphate transporter (Garg et al., 2018a). It has been reported that the degradation of phosphate transporters is also involved in the maintenance of intracellular phosphate homeostasis, e.g. SPX-RING-type ubiquitin ligase Pqr1 regulates phosphate uptake through the ubiquitination and degradation of phosphate transporters, including Pho84 (Sawada et al., 2021). According to studies, phosphate starvation induces the phosphorylation of the stress-response mitogen-activated kinase Sty1 and causes autophagy (Zuin et al., 2010; Corral-Ramos et al., 2022).

The Zn₂Cys₆ transcription factor Pho7 is a crucial factor acting on the response to phosphate starvation (Estill et al., 2015; Schwer et al., 2017; Garg et al., 2019). Pho7 binds to the 5′-TCG(G/C)(A/T)xxTTxAA-3′ DNA motif, but no ortholog has been identified in S. cerevisiae (Vishwanatha et al., 2016; Schwer et al., 2017). Pho7 regulates the expression of genes involved in the PHO pathway, including pho1⁺, pho84⁺ and tgp1⁺ (Carter-O'Connell et al., 2012; Schwer et al., 2017). The expression of pho1⁺, pho84⁺ and tgp1⁺ is regulated by the PHO regulon, and their transcription is negatively regulated by 5′-flanking interfering upstream long noncoding RNAs (lncRNAs), namely, nc-pho1/prt (pho-repressive transcript), prt2 and nc-tgp1, respectively (Shah et al., 2014; Sanchez et al., 2018; Garg et al., 2020; Shuman, 2020). The PHO regulon is activated by slowly transcribing RNA polymerase II (Pol II) through a mechanism involving alternative polyadenylation and premature transcription termination of lncRNAs (Yague-Sanz et al., 2020). Additionally, it has been suggested that the expression of lncRNAs that regulate PHO genes is regulated by the HomolD box (Garg et al., 2018b). The HomolD-binding protein Rrn7 forms a complex with Pol II, and the phosphorylation of Thr67 of Rrn7 is regulated by casein kinase 2; the phosphorylation of Rrn7 by casein kinase 2 suppresses transcriptional activity regulated by HomolD (Moreira-Ramos et al., 2015; Montes et al., 2017).

In this study, we sought to further understand the relationship between the ecl family genes and starvation response; in addition, we aimed to demonstrate that ecl3⁺ is induced by phosphate starvation with a mechanism similar to that of the PHO regulon and contributes to the sexual differentiation in S. pombe. The ecl family genes regulate appropriate G1 cell cycle arrest by stabilizing the cyclin-dependent kinase (CDK) inhibitor Rum1 and then induce sexual differentiation under phosphate starvation. Three genes, pho1⁺, pho84⁺ and tgp1⁺, are regulated by the PHO regulon and involved in the cellular response to phosphate starvation. Our data suggest that, similar to these genes, ecl3⁺ is also integrated as a member of the PHO regulon and contributes to appropriate cellular responses to phosphate starvation in S. pombe.

It has been reported that ecl3⁺ is upregulated in the asp1 mutant asp1-H397A, which encodes a kinase synthesizing inositol pyrophosphate (IPP) (Sanchez et al., 2019). IPP is a signal molecule regulating phosphate homeostasis in S. cerevisiae and Arabidopsis thaliana, and probably also in S. pombe (Sanchez et al., 2019; Shuman, 2020; Benjamin et al., 2022). pho84⁺ and pho1⁺ are present on chromosome II in order, immediately adjacent to the 5′ end of ecl3⁺. It has been proposed that the upregulation of ecl3⁺ in the asp1 mutant reflects its proximity to prt2, which regulates pho84⁺; the prt2 promoter might be bidirectional, and lncRNAs that are transcribed in the opposite direction to prt2 might interfere with ecl3⁺ transcription (Sanchez et al., 2019). However, it is unclear whether the induction of ecl3⁺ is physiologically significant or merely a secondary effect for pho84⁺ regulation, and whether phosphate starvation itself also induces ecl3⁺. We first examined the expression level of ecl3⁺ to confirm whether phosphate depletion itself regulates ecl3⁺ expression. Results showed that phosphate starvation itself also increased ecl3⁺ expression (Fig. 1A). Moreover, we observed the induction of ecl3⁺ irrespective of the presence or absence of auxotrophy.

We next investigated the mechanism that regulates ecl3⁺ induction. According to studies, genes adjacent to ecl3⁺ on chromosome II, namely pho1⁺ and pho84⁺, are regulated by the transcription factor Pho7 and the CDK-activating protein Csk1 (Carter-O'Connell et al., 2012; Schwer et al., 2017). Furthermore, a previous chromatin immunoprecipitation experiment coupled with high-throughput sequencing analysis demonstrated that Pho7 binds upstream of not only pho1⁺ and pho84⁺ but also ecl3⁺ (Carter-O'Connell et al., 2012). Therefore, we explored whether ecl3⁺ expression is regulated by Pho7 and Csk1. Phosphate starvation induced the expression of ecl3⁺, pho1⁺ and pho84⁺ in a Pho7-dependent manner (Fig. 1B), suggesting that Pho7 directly regulates ecl3⁺ expression under phosphate starvation. Although studies have also reported that Csk1 functions upstream of Pho7 and negatively regulates the expression of pho1⁺ and pho84⁺ (Henry et al., 2011; Carter-O'Connell et al., 2012; Estill et al., 2015), the deletion of csk1⁺ increased the expression of ecl3⁺ as well as that of pho1⁺ and pho84⁺ (Fig. 1C). These data suggest that ecl3⁺, similar to its neighboring genes, pho1⁺ and pho84⁺, was also subjected to transcriptional control by Pho7 and Csk1.

Previous studies have reported that the mRNA level of ecl3⁺ is also negatively regulated by the zinc finger protein Zfs1, which regulates mRNA catabolism and sexual differentiation (Goldar et al., 2005; Wells et al., 2012; Navarro et al., 2017). We next investigated the relationship between Zfs1 and ecl3⁺ induction by phosphate starvation. Although Zfs1 is expected to bind ecl3⁺ transcripts and regulate their levels, the deletion of zfs1⁺ increased the mRNA level of ecl3⁺ only in yeast extract (YE) complete medium and not in Edinburgh minimal medium (EMM) (Fig. S1A,B). In addition, phosphate starvation increased the mRNA level of ecl3⁺ with or without Zfs1 (Fig. S1B). These results suggest that Zfs1 is involved in the repression of the ecl3⁺ transcript levels in a nutrient-rich environment but is not required for induction by phosphate starvation.

Next, we examined how phosphate starvation would affect the expression of other ecl family genes. We measured the expression of ecl1⁺ and ecl2⁺ during phosphate starvation using pho7⁺- or csk1⁺-deficient strains (Fig. S2A,B). In S. pombe wild-type strains, the expression of ecl1⁺ and ecl2⁺ was not dramatically induced by phosphate starvation, as seen for ecl3⁺. In contrast, the deletion of pho7⁺ or csk1⁺ significantly increased the expression of ecl1⁺. Studies have reported that ecl1⁺ is expressed in response to various stimuli by multiple transcription factors, including the transcription factors Atf1, Zip1 and Fil1 (Shimasaki et al., 2014; Ohtsuka et al., 2017, 2019). It has also been reported that ecl1⁺ is subject to gene silencing by the Erh1–Mmi1 complex (Xie et al., 2019), which regulates meiotic mRNA decay and heterochromatin assembly (Sugiyama et al., 2016; Wei et al., 2021). Consistent with this report, a temperature-sensitive mutant of mmi1⁺ exhibited high expression of ecl1⁺ under restrictive temperature (Fig. S2C). It is currently unknown whether the upregulation of ecl1⁺ by pho7⁺ or csk1⁺ deficiency is direct or indirect through other transcription factors or gene silencing. Taken together, these data indicate that phosphate starvation primarily induces the expression of ecl3⁺ among the ecl family genes in a Pho7-dependent manner.

Previous studies have reported that the Δpho7 strain exhibits temperature sensitivity, cold sensitivity, and resistance to 2-deoxyglucose (2-DG) and H₂O₂ (Vishwanatha et al., 2016, 2017; Schwer et al., 2017). We investigated the growth of the Δpho7 and Δecl1/2/3 (Δecls) strains under these stress conditions. Results showed that Δpho7 cells exhibited weak growth retardation under high temperature and strong 2-DG resistance, whereas Δecls cells showed no significant differences from wild-type cells (Fig. S3A,B). However, both Δpho7 and Δecls cells exhibited growth retardation under low temperature conditions. Nevertheless, these sensitivities of Δpho7 cells to high or low temperature were not suppressed by ecl3⁺ expression in cells carrying the pEcl3 plasmid, which expressed ecl3⁺ at a level equal to or higher than that in wild-type cells, even in Δpho7 cells (Fig. S3A,C). This result suggests that these temperature sensitivities of Δpho7 cells are not due to reduced ecl3⁺ expression. Δpho7 cells were resistant to H₂O₂ stress, whereas Δecls cells were sensitive, consistent with a previous report (Shimasaki et al., 2014). Overexpression of ecl3⁺ resulted in sensitivity to 2-DG, but ecl3⁺ expression in Δpho7 cells did not suppress the resistance to 2-DG (Fig. S3B). Based on these results, we could not identify any strong connection between the ecl family gene and pho7⁺ in these stress responses.

lncRNAs that regulate the PHO regulon are believed to be regulated by the HomolD box (Garg et al., 2018b) (Fig. 2A). Casein kinase 2 phosphorylates the HomolD-binding protein Rrn7 at Thr67, inhibiting the binding of Rrn7 to the HomolD box, causing the suppression of lncRNA transcriptional activity (Moreira-Ramos et al., 2015; Shuman, 2020). Casein kinase 2, which consists of the catalytic subunit α and the regulatory subunit β, is a widely conserved serine/threonine kinase essential for survival and proliferation and has several intracellular substrates (Homma, 2008). In S. pombe, cka1⁺ has been identified as encoding a catalytic subunit, and ckb1⁺ and ckb2⁺ have been identified as encoding regulatory subunits (Moreira-Ramos et al., 2015; Nakazawa et al., 2019; Romila et al., 2021). Although there are limited reports about the role of casein kinase 2 in fission yeast, Ckb1 has been suggested to be involved in regulating the Ca²⁺/calcineurin/Prz1 pathway (Ma et al., 2011). To explore whether ecl3⁺ expression is affected by lncRNAs, we investigated the level of ecl3⁺ during phosphate starvation in Δckb1 cells. Phosphate starvation did not induce ecl3⁺ expression in Δckb1 cells, indicating that the induction was dependent on Ckb1 (Fig. 2B). This supports the idea that the lncRNA-mediated PHO regulon, which is regulated through the HomolD box, also regulates the expression of ecl3⁺. In other words, Ckb1 deletion might suppress the phosphorylation of Rrn7 and promote the expression of lncRNAs, which in turn suppresses ecl3⁺ expression even under phosphate starvation.

It has been reported that Δckb1 cells are sensitive to 5-fluorouracil (5-FU) (Mojardín et al., 2015). Unlike Δckb1 cells, Δecls cells showed no significant 5-FU sensitivity (Fig. 2C). Moreover, consistent with previous reports, Δckb1 cells showed poor growth even under non-stress conditions (Roussou and Draetta, 1994; Ma et al., 2011), but Δecls cells revealed no such growth retardation (Fig. 2C).

Given that phosphate starvation induced the ecl family gene that regulates CLS, we investigated the effect of phosphate starvation on CLS. Both the prototroph JY1 and auxotroph JY333 strains exhibited much weaker, if any, CLS extension under phosphate starvation, which was significantly different from glucose restriction (calorie restriction), nitrogen starvation, leucine starvation and sulfur starvation (Fig. S4) (Ohtsuka et al., 2017, 2019; Imai et al., 2020). This might be attributable to the fact that phosphate starvation causes reduced induction of ecl family genes than other nutrient starvation conditions, such as leucine and sulfur starvation, which extend CLS in an ecl family gene-dependent manner; e.g. leucine starvation or sulfur starvation induces ecl1⁺ by approximately 100-fold (Ohtsuka et al., 2017, 2019, 2022d), whereas phosphate starvation induces ecl3⁺ by only approximately 20-fold, which is generally expressed at a lower level in comparison with ecl1⁺.

The ecl family genes are also involved in sexual differentiation, that is, an overexpression of all ecl family genes including ecl3⁺ contributes to the induction of sexual differentiation (Ohtsuka et al., 2015, 2021d). Hence, we examined how ecl family genes affect sexual differentiation during phosphate starvation. The wild-type strain showed increased mating and sporulation rates under phosphate as well as nitrogen starvation (Fig. 3A,B; Fig. S5). Under nitrogen starvation, Δecls cells exhibited a slightly delayed mating rate and similar sporulation rate to that of wild-type cells. However, phosphate starvation led to dramatically decreased mating and delayed sporulation rates. This indicates that ecl family genes are essential for proper sexual differentiation under phosphate starvation.

It has been reported that the transcription factor Ste11 and RNA-binding protein Mei2 play crucial roles in sexual differentiation in S. pombe (Oowatari et al., 2011; Yamashita et al., 2017). As ecl family genes contribute to the expression of these factors (Ohtsuka et al., 2015), we examined their expression during phosphate starvation (Fig. 3C). Phosphate starvation altered the expression levels of mel1⁺, encoding a secreted α-galactosidase (Goddard et al., 2005), and adh4⁺, encoding an alcohol dehydrogenase (Sakurai et al., 2004), partially in an ecl family gene-dependent manner. However, the expression of ste11⁺ and mei2⁺ was induced irrespective of the presence or absence of ecl family genes. This result suggests that the essential function of ecl family genes in sexual differentiation during phosphate starvation is not the contribution to inducing the expression of genes such as ste11⁺ and mei2⁺.

The phosphorylation of Ste11 by Cdc2 (CDK1) decreases its DNA-binding ability and then suppresses mating outside of the G1 phase; only G1 phase cells perform mating (Kjaerulff et al., 2007; Martín et al., 2020). Furthermore, the transcription factor Fkh2, which induces ste11⁺, is repressively phosphorylated by Cdc2, reinforcing the cell cycle regulation of mating (Shimada et al., 2008). We explored the possibility that ecl family genes affect mating through cell cycle control during phosphate starvation. To confirm this possibility, we observed the cell cycle during phosphate starvation using a flow cytometer. Phosphate starvation induced G1 arrest in wild-type cells, but the arrest was significantly delayed in Δecls cells, suggesting that ecl family genes are required for proper G1 arrest under phosphate starvation (Fig. 4). As reported previously (Ohtsuka et al., 2017), also under nitrogen starvation, the G1 arrest was slightly delayed in Δecls cells. These results support the notion that the decrease in mating rate in Δecls cells is due to the failure of proper G1 arrest under phosphate starvation.

To elucidate how ecl family genes cause G1 arrest under phosphate starvation, we measured the protein amount of the CDK inhibitor Rum1, a regulator of Cdc2–cyclin B complexes that maintains low CDK activity in G1 and is required for G1 arrest under nitrogen starvation (Benito et al., 1998; Rubio et al., 2018; García-Blanco and Moreno, 2019). Moreover, under nutrient-rich conditions, Rum1 is phosphorylated and then degraded by SCF^Pop1/Pop2 (Skp1–Cullin1–F-box), causing the increase of CDK activity and S-phase progression (MacKenzie and Lacefield, 2020; Stonyte et al., 2020). Under nitrogen starvation, Rum1 is stabilized, suppresses CDK activity and promotes G1 arrest.

Similar to nitrogen starvation, phosphate starvation also resulted in the accumulation of Rum1 in wild-type cells (Fig. 5A; Fig. S6), suggesting that Rum1 contributes to G1 arrest not only under nitrogen starvation but also under phosphate starvation. However, in the absence of ecl family genes, Rum1 accumulation was not detected even after 1 day of phosphate starvation (Fig. 5A), suggesting that ecl family genes contribute to Rum1 accumulation during phosphate starvation.

We next investigated whether overexpression of ecl family genes would promote Rum1 accumulation under phosphate starvation. The induction of each ecl family gene, namely ecl1⁺, ecl2⁺ and ecl3⁺, using plasmids with the nmt1 promoter, resulted in Rum1 accumulation (Fig. 5B). This result supports the assertion that ecl family genes promote Rum1 accumulation.

Because the overexpression of ecl family genes resulted in Rum1 accumulation, we next investigated the cell cycle during phosphate starvation in cells with overexpression of ecl family genes. In control cells, G1-arrested cells were observed after 6 h of phosphate starvation, whereas in the ecl family gene-overexpressing cells, G1-arrested cells were observed from 4 h onward, indicating that ecl family genes promote G1 arrest (Fig. 5C). However, because G1-arrested cells were not detected before phosphate starvation in the ecl family gene-overexpressing cells, overexpression of ecl family genes alone was not sufficient to induce G1 arrest, suggesting that both ecl family genes and the starvation signal are required for G1 arrest. These data indicate that ecl family genes can promote G1 arrest by inducing the accumulation of Rum1 during phosphate starvation.

Because Ecl family proteins reportedly suppress the activity of TORC1 under sulfur starvation (Ohtsuka et al., 2022a), the function of ecl family genes might be partially dependent on TORC1. Similar to the overexpression of ecl family genes, the suppression of TORC1 induces G1 arrest and sexual differentiation (Álvarez and Moreno, 2006). Hence, we investigated their involvement under phosphate starvation.

First, to examine TORC1 activity during phosphate starvation, we measured the level of TORC1-added phosphorylation of Psk1, a TORC1 direct targets (Nakashima et al., 2012; Otsubo et al., 2017). We found that in the presence of ecl family genes, phosphate starvation decreased the phosphorylation of Psk1, suggesting the decreased activity of TORC1 (Fig. 6A). In contrast, the decrease in Psk1 phosphorylation was partially suppressed in the absence of ecl family genes. This indicates that ecl family genes can repress TORC1 under phosphate starvation. However, the decrease of phosphorylated Psk1 level appeared to be less under phosphate starvation than under sulfur starvation (Ohtsuka et al., 2022a). This might be because the expression of ecl family genes is less increased under phosphate starvation than under sulfur starvation (Ohtsuka et al., 2017). Moreover, it appears that the increase in total Psk1 protein level was observed under phosphate starvation (Fig. 6A), which might contribute to the increase in phosphorylated Psk1 level.

Because ecl family genes also repressed TORC1 under phosphate starvation, we next investigated the involvement of TORC1 in cellular responses that require ecl family genes, under phosphate starvation. As the CDK inhibitor Rum1 was stabilized in an ecl family gene-dependent manner under phosphate starvation (Fig. 5), we expected that TORC1 suppression could stabilize Rum1 even in Δecls cells, if TORC1 acts downstream of ecl family genes. Treatment with rapamycin and caffeine is known to inhibit TORC1 in fission yeast (Takahara and Maeda, 2012; Rallis et al., 2013). However, in our study, the reduction of Rum1 level in Δecls cells during phosphate starvation was not restored by treatment with rapamycin and caffeine (Fig. 6B). This implied that TORC1 suppression was not sufficient for Rum1 stabilization under phosphate starvation.

Next, we explored the involvement of TORC1 in sexual differentiation under phosphate starvation. Similar to the results shown in Fig. 3A, the ecl family genes were required for mating during phosphate starvation (Fig. 6C). Treatment with rapamycin and caffeine did not restore the mating defect of Δecls cells, suggesting that the mating defect of Δecls cells is not restored by TORC1 suppression and that ecl family genes are essential for mating under phosphate starvation. In contrast, Δecls cells showed a low sporulation rate as depicted in Fig. 3B, which was recovered by drug treatment (Fig. 6C). The reduction of the sporulation rate in Δecls cells might be partially caused by the defect of TORC1 suppression.

The ecl family genes might be required for G1 arrest under phosphate starvation, and, thus, although Δecls cells can sporulate from diploid cells, they cannot conjugate from haploid cells. Finally, we examined whether the defect of G1 arrest in Δecls cells could be restored by TORC1 suppression. We observed the cell cycles during phosphate starvation with rapamycin and caffeine treatment by flow cytometry (Fig. 6D). The defective G1 arrest in Δecls cells was not restored by treatment with these drugs. Conversely, these drugs hastened G1 arrest only in the presence of ecl family genes, suggesting that these genes are required for the promotion of G1 arrest by TORC1 suppression. Our data suggested that, during phosphate starvation, ecl family genes suppress TORC1 activity and are also required for appropriate cellular responses induced by TORC1 suppression, including G1 arrest.

This study demonstrated that phosphate starvation induces ecl3⁺ expression and ecl family genes contribute to appropriate G1 arrest and sexual differentiation under phosphate starvation. Moreover, the induction of ecl3⁺ was significantly dependent on the transcription factor Pho7 (Fig. 1B). ecl3⁺ is located at the bases 4,441,609–4,441,340 of chromosome II, and pho84⁺ and pho1⁺, which act in the PHO pathway, are located upstream of its 5′ side (Fig. 7A). All these genes are regulated by Pho7, and there are multiple Pho7-binding sites in the chromosome region from ecl3⁺ to pho1⁺ (Carter-O'Connell et al., 2012). Because there are also Pho7-binding sites upstream of ecl3⁺, Pho7 is considered to directly induce the transcription of ecl3⁺ in the same manner as that by pho1⁺ and pho84⁺.

In S. pombe, ecl family genes are induced in response to various starvation conditions and environmental stresses (Ohtsuka and Aiba, 2017) (Fig. 7B). Because Ecl1, Ecl2 and Ecl3 have similar amino acid sequences, all of them can complement the Δecls phenotype, and each elicits similar intracellular responses, so their molecular functions are also probably similar (Ohtsuka et al., 2015; Ohtsuka and Aiba, 2017). Using these three ecl family genes, S. pombe might induce appropriate intracellular responses to various environmental stresses and nutritional starvations, leading to the maintenance of cell survival, including CLS extension and induction of sexual differentiation. Although environmental stresses and nutritional starvation are extremely diverse, and because some of these stimuli are transmitted and converged to ecl family genes, it is expected that the S. pombe cell can exhibit similar cellular responses to various environmental stimuli.

Phosphate starvation results in a G1 arrest that is mediated by ecl family genes (Figs 4 and 5C). Similar to phosphate starvation, amino acid starvation causes cells to undergo G1 arrest, and this arrest also depends on ecl family genes (Ohtsuka et al., 2019). However, sulfur starvation, as well as phosphate or amino acid starvation, also induces ecl family genes, but cells undergo arrest in the G2 phase (Ohtsuka et al., 2017). The CDK inhibitor Rum1 plays a vital role in the G1 arrest of cells (Stonyte et al., 2020). In this study, we found that Rum1 stabilization is essential for proper G1 arrest during phosphate starvation and that ecl family genes contribute to this (Fig. 5). In contrast, during sulfur starvation, Rum1 is degraded even in the presence of ecl family genes (Ohtsuka et al., 2022a). This indicates that ecl family genes contribute to the stabilization of Rum1, but the induction of ecl family genes alone is not sufficient to stabilize Rum1. This does not contradict the fact that the overexpression of ecl family genes alone did not cause either Rum1 stabilization or G1 arrest (0 h in Fig. 5B,C), and both the presence of ecl family genes and phosphate starvation were required for their responses. It is not yet known how Ecl family proteins stabilize Rum1 under phosphate starvation.

Similar to other PHO regulon genes, ecl3⁺ expression was upregulated in Δcsk1 cells even under phosphate-replete conditions and was induced to a lower degree by phosphate starvation (Fig. 1C). Moreover, ecl3⁺ is located adjacent to a PHO gene, pho84⁺, suggesting that the mechanism of ecl3⁺ induction is extremely similar to that of other PHO regulon genes (Sanchez et al., 2019; Garg et al., 2020; Schwer et al., 2021). Studies have also reported that regulation of the PHO regulon in S. pombe is mediated by upstream lncRNAs regulated by C-terminal domain (CTD) phosphorylation of Pol II, which is partially regulated by Csk1 via Cdk9, and by silencing through the methylation of histone H3 Lys9 at PHO gene regions, causing the inhibition of Pho7-binding to the promoter site of PHO genes (Shah et al., 2014; Yague-Sanz et al., 2020). The result that ecl3⁺ expression was regulated not only by Pho7 but also by Csk1 supports the idea that ecl3⁺ is regulated by the PHO regulon. In addition, consistent with this idea, the deletion of Ckb1, the regulatory subunit of casein kinase 2, which mediates the HomolD box-regulated transcription, including the lncRNA of PHO regulon, did not induce ecl3⁺ under phosphate starvation.

Various factors, such as Csk1, Pho7, Ecl family proteins, including Ecl3, and Rum1, are considered to be involved in sequential cellular processes, ranging from perception of phosphate starvation to conjugation (Fig. 7C). Csk1 might recruit Cdk9 to the Pol II complex through the phosphorylation of Mcs6; Cdk9 phosphorylates the Pol II CTD and the Spt5–Spt4 complex, resulting in the alleviation of Pol II elongation arrest at the promoter proximal position (Hermand et al., 1998; Pei and Shuman, 2003; Viladevall et al., 2009; Schwer et al., 2009). Thus, it has been suggested that transcriptional elongation is stimulated through reactions involving the phosphorylation of Pol II CTD by Mcs6 and Cdk9 (Coudreuse et al., 2010). In contrast, a non-Mcs6-mediated pathway has also been proposed for the action of Csk1 on Cdk9 (Gerber et al., 2008). It is believed that the Csk1-mediated regulation of transcriptional elongation also targets the lncRNA located upstream of PHO genes, and the inhibition of Csk1 suppresses lncRNA elongation (Carter-O'Connell et al., 2012; Estill et al., 2015; Sanchez et al., 2018). Insufficient elongation of lncRNAs might allow the binding of Pho7 to the upstream sequences of PHO genes, including ecl3⁺. Our data suggested that the induction of ecl3⁺ by Pho7 causes stabilization of Rum1 and suppression of Cdc2 activity, promoting G1 arrest. Simultaneously, the suppression of Cdc2, which phosphorylates and inhibits factors related to sexual differentiation, such as Ste11 and Fkh2 (Kjaerulff et al., 2007; Shimada et al., 2008), is also considered to cause sexual differentiation in fission yeast.

S. pombe strains are listed in Table S1. The S. pombe Deletion Mutant Library from Bioneer (http://us.bioneer.com/products/spombe/spombeoverview.aspx) was also used. The correct genotypes of the deletion mutants have been tested by using appropriate primers with PCR. Cells were grown in EMM supplemented with essential nutrients (Takuma et al., 2013). For nitrogen- and phosphate-depleted media, NH₄Cl and Na₂HPO₄ were omitted from EMM, respectively. The amounts of supplemental nutrients were as follows: 40 µg ml⁻¹ adenine, 60 µg ml⁻¹ leucine and 20 µg ml⁻¹ uracil. Cells were grown at 30°C unless otherwise stated.

The following plasmids were used: pLB-Dblet (Ohtsuka et al., 2009), pEcl3 (Ohtsuka et al., 2009), pREP1 (Ohtsuka et al., 2008), pREP1-Ecl1 (Ohtsuka et al., 2008), pREP1-Ecl2 (Ohtsuka et al., 2009) and pREP1-Ecl3 (Ohtsuka et al., 2009).

The FY7288Δecl1/2/3 and JY333 Rum1HA h⁹⁰ strains were generated by mating cells of the FY7288 strain from the National Bio-Resource Project, Japan, with cells of the JY333Δecl1/2/3 strain (Ohtsuka et al., 2015). The ED668Δecl1/2/3 h⁺ strain was generated by mating cells of the ED668 strain with cells of the JY333Δecl1/2/3 strain (Ohtsuka et al., 2015). The JY333Δcsk1 and JY333Δpho7 strains were generated by mating cells of the JY333 strain with cells of Δcsk1 and Δpho7 strains from the S. pombe Deletion Mutant Library from Bioneer, respectively. The correct genotypes of these strains have been tested using appropriate primers with PCR.

To construct JS183, we fused a 13MYC-tag to the C-terminus of the Psk1 protein on the chromosomes of JY1 as described previously (Sato et al., 2005). For ecl1⁺, ecl2⁺ and ecl3⁺ disruptions, the ORF regions of Ecl proteins were replaced with kan^r, nat^r and bsd^r cassettes, respectively, using previously described methods (Sato et al., 2005; Fujita et al., 2015). The primers used for this purpose are listed in Table S1. The correct genotypes of these strains have been tested using appropriate primers with PCR.

All data were calculated by counting at least 300 cells. The mating and sporulation rates were calculated as described previously (Ohtsuka et al., 2015).

Real-time PCR analysis was performed as described previously (Hibi et al., 2018) using the housekeeping gene cdc2⁺ as a control. The primers are listed in Table S1.

Western blotting was performed as described previously (Hibi et al., 2018). Rum1–HA was detected with the anti-HA (12CA5) antibody (1:10,000; Roche, 11666606001), Cdc2 with the anti-CDK1 antibody (Y100.4) (1:10,000; Abcam, ab5467), Psk1–myc with the MYC/c-Myc antibody (1:2000; Santa Cruz Biotechnology, sc-40), phospho-Psk1-myc with the anti-phospho-Psk1/RPS6KB1/p70 S6 kinase (Thr389) antibody (1:1000; Cell Signaling Technology, 9206), γ-tubulin with the monoclonal anti-Gtb1/γ-tubulin antibody (1:2000; Sigma-Aldrich, T5326), and α-tubulin with the monoclonal anti-α-tubulin antibody (1:10,000; Sigma-Aldrich, T6074).

The cells were fixed with 70% ethanol and treated with an RNase-solution (Sigma-Aldrich) to measure the DNA content for 1 h at 37°C. Nucleic acids were stained with propidium iodide (PI). Flow cytometry analysis was performed using the Attune Acoustic Focusing Cytometer (Life Technologies). More than 4000 cells were examined for each sample; the vertical axis of the graph indicates the maximum percentage, and the maximum values of the peak are 100%. The PI intensity was detected at a voltage of 2.2 V on the BL2 channel.

Quantitative data shown in figures represent the average of at least three independent experiments ±s.d. Statistical analyses were performed by two-tailed unpaired Student's t-test.

The authors thank Enago for editing the manuscript. The FY7288 strain was provided by the National Bio-Resource Project (Yeast Genetic Resource Center) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.