The Greatwall–Endosulfine Switch Accelerates Autophagic Flux during the Cell Divisions Leading to G1 Arrest and Entry into Quiescence in Fission Yeast

Entry into quiescence in the fission yeast Schizosaccharomyces pombe is induced by nitrogen starvation. In the absence of nitrogen, proliferating fission yeast cells divide twice without cell growth and undergo cell cycle arrest in G1 before becoming G0 quiescent cells. Under these conditions, autophagy is induced to produce enough nitrogen for the two successive cell divisions that take place before the G1 arrest. In parallel to the induction of autophagy, the Greatwall–Endosulfine switch is activated upon nitrogen starvation to down-regulate protein phosphatase PP2A/B55 activity, which is essential for cell cycle arrest in G1 and implementation of the quiescent program. Here we show that, although inactivation of PP2A/B55 by the Greatwall–Endosulfine switch is not required to promote autophagy initiation, it increases autophagic flux at least in part by upregulating the expression of a number of autophagy-related genes.


Introduction
Autophagy is a highly conserved catabolic process, from yeast to humans, that helps maintain cellular homeostasis by recycling cellular components in nutrient-or energydeficient conditions [1][2][3]. During autophagy, cytoplasmic components and organelles (ribosomes, mitochondria, etc.) are incorporated into double-membrane vesicles or autophagosomes [3,4] and delivered to the degradative compartment of the cell, the vacuole in yeast, or the lysosome in animal cells, where acid hydrolases degrade the autophagosome contents before being recycled to the cytoplasm and used in biosynthetic pathways.
Autophagosome formation is a multi-step process mediated by a conserved group of proteins encoded by autophagy-related genes (ATG), first discovered in the budding yeast Saccharomyces cerevisiae [1,5]. The initial step of this pathway is regulated by the Atg1 complex in yeast [6,7] or the Ulk1 complex in mammalian cells [8][9][10]. Autophagy is normally activated by nutritional stress signals. The Target of Rapamycin (TOR) kinase, which plays a key role in promoting cell growth, inhibits autophagy. TOR is a serine/threonine protein kinase that is part of two complexes, TOR complex 1 (TORC1) and TOR complex 2 (TORC2), with conserved functions throughout evolution. Phosphorylation of Atg13, part of the Atg1 complex, by TORC1 prevents the activation of the Atg1 complex in budding and fission yeasts [11,12] and animal cells [13,14], thus blocking autophagy under nutrient-rich conditions.

TORC1 and Greatwall Protein Kinase Activity Oscillates during Nitrogen Starvation
Nitrogen-starved fission yeast cells rapidly inactivate TORC1. Therefore, Greatwall is activated, promoting the phosphorylation of Igo1 at Ser64, which can be monitored using specific antibodies that recognize the conserved phospho-Ser64 epitope. Figure 2 shows that the levels of phosphorylated Igo1 at Ser64 increased in wild type cells, reaching a maximum in 30 min before decreasing after 4 h of nitrogen starvation. This oscillation in Greatwall activity mirrored TORC1 activity (Figure 2, Rps6-P signal). This result clearly indicates that both TORC1 and Greatwall activity oscillate out of phase during the two cell divisions that occur after nitrogen starvation. The transient reactivation of TORC1 after 4 h of nitrogen starvation could be explained by the generation of recycled nitrogen, mainly amino acids, in the vacuole by autophagy (see Section 2.3, Figure 3). Reactivation of TORC1 inactivated Greatwall (Figure 2, time 4 h). TORC1 reactivation also occurred in mutants lacking Endosulfine (igo1∆), but in this case, with a delay of about 4 h ( Figure 2, time 8 h).
A possible interpretation of this result could be that the onset of autophagy is delayed in the igo1∆ mutant compared to the wild type (see Section 2.3, Figure 3). This experiment clearly indicates that in fission yeast, TORC1 and Greatwall protein kinase activities oscillate out of phase during the two cell divisions that precede G1 arrest after nitrogen starvation. mainly amino acids, in the vacuole by autophagy (see Section 2.3, Figure 3). Reactivati of TORC1 inactivated Greatwall (Figure 2, time 4 h). TORC1 reactivation also occurred mutants lacking Endosulfine (igo1∆), but in this case, with a delay of about 4 h (Figure time 8 h). A possible interpretation of this result could be that the onset of autophagy delayed in the igo1∆ mutant compared to the wild type (see Section 2.3, Figure 3). Th experiment clearly indicates that in fission yeast, TORC1 and Greatwall protein kina activities oscillate out of phase during the two cell divisions that precede G1 arrest aft nitrogen starvation. Figure 2. TORC1 and Greatwall activity oscillate out of phase during nitrogen starvation. Wild ty and igo1∆ cells grown in minimal medium at 25 °C until mid-exponential phase were transferred nitrogen-free minimal medium (MM-N) for 8 h at 25 °C. Samples were taken at 0, 0.5, 1, 2, 4, and h in MM-N. Greatwall and TORC1 activities were determined using specific anti-phospho-Ser Igo1 and anti-Rps6-Ser235/236 antibodies. Igo1 protein levels were determined using anti-Igo1 a tibodies. Tubulin was used as a loading control. The Igo1-P/Igo1 and Rps6-P/Tubulin ratios are dicated under the gels. Underlined numbers correspond to the peaks of activity.

The Greatwall (Ppk18, Cek1 and Ppk31)-Endosulfine (Igo1) Switch Accelerates Autophag Flux
To test whether the Ppk18, Cek1, and Ppk31-Igo1 switch play a role in the regulati of autophagy, we decided to assess autophagic flux in different mutants of this pathw using the processing of N-terminally-tagged Atg8 with CFP as a readout of autophag CFP is resistant to protein degradation in the vacuole, and thus, when autophagy is duced, a free CFP band can be detected with Western blot [20][21][22].
Autophagy was induced by transferring the fission yeast cells from minimal mediu (MM) to minimal medium without nitrogen (MM-N) [21,22]. Samples were collected different times for Western blot analysis. CFP:Atg8 was processed to free CFP in CFP:At wild type cells after 2-4 h in MM-N ( Figure 3, wt). In cells lacking Endosulfine (igo1∆), t appearance of the free CFP band was delayed and less intense compared to the wild ty ( Figure 3, wt vs. igo1∆). The intensity of the CFP band in the igo1∆ mutant increased w time but did not reach the levels of wild type cells, suggesting that Igo1 is required promote a normal autophagic flux under nitrogen starvation conditions. As a negati control of autophagy, we used the atg13∆ mutant, in which the formation of the Atg1 i tiation complex is blocked ( Figure 3). The delay in autophagy flux in the igo1∆ mutant w Figure 2. TORC1 and Greatwall activity oscillate out of phase during nitrogen starvation. Wild type and igo1∆ cells grown in minimal medium at 25 • C until mid-exponential phase were transferred to nitrogen-free minimal medium (MM-N) for 8 h at 25 • C. Samples were taken at 0, 0.5, 1, 2, 4, and 8 h in MM-N. Greatwall and TORC1 activities were determined using specific anti-phospho-Ser64-Igo1 and anti-Rps6-Ser235/236 antibodies. Igo1 protein levels were determined using anti-Igo1 antibodies. Tubulin was used as a loading control. The Igo1-P/Igo1 and Rps6-P/Tubulin ratios are indicated under the gels. Underlined numbers correspond to the peaks of activity.

The Greatwall (Ppk18, Cek1 and Ppk31)-Endosulfine (Igo1) Switch Accelerates Autophagic Flux
To test whether the Ppk18, Cek1, and Ppk31-Igo1 switch play a role in the regulation of autophagy, we decided to assess autophagic flux in different mutants of this pathway using the processing of N-terminally-tagged Atg8 with CFP as a readout of autophagy. CFP is resistant to protein degradation in the vacuole, and thus, when autophagy is induced, a free CFP band can be detected with Western blot [20][21][22].
Autophagy was induced by transferring the fission yeast cells from minimal medium (MM) to minimal medium without nitrogen (MM-N) [21,22]. Samples were collected at different times for Western blot analysis. CFP:Atg8 was processed to free CFP in CFP:Atg8 wild type cells after 2-4 h in MM-N (Figure 3, wt). In cells lacking Endosulfine (igo1∆), the appearance of the free CFP band was delayed and less intense compared to the wild type ( Figure 3, wt vs. igo1∆). The intensity of the CFP band in the igo1∆ mutant increased with time but did not reach the levels of wild type cells, suggesting that Igo1 is required to promote a normal autophagic flux under nitrogen starvation conditions. As a negative control of autophagy, we used the atg13∆ mutant, in which the formation of the Atg1 initiation complex is blocked (Figure 3). The delay in autophagy flux in the igo1∆ mutant was also observed in cells expressing Pgk1:GFP (Supplementary Figure S2), which has been used as an alternative readout of autophagy in fission yeast [23].
To establish whether Greatwall is also involved in the regulation of autophagy, we assessed the CFP:Atg8 processing in mutants of the fission yeast Greatwall orthologues Ppk18, Cek1, and Ppk31. Ppk18 is the main kinase responsible for Igo1 Ser64 phosphorylation and activation, while Cek1 plays a secondary and redundant role [15], and Ppk31 appears to play no role [16]. When autophagy is induced in these mutants, we could observe that only ppk18∆ showed a delay in the appearance of free CFP, a similar but less severe phenotype than the igo1∆ mutant ( Figure 4A). In the ppk18∆ cek1∆ double mutant that completely lacks Greatwall activity [15], we observed a severe reduction in CFP:Atg8 processing comparable to that of igo1∆ ( Figure 4B). The cek1∆ ppk31∆ double mutant also showed a slight delay in the generation of free CFP ( Figure 4B). Together, these results suggest that the Greatwall kinases Ppk18 and, to a lesser extent, Cek1 and Ppk31 play a role in nitrogen starvation-induced autophagy.  Figure S2), which has been used as an alternative readout of autophagy in fission yeast [23]. To establish whether Greatwall is also involved in the regulation of autophagy, we assessed the CFP:Atg8 processing in mutants of the fission yeast Greatwall orthologues Ppk18, Cek1, and Ppk31. Ppk18 is the main kinase responsible for Igo1 Ser64 phosphorylation and activation, while Cek1 plays a secondary and redundant role [15], and Ppk31 appears to play no role [16]. When autophagy is induced in these mutants, we could observe that only ppk18Δ showed a delay in the appearance of free CFP, a similar but less severe phenotype than the igo1Δ mutant ( Figure 4A). In the ppk18Δ cek1Δ double mutant that completely lacks Greatwall activity [15], we observed a severe reduction in CFP:Atg8 processing comparable to that of igo1∆ ( Figure 4B). The cek1Δ ppk31∆ double mutant also showed a slight delay in the generation of free CFP ( Figure 4B). Together, these results suggest that the Greatwall kinases Ppk18 and, to a lesser extent, Cek1 and Ppk31 play a role in nitrogen starvation-induced autophagy.  Figure S2), which has been used as an alternative readout of autophagy in fission yeast [23]. To establish whether Greatwall is also involved in the regulation of autophagy, we assessed the CFP:Atg8 processing in mutants of the fission yeast Greatwall orthologues Ppk18, Cek1, and Ppk31. Ppk18 is the main kinase responsible for Igo1 Ser64 phosphorylation and activation, while Cek1 plays a secondary and redundant role [15], and Ppk31 appears to play no role [16]. When autophagy is induced in these mutants, we could observe that only ppk18Δ showed a delay in the appearance of free CFP, a similar but less severe phenotype than the igo1Δ mutant ( Figure 4A). In the ppk18Δ cek1Δ double mutant that completely lacks Greatwall activity [15], we observed a severe reduction in CFP:Atg8 processing comparable to that of igo1∆ ( Figure 4B). The cek1Δ ppk31∆ double mutant also showed a slight delay in the generation of free CFP ( Figure 4B). Together, these results suggest that the Greatwall kinases Ppk18 and, to a lesser extent, Cek1 and Ppk31 play a role in nitrogen starvation-induced autophagy.  To determine whether Greatwall activation by itself is capable of inducing autophagy, we monitored autophagy in cells overexpressing ppk18 + . Moderate overexpression of ppk18 + from the nmt41 promoter generates small cells that show an extended G1 phase ( [15]; Figure 1C). Strong overexpression of ppk18 + from the nmt1 promoter also generates small cells that fail to divide and show a phenotype similar to quiescent cells (Pérez-Hidalgo, unpublished results). Cells overexpressing ppk18 + for up to 46 h in MM showed limited signs of CFP:Atg8 processing ( Figure 5, left panel), indicating that strong overexpression of ppk18 + in nitrogenrich medium is not sufficient to induce autophagy. However, these cells overexpressing ppk18 + for 22 h showed a slight acceleration of the autophagic flux when deprived of nitrogen ( Figure 5, middle and right panels), indicating that although Greatwall activation is unable to induce autophagy in nitrogen-rich medium, it can promote an increase in autophagy flux once autophagy is induced by nitrogen starvation. Similar results were obtained when Pgk1:GFP was used as an autophagy readout (Supplementary Figure S3). cleavage is indicated under the gels.
To determine whether Greatwall activation by itself is capable of inducing autophagy, we monitored autophagy in cells overexpressing ppk18 + . Moderate overexpression of ppk18 + from the nmt41 promoter generates small cells that show an extended G1 phase ( [15]; Figure 1C). Strong overexpression of ppk18 + from the nmt1 promoter also generates small cells that fail to divide and show a phenotype similar to quiescent cells (Pérez-Hidalgo, unpublished results). Cells overexpressing ppk18 + for up to 46 h in MM showed limited signs of CFP:Atg8 processing ( Figure 5, left panel), indicating that strong overexpression of ppk18 + in nitrogen-rich medium is not sufficient to induce autophagy. However, these cells overexpressing ppk18 + for 22 h showed a slight acceleration of the autophagic flux when deprived of nitrogen ( Figure 5, middle and right panels), indicating that although Greatwall activation is unable to induce autophagy in nitrogen-rich medium, it can promote an increase in autophagy flux once autophagy is induced by nitrogen starvation. Similar results were obtained when Pgk1:GFP was used as an autophagy readout (Supplementary Figure S3). Interestingly, ppk18 + overexpression upregulates the expression of a number of autophagy-related genes, including cpy1 + , atg2402 + , atg43 + , atg1 + , atg13 + , atg1801 + , atg4 + , atg20 + and atg7 + ( Figure 6A and Supplementary Table S1), indicating that the Greatwall-Endosulfine-PP2A/B55 pathway may be upregulating the expression of a subset of genes encoding proteins involved in autophagy. Consistent with this idea, Greatwall (ppk18∆ cek1∆) and Endosulfine (igo1∆) mutants showed a reduction in the expression of autophagy response genes compared to the wild type after 1 h or 4 h of nitrogen starvation ( Figure 6B and Supplementary Tables S2 and S3). These results suggest that increased expression of autophagy-related genes through Greatwall activation may precondition the autophagy response to nitrogen starvation. Interestingly, ppk18 + overexpression upregulates the expression of a number of autophagyrelated genes, including cpy1 + , atg2402 + , atg43 + , atg1 + , atg13 + , atg1801 + , atg4 + , atg20 + and atg7 + ( Figure 6A and Supplementary Table S1), indicating that the Greatwall-Endosulfine-PP2A/B55 pathway may be upregulating the expression of a subset of genes encoding proteins involved in autophagy. Consistent with this idea, Greatwall (ppk18∆ cek1∆) and Endosulfine (igo1∆) mutants showed a reduction in the expression of autophagy response genes compared to the wild type after 1 h or 4 h of nitrogen starvation ( Figure 6B and Supplementary Tables S2 and S3). These results suggest that increased expression of autophagy-related genes through Greatwall activation may precondition the autophagy response to nitrogen starvation.

S6 Kinases Are Negative Regulators of Autophagy
Ppk18 activity is negatively regulated by TORC1 and S6K [15]. In mammalian cells, TORC1 phosphorylates and activates S6 Kinase [24]. In S. pombe, three proteins show homology to S6 kinases: Sck1, Sck2, and Psk1 [25]. To test whether S6 kinases regulate autophagy, we examined the processing of CFP:Atg8 in S6 kinase deletion mutants (Figure 7). The sck1∆, sck2∆, and psk1∆ strains showed more free CFP compared to the wild type strain (Figure 7). This result suggests that the fission yeast S6 kinases negatively modulate autophagic flux, probably by negatively regulating Ppk18 and Cek1 kinase activity.

S6 Kinases Are Negative Regulators of Autophagy
Ppk18 activity is negatively regulated by TORC1 and S6K [15]. In mammalian cells, TORC1 phosphorylates and activates S6 Kinase [24]. In S. pombe, three proteins show homology to S6 kinases: Sck1, Sck2, and Psk1 [25]. To test whether S6 kinases regulate autophagy, we examined the processing of CFP:Atg8 in S6 kinase deletion mutants ( Figure  7). The sck1∆, sck2Δ, and psk1Δ strains showed more free CFP compared to the wild type strain (Figure 7). This result suggests that the fission yeast S6 kinases negatively modulate autophagic flux, probably by negatively regulating Ppk18 and Cek1 kinase activity.

S6 Kinases Are Negative Regulators of Autophagy
Ppk18 activity is negatively regulated by TORC1 and S6K [15]. In mammalian cells TORC1 phosphorylates and activates S6 Kinase [24]. In S. pombe, three proteins show homology to S6 kinases: Sck1, Sck2, and Psk1 [25]. To test whether S6 kinases regulate autophagy, we examined the processing of CFP:Atg8 in S6 kinase deletion mutants ( Figure  7). The sck1∆, sck2Δ, and psk1Δ strains showed more free CFP compared to the wild type strain (Figure 7). This result suggests that the fission yeast S6 kinases negatively modulate autophagic flux, probably by negatively regulating Ppk18 and Cek1 kinase activity.

PP2A/Pab1 Negatively Regulates Nitrogen Starvation-Induced Autophagy
Finally, we tested the possible role of PP2A/B55 in autophagy as the element acting downstream in the pathway. The PP2A/Pab1 phosphatase complex consists of three subunits: the structural subunit Paa1, the catalytic subunits Ppa1 or Ppa2, and the regulatory subunit Pab1. One possible explanation for slow autophagic flux in Greatwall (ppk18∆ cek1∆) and Endosulfine (igo1∆) mutants could be that PP2A/Pab1 activity is not downregulated during nitrogen starvation in these mutants. High levels of PP2A/B55 activity in nitrogen-starved cells could inhibit autophagy. To test this hypothesis, we reduced the levels of PP2A/Pab1 activity either by repressing pab1 + gene expression, using cells expressing the pab1 + gene from the nmt41 promoter, or by deleting the pab1 + gene. In both cases, autophagy flux increased slightly ( Figure 8A), indicating that PP2A/Pab1 activity inhibits autophagy.

PP2A/Pab1 Negatively Regulates Nitrogen Starvation-Induced Autophagy
Finally, we tested the possible role of PP2A/B55 in autophagy as the element acting downstream in the pathway. The PP2A/Pab1 phosphatase complex consists of three subunits: the structural subunit Paa1, the catalytic subunits Ppa1 or Ppa2, and the regulatory subunit Pab1. One possible explanation for slow autophagic flux in Greatwall (ppk18∆ cek1∆) and Endosulfine (igo1∆) mutants could be that PP2A/Pab1 activity is not downregulated during nitrogen starvation in these mutants. High levels of PP2A/B55 activity in nitrogen-starved cells could inhibit autophagy. To test this hypothesis, we reduced the levels of PP2A/Pab1 activity either by repressing pab1 + gene expression, using cells expressing the pab1 + gene from the nmt41 promoter, or by deleting the pab1 + gene. In both cases, autophagy flux increased slightly ( Figure 8A), indicating that PP2A/Pab1 activity inhibits autophagy. To confirm this result, we deleted the ppa2 + gene, which encodes the main catalytic subunit of PP2A that contributes to 80% of the phosphatase activity [26]. Reduction of PP2A activity in ppa2∆ cells also increased autophagy flux ( Figure 8B). Furthermore, deletion of ppa2 + rescued the igo1∆ phenotype, suggesting that PP2A/Pab1 is acting downstream of Igo1 and that high levels of PP2A/Pab1 are indeed responsible for the inhibition of autophagy in the igo1∆ mutant ( Figure 8B).
Taken together, these results point to a model in which PP2A/Pab1 protein phosphatase is acting as an inhibitor of nitrogen starvation-induced autophagy. To confirm this result, we deleted the ppa2 + gene, which encodes the main catalytic subunit of PP2A that contributes to 80% of the phosphatase activity [26]. Reduction of PP2A activity in ppa2∆ cells also increased autophagy flux ( Figure 8B). Furthermore, deletion of ppa2 + rescued the igo1∆ phenotype, suggesting that PP2A/Pab1 is acting downstream of Igo1 and that high levels of PP2A/Pab1 are indeed responsible for the inhibition of autophagy in the igo1∆ mutant ( Figure 8B).
Taken together, these results point to a model in which PP2A/Pab1 protein phosphatase is acting as an inhibitor of nitrogen starvation-induced autophagy.
In addition to studying the link between autophagy and the Greatwall-Endosulfine-PP2A/B55 (Ppk18,Cek1-Igo1-PP2A/Pab1) pathway, we tested the effect on autophagy of different elements acting upstream and downstream of the pathway. In the presence of nitrogen, the S6 kinase orthologues Sck1, Sck2, and Psk1 are phosphorylated and activated by TORC1, similar to what occurs in mammalian cells. Although in mammalian cells, a relationship between S6K1 and Mastl, a Greatwall orthologue, has not been described, in S. pombe, Sck2 negatively regulates Ppk18 [15]. In mammalian cells, mTORC1 inhibits autophagy by phosphorylating the Ulk1/Ulk2 initiation complex, orthologue of Atg1 in yeast. An increase in autophagy has been described with the use of S6K1 inhibitors [39]. On the other hand, it appears that S6K1 might be required for autophagosome maturation when autophagy is induced by serum starvation. A lower level of S6K1 results in an altered autophagic flux in which autophagosomes are accumulated and there is a reduction in the number of lysosomes [40]. Our results point to the S6 kinase orthologues as inhibitors of the autophagic flux, as cells carrying a deletion of sck2 + or psk1 + , and, to a lesser extent, sck1 + showed more CFP:Atg8 processing.

Fission Yeast Strains and Methods
The fission yeast strains used in this study are listed in Supplementary Table S4. Fission yeast cells were grown and manipulated genetically according to standard protocols [41]. Genetic crosses were performed on malt extract agar plates. Cells were typically grown overnight in yeast extract supplemented with adenine, leucine, histidine, lysine, and uracil (YES) and then transferred to Edinburgh minimal medium containing 93.5 mM ammonium chloride (MM). Nitrogen starvation experiments were performed by transferring cells to minimal medium without nitrogen (MM-N). Cells were grown to mid-exponential phase, centrifuged, and washed three times in MM-N. Experiments were performed at 25 • C.
For the overexpression of ppk18+ from the nmt1 + promoter, cells were grown to midexponential phase in MM containing 5 µg/mL thiamine, harvested, washed twice with MM, and grown in fresh MM without thiamine at 25 • C for the times indicated in the figures. The maximum level of ppk18+ overexpression was reached after 12-14 h of culture in MM.

Strain Construction
A PCR-based strategy was used to insert the GFP tag to express Pgk1 fused to GFP at the pgk1 + genomic locus. Oligonucleotides with 80 bases of homology to the sequence upstream and downstream of the stop codon of the target gene were used to amplify the tag and resistance cassette sequence from plasmid pFA6a-GFP-kanMX6 and pFA6a-GFP-hphMX6. PCR products obtained using High Fidelity DNA polymerase (Roche, Basel, Switzerland) were used to transform a wild type strain h − 972 (S2726).

Flow Cytometry
Samples of 10 7 cells were fixed in 70% (v/v) ethanol and then washed with a solution of 50 mM sodium citrate and resuspended in 0.5 mL of 50 mM of sodium citrate containing 0.1 mg/mL RNase A and incubated overnight at 37 • C. Subsequently, 0.5 mL of 50 mM of sodium citrate containing 4 µg/mL of propidium iodide was added. Cell suspensions were sonicated before analysis in a BD FACSCalibur Flow Cytometer (BD Biosciences, Franklin Lakes, NJ, USA). Analysis was performed using BD Cell Quest ProTM 6.0.3 (BD Biosciences, Franklin Lakes, NJ, USA).

Microscopy
Images were acquired with a Nikon Eclipse 90i microscope coupled to an Orca ER camera and equipped with MetaMorph software (Molecular Devices, San José, CA, USA). Images were processed and assembled with ImageJ software v1.53k.