The AMPK-TORC1 signalling axis regulates caffeine-mediated DNA damage checkpoint override and cell cycle effects in fission yeast

  1. John. P. Alao
  2. Charalampos Rallis

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Caffeine, a widely consumed neuroactive compound, induces DNA damage checkpoint signalling override, and enhances sensitivity to DNA damaging agents. However, the precise underlying mechanisms have remained elusive. In fission yeast S. pombe , the Ataxia Telangiectasia Mutated (ATM) and Ataxia Telangiectasia mutated Related (ATR) orthologue Rad3 has been proposed as the cellular target of caffeine. Nevertheless, recent studies suggest that the Target of Rapamycin Complex 1 (TORC1) might be the main target. Caffeine mimics the effects of activating the Sty1-regulated stress response and the AMP-Activated Protein Kinase (AMPK) homologue Ssp1-Ssp2 pathways on cell cycle progression. Direct inhibition of TORC1 with the ATP-competitive inhibitor torin1, is sufficient to override DNA damage checkpoint signalling. It is, therefore, plausible, that caffeine modulates cell cycle kinetics by indirectly suppressing TORC1 through Ssp2 activation. S sp1 and ssp2 deletion suppresses the effects of caffeine on cell cycle progression. In contrast, direct inhibition of TORC1 advances cell division in these mutants. These observations suggest that caffeine overrides DNA damage signalling, in part, via the indirect inhibition of TORC1 through Ssp2 activation. Alternatively, Ssp1 and Ssp2 may potentiate the effect of caffeine on Cdc25 activity. The AMPK-mTORC1 signalling axis plays an important role in aging and disease and presents a potential target for chemo- and radio-sensitization. Our results provide further insights of the underlying mechanisms by which caffeine modulates cell cycle progression in the context of Ssp1-AMPKα Ssp2 -TORC1 signalling activities and can potentially aid in the development of novel dietary regimens, therapeutics, and chemo-sensitizing agents.

Article activity feed

  1. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    1. General Statements [optional]

    We would like to thank all reviewers for their constructive feedback and for raising specific points that have helped to improve our manuscript. We accept that the initial submission did not include some quantitative aspects of the observed effects. These are now included together with all the suggested experiments from the reviewers with the use of additional mutants and appropriate protein markers. We believe that the manuscript offers a conceptual advance and a molecular mechanism for the effects of caffeine on cell cycle progression of eukaryotic cells and is of interest to geneticists working on cell cycle, cancer and biogerontology.

    Reply to the reviewers

    Reviewer #1 (Evidence, reproducibility and clarity):

    Summary:

    In the manuscript “The AMPK-TORC1 signaling axis regulates caffeine-mediated DNA damage checkpoint override and cell cycle effects in fission yeast,” the authors studied the role of genes that are potentially involved in the caffeine-mediated override of a cell cycle arrest caused by activation of the DNA damage checkpoint. The methylxanthine substance caffeine has been known to override the DNA damage checkpoint arrest and enhance sensitivity to DNA damaging agents. While caffeine was reported to target the ATM ortholog Rad3, the authors previously reported that caffeine targets TORC1 (Rallis et al, Aging Cell, 2013). Inhibition of TORC1, like caffeine, was also reported to override DNA damage checkpoint signaling. Therefore, in the present study, the authors compared the effects of caffeine and torin1 (a potent inhibitor for TORC1 and TORC2) on cell cycle arrest caused by phleomycin, a DNA damaging agent, using various gene deletion S. pombe mutants.

    The authors concluded that they identified a novel role of Ssp1 (calcium/calmodulin-dependent protein kinase) and Ssp2 (catalytic subunit of AMP-activated kinase) in the cell cycle effects caused by caffeine, based on the following findings; (1) the caffeine-mediated DNA damage checkpoint override requires Ssp1 and Ssp2; (2) Ssp1 and Ssp2 are required for caffeine-induced hypersensitivity against phleomycin; (3) under normal growth conditions, caffeine leads to a sustained increase of the septation index in a Ssp2-dependent manner; (4) Caffeine activates Ssp2 and partially inhibits TORC1.

    Major comments:

    I do not think that many of the authors’ claims are supported by the results of the present study. The corresponding parts are detailed below.

    1. The conclusion of the first paragraph in the Results (top in page 6; Our findings indicate that caffeine and torin1 indirectly and directly inhibit TORC1 activity respectively.) is not supported by the data in Figure 1. The result that caffeine, but not torin1, requires Ssp1 and Ssp2 to override the phleomycin-induced cell cycle arrest does not necessarily indicate that caffeine indirectly inhibits TORC1 via Ssp1 and Ssp2. Rather, the authors should mention that this conclusion is based on the authors’ previous reports by citing them (e.g., Rallis et al, Sci Rep, 2017). To add to Figure 1, an additional experiment using a constitutively active AMPK mutant, a temperature-sensitive TORC1 mutant, and a srk1 deletion mutant will help the authors claim their original conclusion as one possibility.

    Torin1 inhibits TORC1 and 2 leading to G2 cell cycle arrest following accelerated mitosis. In contrast, caffeine has been reported to enhance the inhibitory effect of rapamycin on TORC1 signaling but does not inhibit growth. It has not been reported that TORC1 is a direct target of rapamycin. We previously demonstrated that caffeine induces Srk1 in a Sty1 dependent manner (Alao et al., 2014). Furthermore, Ssp1 plays a role in regulating Srk1/ Cdc25 activity. It is therefore possible, that Ssp1 influences the ability of caffeine to promote mitotic progression as part of the stress response while also affecting TORC1 activity via Ssp2. As ssp2∆ cells have higher intrinsic TORC1 activity, this could also attenuate the effect of caffeine on mitosis.

    We have modified the first paragraph of the results section to address the reviewer’s concerns.

    We have previously reported that Srk1 modulates the ability of caffeine to drive cells into mitosis (Alao et al., 2014).

    1. The conclusion of the second paragraph in the Results (lower-middle in page 6; Our results indicate that caffeine induces the activation of Ssp2.) is not based on the results of Figure 2. Figure 2 simply illustrates that both caffeine and torin1 cause hypersensitivity to phleomycin dependent on Ssp1 and Ssp2.

    We appreciate the reviewer’s contention and have modified the text.

    1. The conclusion of the fourth paragraph in the Results (middle in page 7) is not clearly supported by the result, due to an insufficient data analysis. As the cell length and the progress through mitosis are the key assay parameters in Figure 3, the average cell length should be shown next to each micrograph of Figure 3A and 3B. In Figure 3C, a mitotic index and the average cell length should be shown next to each micrograph. A statistical analysis is necessary for the authors to compare the measurements and to claim as the headline (Caffeine exacerbates the ssp1D phenotype under environmental stress conditions), as the effect of caffeine was not evident._

    We have conducted additional experiments to measure cell length and modified the figure to include this data. We believe our observation that caffeine alone induces increased cell length in ssp1 mutants, confirms a role for the Ssp1 protein in modulating the effects of caffeine. We previously showed that Caffeine activates Srk1 which in turn inhibits Cdc25 activity similar to other environmental stresses (Alao et al., 2014). Ssp1 negatively regulates Srk1 following exposure to stress. In contrast, caffeine advances mitosis in wt cells and thus does not result in increased cell length. We also demonstrate that caffeine greatly enhances cell length in ssp1 mutants exposed to heat stress in marked contrast to rapamycin and torin1. These findings indicate that Ssp1 mediates the effect of caffeine on mitosis.

    1. In the middle of page 8, the statement “Accordingly, the effect of caffeine and torin1 on DNA damage sensitivity was attenuated in gsk3D mutants (Figure 5C and 5D).” is not supported by the corresponding results. Rather, Figure 5C and 5D look almost the same.

    We agree with this and other reviewers that demonstrating enhanced sensitivity to caffeine is problematic. Nonetheless, our cell cycle data clearly indicate a differential role for Gsk3 in mediating the cell cycle effects of caffeine and torin1. In terms of DNA damage sensitivity, we have reproducibly observed a lower degree of DNA damage sensitivity in gsk3 mutants relative to wt cells. Hence, while caffeine is less effective at enhancing DNA damage sensitivity relative to torin1 in wt cells; we observed that caffeine and torin1 increase DNA damage sensitivity to a similar degree in gsk3 mutants.

    1. The description and the conclusion of the last paragraph in the Results (bottom in page 8 – page 9) are not supported by the results of Figure 6, due to an insufficient data analysis. The extent of phosphorylation must be quantified as a ratio of the phosphorylated species (e.g., pSsp2) to all species of the protein (e.g., Ssp2).

    We have carefully repeated our experiments under various conditions. Our results clearly indicate caffeine induced Ssp2 phosphorylation. These observations have not been reported previously.

    From Figure 6, the authors claim that caffeine (10 mM) partially inhibits TORC1 signaling. However, the authors previously showed that the same concentration of caffeine inhibited phosphorylation of ribosome S6 kinase as strongly as rapamycin, the potent TOR inhibitor (Rallis et al, Aging Cell, 2013). The authors are advised to assess phosphorylation of S6 kinase again in the present study and compare to the results of the present results in Figure 6, because addition of that data may allow the authors to discuss that caffeine affects TORC1 downstream pathways at different intensities.

    While rapamycin is a strong inhibitor of TORC1 in budding yeast, this is not the case in fission yeast. Our previous assessments of p-S6 levels and polysomal profiles as well as cell-cycle progression kinetics have shown this (Rallis et al, Aging Cell, 2013). In addition, gene expression analysis from our previous studies have shown that caffeine treatment results in a gene expression profile similar to that of cells in nitrogen starvation (TORC1 inhibition).

    We have now used an Sck1-HA strain to further enhance our study and address the reviewer’s concerns. Previous studies have shown that 100 ng/mL rapamycin does not affect Sck1 phosphorylation. We demonstrate that in contrast to rapamycin (100 ng/ mL) 10 mM caffeine affects Sck1-HA expression and or phosphorylation. This effect was also observed with 5 µM torin1 albeit to a greater degree.

    Also, immunoblotting of the same proteins looks somehow different from panel to panel (e.g., pSsp2 in panel A and D; Actin in panel A, C, and D). Therefore, the blotting result before clipping had better be shown as a supplementary material.

    We repeated the blots were necessary and used ponceau S as a loading control. The original blots can be made available to all.

    Minor comments:

    1. (Figure 1) The septation index of the phleomycin-treated cells (without any further additional drugs) should be shown, as a baseline.

    We have included data for untreated cultures and phleomycin-only treated cultures.

    1. (Figure 1D, Optional) As a ppk18D cek1D double deletion mutant is reported, the authors are advised to add and test that mutant in this experiment.

    We have added the related data for the ppk18Δ cek1Δ double mutant.

    1. (Figure 2) The authors need to clarify the number of cell bodies spotted (e.g., in the Figure legend).

    We have modified the figure legend accordingly.

    1. (Figure 3) The different number of cells in micrographs may give an (wrong) impression on the cell proliferation rate. Therefore, it is advisable to use the micrographs in which the similar number of cells are shown for conditions with the similar cell proliferation rates.

    We have included data to show the cell lengths under different conditions. We find that different conditions greatly affect proliferation rates. For instance, cells do not proliferate in the presence of torin1. We initially sought to investigate if caffeine induces a phenotype in ssp1 mutants by virtue of its interaction with the DNA damage response. The micrographs were included as representative examples and have been now complemented with cell length data.

    1. (Figure 4B) ssp2D, not spp2D.

    The figure legend has been edited.

    1. (Figure 4) The septation index of the none-treated cells should be shown as a baseline.

    We have included base line data for untreated wt cells in figure 1. We have no reason to suspect any of the mutants would provide different results over the time investigated.

    1. (Figure 6B, 6E) What do the black arrows indicate? Figure Legend does not seem to explain them.

    The legend has been modified to indicate what the arrows refer to.

    1. (Figure 6C) Indicate which part of the Maf1-PK blot corresponds to the phosphorylated species, because Maf1-PK is probed with an anti-V5 (not a phosphorylation-specific) antibody.

    These experiments have been carefully repeated under different conditions and the figure is now modified accordingly.

    1. (Figure 6D) gsk3Dssp1D, not gs3Dssp1D.

    We have deleted this figure and have now replaced it with data we believe is more appropriate.

    Reviewer #1 (Significance):

    As caffeine is implicated in protective effects against diseases including cancer and improved responses to clinical therapies, the topic of the present study is of interest and importance to the broad audience.

    In the present study, the most significant finding is that caffeine- and torin1-induced hypersensitivity to phleomycin is dependent on Ssp1 and Ssp2 (Figure 2). This result may be important in chemotherapy against cancers. On the other hand, caffeine is known to activate AMPK (e.g., Jensen Am J Physiol Endocrinol, 2007). Besides, as detailed in the Major comments, many of the major conclusions are not supported by the present results. Therefore, based on my field of expertise (cell cycle, cell proliferation, and TOR signaling), I conclude that the present study hardly extends the knowledge in the field of "the cell biology of caffeine."_

    We thank the reviewer for their helpful comments. We accept the constructive criticisms and have carried out extensive additional experiments to provide further roles for Ssp2 and TORC1, in mediating the cell cycle effects of caffeine. We stress that caffeine has previously been proposed its effects via inhibition of Rad3 activity. Our previous work showed that caffeine did not inhibit Rad3 mediated checkpoint signaling. As later studies suggested caffeine inhibited TORC1 activity, the major goal was to investigate if caffeine is an indirect inhibitor of TORC1 via Ssp2 which is activated by several stresses. It has never been demonstrated that caffeine signals via Ssp2. This study provides the first evidence that caffeine modulates cell cycle progression by at least partially signaling via Ssp2 and TORC1. After nearly 30 years, it is vital that its precise activity, in particular enhancing DNA damage sensitivity is properly characterized. Such work woold open the way for additional studies on how caffeine activates cell physiology. For instance, we show that caffeine at 10 mM is more effective at inhibiting Sck1 activity than Rapamycin at 100 ng/ ml. In contrast, rapamycin at this concentration is more effective at inhibiting Maf1 activity. Hence further studies on how exactly the combination of caffeine and rapamycin influences their effect on ageing and other TORC1 regulated processes.

    Reviewer #2 (Evidence, reproducibility and clarity):

    Summary: In this paper, Alao and Rallis analyze the role of AMPK and TORC1 pathways, and the respective crosstalk, in regulating cell cycle progression in the presence of DNA damage in S. pombe. The authors show, almost exclusively through chemo-genetic epistasis assays, that caffeine inhibits TORC1 indirectly activating AMPK, in contrast to the specific ATP-competitive TORC1 inhibitor torin1. Specifically, it is shown that in the absence of a functional AMPK pathway caffeine is unable to revert the TORC1-inhibition-dependent override of cell-cycle arrest caused by the DNA-damaging agent phleomycin, henceforth partially suppressing the growth inhibition caused by the co-treatment.

    Major comments: The overall story of the paper is convincing. However, the choice of an almost exclusively chemo-genetic approach, lack of controls in some experiments and some discrepancy in data presentation suggest that the manuscript undergoes revision before the authors claim that their conclusions are fully supported by the results. In detail:

    In Figure 1, graphs of septation indexes are presented separately for each strain. This presentation prevents the reader from clearly comparing the differences of septation caused by genetic background rather than the treatment, i.e. the septation happening by treatment with torin1. I feel it would be better to group the results by drug rather than by strain/mutant. If the results are presented this way because the experiments on different strains were run separately, I further suggest that they are re-run so to always include at least the wt in every run._

    We have included data for untreated and phleomycin only treated wt cells as a reference. Additionally, all experiments were repeated at least 2 times. We have used this assay for over 10 years and have found it to be reproducible and reliable. We are not able to include wt cells in every run as this would be beyond the manpower capacity and time constraints involved. It is also likely that torin1 activity is influenced by the ssp1/ 2 backgrounds due to increased basal TORC1 activity as previously reported. The main goal was to illustrate that caffeine differs from a direct inhibitor such as torin1.

    Furthermore, torin1 inhibits both TORC1 and TORC2 and thus cannot be directly compared to caffeine. We do prove however, in this and other figures that in contrast to torin1 and rapamycin that caffeine signals via targets upstream of TORC1. We can therefore deduce that it functions in a manner similar to other environmental and nutrient stresses, which require with the Ssp1 and Sty1 regulated pathways to advance mitosis and other processes such as autophagy induction.

    In Figure 2C-D, an inconsistency is observable between the phleo+caffeine sensitivity of ssp1Δ and ssp2Δ, the latter retaining a higher sensitivity. Provided that this is not only due to this specific replicate, how would the authors explain such a difference and fit it into their conclusion of a "cascade" signaling with Ssp1 acting upstream of Ssp2?

    We agree that analyzing the different interacting pathways involved, is complex. For instance, Ssp1 is required for suppressing Srk1 following Sty1 activation independently of its effects on Ssp2 and TORC1. Furthermore, basal TORC1 activity is higher in Ssp2 mutants as previously reported. It is likely that Ssp1 exerts a more definitive role as it is required to directly reactivate Cdc25 activity following exposure to stress. In contrast Ssp2 activation eventually results in increased Cdc25 activity via inhibition of PP2A (Figure 8). These experiments are, thus, intended to compliment those in figure1 but the DNA damaging effects of caffeine must also be taken into account.

    In Figure 2I, a huge discrepancy is observable compared to panel 2A in terms of phleo+caffeine (no ATP) sensitivity of wt cells. Here, cells seem to cope well with the phleomycin treatment even if co-treated with caffeine. This renders the main finding of the panel (the effect of phelo+caffeine+ATP) rather uninterpretable.

    We have noted that relevant assays, at least in fission yeast, are influenced by the culture vessels (e.g., plastic type/ glass) as well as the vessel volume (probably due to different aeration, oxygen availability that affects growth and metabolism parameters). We have corrected figure 1a. In terms of ATP, these experiments are highly reproducible even if the exact mechanism remains unclear.

    In Figure 3A, the simple observation of elongation is sometimes hard to assess, for example in the ATP-caused suppression of the effect of torin 1, as also acknowledge by the authors in the text. I feel it would be really necessary to quantify such results on an adequate number of cells.

    We have reproducibly observed this uncharacterized effect of ATP. We have analysed the cell length in additional experiments to show that ATP influences average cell length under these conditions. It is important to note that the effects of phleomycin are pleotropic. For instance, it likely induces cell cycle arrest at various cell cycle phases as well as in early and late G2. Additionally, it may influence other cellular processes such as DNA or compete with drug targets such as TORC1 which is influenced by ATP.

    In Figure 3B,C wt is missing to compare the results in the presence of the same treatments. I understand the focus on Ssp1, but the authors should show the same treatments on wt cells. Similarly, it would be better to show the drug treatments in panel C also at 30{degree sign}C. For the same reasons as in the previous point, quantifications would greatly enhance the credibility of the claims here.

    Previous work by other investigators have shown that wt cells proliferate normally under these conditions. We also show in figure 1 that cell proliferation is not affected under nor cycling conditions in these assays. We have added cell length data that convincingly prove that Ssp1 is required to mediate the mitotic effects of caffeine. It appears that caffeine induces a cell cycle delay that requires Ssp1 to suppress Srk1- mediated Cdc25 inhibition. Furthermore, recent studies have demonstrated that rapamycin (which targets TORC1 downstream of Ssp1) allows cell proliferation at higher temperatures in S. pombe.

    A major point is the almost complete absence of molecular data. Except for Figure 6, the data do not include a detection of the relative activation of the relevant pathways. Figure 6 could hardly fill this gap, since the samples therein analyzed are not the ones utilized in most of the other figures, but simple, single time-point treatment with a single drug. The authors usually refer in the text to previous knowledge about how a treatment influences a pathway. However, they should show it here in their experimental conditions.

    We have performed extensive additional experiments including those suggested by the reviewer. These experiments conclusively show caffeine induces Ssp2 phosphorylation in an Ssp1- dependent manner. We also demonstrate that caffeine attenuates TORC1 signaling. Together with the cell cycle data, our findings strongly suggest caffeine indirectly inhibits TORC1 signaling a manner analogous to other environmental stresses. We also note that the inhibitory effect of caffeine on TORC1 has been demonstrated in several studies. What have provided further evidence for this but have for the first time demonstrated, that caffeine affects Ssp2.

    Minor comments:
    • A different grouping of the experiments/panels would help the reader. For example, Fig. 2I would fit better together with Fig. 3A, to match the composition of the various chapters of the results.

    We have performed additional experiments as suggested by the other reviewers. We believe the data is now easier to understand.

    Torin 1 is sometimes referred to with a capital T or with a lowercase t, especially in the Figures. I suggest to uniform the nomenclature.

    We have edited the text.

    In the results, the authors state that "ATP may increase TORC1 activity or act as a competitive inhibitor towards both compounds.". It's a little bit odd to refer to ATP as a competitive inhibitor of drugs. I would rather be ATP, the physiological agonist, outcompeting two compounds which are working as ATP-competitive inhibitors.

    We have modified the text accordingly.

    Reviewer #2 (Significance):

    The interplay between TORC1 and AMPK is of great interest in the cell signaling field, basically in every model organism.

    The paper provides a conceptual advance in the field showing a genetic interaction between the two pathways using a model organism which has probably been overlooked so far, which is a pity because S. pombe is the best organism to study G2/M cell cycle/size regulation. The story would be of interest especially for an audience working in cell signaling in microorganisms, but not so much (at least at this stage) for the community working on aging, disease and chemo-/radio-sensitization, contrary to what the authors claim. Furthermore, for the above-mentioned reasons, I feel like the authors are a little bit overshooting when claiming (for example in the abstract and in the discussion), that their work provides a clear understanding of the mechanism.
    As requested by Review Commons, I specify that my expertise is on TORC1/AMPK/PKA pathways, on their crosstalk and their regulation by metabolic intermediates.

    We believe that the additional requested experiments have adequately improved the manuscript and support our presented mechanistic model.

    Caffeine is interest in cancer biology and the biogerontology field proven by recent reports on metabolic phenotyping, liver function testing, induction of autophagy and interplay with HIF-1, just to mention a few.

    Reviewer #3 (Evidence, reproducibility and clarity):

    Summary
    This manuscript examines the genetic requirements for checkpoint override by caffeine in the fission yeast model organism. The main outcome is to show that checkpoint override, which has previously been linked to the downregulation of TORC1, is dependent on on the AMPK pathway (Ssp1/Ssp2). Additional analysis of downstream factors and the cross-talking Sty1 pathway implicates Greatwall kinases and Igo1 (PP2A inhibitor - endosulfine analogue) although the pleiotropic nature of these pathways and the rather blunt endpoints of septation index and phleomycin sensitivity makes robust data interpretation difficult.

    Major comments
    For clarity the manuscript would benefit from some restructuring. In particular it would help the reader if the diagram presented in figure 7 was presented first as this would help orientate the reader with the pathways. The mammalian equivalents should be indicated.

    Figure 8 (previously figure 7) summarizes our findings schematically. We believe that it works well at the end as a conclusion to the work and the discussion. Wherever appropriate we have mentioned the mammalian equivalent (e.g., for Rad3).

    For scientific accuracy and clarity the manuscript requires significant attention. For example in the abstract where Rad3 is introduced it is not made clear that this is the fission yeast gene. It would be better to introduce ATR at this point? Anther example in the abstract: 'Deletion of ssp1 and ssp2 suppresses...' should read 'Deletion of ssp1 or ssp2 suppresses...' as the two genes are not deleted in the same strain. I would recommend that the authors carefully revise the manuscript paying close attention to each statement. Fore example on page 4: 'Downstream of TORC1, caffeine failed to accelerate ppk18D but not igo1D and partially overrode DNA damage checkpoint signalling'. It is unclear what the authors mean by accelerate. I assume they mean accelerate cell cycle progression, but there is no direct analysis of cell cycle kinetics in the results. Similarly on page 5: '... ppk18D mutant displayed slower cell cycle kinetics than wild type cells exposed to phleomycin and caffeine or torin1 (Figuer 1D)'. However, the figure shows no cell cycle kinetic analysis.

    We have modified the wording of the abstract according to the reviewer’s suggestions.

    We refer to accelerated progression into mitosis and have edited the text where appropriate. Depending on the type of DNA damage, S. pombe cells transiently or permanently arrest cell cycle progression. It is well known that caffeine overrides these cell cycle DNA damage checkpoints. We previously proved that this was not due to Rad3 inhibition. Additionally, TORC1 (which controls the timing of mitosis) inhibition overrides checkpoint signaling. Our aim was to investigate if caffeine mimics this effect at least partially, via activation of Ssp2. We have demonstrated this is the case, although the basal state of the various mutants can complicate the data analysis in terms of cell cycle progression. Following exposure to phleomycin, this septation index peaks at 60 minutes following exposure to caffeine. In ppk18 mutants this peak was delayed by 30 minutes. Thus, wt and ppk18 mutants proceed through mitosis and cytokinesis at different rates (as determined by measuring the septation index).

    The authors appear to make the assumption that 'Inhibition of DNA damage signalling by caffeine and torin1 enhanced phleomycin sensitivity...' (page 6) but then clearly go on to show that the mutants used are sensitive for other unknown reasons. To make this link it would be necessary to artificially impose a G2 delay and show how much and in which circumstances this reverses the effect on sensitivity of caffeine/torin1. The authors should thus be very clear that they cannot equate sensitivity to 'checkpoint over-ride' and adjust their wording and assumptions accordingly. Assumptions on epistasis need to use the same assay and not equate between assays. As an example F1C and F2D do not equate as phleo+caffeine would be expected to be sensitised above phleo+torin1. This is not commented on in the text. Also on page 7 '... ATP also suppressed the ability of torin1 to override DNA damage checkpoint signalling albeit to a lesser degree (Figure 2I).' However, this figure only shows sensitivity, not septation index.

    We accept that these results can be difficult to interpret. Firstly, caffeine appears to modulate cell cycle progression by various means. We previously demonstrated that it stabilizes Cdc25 independently of checkpoint signaling. However, it also activates Ssp2 which subsequently affects Cdc25 activity via PP2A. Its effect on mitosis can thus differ depending on the context. For instance, igo1 mutants already have high PP2A activity which would affect the subsequent effect of caffeine on Cdc25 activity. Ssp2 on the other hand appears to regulate cell fate according to the nutritional state. Its sensing of nutritional cues is not limited to ATP/ AMP levels as it also regulates the response to amino acid quality (e.g., glutamate versus torin1).

    We have carried out additional experiments on the effect of ATP. While it did affect progression into mitosis, the results were complicated and have not been shown. Instead, we have provided additional data to show that it affects cell length which is an indicator of G2 cell length. In other words, longer cells spend more time in G2 prior to septation.

    We also suspect that caffeine is itself a DNA damaging agent as previously reported in the early 1970s. More recent studies have also indicated a role for Rad3 and DNA repair proteins for tolerance to caffeine. In fact, TORC1 itself has been reported to be required for DNA damage repair. Thus, TORC1 inhibition could potentially enhance DNA damage sensitivity independently of mitotic progression as shown in some of our experiments.

    While we have clearly identified a role for Ssp2 in mediating the cell cycle effects of caffeine, we accept that these findings will require further studies (beyond the scope of this one); to give more insights on how these caffeine- mediated effects occur. What is clear is that caffeine overrides DNA damage checkpoint signaling by at least partially inhibiting TORC1 signaling.

    All the septation index graphs require an untreated (I.e no caffeine or torin1) control.

    We now show in figure 1a, that the septation index does not change over the time period studied, when cells were left untreated. These assays have been routinely used for many years now and are very reproducible. The graphs clearly show the differential effects caffeine and torin1 exert on cell cycle progression in wt and mutant strains exposed to phleomycin.

    Figure 3 is not quantitative and cannot support the conclusions drawn from it. If, for example, the authors wish to demonstrate ATP can suppress checkpoint override (Figure 3A) they should use the same septation assay used before. If this is not possible, then it should be explained why not and an alternative quantitative assay should be developed. It is unclear why the authors include Figure 3B,C at all.

    Ssp2, on the other hand, appears to regulate cell fate according to the nutritional state. Its sensing of nutritional cues is not limited to ATP/AMP levels as it also regulates the response to amino acid quality (e.g., glutamate versus torin1). Additionally, exposure to stress may induce a transient decline in ATP levels. We thus investigated how ATP might affect caffeine or torin1. We could not detect any major changes in the septation index (not shown). Cells exposed to ATP in the presence of caffeine and phleomycin were shorter. We cannot tell how exactly suppresses the effect of caffeine and torin1 on DNA damage sensitivity.

    It is unclear to this reviewer what the significance of the data with gsk3D cells is (Figure 5). The authors should introduce the protein, why there is an expectation that it would have a role in the pathway and explain its relevance. Similarly when discussing the resulting data.

    Gsk3 lies downstream of TORC2 which is inhibited by torin1 but not caffeine. Gsk3 regulates Pub1 stability which is the E3 ligase for Cdc25. We showed previously that caffeine stabilizes Cdc25, suggesting it might interfere with Pub1 activity. Additionally, we are investigating caffeine as an indirect inhibitor of TORC1 with torin1 that directly inhibits both complexes. Our data provide further evidence for a differential effect of caffeine and torin1 on TORC1 signaling. We have modified the text accordingly.

    Figure 5A shows a similar response of wild type cells to phleomycin regarding checkpoint override as was shown in Figure 1A. However Figure 5C is not recognisable as equivalent to Figure 2A, yet both report sensitivity to phleomycin od wild type cells under equivalent circumstances. This is a major concern as to reproducibility of these data. It is also not possible to conclude from either Figure 5C or 5D that caffeine or torin1 treatment is, or is not, sensitising cells to phleomycin treatment, yet this conclusion is made when discussing the data.

    We agree with this and other reviewers that demonstrating enhanced sensitivity to caffeine is problematic. Nonetheless, our cell cycle data clearly indicate a differential role for Gsk3 in mediating the cell cycle effects of caffeine and torin1. In terms of DNA damage sensitivity, we have reproducibly observed a lower degree of DNA damage sensitivity in gsk3 mutants relative to wt cells. Hence, while caffeine is less effective at enhancing DNA damage sensitivity relative to torin1 in wt cells; we observed that caffeine and torin1 increase DNA damage sensitivity to a similar degree in gsk3 mutants.

    Figure 6A shows that caffeine, but not torin1 results in Ssp2 phosphorylation. Is this experiment reproducible and does the total level of Ssp2 increase reproducibly? This should be doe ae and the results discussed. Ideally, the bands would be quantified against actin intensity and presented as a bar graph with standard deviation.

    We have repeated these experiments alone and in combination with phleomycin. This data convincingly show that caffeine but not torin1 induces Ssp2 phosphorylation. In fact, torin1 suppresses Ssp2 phosphorylation, likely due to inhibition of a feedback mechanism resulting from TORC1 inhibition. In contrast, caffeine likely activates Ssp1 via the stress response, which in turn phosphorylates Ssp2.

    Figure 6B, when introduced should explain the background as to why eIF2alpha phosphorylation is a readout of TORC1 activity. Importantly, the figure should be supported by an actin control and 3 repeats quantified. Figure 6C purports to establish that caffeine moderately attenuates Maf1 phosphorylation. To be able to state this, it would be essential to quantify the gel and report repeated results relative to actin and the total levels of Maf1. Similarly Figure6D and 6E require an actin control and would benefit from proper quantification.

    We have repeated the Maf1 experiments to clarify the data and show that caffeine suppresses Sck1 an additional TORC1 phosphorylation target.

    Minor comments
    p3 'cigarette smoke and other gases'?

    We have edited the statement.

    P4 torin1 was dissolved in DMSO (not were)

    We have edited the text.

    p5 phospho not phosphor Ssp2

    We have edited the text.

    p6 exlpain why ppk18 deletion results are surprising. Also this result could be discussed.

    It had been proposed previously, that Ppk18 is the Greatwall homologue in S. pombe and thus the major regulator of PP2A and mitosis downstream of TOCR1. Later studies suggested a redundant role for Cek1 in this pathway. While deletion of cek1 in a ppk18 background modulated the effect of torin1 on cell cycle progression, it did not interfere with the effects of caffeine. At present we cannot account for this observation. We cannot rule out that caffeine activates an additional kinase that regulates Igo1 activity.

    Together our data show that caffeine advances progression into mitosis in a manner that differs from direct inhibition of TORC1 by torin1.

    We have now added the relevant comments on this unexpected observation within the discussion.

    Explain why Cek1 is not tested

    We have now tested a ppk18 cek1 double mutant.

    p6 introduce what pap1 is when first mentioned

    We have introduced PP2APab1 as requested.

    Reviewer #3 (Significance):

    The data show that fission yeast Ssp1/2 has a role in inhibiting TORC1 in response to caffeine and this influences checkpoint override. This is an incremental, but potentially interesting, observation contributing to understanding mechanism(s) of caffeine action. The lack of quantification, the pleiotropic nature of the mutants used and the rather blunt endpoints assayed make it hard to establish to what extent the direct TORC1 inhibition by Ssp2 causes the checkpoint override, which limits is potential impact. The core observation may, however, be of interest to the wider caffeine field. The referee has the perspective of a yeast cell cycle geneticist.

    We thank the reviewer for identifying the significance of the study in understanding the mechanisms of caffeine effects on the cell cycle. We have added all the suggested experiments with additional mutants and protein markers as well quantitative approaches that have appropriately improved the manuscript. We believe that the mechanism provided is of more general interest and not limited to the caffeine field: manipulating the cell cycle and understanding the interplays between growth and stress are of general interest and importance.

    Reviewer #4 (Evidence, reproducibility and clarity):

    The authors provide a series of genetic studies identifying a role for Ssp1-Ssp2 signaling in TORC1-dependent responses to DNA damage. The main assays are cell division (i.e. septation index) and cell viability (i.e. serial dilution spot assays) following treatment with the DNA damaging agent phleomycin. The authors perform these assays in a number of genetic mutant backgrounds to determine which genes and pathways are required for the relevant cellular response. Supporting data also include microscopy images and western blots to test protein phosphorylation. In general, the results support a role for Ssp1-Ssp2 acting upstream of TORC1. However, in several cases the data do not support a straightforward relationship, and it is confusing to parse through a number of intermediate effects, which often vary between different assays. I have provided some specific comments below that might be addressed to strengthen the technical aspects of the manuscript.

    Major

    1. The authors conclude "that caffeine and torin1 indirectly and directly inhibit TORC1 activity respectively" based on Figure 1. This conclusion seems quite strong given the indirect nature of assays in Figure 1, which test septation in the presence of DNA damage. The conclusion would require experiments that assay TORC1 activity itself.

    Both caffeine and torin1 have previously been reported to inhibit TORC1 which controls the timing of mitosis. We sought to investigate if caffeine mediates its effects via the stress response pathway. We have conducted additional experiments which clearly demonstrate that caffeine inhibits TORC1 at least partially via the activation of Ssp2. These observations make sense as we have previously shown that caffeine actives the stress response pathway to activate Srk1 which inhibits Cdc25. More recent studies my others indicate that Ssp1 is required to suppress Srk1 to allow progression into mitosis. This accounts for the failure of ssp1 mutants to advance mitosis under stress conditions. Additionally, Ssp1 activates Ssp2 which leads to the downstream inhibition of TORC1.

    1. Figure 2 needs some explanation to introduce the idea that cell growth reflects an intact DNA damage response that prevented division in the presence of phleomycin. I also felt that the conclusions were very strong given the data, and the authors should discuss each case more carefully. For example, deletion of ssp1 does not really suppress the ability of torin1 to enhance phleo sensitivity (Figure 2C).

    We would not expect the deletion of ssp1 to suppress the effect of torin1 under stress conditions. We have provided further evidence to show that Ssp1 is required to facilitate progression into mitosis at least in the presence of phleomycin or heat stress.

    1. Microscopy imaging in Figure 3 nicely complements some of the other assays. However, it seems important to know if the cells are actively growing in each of these cases. An example is torin and rapamycin shortening ssp1 mutants at 35 degrees: are these cells actively cycling?

    Our aim was to demonstrate that caffeine exacerbates the ssp1 phenotype. This would provide further evidence to show that caffeine exerts its effects at least in part by activating Ssp1. Cells do not cycle in the presence of torin1 as it inhibits both TORC complexes. We have provided additional evidence to show that caffeine does indeed interact with Ssp1. As the primary aim of the study was to determine is caffeine overrides DNA damage via Ssp1 we have not investigated if they are cycling. Their shortened size suggests that rapamycin and torin1 affect cell division in a different manner from caffeine.

    1. From Figure 6A, the authors conclude that caffeine induces phosphorylation of Ssp2. However, it appears that both Ssp2 protein levels and its phosphorylation levels are both increased, which seems an important distinction.

    We have repeated these experiments several times under different conditions. Some proteins become more stable when phosphorylated as has been previously demonstrated for Srk1 for instance.

    1. In Figure 6D, the authors should show separate gsk3 and ssp1 mutants. It seems likely that all phosphorylation of Ssp2 is due to Ssp1, but this should be shown.

    We have replaced the figure with a ssp1 single mutant.

    1. I am confused about Maf1 phosphorylation in Figure 6C. It is increased upon torin1 treatment, but it is discussed as an indicator or TORC1 activity. Does that mean that loss of its phosphorylation correlates with increased TORC1 activity? As written, I thought it was a TORC1 substrate, which led to confusion about its increased phosphorylation upon torin1 treatment.

    Maf1 is phosphorylated by TORC1. Inhibition of TORC1 would thus lead to a loss of phospho-Maf1 moieties and the accumulation of the unphosphorylated form. We have conducted additional experiments and under various conditions to show that caffeine weakly inhibits Maf1 phosphorylation. We note however, that different stresses result in differential outcomes following TORC1 inhibition. As such we have included new data to show that caffeine suppresses the TORC1 target Sck1. In S. pombe Sck1 and Sck2 regulate progression into mitosis.

    Minor

    1. An untreated control should be shown for assays in Figure 1.

    We have included this data for figure 1a.

    1. An untreated control should be shown for assays in Figure 4.

    We have noted in the results for figure 1, that untreated cells and phleomycin only treated cells do not show any changes in septation index over the time course studied in these experiments.

    Reviewer #4 (Significance):

    The study has significance in connecting several conserved and central signaling pathways including TORC1, AMPK, and PP2A. Also, the study uses caffeine and torin1 that have effects in many different cell types. The connection between caffeine and torin1 effects on phleomycin-treated cells was previously established by these researchers. The significance of the current study is providing a genetic pathway for this connection. The significance is partly limited by some of the technical points raised in the previous section, such as some inconsistencies in the strength of results from different assays. Also, the role of these pathways in DNA damage response signaling is not new. While the main significance of this work might relate to a more specialized audience, it does add to a broader body of literature regarding these conserved pathways and processes.

    My expertise is yeast cell biology.

    While the roles of the pathways in DNA damage has been reported usinbg genetic and pharmacological combinations we dissect their relationships and provide mechanistic connections.

    We thank the reviewer for identifying the significance of this study. We believe we have now addressed the technical issues raised.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #4

    Evidence, reproducibility and clarity

    The authors provide a series of genetic studies identifying a role for Ssp1-Ssp2 signaling in TORC1-dependent responses to DNA damage. The main assays are cell division (i.e. septation index) and cell viability (i.e. serial dilution spot assays) following treatment with the DNA damaging agent phleomycin. The authors perform these assays in a number of genetic mutant backgrounds to determine which genes and pathways are required for the relevant cellular response. Supporting data also include microscopy images and western blots to test protein phosphorylation. In general, the results support a role for Ssp1-Ssp2 acting upstream of TORC1. However, in several cases the data do not support a straightforward relationship, and it is confusing to parse through a number of intermediate effects, which often vary between different assays. I have provided some specific comments below that might be addressed to strengthen the technical aspects of the manuscript.

    Major

    1. The authors conclude "that caffeine and torin1 indirectly and directly inhibit TORC1 activity respectively" based on Figure 1. This conclusion seems quite strong given the indirect nature of assays in Figure 1, which test septation in the presence of DNA damage. The conclusion would require experiments that assay TORC1 activity itself.
    2. Figure 2 needs some explanation to introduce the idea that cell growth reflects an intact DNA damage response that prevented division in the presence of phleomycin. I also felt that the conclusions were very strong given the data, and the authors should discuss each case more carefully. For example, deletion of ssp1 does not really suppress the ability of torin1 to enhance phleo sensitivity (Figure 2C).
    3. Microscopy imaging in Figure 3 nicely complements some of the other assays. However, it seems important to know if the cells are actively growing in each of these cases. An example is torin and rapamycin shortening ssp1 mutants at 35 degrees: are these cells actively cycling?
    4. From Figure 6A, the authors conclude that caffeine induces phosphorylation of Ssp2. However, it appears that both Ssp2 protein levels and its phosphorylation levels are both increased, which seems an important distinction.
    5. In Figure 6D, the authors should show separate gsk3 and ssp1 mutants. It seems likely that all phosphorylation of Ssp2 is due to Ssp1, but this should be shown.
    6. I am confused about Maf1 phosphorylation in Figure 6C. It is increased upon torin1 treatment, but it is discussed as an indicator or TORC1 activity. Does that mean that loss of its phosphorylation correlates with increased TORC1 activity? As written, I thought it was a TORC1 substrate, which led to confusion about its increased phosphorylation upon torin1 treatment.

    Minor

    1. An untreated control should be shown for assays in Figure 1.
    2. An untreated control should be shown for assays in Figure 4.

    Significance

    The study has significance in connecting several conserved and central signaling pathways including TORC1, AMPK, and PP2A. Also, the study uses caffeine and torin1 that have effects in many different cell types. The connection between caffeine and torin1 effects on phleomycin-treated cells was previously established by these researchers. The significance of the current study is providing a genetic pathway for this connection. The significance is partly limited by some of the technical points raised in the previous section, such as some inconsistencies in the strength of results from different assays. Also, the role of these pathways in DNA damage response signaling is not new. While the main significance of this work might relate to a more specialized audience, it does add to a broader body of literature regarding these conserved pathways and processes.

    My expertise is yeast cell biology.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary

    This manuscript examines the genetic requirements for checkpoint override by caffeine in the fission yeast model organism. The main outcome is to show that checkpoint override, which has previously been linked to the downregulation of TORC1, is dependent on on the AMPK pathway (Ssp1/Ssp2). Additional analysis of downstream factors and the cross-talking Sty1 pathway implicates Greatwall kinases and Igo1 (PP2A inhibitor - endosulfine analogue) although the pleiotropic nature of these pathways and the rather blunt endpoints of septation index and phleomycin sensitivity makes robust data interpretation difficult.

    Major comments

    For clarity the manuscript would benefit from some restructuring. In particular it would help the reader if the diagram presented in figure 7 was presented first as this would help orientate the reader with the pathways. The mammalian equivalents should be indicated.

    For scientific accuracy and clarity the manuscript requires significant attention. For example in the abstract where Rad3 is introduced it is not made clear that this is the fission yeast gene. It would be better to introduce ATR at this point? Anther example in the abstract: 'Deletion of ssp1 and ssp2 suppresses...' should read 'Deletion of ssp1 or ssp2 suppresses...' as the two genes are not deleted in the same strain. I would recommend that the authors carefully revise the manuscript paying close attention to each statement. Fore example on page 4: 'Downstream of TORC1, caffeine failed to accelerate ppk18D but not igo1D and partially overrode DNA damage checkpoint signalling'. It is unclear what the authors mean by accelerate. I assume they mean accelerate cell cycle progression, but there is no direct analysis of cell cycle kinetics in the results. Similarly on page 5: '... ppk18D mutant displayed slower cell cycle kinetics than wild type cells exposed to phleomycin and caffeine or torin1 (Figuer 1D)'. However, the figure shows no cell cycle kinetic analysis.

    The authors appear to make the assumption that 'Inhibition of DNA damage signalling by caffeine and torin1 enhanced phleomycin sensitivity...' (page 6) but then clearly go on to show that the mutants used are sensitive for other unknown reasons. To make this link it would be necessary to artificially impose a G2 delay and show how much and in which circumstances this reverses the effect on sensitivity of caffeine/torin1. The authors should thus be very clear that they cannot equate sensitivity to 'checkpoint over-ride' and adjust their wording and assumptions accordingly. Assumptions on epistasis need to use the same assay and not equate between assays. As an example F1C and F2D do not equate as phleo+caffeine would be expected to be sensitised above phleo+torin1. This is not commented on in the text. Also on page 7 '... ATP also suppressed the ability of torin1 to override DNA damage checkpoint signalling albeit to a lesser degree (Figure 2I).' However, this figure only shows sensitivity, not septation index.

    All the septation index graphs require an untreated (I.e no caffeine or torin1) control.

    Figure 3 is not quantitative and cannot support the conclusions drawn from it. If, for example, the authors wish to demonstrate ATP can suppress checkpoint override (Figure 3A) they should use the same septation assay used before. If this is not possible, then it should be explained why not and an alternative quantitative assay should be developed. It is unclear why the authors include Figure 3B,C at all.

    It is unclear to this reviewer what the significance of the data with gsk3D cells is (Figure 5). The authors should introduce the protein, why there is an expectation that it would have a role in the pathway and explain its relevance. Similarly when discussing the resulting data.

    Figure 5A shows a similar response of wild type cells to phleomycin regarding checkpoint override as was shown in Figure 1A. However Figure 5C is not recognisable as equivalent to Figure 2A, yet both report sensitivity to phleomycin od wild type cells under equivalent circumstances. This is a major concern as to reproducibility of these data. It is also not possible to conclude from either Figure 5C or 5D that caffeine or torin1 treatment is, or is not, sensitising cells to phleomycin treatment, yet this conclusion is made when discussing the data.

    Figure 6A shows that caffeine, but not torin1 results in Ssp2 phosphorylation. Is this experiment reproducible and does the total level of Ssp2 increase reproducibly? This should be doe ae and the results discussed. Ideally, the bands would be quantified against actin intensity and presented as a bar graph with standard deviation.

    Figure 6B, when introduced should explain the background as to why eIF2alpha phosphorylation is a readout of TORC1 activity. Importantly, the figure should be supported by an actin control and 3 repeats quantified. Figure 6C purports to establish that caffeine moderately attenuates Maf1 phosphorylation. To be able to state this, it would be essential to quantify the gel and report repeated results relative to actin and the total levels of Maf1. Similarly Figure6D and 6E require an actin control and would benefit from proper quantification.

    Minor comments

    p3 'cigarette smoke and other gases'?

    P4 torin1 was dissolved in DMSO (not were)

    p5 phospho not phosphor Ssp2

    p6 exlain why ppk18 deletion results are surprising. Also this result could be discussed.

    Explain why Cek1 is not tested

    p6 introduce what pap1 is when first mentioned

    Significance

    The data show that fission yeast Ssp1/2 has a role in inhibiting TORC1 in response to caffeine and this influences checkpoint override. This is an incremental, but potentially interesting, observation contributing to understanding mechanism(s) of caffeine action. The lack of quantification, the pleiotropic nature of the mutants used and the rather blunt endpoints assayed make it hard to establish to what extent the direct TORC1 inhibition by Ssp2 causes the checkpoint override, which limits is potential impact. The core observation may, however, be of interest to the wider caffeine field. The referee has the perspective of a yeast cell cycle geneticist.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary: In this paper, Alao and Rallis analyze the role of AMPK and TORC1 pathways, and the respective crosstalk, in regulating cell cycle progression in the presence of DNA damage in S. pombe. The authors show, almost exclusively through chemo-genetic epistasis assays, that caffeine inhibits TORC1 indirectly activating AMPK, in contrast to the specific ATP-competitive TORC1 inhibitor torin 1. Specifically, it is shown that in the absence of a functional AMPK pathway caffeine is unable to revert the TORC1-inhibition-dependent override of cell-cycle arrest caused by the DNA-damaging agent phleomycin, henceforth partially suppressing the growth inhibition caused by the co-treatment.

    Major comments: The overall story of the paper is convincing. However, the choice of an almost exclusively chemo-genetic approach, lack of controls in some experiments and some discrepancy in data presentation suggest that the manuscript undergoes revision before the authors claim that their conclusions are fully supported by the results. In detail:

    • In Figure 1, graphs of septation indexes are presented separately for each strain. This presentation prevents the reader from clearly comparing the differences of septation caused by genetic background rather than the treatment, i.e. the septation happening by treatment with torin 1. I feel it would be better to group the results by drug rather than by strain/mutant. If the results are presented this way because the experiments on different strains were run separately, I further suggest that they are re-run so to always include at least the wt in every run.
    • In Figure 2C-D, an inconsistency is observable between the phleo+caffeine sensitivity of ssp1Δ and ssp2Δ, the latter retaining a higher sensitivity. Provided that this is not only due to this specific replicate, how would the authors explain such a difference and fit it into their conclusion of a "cascade" signaling with Ssp1 acting upstream of Ssp2?
    • In Figure 2I, a huge discrepancy is observable compared to panel 2A in terms of phleo+caffeine (no ATP) sensitivity of wt cells. Here, cells seem to cope well with the phleomycin treatment even if co-treated with caffeine. This renders the main finding of the panel (the effect of phelo+caffeine+ATP) rather uninterpretable.
    • In Figure 3A, the simple observation of elongation is sometimes hard to assess, for example in the ATP-caused suppression of the effect of torin 1, as also acknowledge by the authors in the text. I feel it would be really necessary to quantify such results on an adequate number of cells.
    • In Figure 3B,C wt is missing to compare the results in the presence of the same treatments. I understand the focus on Ssp1, but the authors should show the same treatments on wt cells. Similarly, it would be better to show the drug treatments in panel C also at 30{degree sign}C. For the same reasons as in the previous point, quantifications would greatly enhance the credibility of the claims here.
    • A major point is the almost complete absence of molecular data. Except for Figure 6, the data do not include a detection of the relative activation of the relevant pathways. Figure 6 could hardly fill this gap, since the samples therein analyzed are not the ones utilized in most of the other figures, but simple, single time-point treatment with a single drug. The authors usually refer in the text to previous knowledge about how a treatment influences a pathway. However, they should show it here in their experimental conditions.

    Minor comments:

    • A different grouping of the experiments/panels would help the reader. For example, Fig. 2I would fit better together with Fig. 3A, to match the composition of the various chapters of the results.
    • Torin 1 is sometimes referred to with a capital T or with a lowercase t, especially in the Figures. I suggest to uniform the nomenclature.
    • In the results, the authors state that "ATP may increase TORC1 activity or act as a competitive inhibitor towards both compounds.". It's a little bit odd to refer to ATP as a competitive inhibitor of drugs. I would rather be ATP, the physiological agonist, outcompeting two compounds which are working as ATP-competitive inhibitors.

    Significance

    The interplay between TORC1 and AMPK is of great interest in the cell signaling field, basically in every model organism. The paper provides a conceptual advance in the field showing a genetic interaction between the two pathways using a model organism which has probably been overlooked so far, which is a pity because S. pombe is the best organism to study G2/M cell cycle/size regulation. The story would be of interest especially for an audience working in cell signaling in microorganisms, but not so much (at least at this stage) for the community working on aging, disease and chemo-/radio-sensitization, contrary to what the authors claim. Furthermore, for the above-mentioned reasons, I feel like the authors are a little bit overshooting when claiming (for example in the abstract and in the discussion), that their work provides a clear understanding of the mechanism.

    As requested by Review Commons, I specify that my expertise is on TORC1/AMPK/PKA pathways, on their crosstalk and their regulation my metabolic intermediates.

  5. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary:

    In the manuscript "The AMPK-TORC1 signalling axis regulates caffeine-mediated DNA damage checkpoint override and cell cycle effects in fission yeast," the authors studied the role of genes that are potentially involved in the caffeine-mediated override of a cell cycle arrest caused by activation of the DNA damage checkpoint. The methylxanthine substance caffeine has been known to override the DNA damage checkpoint arrest and enhance sensitivity to DNA damaging agents. While caffeine was reported to target the ATM ortholog Rad3, the authors previously reported that caffeine targets TORC1 (Rallis et al, Aging Cell, 2013). Inhibition of TORC1, like caffeine, was also reported to override DNA damage checkpoint signaling. Therefore, in the present study, the authors compared the effects of caffeine and torin1 (a potent inhibitor for TORC1 and TORC2) on cell cycle arrest caused by phleomycin, a DNA damaging agent, using various gene deletion S. pombe mutants.

    The authors concluded that they identified a novel role of Ssp1 (calcium/calmodulin-dependent protein kinase) and Ssp2 (catalytic subunit of AMP-activated kinase) in the cell cycle effects caused by caffeine, based on the following findings; (1) the caffeine-mediated DNA damage checkpoint override requires Ssp1 and Ssp2; (2) Ssp1 and Ssp2 are required for caffeine-induced hypersensitivity against phleomycin; (3) under normal growth conditions, caffeine leads to a sustained increase of the septation index in a Ssp2-dependent manner; (4) Caffeine activates Ssp2 and partially inhibits TORC1.

    Major comments:

    I do not think that many of the authors' claims are supported by the results of the present study. The corresponding parts are detailed below.

    1. The conclusion of the first paragraph in the Results (top in page 6; Our findings indicate that caffeine and torin1 indirectly and directly inhibit TORC1 activity respectively.) is not supported by the data in Figure 1. The result that caffeine, but not torin1, requires Ssp1 and Ssp2 to override the phleomycin-induced cell cycle arrest does not necessarily indicate that caffeine indirectly inhibits TORC1 via Ssp1 and Ssp2. Rather, the authors should mention that this conclusion is based on the authors' previous reports by citing them (e.g., Rallis et al, Sci Rep, 2017).

    To add to Figure 1, an additional experiment using a constitutively active AMPK mutant, a temperature-sensitive TORC1 mutant, and a srk1 deletion mutant will help the authors claim their original conclusion as one possibility.

    1. The conclusion of the second paragraph in the Results (lower-middle in page 6; Our results indicate that caffeine induces the activation of Ssp2.) is not based on the results of Figure 2. Figure 2 simply illustrates that both caffeine and torin1 cause hypersensitivity to phleomycin dependent on Ssp1 and Ssp2.
    2. The conclusion of the fourth paragraph in the Results (middle in page 7) is not clearly supported by the result, due to an insufficient data analysis. As the cell length and the progress through mitosis are the key assay parameters in Figure 3, the average cell length should be shown next to each micrograph of Figure 3A and 3B. In Figure 3C, a mitotic index and the average cell length should be shown next to each micrograph. A statistical analysis is necessary for the authors to compare the measurements and to claim as the headline (Caffeine exacerbates the ssp1D phenotype under environmental stress conditions), as the effect of caffeine was not evident.
    3. In the middle of page 8, the statement "Accordingly, the effect of caffeine and torin1 on DNA damage sensitivity was attenuated in gsk3D mutants (Figure 5C and 5D)." is not supported by the corresponding results. Rather, Figure 5C and 5D look almost same.
    4. The description and the conclusion of the last paragraph in the Results (bottom in page 8 - page 9) are not supported by the results of Figure 6, due to an insufficient data analysis. The extent of phosphorylation must be quantified as a ratio of the phosphorylated species (e.g., pSsp2) to all species of the protein (e.g., Ssp2).

    From Figure 6, the authors claim that caffeine (10 mM) partially inhibits TORC1 signaling. However, the authors previously showed that the same concentration of caffeine inhibited phosphorylation of ribosome S6 kinase as strongly as rapamycin, the potent TOR inhibitor (Rallis et al, Aging Cell, 2013). The authors are advised to assess phosphorylation of S6 kinase again in the present study and compare to the results of the present results in Figure 6, because addition of that data may allow the authors to discuss that caffeine affects TORC1 downstream pathways at different intensities.

    Also, immunoblotting of the same proteins looks somehow different from panel to panel (e.g., pSsp2 in panel A and D; Actin in panel A, C, and D). Therefore, the blotting result before clipping had better be shown as a supplementary material.

    Minor comments:

    1. (Figure 1) The septation index of the phleomycin-treated cells (without any further additional drugs) should be shown, as a baseline.
    2. (Figure 1D, Optional) As a ppk18D cek1D double deletion mutant is reported, the authors are advised to add and test that mutant in this experiment.
    3. (Figure 2) The authors need to clarify the number of cell bodies spotted (e.g., in the Figure legend).
    4. (Figure 3) The different number of cells in micrographs may give an (wrong) impression on the cell proliferation rate. Therefore, it is advisable to use the micrographs in which the similar number of cells are shown for conditions with the similar cell proliferation rates.
    5. (Figure 4B) ssp2D, not spp2D.
    6. (Figure 4) The septation index of the none-treated cells should be shown as a baseline.
    7. (Figure 6B, 6E) What do the black arrows indicate? Figure Legend does not seem to explain them.
    8. (Figure 6C) Indicate which part of the Maf1-PK blot corresponds to the phosphorylated species, because Maf1-PK is probed with an anti-V5 (not a phosphorylation-specific) antibody.
    9. (Figure 6D) gsk3Dssp1D, not gs3Dssp1D.

    Significance

    As caffeine is implicated in protective effects against diseases including cancer and improved responses to clinical therapies, the topic of the present study is of interest and importance to the broad audience.

    In the present study, the most significant finding is that caffeine- and torin1-induced hypersensitivity to phleomycin is dependent on Ssp1 and Ssp2 (Figure 2). This result may be important in chemotherapy against cancers. On the other hand, caffeine is known to activate AMPK (e.g., Jensen Am J Physiol Endocrinol, 2007). Besides, as detailed in the Major comments, many of the major conclusions are not supported by the present results. Therefore, based on my field of expertise (cell cycle, cell proliferation, and TOR signaling), I conclude that the present study hardly extends the knowledge in the field of "the cell biology of caffeine."

  6. Version published to 10.1101/2022.11.08.515652v2 on bioRxiv
  7. Version published to 10.1101/2022.11.08.515652v1 on bioRxiv