Sulfur starvation-induced autophagy in Saccharomyces cerevisiae involves SAM-dependent signaling and transcription activator Met4

Autophagy is a key lysosomal degradative mechanism allowing a prosurvival response to stresses, especially nutrient starvation. Here we investigate the mechanism of autophagy induction in response to sulfur starvation in Saccharomyces cerevisiae. We found that sulfur deprivation leads to rapid and widespread transcriptional induction of autophagy-related (ATG) genes in ways not seen under nitrogen starvation. This distinctive response depends mainly on the transcription activator of sulfur metabolism Met4. Consistently, Met4 is essential for autophagy under sulfur starvation. Depletion of either cysteine, methionine or SAM induces autophagy flux. However, only SAM depletion can trigger strong transcriptional induction of ATG genes and a fully functional autophagic response. Furthermore, combined inactivation of Met4 and Atg1 causes a dramatic decrease in cell survival under sulfur starvation, highlighting the interplay between sulfur metabolism and autophagy to maintain cell viability. Thus, we describe a pathway of sulfur starvation-induced autophagy depending on Met4 and involving SAM as signaling sulfur metabolite.

1-All the work is done using yeast strains carrying several auxotrophic markers (mutations in amino acids and nucleotide synthesis pathway genes) that could interfere with the observations.I understand that these "classical" lab strains are more convenient to use but it seems to me that some of the key experiments should be repeated in prototrophic yeast strains to ensure that the conclusions are robust and physiologically correct.This is of particular importance for Fig. 1a, Fig. 3c and supplemental Fig. 3.
2-Strains used in each experiment should be strictly isogenic.This is apparently not the case although difficult to figure out in several cases (see point 5 below).Using non isogenic strains may severely affect the conclusions and should be avoided.For example, in supplemental Fig. 3, the two strains BY4741 and BY4742 differ at the LYS2 locus in addition to the MET17 (MET15) locus which is the one tested in the experiment.The lys2 auxotrophy strongly affects yeast physiology.
3-Although elucidating the molecular underlying mechanisms is not in the scope of this work, it would be nice to have some experiments done under sulfur and nitrogen starvation conditions in parallel, to establish which observations are specific to sulfur starvation.This is of particular importance for Fig. 1c, Fig. 3b, supplementary Fig. 2a.Specifically, the authors should address the following questions: do the responses to these two nutritional signals share some pathway components?For example, do the sulfur regulation signals need a functional TOR pathway?Is autophagy normally induced in a met4 mutant in response to nitrogen starvation?A few simple experiments would allow to widen the conclusions and give a preview of what is taking place (or not) at the molecular level.
4-The role of Atg41 in the response to sulfur starvation is unclear.Data in supplementary Fig. 2a show that the atg41 mutation blocks GFP-Atg8 processing indicating a role of Atg41 in autophagy (as expected), but also blocks GFP-Atg8 expression suggesting that a functional autophagy pathway might be required for the transcriptional response to sulfur limitation to occur.This should be clarified by investigating whether other Atg mutants (affecting different steps of the autophagy process) have the same effect and whether they affect expression of other Met4 target genes (MET genes).
Minor points: 5-The name of the strains used in each figure should be clearly specified.For example, in Fig. 1c, "WT" could refer to several strains in supplementary Table1.
Reviewer #2 (Remarks to the Author): In the current manuscript, the authors analyze the autophagy response of budding yeast upon sulfur starvation.They identify a strong transcriptional response for autophagy genes mediated by the known central transcription factor Met4 in the absence of sulfur.Atg41, a factor highly upregulated upon sulfur starvation in a Met4-dependent manner, plays a critical role for Atg8 expression and autophagy flux.The authors convincingly dissect the possible metabolic inputs for autophagy regulation and identify S-adenosylmethionine as a key metabolite.
The authors describe an interesting dataset and use systematic yeast genetics to identify the key metabolite SAM for sulfur starvation-induced autophagy.However, the study falls short in providing initial mechanistic insights into how SAM levels might be connected to the regulation of Met4 and/or autophagy and, importantly, ignores published data that have linked SAM to autophagy regulation, as described below.

Critical points:
(1) SAM has been linked to the regulation of autophagy in a previous publication at mechanistic level; SAM is a substrate for the methyltransferase Ppm1, which regulates the protein phosphates PPA2; methylated PPA2 promotes dephosphorylation of Npr2, a negative regulator of TORC1 and positive regulator of autophagy (Sutter et al. Cell 2013).It is inappropriate that the authors cite this article, when speculating about a link of TORC1 and Met4 regulation, but do not mention the described role of SAM for autophagy regulation.Clearly, these mechanisms might also underly autophagy and/or Met4 regulation during sulfur starvation-induced autophagy.
(2) Overall, the manuscript lacks insight into how Atg8 protein levels, autophagy flux and the transcriptional program are linked and/or their mechanistic bases.How SAM levels are translated into Met4 and/or autophagy flux are key questions.
(3) The essential role for Atg41 for sulfur starvation-induced autophagy is really interesting.In particular, it is interesting that Atg8 levels and turnover are significantly impaired in the absence of Atg41.Is the effect on Atg8 protein levels specific to the absence of Atg41 or does a block in any core ATGs cause these effects?
Reviewer #3 (Remarks to the Author): Summary Prigent et al. provide evidence that the master regulator of sulfur metabolism in yeast, Met4, activates the transcription of ATG genes in response to sulfur starvation, and suggest that this signaling is specifically in response to SAM depletion.The authors also demonstrate the importance of sulfur starvation-induced autophagy for cellular viability after prolonged sulfur deprivation.

Validity
In general, the interpretation of data and conclusions drawn by the authors are reasonable.While the ChIP experiments demonstrate that Met4 is recruited to the ATG genes examined, it is less obvious that Met4 is affecting the mRNA levels of these genes.There appears to be significant alteration in the levels of ATG mRNA during the time course (e.g., up 40 min, down 80 min, up again 120 min) in Fig 3A for the WT strain that make data interpretation and significance difficult to assess, in particular with respect to the met4 and gcn4 single deletions.Can the authors explain this?

Significance
The work done by Prigent et al. potentially represent a significant advance in the understanding of the mechanisms governing transcriptional regulation of autophagy.However, some of the data are not compelling and could be supported by additional experiments.Perhaps the authors might be overlooking some alternative mechanisms.

Data and Methodology
The figures and data are organized and presented clearly.The utilization of specially formulated "sulfur free" media is a strength of their methodology, as well as the use of specific sulfur metabolism mutants to identify the responsible metabolite.However, one concern the authors should look into is the use of BY strain background that may be auxotrophic for met15∆, which could alter the response to sulfur and methionine metabolism or synergize with met4∆ deletion.They should consider validating key results using a prototrophic strain or other strain background such as W303.

Analytical Approach
The authors should state the n for their RT-qPCR experiments, unless I have missed it.

Suggested Improvements
Major points: The authors rely heavily on the GFP-Atg8 cleavage assay to monitor autophagy.While some results obtained with this assay are consistent with their model, the marked reduction in the levels of GFP-Atg8 protein in the met4 and gcn4 single and double mutants compared to WT makes it difficult to assess levels of autophagy by the very nature of the mutants they are testing.This also applies to their microscopy data and other data where the growth condition alters the abundance of GFP-Atg8.While this reduction in GFP-Atg8 is informative in and of itself, the use of a complementary and more quantitative method to monitor autophagy such as the Pho860 alkaline phosphatase assay would strengthen the work.After all, the induction of ATG genes and protein amounts may not matter in the end if ATG8 is not processed, etc. and there is comparable delivery of autophagosomes to the vacuole.To this point, the extent of processing of ATG8 for various mutants does not appear that different, when it might be expected that methionine addition should repress actual autophagy (e.g., Fig 5b , 5e) -these observations would appear inconsistent with the authors initial finding that sulfur starvation induces autophagy.
The authors provide evidence that Met4 may bind several ATG genes using ChIP, however the prominent induction of ATG genes can still be observed in the absence of Met4.This is quite confusing and suggests that the observed Met4 binding may not be functionally relevant.If Met4 were indeed a regulator of ATG genes, one might expect effects on mRNA levels similar to the extent observed with the MET3 transcript.Can the authors also observe binding of Cbf1, Met31 or Met32?Although significant effects were observed in the double gcn4 met4 knockout, there is the concern that this strain could be quite sick as Gcn4 is generally required for transcriptional response to amino acid starvation.

Minor points:
The authors should consider including growth rates of various mutants under normal versus starvation conditions.

Clarity and Context
The text was written with clarity and context. 1

Responses to reviewers' comments on manuscript NCOMMS-22-17321
Reviewers' comments are in blue; responses are in black; modified figures are indicated in bold.
Addition introduced in the manuscript itself are marked in dark red and rephrasing in green.

Reviewer #1:
The manuscript entitled « Sulfur starvation-induced autophagy in Saccharomyces cerevisiae involves SAM-dependent signaling and transcription activator Met4" by Prigent and co-workers reports the transcriptional induction of autophagy-genes in response to sulfur limitation.They show an important role for the transcription factor Met4 in this process and for the metabolite S-adenosyl methionine.Depletion of methionine or cysteine also had an additional post-transcriptional effect, indicating that sulfur limitation results both in an increase of the autophagy flow rate through posttranscriptional activation, but also an increase of the autophagy flux capacity through transcriptional activation of several ATG genes.The underlying molecular mechanisms remain to be elucidated.This is a nice and globally solid piece of work.The experiments are sound, the results are clear and support the conclusions.I have several concerns which are detailed below.Addressing experimentally these various points would, in my opinion, strengthen the conclusions.
We thank the reviewer for the positive assessment.

Major points:
1-All the work is done using yeast strains carrying several auxotrophic markers (mutations in amino acids and nucleotide synthesis pathway genes) that could interfere with the observations.I understand that these "classical" lab strains are more convenient to use but it seems to me that some of the key experiments should be repeated in prototrophic yeast strains to ensure that the conclusions are robust and physiologically correct.This is of particular importance for Fig. 1a, Fig. 3c and supplemental Fig. 3.
Response: Most published studies on autophagy in yeast used strains with auxotrophic mutations.It may not be a problem in the case of studies focusing on the core autophagic process, but it might in the case of those interested in regulatory aspects.A strain containing auxotrophic mutations will surely not sense nitrogen starvation in the same manner as a prototrophic strain.The former should be more sensitive to the amino acids that it cannot synthesize and trigger additional signaling pathways in their absence.We believe that it should be less of a concern in the case of sulfur depletion from a medium containing all amino acids, but the concern still exists considering how metabolic pathways are intertwined.Therefore, we have repeated several experiments in prototrophic strains derived from BY4742.We have performed autophagy flux (GFP-Atg8 assay), transcriptional and viability analyses following sulfur depletion (Fig. 2 and supplemental Fig. 3).We have obtained similar results as with BY4742, indicating that the auxotrophic mutations present in BY4742do not affect the autophagic response to sulfur depletion.We have also repeated autophagy flux and transcriptional analyses in prototrophic strains containing gcn4 and met4-null alleles (supplemental Fig. 3).The results of these experiments support our conclusion that Met4 is essential for ATG gene induction upon sulfur starvation.Viewing these new results, we have not considered that reproducing the experiments using the met17Δ mutation (supplementary fig. 3 in the initial manuscript) in a prototrophic background was a priority.However, we have repeated the experiments using a strain containing met17Δ in BY4742 background (see comment#2 below).
2-Strains used in each experiment should be strictly isogenic.This is apparently not the case although difficult to figure out in several cases (see point 5 below).Using non isogenic strains may severely affect the conclusions and should be avoided.For example, in supplemental Fig. 3, the two strains BY4741 and BY4742 differ at the LYS2 locus in addition to the MET17 (MET15) locus which is the one tested in the experiment.The lys2 auxotrophy strongly affects yeast physiology.
Response: We have repeated the experiment of supplemental Fig. 3 with a met17Δ mutant derived from BY4742 instead of using BY4741 (supplemental Fig. 5).We have also added the name of strains in the legends of figures.In the process, we have noticed a few errors and omissions in the Table listing the strains that might have led to confusion.We apologize for that.
3-Although elucidating the molecular underlying mechanisms is not in the scope of this work, it would be nice to have some experiments done under sulfur and nitrogen starvation conditions in parallel, to establish which observations are specific to sulfur starvation.This is of particular importance for Fig. 1c, Fig. 3b, supplementary Fig. 2a.Specifically, the authors should address the following questions: do the responses to these two nutritional signals share some pathway components?For example, do the sulfur regulation signals need a functional TOR pathway?Is autophagy normally induced in a met4 mutant in response to nitrogen starvation?A few simple experiments would allow to widen the conclusions and give a preview of what is taking place (or not) at the molecular level.
Response: The mechanisms underlying autophagy induction under sulfur starvation condition were completely unknown when we started this work.Trying to unveil the complete regulatory cascade in one study would not have been realistic; therefore, after observing that ATG genes were upregulated upon sulfur depletion, we chose to focus first on the downstream players.We believe that the set of data that we present in this paper constitutes a consistent and coherent story already providing new mechanistic insights.We intend to focus on the upstream players in a following study that will be presented separately.Yet, we have added experiments done under nitrogen starvation conditions to emphasize the specificity of the mechanisms involved under sulfur starvation conditions, as suggested by the reviewer.In Fig. 2, we show that autophagy flux dynamics following sulfur and nitrogen depletion are comparable, although not similar.More importantly, we show that ATG gene transcription is differentially induced upon sulfur and nitrogen depletion.In fact, only ATG8 is strongly induced upon nitrogen depletion, the other ATG genes are not induced, or only weakly.
Moreover in supplemental Fig. 4, we show that inactivation of Gcn4 or Met4 does not affect autophagy flux under nitrogen starvation.We believe that these additional experiments strengthen our study by highlighting the very specific and distinctive role of Met4 and transcription regulation under sulfur starvation.We are thankful to the reviewer for the suggestion.We have not compared the survival capacity of autophagy-deficient mutants under sulfur and nitrogen starvation.This would certainly be an interesting experiment, but difficult to interpret and we were not convinced that this comparison would provide useful information regarding the mechanisms of regulation of autophagy.
4-The role of Atg41 in the response to sulfur starvation is unclear.Data in supplementary Fig. 2a show that the atg41 mutation blocks GFP-Atg8 processing indicating a role of Atg41 in autophagy (as expected), but also blocks GFP-Atg8 expression suggesting that a functional autophagy pathway might be required for the transcriptional response to sulfur limitation to occur.This should be clarified by investigating whether other Atg mutants (affecting different steps of the autophagy process) have the same effect and whether they affect expression of other Met4 target genes (MET genes).
Response: We have added two sets of data in supplemental Fig. 2. First, we have used the Pho8Δ60 assay to monitor autophagy flux/activity.The results consolidate those obtained with the GFP-Atg8 processing assay and confirm that Atg41 is required for autophagy induction under sulfur starvation.
Second, we have performed transcriptional analyses.We show that upon sulfur depletion ATG8 transcription induction is diminished in atg41Δ compared to WT, but not in atg1Δ, which indicates that ATG8 induction upon sulfur depletion does require Atg41 but not a functional autophagy pathway.In contrast, ATG1 is induced at a slightly higher level in atgΔ41, and ATG9 or MET3 are induced at similar levels.These new data emphasize Atg41 importance for the autophagic response under sulfur starvation; however, we believe that a more detailed analysis of its precis molecular function is out of the scope of the present work and should be addressed in another study.
Minor points: 5-The name of the strains used in each figure should be clearly specified.For example, in Fig. 1c, "WT" could refer to several strains in supplementary Table1.
Response: Please see our reply to point #2.
signaling is specifically in response to SAM depletion.The authors also demonstrate the importance of sulfur starvation-induced autophagy for cellular viability after prolonged sulfur deprivation.

Validity
In general, the interpretation of data and conclusions drawn by the authors are reasonable.While the ChIP experiments demonstrate that Met4 is recruited to the ATG genes examined, it is less obvious that Met4 is affecting the mRNA levels of these genes.There appears to be significant alteration in the levels of ATG mRNA during the time course (e.g., up 40 min, down 80 min, up again 120 min) in Fig 3A for the WT strain that make data interpretation and significance difficult to assess, in particular with respect to the met4 and gcn4 single deletions.Can the authors explain this?
Response: The "up-down-up" transcription pattern observed in the WT strain in the case of ATG and MET3 genes is not a typical pattern (see for instance Fig. 1) but is due to the SAM supplement.To address the concern about the difficulty to interpret the results owing to this pattern, we have repeated the transcription analyses and the GFP-Atg8 processing experiments of Fig. 3 using media containing 0.01 mM SAM instead of 0.05 mM (Fig. 4 in the revised version).We find that at lower SAM concentration, transcription of ATG and MET3 genes do not go down at 80 min.Moreover, ATG gene transcription and GFP-Atg8 expression are more diminished in met4Δ (comparatively to the WT cells) when the growth medium contains 0.01 mM SAM instead of 0.05 mM, which highlights better Met4 requirement.One possibility to explain the difference could be that Gcn4 expression and/or access to ATG gene promoters depends on SAM concentration and is higher when SAM levels are increased, resulting in higher Gcn4 occupancy at ATG promoter in Met4 absence of.We also observed that GFP-Atg8 processing appears more efficient in met4Δ grown in 0.01 mM SAM than in 0.05 mM.It could be because the impaired autophagic capacity of met4Δ (as shown by the Pho8Δ60 essay in Fig. 4c) is still sufficient to process efficiently the low GFP-Atg8 amounts in the cells grown in 0.01 mM SAM, but not when the cells are grown in 0.05 mM SAM.Since the purpose of this figure was to show the role of Met4 in autophagy regulation and not study the effect of variations in SAM concentration in the growth medium, we have chosen to keep only the results obtained in 0.01 mM SAM and remove those obtained in 0.05 mM.

Significance
The work done by Prigent et al. potentially represent a significant advance in the understanding of the mechanisms governing transcriptional regulation of autophagy.However, some of the data are not compelling and could be supported by additional experiments.Perhaps the authors might be overlooking some alternative mechanisms.the end if ATG8 is not processed, etc. and there is comparable delivery of autophagosomes to the vacuole.To this point, the extent of processing of ATG8 for various mutants does not appear that different, when it might be expected that methionine addition should repress actual autophagy (e.g., Fig 5b , 5e) -these observations would appear inconsistent with the authors initial finding that sulfur starvation induces autophagy.
Response: As suggested by the reviewer, we have added Pho8Δ60 activity measures in two figures: the figures with the met4Δ and gcn4Δ mutants, and the figure with the atg41Δ mutant (Fig. 4 and Supplementary Fig. 2, respectively).The results showed in both cases a decreased of Pho8Δ60 activity in the mutants compared to WT, confirming that autophagic activity is strongly reduced, which correlates with the decrease of GFP-Atg8 and free GFP levels.
Concerning the GFP-Atg8 assay, we agree that the percentage of free GFP over total GFP may be more difficult to interpret.To highlight the differences in the amounts of free GFP between the strains and starvation conditions, graphs giving the levels of GFP release were systematically added alongside the graphs giving the levels of GFP-Atg8 expression and processing.In Fig. 5b and 5e (Fig. 6b and 6e in the revised version) we observed high expression and efficient processing of GFP-Atg8 in the two mutants upon transfer into the medium completely deprived of sulfur, demonstrating strong autophagy induction.However, when the mutants are deprived of cysteine only (in the case of cys4Δ) or methionine only (in the case of met6Δ mht1Δ sam4Δ), GFP-Atg8 expression is lower compared to total sulfur depletion, but not completely shut down, and the low amount of GFP-Atg8 expressed can be processed fairly efficiently.However, the levels of free GFP release are significantly reduced, meaning lower autophagy levels, in the conditions where only cysteine or methionine are depleted.The Discussion section has also been rewritten (see the second paragraph in particular) to better highlight this point.Altogether, the two assays support the model that autophagy induction in response to sulfur depletion results mainly from SAM depletion.
The authors provide evidence that Met4 may bind several ATG genes using ChIP, however the prominent induction of ATG genes can still be observed in the absence of Met4.This is quite confusing and suggests that the observed Met4 binding may not be functionally relevant.If Met4 were indeed a regulator of ATG genes, one might expect effects on mRNA levels similar to the extent observed with the MET3 transcript.
Response: The experiments with the met4Δ and gcn4Δ single and double mutants were repeated in slightly different growth conditions (see the reply to the Validity point) and we believe that the results demonstrate more convincingly the role of Met4 in the transcriptional regulation of ATG genes.Even if Met4 is a regulator of ATG genes, we do not think that ATG genes should obligatory be GFP-Atg8 assays: The authors have addressed the concern over the GFP-Atg8 assays by performing two alkaline phosphatase assays and adding graphs to accompany the GFP-Atg8 assays.I believe the GFP-Atg8 assay remains a problematic assay for monitoring autophagy flux but can represent the abundance of Atg8 which the authors propose is dependent on Met4 and SAM levels along with the levels of other Atg proteins.In my opinion, it remains incumbent on the authors to better shown how autophagy flux can change in their mutant cells and various media conditions.The authors could perform additional alkaline phosphatase assays and/or perform additional GFP cleavage assays (ie Idh1-GFP cleavage assays) to better assess autophagy flux.
Re: We have added an additional GFP cleavage assay, the Pgk1-GFP cleavage assay, to complement the experiments with the metabolic mutants cys4, met6 mht1 sam4 and sam1 sam2 (new Supplementary Fig. 8).The new data consolidate and extend our previous results.They first show that Pgk1-GFP is cleaved in an autophagy-dependent manner following sulfur depletion.They also show that methionine depletion in met6 mht1 sam4 induces very weakly Pgk1-GFP cleavage, whereas depletion of SAM in sam1 sam2 induced Pgk1-GFP cleavage almost as efficiently as sulfur depletion.These two results are in line with the results of the GFP-Atg8 cleavage assay.By contrast, cysteine depletion in cys4 induces higher levels of cleaved Pgk1-GFP than might be expected based on the GFP-Atg8 cleavage assay.In fact, in this mutant, levels of cleaved Pgk1-GFP are similar upon cysteine and sulfur depletion while levels of GFP-Atg8 expression and GFP release are lower upon cysteine depletion compared to sulfur depletion.One hypothesis could explain these seemingly contradictory results: cysteine depletion signals the induction of bulk autophagy, not induction of selective autophagy.By contrast, sulfur depletion, which also causes SAM depletion, would signal the induction of both types of autophagy.One functional consequence would be that the type of cargoes sequestered by autophagosomes might vary depending on the sulfur compound.
The text of the manuscript was modified in the "Results" and "Discussion" sections to integrate the new results.ChIP issues: The authors reasonably propose that ATG transcription may not be entirely dependent on Met4 but could be regulated by other factors.It remains poorly understood whether Met4 is actively inducing ATG gene transcription in a meaningful way to regulate autophagy or if the effects of Met4 on autophagy are primarily posttranscriptional.After all, autophagy is apparently highly repressed in met4-ko cells while ATG transcript levels remain elevated.The authors declined to further assess Cbf1, Met31, Met32 binding because it would be a lot of work.
Re: We believe that our ChIP experiments provide good evidence that Met4 associates with ATG genes.The fact that the promoters contain binding sites for Met31/Met32 (and Cbf1 in some cases) gives additional support to the hypothesis that Met4 actively induces their transcription.Met4 is a transcriptional activator possessing a characteristic activation domain and a bZIP DNA-binding domain.There is no data suggesting that Met4 might have other functions that would support the possibility of a posttranscriptional role.Nevertheless, we cannot completely rule out the possibility of an indirect role involving its transcriptional activation and DNA-binding capacities, but performing only additional ChIP with Cbf1, Met31 or Met32 will not answer the question.

Specific comments
The use of strains of different backgrounds makes interpretation of the data confusing.Genetic backgrounds could be depicted in the figures to make interpretation easier.
The GFP-Atg8 assays are difficult to interpret in certain cases in which the GFP-Atg8 abundance is diminished in the mutants.The use of another GFP cleavage assays (e.g., Idh1-GFP cleavage) could