Regulation of TFEB and V-ATPases by mTORC1 - Scolaris

The EMBO Journal Peer Review Process File - EMBO-2011-76845

Manuscript EMBO-2010-76845

Regulation of TFEB and V-ATPases by mTORC1 Samuel Peña-Llopis, Silvia Vega-Rubin-de-Celis, Jacob C. Schwartz, Nicholas Wolff, Tram Anh T. Tran, Lihua Zou, Xin-Jin Xie, David R. Corey and James Brugarolas Corresponding author: James Brugarolas, University of Texas

Review timeline:

Submission date: Editorial Decision: Resubmission: Editorial Decision: Revision: Editorial Decision: Revision Editorial Decision: Accepted:

20 September 2010 03 November 2010 21 December 2010 01 February 2011 21 April 2011 03 June 2011 22 June 2011 24 June 2011 24 June 2011

Transaction Report: (Note: With the exception of the correction of typographical or spelling errors that could be a source of ambiguity, letters and reports are not edited. The original formatting of letters and referee reports may not be reflected in this compilation.)

1st Editorial Decision

03 November 2010

Thank you for submitting your manuscript for consideration by The EMBO Journal. Let me first of all apologise for the delay in getting back to you with a decision. Unfortunately, we experienced severe difficulties in finding suitable and willing referees for this manuscript. In addition, one of the referees was not able to return his/her report a quickly as initially expected. In the meantime your manuscript has now finally been seen by three referees whose comments are shown below. As you will see all three referees consider the study as interesting in principle, but also think that the mechanistic depth of your study is not sufficient to justify the conclusions drawn. One major aspect is that the functional link between mTORC1-dependent upregulation of TFEB and the downstream effects on V-ATPase and in turn on mTORC1 activity (positive feedback regulation) is not backed up sufficiently by a causal line of evidence. Also, the data on the integration of this line of events in a physiological stimulus, i.e. in the context of nutrient regulation of mTORC1, are not considered as sufficient at this point. An important point that refers to the conclusiveness of the data more directly is that there are discrepancies between effects on the mRNA and the protein levels of V-ATPase subunits as put forward by referees 2 and 3. So, even if deeper detailed mechanistic insight into how mTORC1 affects TFEB levels is not required, it becomes clear that the study is not sufficiently developed to be conclusive and convincing and thus publishable here in its present form. As the outcome of the required substantial additional experimentation cannot be predicted at this point, I am afraid to say that we cannot consider or commit to a revision.

© European Molecular Biology Organization

1


Still, the referees consider the study as interesting in principle and put forward a number of suggestions how the study could be improved by extensive further experimentation. We would thus be able to consider a new version of the manuscript that is developed further considerably along the lines put forward by the referees as a new submission rather than as a revision. However, all issues raised by the referee need to be addressed to their satisfaction, even though deeper mechanistic insight into how exactly mTORC1 regulates TFEB will not be required. Still, I would like to stress that the causality of all the mentioned functional links themselves need to be strengthened. To be completely clear, however, I would like to point out that if you wish to send a new manuscript this will be treated as a new submission rather than a revision and will be evaluated again at the editorial level and reviewed afresh involving the same referees if available at the time of resubmission, also with respect to the literature and the novelty of your findings at the time of resubmission. At this stage of analysis, though, I am sorry to have to disappoint you. Thank you for the opportunity to consider your work for publication. I am sorry we cannot be more positive at this point, but we hope nevertheless that you will find our referees' comments helpful. Please do not hesitate to get in touch to consult with us on details of a possible resubmission. Yours sincerely, Editor The EMBO Journal -----------------------------------------------REFEREE COMMENTS Referee #1: Brugarolas and colleagues report that a component of the endosome, the vacuolar H+ATPase is a target of mTORC1 and in a regulatory loop helps to maintain mTORC1 activity. Using a series of public and their own array analysis of TSC -/- cells and rapamycin treatment the authors identify VATPases as mTOR targets. It appears that this control is through transcription mediated by the TFEB transcription factor which mRNA is also controlled by mTORC1. Pharmacological inhibition of V-ATPases seems to also affect mTORC1 activity as measured by P-S6K1 and 4E-BP1. Overall Evaluation: Although the results of this manuscript are of potential interest the authors embrace a large number of different conclusions that are not supported by the data. Based on the provided experimental data is difficult to fully accept the model. In my view, this manuscript needs as it largely lacks complete mechanistic depth of the different conclusions is more tailored for a more specialized journal, it is not at the level of EMBO Journal. Specific Evaluations: 1- Most of the experiments that determine the role of v-ATPases on mTORC1 function are using a pharmacological inhibitor. It is unclear the off target effects of this inhibitor and a genetically approach is encouraged. 2- Along the same lines, it is unclear how this inhibitor might affect the energetic state of the cell and AMPK status, a well known regulator of mTORC1. 3- It is impossible to assess the transcriptional effects of rapamycin using a time point of 24 hours. Since translational effects also occurred it is unclear the overall direct transcriptional effects. 4- Along point 3, all the experiments of loss of mTORC1 are performed using rapamycin, though this is a fairly specific drug other described mTORC1 inhibitors and genetic approaches should be carried out. 5- One of the weakest points of the manuscript is the transcriptional involvement of TFEB transcription factor. The overall contribution to the mTORC1 direct transcriptional response is


2


completely unclear. Though this transcription factor targets V-ATPases, the expression of these genes is still sensitive to rapamycin when TFEB knock down is achieved. 6- Importantly, there is no mechanism provided of how mTORC1 controls TFEB, this is a critical point. 7- There are no experiments to support the model that TFEB affects mTORC1 signaling in gain or loss-of-function.

Referee #2: These studies identify V-ATPases to be regulated by mTORC1. By microarray analyses, the authors demonstrate that several genes that include a number of V-ATPases become upregulated in Tsc2-/cells and that rapamycin can suppress this effect in these cells. Based on the known function of VATPases in mediating acidification of endosomes and recent findings on the role of the protein TFEB in lysosomal biogenesis, they then examined whether TFEB could mediate the regulation of V-ATPases by mTORC1. They then investigated a possible role of mTORC1 in endocytosis and how pharmacological inhibition of V-ATPases could affect mTORC1 activation. Based on their findings, they propose a feedback model whereby mTORC1 controls V-ATPase expression and the increased V-ATPase expression reciprocally promotes mTORC1 activation. The finding that mTORC1 regulates expression of V-ATPases is novel although the role of TORC1 in endocytosis has been described in several previous studies including those from other model organisms. Several genes have also been recently reported to be mTORC1-regulated using the Tsc1/2 null MEF models. Nevertheless, the current studies examine more closely how mTORC1 may regulate V-ATPase expression both in cellular and mouse models. While the microarray data are quite convincing, the role of TFEB is still unclear and more evidence is needed to support the role of V-ATPase and TFEB in endocytosis via mTORC1. Specific comments: 1. Some negative controls should be added to show that not all mRNA levels behave in a LLHL pattern (eg Fig 1E, 2B, 5B). 2. While the mRNA levels are clearly following a LLHL response in parallel to mTORC1 activation, the protein expression does not seem to correlate with this response. How do the authors explain this? 3. In Fig. 4B, phosphorylation of T389-S6K should be included as well. 4. The inclusion of myr-Akt-expressing cells in Fig. 4D does not add much to the paper. ATPase expression does not seem to behave in a LLHL manner. Instead, these results suggest that increased ATPase expression occurs in an mTORC1-independent fashion. 5. Fig. 5C and 5D are in essence good negative controls as they show no significant effect of Hifalpha or hypoxia on ATPase expression in Tsc2-/- cells. But again, here is another case wherein the mRNA expression does not correlate with the protein levels. 6. The authors conclude in Fig. 6 that Tfeb levels parallel mTORC1 activation. However, it seems that as far as protein levels (Fig 6A), it is not the amount but there may be modification changes instead (based on mobility). In this regard, Tfeb expression in Fig 6D is hard to evaluate. 7. In Fig. 7B, do the genes/probes that show dramatic response to Tfeb knockdown (boxed) include V-ATPases? This should be mentioned in the text since the Supplementary Table contains 100 pages! 8. Figure 8 is a good initial attempt to address the role of mTORC1 activation in endocytosis. However, the effects seen in TSC2-/- cells are very subtle. Other assays to link mTORC1 activation


3


with V-ATPase function should be performed. For example, how does nutrient regulation of mTORC1 affect V-ATPase expression/assembly and function? Could V-ATPase play a role in autophagy? 9. Since the paper proposes a model of reciprocal regulation of mTORC1 and V-ATPase, the effect of V-ATPase expression and inhibition on mTORC1 activity should be addressed further other than examining S6K and 4E-BP phosphorylation. V-ATPase and Tfeb expression (protein, mRNA levels) should also be analyzed in Fig 9 in response to amino acids/serum. Could localization of mTORC1 in the endosomes or phosphorylation of mTORC1 components be affected upon VATPase inhibition? Other minor comments: 1. The legend in Supplementary Figure 3 referring to Figure 2E did not make sense. 2. The data on Hif and hypoxia may be moved to supplementary figures since they do not directly relate to the main point of the paper (although it was a good negative control). Referee #3: The manuscript titled, "Reciprocal regulation of V-ATPase and mTORC1" by Pena-Llopis et al. describes a novel role of mTORC1 in V-ATPase expression. The authors identified V-ATPase among mTORC1-regulted genes through an unbiased microarray screen. They demonstrated that in TSC deficient cells V-ATPase expression is elevated in a rapamycin-sensitive manner and is transcriptionally regulated by mTORC1 in a TFEB-dependent manner. They also demonstrated the requirement of V-ATPase activity for AA/GF-mediated mTORC1 activation. So, they proposed that mTORC1 promotes its own activity through its positive feed-forward regulation of V-ATPase expression. Overall, this study was presented logically and adds more insight not only into the dynamics of mTORC1-regulated gene expression but also into the regulation of mTORC1 signaling at the lysosome. However, there are some shortcomings that need to be addressed; In particular, the authors need to address better the importance of TFEB-dependent V-ATPase expression in mTORC1 activation. The following are my major and minor concerns that need to be addressed. Major Comments: 1. It is somewhat premature for the authors to claim that V-ATPase expression by mTORC1 in a TFEB-dependent manner is necessary for the promotion of its own activity. Although the requirement of V-ATPase activity for mTORC1 activation seems true due to the inhibitory effect of bafilomycin A1, this is insufficient to satisfy the above claim. Thus, they should address 1) whether in TSC deficient cells, a decrease in V-ATPase expression by TFEB knockdown indeed causes a decrease in basal mTORC1 activity or inhibition of AA-mediated mTORC1 activation, and 2) whether overexpression of TFEB in TSC2+/+ cells is sufficient to induce V-ATPase expression, thereby leading to activation of mTORC1. 2. It will be much more convincing if the authors could address whether disruption of V-ATPase activity by knocking down its V0 domain c subunit (a target of bafilomycin A1) or multiple subunits indeed results in mTORC1 suppression. 3. As the authors cited, a recent work from Sardiello et al. (2009, Science, 325:473-7) discovered that most lysosomal genes including V-ATPase exhibit coordinated behavior and are regulated by TFEB. They showed that TFEB expression not only enhances expression of lysosomal genes, but also increases the number of lysosomes in the cell, leading to improvement of the lysosomemediated degradative pathways. Based on this work, therefore, several interesting questions could be asked. First, are other lysosomal genes such as cathepsin D also regulated by mTORC1 like VATPase? What is the expression profile of the overall lysosomal genes in the data set of the authors' microarray? Do they also show the LLHL pattern similar to V-ATPase expression? Second, is the number of lysosomes (primary location for mTORC1 activation) elevated in TSC deficient cells


4


relative to TSC2+/+ cells due to the increased TFEB levels by hyperactive mTORC1? Can rapamycin treatment also decrease the number of lysosomes in TSC-/- cells? 4. In figure 6 and 7, it is not clear how TFEB expression itself is regulated by mTORC1. Is it regulated through transcriptional activation like V-ATPase or through increased translational rate like HIF-1a? 5. There appear some discrepancies between mRNA and protein expression in terms of correlation of their fold increases. For instance, while increase in protein expression of V-ATPase subunits (figure 4b) is quite dramatic compared to the small increase in their mRNA levels (less than 2 fold; figure 4c), an increase of over 5 times in mRNA expression (figure 2b) seems to correlate with only a minor increase in protein expression (figure 2c). 6. Again, in figure 6a, b, and c, a dramatic increase (over 6 times) in TFEB mRNA expression in TSC-/- cells relative to TSC+/+ cells did not show a corresponding increase in its protein expression. Rather, TFEB protein displayed a band shift up/down pattern, perhaps representing a sign of protein modification such as phosphorylation instead of changes in protein levels per se. The authors should provide the re-evaluated data or some explanation for this significant discrepancy. Minor Comments: 1. In page 3, line 65, Rab4 should be corrected by Rab5. In addition, the authors should cite a recent work from Guan's group (Li et al., 2010, J. Biol. Chem., 285:19705-9) as this work also identified the role of Rab5 in mTORC1 signaling. 2. In page 8, line 209, TSC2-/- should be changed into TSC2+/+. 3. Along with figure 4e, the authors should provide the corresponding Western data as well. 4. In figure 9, the authors should also provide the Western data of V-ATPase and TFEB expressions in order to reversely show that mTORC1 activation positively regulate the expression of these proteins.

Resubmission

21 December 2010

Referee #1: Brugarolas and colleagues report that a component of the endosome, the vacuolar H+ATPase is a target of mTORC1 and in a regulatory loop helps to maintain mTORC1 activity. Using a series of public and their own array analysis of TSC -/- cells and rapamycin treatment the authors identify V-ATPases as mTOR targets. It appears that this control is through transcription mediated by the TFEB transcription factor which mRNA is also controlled by mTORC1. Pharmacological inhibition of V-ATPases seems to also affect mTORC1 activity as measured by P-S6K1 and 4E-BP1. Overall Evaluation: Although the results of this manuscript are of potential interest the authors embrace a large number of different conclusions that are not supported by the data. Based on the provided experimental data is difficult to fully accept the model. In my view, this manuscript needs as it largely lacks complete mechanistic depth of the different conclusions is more tailored for a more specialized journal, it is not at the level of EMBO Journal. Specific Evaluations: 1- Most of the experiments that determine the role of v-ATPases on mTORC1 function are using a pharmacological inhibitor. It is unclear the off target effects of this inhibitor and a genetically approach is encouraged. Point is addressed together with the next one to which it relates.


5


2- Along the same lines, it is unclear how this inhibitor might affect the energetic state of the cell and AMPK status, a well known regulator of mTORC1. While we thought that data from N. crassa unequivocally established the specificity of bafilomycin A1 in that mutations in the V0 domain C subunit cause resistance to the drug (Bowman et al., JBC 2002), the questions raised by the reviewer made us look into this further, and I have to say that I’m glad that we did. The reviewer raised a very specific issue, and that was whether bafilomycin activated AMPK. I’m unaware of literature supporting this, but it was easy enough to test and if that were the case, it was important that we found out. While not strongly (at least under the experimental conditions tested), we did find evidence that bafilomycin treatment was associated with the phosphorylation of raptor at an AMPK site. How bafilomycin activates AMPK and whether V-ATPases are involved in this process, we don’t know. We followed-up with additional experiments using Lkb1-deficient and wildtype littermate control MEFs and then by trying to establish a more direct link between V-ATPases and mTORC1 through (as also suggested by another reviewer) knockdown experiments of V-ATPase subunits. Knockdown of either V0c or V1A in HeLa cells (which incidentally are deficient in LKB1) did not appreciably inhibit mTORC1. The data are negative, but we did find, and have added onto the paper, that the same knockdowns downregulated endocytosis. While due to the negative nature of the results, their interpretation is difficult, and while we think that the inhibition of amino acid-induced mTORC1 activation by bafilomycin is very interesting, the data has been moved to the supplementary section. 3- It is impossible to assess the transcriptional effects of rapamycin using a time point of 24 hours. Since translational effects also occurred it is unclear the overall direct transcriptional effects. We sought to obtain insight into mTORC1 function through the identification of genes whose expression was stably affected by changes in mTORC1. We reasoned that earlier effects may not persist and may be of questionable biological significance. As shown through the analyses of publicly available array data, changes in V-ATPase expression by mTORC1 can be detected within a few hours. More importantly, the validity of the approach has been established by the discovery of a transcription factor, TFEB, which as we show, is clearly regulated by mTORC1. 4- Along point 3, all the experiments of loss of mTORC1 are performed using rapamycin, though this is a fairly specific drug other described mTORC1 inhibitors and genetic approaches should be carried out. As the reviewer points out, rapamycin is an allosteric inhibitor and is fairly specific. In addition, the same system was used by Duvel et al., in their Mol Cell 2010 paper where it was validated. Nevertheless, raptor knockdown experiments have been done and we show that their effect on TFEB is undistinguishable from that of rapamycin (please see Fig. 10 F and G). 5- One of the weakest points of the manuscript is the transcriptional involvement of TFEB transcription factor. The overall contribution to the mTORC1 direct transcriptional response is completely unclear. Though this transcription factor targets V-ATPases, the expression of these genes is still sensitive to rapamycin when TFEB knock down is achieved. As now shown in Figs 9 and 10, mTORC1 coordinately regulates the phosphorylation and nuclear localization of TFEB. In addition, while it is true that the expression of V-ATPases remains somewhat sensitive to rapamycin, as shown in Fig. 7, the effects are significantly blunted by TFEB shRNA and the residual effect may result from residual levels of TFEB in the knockdown cells. 6- Importantly, there is no mechanism provided of how mTORC1 controls TFEB, this is a critical point. As alluded to, this is addressed in the new Figs 9 and 10.


6


7- There are no experiments to support the model that TFEB affects mTORC1 signaling in gain or loss-of-function. The foundation for implicating TFEB in mTORC1 activation was the bafilomycin experiments and they have been moved to the supplement due to the aforementioned reservations. Referee #2: These studies identify V-ATPases to be regulated by mTORC1. By microarray analyses, the authors demonstrate that several genes that include a number of V-ATPases become upregulated in Tsc2-/- cells and that rapamycin can suppress this effect in these cells. Based on the known function of V-ATPases in mediating acidification of endosomes and recent findings on the role of the protein TFEB in lysosomal biogenesis, they then examined whether TFEB could mediate the regulation of V-ATPases by mTORC1. They then investigated a possible role of mTORC1 in endocytosis and how pharmacological inhibition of V-ATPases could affect mTORC1 activation. Based on their findings, they propose a feedback model whereby mTORC1 controls V-ATPase expression and the increased V-ATPase expression reciprocally promotes mTORC1 activation. The finding that mTORC1 regulates expression of V-ATPases is novel although the role of TORC1 in endocytosis has been described in several previous studies including those from other model organisms. Several genes have also been recently reported to be mTORC1-regulated using the Tsc1/2 null MEF models. Nevertheless, the current studies examine more closely how mTORC1 may regulate V-ATPase expression both in cellular and mouse models. While the microarray data are quite convincing, the role of TFEB is still unclear and more evidence is needed to support the role of V-ATPase and TFEB in endocytosis via mTORC1. Specific comments: 1. Some negative controls should be added to show that not all mRNA levels behave in a LLHL pattern (eg Fig 1E, 2B, 5B). The data in 1E and 2B pertain to similar MEF experiments and were normalized to actin. We have checked other genes, cyclophilin and ß2-microglobulin, and they similarly do not show a LLHL pattern. An IgG control for the ChIP experiments in 5B fails to show a LLHL pattern, and the results were normalized to actin, which also does not exhibit a LLHL pattern. 2. While the mRNA levels are clearly following a LLHL response in parallel to mTORC1 activation, the protein expression does not seem to correlate with this response. How do the authors explain this? We agree with the reviewer. The data between mRNA and protein do not always correlate perfectly and while we typically find a very consistent upregulation in V-ATPase mRNA levels, in the liver, by contrast to other cells in culture, the mRNA upregulation was modest by comparison to a marked induction in protein. The regulation is likely to be more complex and there may be post-translational mechanisms. 3. In Fig. 4B, phosphorylation of T389-S6K should be included as well. The data has been added. 4. The inclusion of myr-Akt-expressing cells in Fig. 4D does not add much to the paper. ATPase expression does not seem to behave in a LLHL manner. Instead, these results suggest that increased ATPase expression occurs in an mTORC1-independent fashion. These are results from transgenic mice expressing myr-Akt in the prostate and treated systemically with everolimus, and while this is a very different system that the cell culture system, the results are not that dissimilar as there is a statistically significant LLHL pattern for 13 probes, corresponding to 9 V-ATPase genes. While not at a statistically significant level, this pattern is also observed for other V-ATPase probes. The text has been clarified accordingly.


7


5. Fig. 5C and 5D are in essence good negative controls as they show no significant effect of Hifalpha or hypoxia on ATPase expression in Tsc2-/- cells. But again, here is another case wherein the mRNA expression does not correlate with the protein levels. Since the regulation of Hif-1a by hypoxia is largely post-translational, we would not expect to see a difference in Hif-1a mRNA levels. With respect to V-ATPases, while as pointed out by the reviewer, there are some differences, overall, both mRNA and protein are upregulated in Tsc2-deficient cells compared to the wild-type as would be expected. 6. The authors conclude in Fig. 6 that Tfeb levels parallel mTORC1 activation. However, it seems that as far as protein levels (Fig 6A), it is not the amount but there may be modification changes instead (based on mobility). In this regard, Tfeb expression in Fig 6D is hard to evaluate. The comment is quite perceptive and the reviewer is indeed correct. The mobility of TFEB is clearly affected by mTORC1. We now show in Figs 9 and 10 that TFEB is found in a hyperand hypo-phosphorylated forms and that the transition is regulated by mTORC1. 7. In Fig. 7B, do the genes/probes that show dramatic response to Tfeb knockdown (boxed) include V-ATPases? This should be mentioned in the text since the Supplementary Table contains 100 pages! Only a handful of V-ATPases were in the box, and a few more failed to reach statistical significance. The array platforms are different and this may be a factor. Importantly, as demonstrated by qRT-PCR, which is a lot more sensitive and specific than microarrays, V-ATPase expression is markedly downregulated by TFEB knockdown as would be expected. 8. Figure 8 is a good initial attempt to address the role of mTORC1 activation in endocytosis. However, the effects seen in TSC2-/- cells are very subtle. Other assays to link mTORC1 activation with V-ATPase function should be performed. For example, how does nutrient regulation of mTORC1 affect V-ATPase expression/assembly and function? Could V-ATPase play a role in autophagy? We agree with the reviewer. However, while the effects are not profound, they are consistent and the results are statistically significant. While it would be very interesting to determine the effects of nutrients on V-ATPase expression/assembly and function and the effects of V-ATPases on autophagy, these experiments would not be trivial to do. 9. Since the paper proposes a model of reciprocal regulation of mTORC1 and V-ATPase, the effect of V-ATPase expression and inhibition on mTORC1 activity should be addressed further other than examining S6K and 4E-BP phosphorylation. V-ATPase and Tfeb expression (protein, mRNA levels) should also be analyzed in Fig 9 in response to amino acids/serum. Could localization of mTORC1 in the endosomes or phosphorylation of mTORC1 components be affected upon V-ATPase inhibition? The idea that V-ATPase inhibition may affect other aspects of mTORC1 is quite interesting, however, due to the limitations pointed out by the reviewers associated with the use of bafilomycin, the figure has been moved to the supplementary section and has been substituted for two figures on the regulation of TFEB by mTORC1. Other minor comments: 1. The legend in Supplementary Figure 3 referring to Figure 2E did not make sense. Thank you. It has been corrected. 2. The data on Hif and hypoxia may be moved to supplementary figures since they do not directly relate to the main point of the paper (although it was a good negative control). As the reviewer points out, it is a good negative control, of which we didn’t have many and


8


given that Hif is an established transcriptional regulator downstream of mTORC1, we believe it adequately sets the background for looking at TFEB and we would prefer to leave it, but will leave this to the editor’s discretion. Referee #3 (Remarks to the Author): The manuscript titled, "Reciprocal regulation of V-ATPase and mTORC1" by Pena-Llopis et al. describes a novel role of mTORC1 in V-ATPase expression. The authors identified V-ATPase among mTORC1-regulted genes through an unbiased microarray screen. They demonstrated that in TSC deficient cells V-ATPase expression is elevated in a rapamycin-sensitive manner and is transcriptionally regulated by mTORC1 in a TFEB-dependent manner. They also demonstrated the requirement of V-ATPase activity for AA/GF-mediated mTORC1 activation. So, they proposed that mTORC1 promotes its own activity through its positive feed-forward regulation of V-ATPase expression. Overall, this study was presented logically and adds more insight not only into the dynamics of mTORC1-regulated gene expression but also into the regulation of mTORC1 signaling at the lysosome. However, there are some shortcomings that need to be addressed; In particular, the authors need to address better the importance of TFEB-dependent V-ATPase expression in mTORC1 activation. The following are my major and minor concerns that need to be addressed. Major Comments: 1. It is somewhat premature for the authors to claim that V-ATPase expression by mTORC1 in a TFEB-dependent manner is necessary for the promotion of its own activity. Although the requirement of V-ATPase activity for mTORC1 activation seems true due to the inhibitory effect of bafilomycin A1, this is insufficient to satisfy the above claim. Thus, they should address 1) whether in TSC deficient cells, a decrease in V-ATPase expression by TFEB knockdown indeed causes a decrease in basal mTORC1 activity or inhibition of AA-mediated mTORC1 activation, and 2) whether overexpression of TFEB in TSC2+/+ cells is sufficient to induce V-ATPase expression, thereby leading to activation of mTORC1. As alluded to earlier, we have not been able to obtain evidence independently of that generated with bafilomycin and have replaced this figure with two figures showing that mTORC1 coordinately regulates TFEB phosphorylation and nuclear localization. 2. It will be much more convincing if the authors could address whether disruption of V-ATPase activity by knocking down its V0 domain c subunit (a target of bafilomycin A1) or multiple subunits indeed results in mTORC1 suppression. As requested by the reviewer, we have knocked-down V0c as well as V1A and also in combination. We did not find that these knockdowns affected mTORC1 activity. While the results are negative, we did find that knockdown of these subunits diminished endocytosis, and the data has been added onto the paper. Despite this positive control, it is difficult to draw conclusions from negative results… Nonetheless, the bafilomycin figure has now been moved to the supplementary section. 3. As the authors cited, a recent work from Sardiello et al. (2009, Science, 325:473-7) discovered that most lysosomal genes including V-ATPase exhibit coordinated behavior and are regulated by TFEB. They showed that TFEB expression not only enhances expression of lysosomal genes, but also increases the number of lysosomes in the cell, leading to improvement of the lysosomemediated degradative pathways. Based on this work, therefore, several interesting questions could be asked. First, are other lysosomal genes such as cathepsin D also regulated by mTORC1 like V-ATPase? What is the expression profile of the overall lysosomal genes in the data set of the authors' microarray? Do they also show the LLHL pattern similar to V-ATPase expression? Second, is the number of lysosomes (primary location for mTORC1 activation) elevated in TSC deficient cells relative to TSC2+/+ cells due to the increased TFEB levels by hyperactive mTORC1? Can rapamycin treatment also decrease the number of lysosomes in TSC-/- cells? We had examined the expression of lysosomal genes in our arrays and indeed found that


9


many lysosomal genes exhibited a LLHL pattern. We have highlighted the specific genes in Fig 1D. We also performed some experiments using lysotracker to evaluate lysosomes in Tsc2-deficient cells and wild-type cells, and while we cannot rule out subtle differences, no striking differences were observed. 4. In figure 6 and 7, it is not clear how TFEB expression itself is regulated by mTORC1. Is it regulated through transcriptional activation like V-ATPase or through increased translational rate like HIF-1a? We have added two new figures, 9 and 10, showing that mTORC1 regulates TFEB phosphorylation and nuclear localization. 5. There appear some discrepancies between mRNA and protein expression in terms of correlation of their fold increases. For instance, while increase in protein expression of V-ATPase subunits (figure 4b) is quite dramatic compared to the small increase in their mRNA levels (less than 2 fold; figure 4c), an increase of over 5 times in mRNA expression (figure 2b) seems to correlate with only a minor increase in protein expression (figure 2c). These differences may reflect differences across tissues. Fig. 4b, which pertains to the liver shows a very striking upregulation in V-ATPase protein levels. By contrast, in MEFs the protein upregulation is less (see also Fig. 5c). 6. Again, in figure 6a, b, and c, a dramatic increase (over 6 times) in TFEB mRNA expression in TSC-/- cells relative to TSC+/+ cells did not show a corresponding increase in its protein expression. Rather, TFEB protein displayed a band shift up/down pattern, perhaps representing a sign of protein modification such as phosphorylation instead of changes in protein levels per se. The authors should provide the re-evaluated data or some explanation for this significant discrepancy. Indeed, as the reviewer points out, and we now show in Figs 9 and 10, mTORC1 induces changes in TFEB mobility on SDS-PAGE and TFEB mobility is markedly affected by phosphatase treatment suggesting that the changes in mobility are induced, at least in part, by changes in phosphorylation. Minor Comments: 1. In page 3, line 65, Rab4 should be corrected by Rab5. In addition, the authors should cite a recent work from Guan's group (Li et al., 2010, J. Biol. Chem., 285:19705-9) as this work also identified the role of Rab5 in mTORC1 signaling. Thank you, it has been corrected, and the reference added. 2. In page 8, line 209, TSC2-/- should be changed into TSC2+/+. Thank you, it has been corrected. 3. Along with figure 4e, the authors should provide the corresponding Western data as well. We are not sure what figure this is. 4. In figure 9, the authors should also provide the Western data of V-ATPase and TFEB expressions in order to reversely show that mTORC1 activation positively regulate the expression of these proteins. As alluded to, figure 9 has now been moved to the supplementary section.


10


2nd Editorial Decision

01 February 2011

Thank you for sending us a new version of your manuscript as a new submission. After some delay due to the past Christmas holiday season our original referees have now seen it again. As you will see while referee 1 is more critical and thinks that the functional link between TORC1 signalling and TFEB is not established convincingly enough the other two referees are more positive and would support publication here after some more revisions. Now, given that you have shifted the focus of the study towards the TORC1/TBEB link I think that it will be required to address the points raised by referee 1 with some more data. Also, the remaining issues put forward by the other referees - in particular the physiological significance of the findings - need to be taken care of in a re-revised version of the manuscript. In addition there are a number of editorial issues that need further attention. First I need to ask you to submit your microarray data to one of the public databases and to include the accession details into the manuscript text. Second, please note that at least three independent experiments are required to calculate average values and standard deviations (please see figure 4C and supplementary figure S2). Please include the number of independent experiments performed into the legend of figure 6. When preparing your letter of response to the referees' comments, please bear in mind that this will form part of the Peer Review Process File, and will therefore be available online to the community. For more details on our Transparent Editorial Process initiative, please visit our website: http://www.nature.com/emboj/about/process.html Thank you for the opportunity to consider your work for publication. I look forward to your rerevised manuscript. Yours sincerely, Editor The EMBO Journal -----------------------------------------------REFEREE COMMENTS Referee #1: This resubmitted manuscript version is improved. While it must be credited that some of the points raised in the previous review have been either discussed or satisfactorily answered, some other important issues still remain very cloudy. This is particularly critical, when it comes to TFEB involvement, especially the mechanisms by which mTOR might control this transcription factor. 1.- As indicated in the previous review the fact that bafilomycin A1 affects AMPK complicates the interpretation of the results using this drug; especially with the negative effects of knockdowns of VATPase, paradoxically performed in HeLa cells that the authors admit don't express LKB. 1.- The critical point, however, lies in the TFEB involvement and the biochemical data presented fails to present a cohesive and believable mechanism by which mTOR controls TFEB and its transcriptional implications. 1) In Figs. 9 and 10, it appears that the main fraction that is phosphorylated upon rapamycin treatment is the cytoplasmic fraction, however, the critical pool, at least for transcription, remains unaffected by rapamycin. In addition, the levels of nuclear TFEB are very, very minor suggesting that the authors are missing a key signaling that leads to the translocation of this transcription factor to the nucleus. 2) It is surprising and very concerning that all the phosphorylation changes of TFEB are only observed in TSC2 deficient cells, this raises important issues about the physiological relevance of this pathway in normal cells; why TSC2 needs to be inactivated to see cytoplasmic changes in phosphorylation? 3) Importantly, the model implies


11


that dephosphorylation promotes translocation of TFEB to the nucleus, maybe I missed this data, but I couldn't find convincing data the support this mechanism. In my view, these are critical points that remain unsolved at this point and are essential to support the conclusions of this manuscript.

Referee #2: In the present study, the authors describe a novel role of mTORC1 in regulating the expression of a component of the endosome, the vacuolar H+-ATPase (V-ATPase). Using microarray analyses, the authors demonstrated that the expression of V-ATPase correlates with the activity of mTORC1. They further showed that the regulation of V-ATPase by mTORC1 requires the transcription factor TFEB. They demonstrate that mTORC1 can regulate TFEB phosphorylation and nuclear localization. Specifically, they found that mTORC1 activation correlates with the presence of a hypophosphorylated form of TFEB and that this form is present in the nucleus. mTORC1 also promotes endocytosis in a TFEB- and V-ATPase-dependent manner. Based on their results, the authors conclude that mTORC1 regulates TFEB and V-ATPase and propose a model whereby mTORC1 can control V-ATPases and endocytosis via the oncogenic transcription factor TFEB. The microarray data are extensive and support the involvement of mTORC1 in V-ATPase expression. The involvement of mTORC1 in TFEB regulation is also very interesting and convincing. The authors attempt to link the function of mTORC1 in endocytosis via V-ATPase and TFEB (Fig 8), but the results need some clarification (see below). Some evidence to directly support the regulation of V-ATPase transcription by TFEB (as their model suggests) would significantly strengthen the conclusions of these studies. Specific comments: 1. Figure 4A is confusing as there seems to be some discrepancy with Fig. 4B. In Fig 4A, the nonrecombined Tsc1Flox allele reappeared 5 days after Ad-Cre injection, which means after 5 days (probably at 7 or 10 day time points) the knockdown effect would be attenuated. However, the results in Fig 4B did not reflect that. 2. In Fig 6, the authors showed that the regulation of V-ATPase expression by mTORC1 is Tfebdependent. Have the authors tried to investigate if Tfeb directly regulates the transcription of VATPase genes (perhaps by ChIp assay)? 3. In Fig 8, FACS analysis was performed to see the BSA uptake by Tsc2+/+ and Tsc2-/- cells. As described in the Material and Method, the staining strategy used here would only detect surface proteins. However, the endocytosed BSA would be intracellular. Can the authors please clarify the methodology? 4. In Fig 10F and G, the inactivation of mTORC1 in control scrambled siRNA cells did not diminish S6K1 phosphorylation as one would have expected. Referee #3: In the revised manuscript titled, "Regulation of TFEB and V-ATPase by mTORC1" by Pena-Llopis et al., have completed a significant amount of work to address some of the issues raised from the previous review. Since the authors have failed to verify the role of V-ATPase in mTORC1 regulation thoroughly, they have changed the emphasis of the manuscript to the link between mTORC1 and TFEB activation. They provide an interesting mechanism by which mTORC1 regulates phosphorylation state of TFEB and its nuclear localization. While this could be one of the major mechanisms responsible for transcriptional regulation of the genes by mTORC1 such as the V-ATPase gene, I still have some concerns with the manuscript as follows, and these concerns need to be addressed before being suitable for publication.


12


1. It is not clear what physiological signals may actually cause mTORC1-dependent transcriptional regulation of V-ATPase genes via TFEB. Therefore, in addition to TSC depletion, the authors need to address whether signals capable of activating mTORC1 such as from growth factors, amino acids and/or glucose are indeed able to trigger changes in phosphorylation and nuclear localization of TFEB and the subsequent expression of V-ATPase. These mTORC1-specific changes should be verified with raptor knockdown and/or rapamycin treatment. 2. In figure 6c, the authors should provide Western data as well in order to show that shTFEB actually induces a decrease in V-ATPase protein expression. 3. Since the authors did not observe any inhibitory effect of V-ATPase knockdown on mTORC1 activity, it is more likely that V-ATPase activity may not be involved in mTORC1 regulation, and that the inhibition of mTORC1 by bafilomycin A1 treatment may be caused by off-target or indirect effects. If the authors insist on leaving this data in the manuscript, they must complete a more thorough analysis, the authors should complete AA stimulation or GF stimulation experiments, in addition to starving and stimulating with both to get an idea if Bafilomycin is inhibiting the AA or GF signal to mTORC1. They should also provide how many times they completed the bafilomycin experiments? Show that knockdown of V-ATPase and other genes linked to lysosome biogenesis identified as regulated by mTORC1, also inhibit mTORC1 activation. Additionally, what is the concentration of bafilomycin and how long is it pre-treated? This information is not provided. Is the inhibitory effect of bafilomycin also reproducible when another V-ATPase-specific inhibitor such as conconamycin A is used? Without substantial characterization, the link between bafilomycin and mTORC1 activation should be removed.

1st Revision

21 April 2011

Response to editor: Thank you for sending us a new version of your manuscript as a new submission. After some delay due to the past Christmas holiday season our original referees have now seen it again. As you will see while referee 1 is more critical and thinks that the functional link between TORC1 signalling and TFEB is not established convincingly enough the other two referees are more positive and would support publication here after some more revisions. Now, given that you have shifted the focus of the study towards the TORC1/TBEB link I think that it will be required to address the points raised by referee 1 with some more data. Also, the remaining issues put forward by the other referees - in particular the physiological significance of the findings - need to be taken care of in a re-revised version of the manuscript. We’re addressing the points raised by reviewer 1 and have added a new figure showing that TFEB is extensively regulated in normal cells and is, in fact, required for cell proliferation (see Fig. 9). In addition there are a number of editorial issues that need further attention. First I need to ask you to submit your microarray data to one of the public databases and to include the accession details into the manuscript text. Transcriptome microarrays were deposited in GEO under accession numbers GSE27982 and GSE28021. Information added onto the manuscript in the section referring to the Supplementary information. Although the data will remain private until the manuscript is accepted, the arrays can be accessed using the following confidential links: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=rtuvnauacqaswfi&acc=GSE27982 http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=lralpicomqyeabo&acc=GSE28021 Second, please note that at least three independent experiments are required to calculate average values and standard deviations (please see figure 4C and supplementary figure S2).


13


Figure 4C (now 3B) and S2 (now S4) included an analysis performed in 20 mice (two per genotype and per timepoint). The bar graph represents means and standard errors. The results have been reviewed by our biostatistician, who is a coauthor of this manuscript, Dr. Xie, who agrees with the analysis. In addition, the number of mice used for the experiment has been indicated in the figure legend. Please include the number of independent experiments performed into the legend of figure 6. The qRT-PCR shown in Fig 6. is now shown in Fig. 5A. It involves 3 independent experiments and the number has been added onto the figure legend. Fig. 5E, which also has additional qRT-PCR data, similarly involves 3 independent experiments and the number is in the figure legend. Once again, we are very grateful to the referees for the time they spent on our manuscript and their insightful comments. Please note that some points raised may be addressed out of order. Referee #1: This resubmitted manuscript version is improved. While it must be credited that some of the points raised in the previous review have been either discussed or satisfactorily answered, some other important issues still remain very cloudy. This is particularly critical, when it comes to TFEB involvement, especially the mechanisms by which mTOR might control this transcription factor. 1.- As indicated in the previous review the fact that bafilomycin A1 affects AMPK complicates the interpretation of the results using this drug; We agree with the reviewer and are in fact quite grateful for letting us know of what we believe is unpublished data, that bafilomycin may have effects other than on V-ATPases and that it may activate AMPK. This was in fact the reason why the focus of the paper shifted from the feedback loop model we had originally proposed to TFEB. Already in the previous revision, the data on bafilomycin was moved to the supplementary. Due to these concerns and the request from reviewer 3, it has now been eliminated altogether from the manuscript. In fact, data with bafilomycin is only shown as a control (see next section). especially with the negative effects of knockdowns of V-ATPase, paradoxically performed in HeLa cells that the authors admit don't express LKB. The reason for the use of HeLa cells is that HeLa cells are deficient for LKB1 (Tiainen et al., PNAS 1999), which is required for AMPK activation, and thus in these cells bafilomycin action is likely to be mediated through V-ATPases. In fact, consistent with this notion, the effects of bafilomycin on endocytosis in HeLa cells can be reproduced by knocking down V-ATPase subunits (V0C, V1A or in combination) (see Fig. 6D and E). Nevertheless, it should be pointed out that in these experiments bafilomycin is simply used as a control. 1.- The critical point, however, lies in the TFEB involvement and the biochemical data presented fails to present a cohesive and believable mechanism by which mTOR controls TFEB and its transcriptional implications. The model that mTORC1 controls TFEB migration and nuclear localization is supported by the following data: i. In low serum conditions, in which mTORC1 is active only in cells deficient for TSC1/TSC2, TFEB is found in a fast migrating form only in Tsc2-deficient MEFs. Furthermore, mTORC1 inhibition in these cells with rapamycin shifts TFEB up giving rise to a pattern undistinguishable from that observed in wild-type cells (Fig. 7A). These results are highly reproducible, and the same pattern is observed in Tsc1-deficient MEFs (Fig. 7B). Furthermore, as shown in Fig. 7C, the shift induced by rapamycin is already observed within 1 hour.


14


ii.

TFEB is found in the nucleus, under serum starvation conditions, only in Tsc2-deficient cells (Fig. 7E) and it is only the fast migrating form of TFEB that is nuclear (Fig. 7E). iii. Inasmuch as phosphatase treatment eliminates the differences in TFEB migration in Tsc2deficient (fast TFEB) and wild-type cells (slow TFEB) the data argue that these differences are due, at least in part, to differences in phosphorylation (See Fig. 7D). Furthermore, since phosphatase treatment not only downshifts slow migrating (cytosolic) TFEB, but also fast migrating (nuclear) TFEB (see Fig. 7F), the data show that not only is slow migrating cytosolic TFEB phosphorylated, but also nuclear TFEB is phosphorylated. Taken together the results show that mTORC1 coordinately regulates TFEB phosphorylation and nuclear localization. Importantly, nearly identical results were observed in a totally unrelated cell type, HeLa cells. In this cells, TSC2 depletion resulted in fast migrating TFEB (Fig. 8A), which was retarded by treatment with rapamycin (Fig. 8A and B) and both fast and slow migrating TFEB were phosphorylated (Fig. 8C). As for MEFs, TFEB was nuclear only in TSC2-deficient cells (Fig. 8D) and treatment with rapamycin excluded TFEB from the nucleus (Fig. 8D). Finally, the effects of rapamycin on TFEB were recapitulated by knockdown of raptor (using either siRNA or shRNA). With respect to the transcriptional implications, stable depletion of TFEB downregulates the expression of ~20% of mTORC1 induced genes (Fig. 5D), which we interpret to mean that TFEB is a critical effector downstream of mTORC1. 1) In Figs. 9 and 10, it appears that the main fraction that is phosphorylated upon rapamycin treatment is the cytoplasmic fraction, however, the critical pool, at least for transcription, remains unaffected by rapamycin. While the reviewer is correct that rapamycin treatment increases the amount of hyperphosphorylated, slow migrating TFEB, rapamycin also excludes TFEB from the nucleus (see Fig. 7E and 7G). Thus, rapamycin increases the amount of slow migrating, cytosolic, TFEB and decreases the amount of fast migrating, nuclear, TFEB. In addition, the levels of nuclear TFEB are very, very minor suggesting that the authors are missing a key signaling that leads to the translocation of this transcription factor to the nucleus. We agree with the reviewer that the amount of nuclear TFEB is relatively small by comparison to the total amount of TFEB. However, even relatively small amounts of a transcription factor in the nucleus can have profound consequences. 2) It is surprising and very concerning that all the phosphorylation changes of TFEB are only observed in TSC2 deficient cells, this raises important issues about the physiological relevance of this pathway in normal cells; why TSC2 needs to be inactivated to see cytoplasmic changes in phosphorylation? In response to the reviewer we have now performed experiments in MEFs and HeLa cells that are wild type for TSC2. As shown in Fig. 9, changes in TFEB phosphorylation occur in wildtype cells in response to multiple stimuli. In fact, the complexity is such (and this is in keeping with the finding that TFEB is phosphorylated in over 10 sites by phosphoproteomic studies) that we believe that the tightness of the experimental system we utilized was instrumental in being able to isolate TFEB regulation by mTORC1 from TFEB regulation by multiple other signals. 3) Importantly, the model implies that dephosphorylation promotes translocation of TFEB to the nucleus, maybe I missed this data, but I couldn't find convincing data the support this mechanism. We are sorry if this point was not clear in the original manuscript. We have now re-written the discussion to clarify this point (see lines 429-452). Briefly, our data show that: i. Nuclear and cytosolic TFEB exhibit different migration patterns on SDS-PAGE (Fig. 7E, F, G and 8E) ii. Nuclear TFEB migration can be accelerated by phosphatase treatment, indicating that nuclear TFEB, despite its fast mobility, remains phosphorylated to some extent (Fig. 7F and 8E)


15


iii. Upon phosphatase treatment, the migration of nuclear (fast) and cytosolic (slow) TFEB becomes undistinguishable (see Fig. 7F and 8E). The simplest explanation for these data is that TFEB is hyperphosphorylated in the cytosol and hypophosphorylated in the nucleus. However, since the sites of phosphorylation are unknown, the data is open to other explanations as well, and we cannot exclude that nuclear TFEB may be phosphorylated at different sites than cytosolic TFEB. Referee #2: In the present study, the authors describe a novel role of mTORC1 in regulating the expression of a component of the endosome, the vacuolar H+-ATPase (V-ATPase). Using microarray analyses, the authors demonstrated that the expression of V-ATPase correlates with the activity of mTORC1. They further showed that the regulation of V-ATPase by mTORC1 requires the transcription factor TFEB. They demonstrate that mTORC1 can regulate TFEB phosphorylation and nuclear localization. Specifically, they found that mTORC1 activation correlates with the presence of a hypophosphorylated form of TFEB and that this form is present in the nucleus. mTORC1 also promotes endocytosis in a TFEB- and V-ATPase-dependent manner. Based on their results, the authors conclude that mTORC1 regulates TFEB and V-ATPase and propose a model whereby mTORC1 can control V-ATPases and endocytosis via the oncogenic transcription factor TFEB. The microarray data are extensive and support the involvement of mTORC1 in V-ATPase expression. The involvement of mTORC1 in TFEB regulation is also very interesting and convincing. The authors attempt to link the function of mTORC1 in endocytosis via V-ATPase and TFEB (Fig 8), but the results need some clarification (see below). Some evidence to directly support the regulation of V-ATPase transcription by TFEB (as their model suggests) would significantly strengthen the conclusions of these studies. Specific comments: 1. Figure 4A is confusing as there seems to be some discrepancy with Fig. 4B. In Fig 4A, the nonrecombined Tsc1Flox allele reappeared 5 days after Ad-Cre injection, which means after 5 days (probably at 7 or 10 day time points) the knockdown effect would be attenuated. However, the results in Fig 4B did not reflect that. We appreciate the reviewer pointing this out, as it is something we should have clarified. As we recently reported in a manuscript using the same Ad-Cre delivery system (Kucejova et al. Oncogene 2011), the delivery of Cre using adenoviruses results in liver inflammation and infiltration by lymphocytes and macrophages. We begin to see this 3-4 days after infection (see Kucejova et al. Oncogene 2011), and because these cells have not had Tsc1 recombined, this results in an apparent increase in the unrecombined Tsc1 allele. The reviewer is also correct that this may be expected to result in an increase in the amount of Tsc1 protein. However, lymphocytes express very low levels of Tsc1 mRNA (data not shown) and thus the increase in unrecombined Tsc1 DNA does not translate in an increase in Tsc1 protein. 2. In Fig 6, the authors showed that the regulation of V-ATPase expression by mTORC1 is Tfebdependent. Have the authors tried to investigate if Tfeb directly regulates the transcription of VATPase genes (perhaps by ChIp assay)? We have performed an extensive evaluation of commercially available antibodies to perform ChIP assays on the endogenous protein and have not found a single antibody that works. On the other hand, Sardiello et al (Science 2009) showed that among other lysosomal genes, V-ATPase gene sequences were immunoprecipitated with antibodies against epitope-tagged TFEB in ChIP assays. 3. In Fig 8, FACS analysis was performed to see the BSA uptake by Tsc2+/+ and Tsc2-/- cells. As described in the Material and Method, the staining strategy used here would only detect surface proteins. However, the endocytosed BSA would be intracellular. Can the authors please clarify the methodology?


16


We should have clarified this as well. The staining strategy allows us to detect intracellular BSA, which is conjugated to FITC. However, as the pH of the endosome decreases, FITC signal will also decrease and it is unlikely that we are able to detect BSA-FITC in very acidic compartments. 4. In Fig 10F and G, the inactivation of mTORC1 in control scrambled siRNA cells did not diminish S6K1 phosphorylation as one would have expected. We did not expect that mTORC1 inactivation would diminish S6K1 phosphorylation in these experiments (now Fig. 8F and G), as the experiments were done in low serum (as in the MEFs), and mTORC1 activity was already inhibited at baseline. We have made this more explicit in the text to avoid confusion (see lines 331-333). Referee #3: In the revised manuscript titled, "Regulation of TFEB and V-ATPase by mTORC1" by Pena-Llopis et al., have completed a significant amount of work to address some of the issues raised from the previous review. Since the authors have failed to verify the role of V-ATPase in mTORC1 regulation thoroughly, they have changed the emphasis of the manuscript to the link between mTORC1 and TFEB activation. They provide an interesting mechanism by which mTORC1 regulates phosphorylation state of TFEB and its nuclear localization. While this could be one of the major mechanisms responsible for transcriptional regulation of the genes by mTORC1 such as the V-ATPase gene, I still have some concerns with the manuscript as follows, and these concerns need to be addressed before being suitable for publication. 1. It is not clear what physiological signals may actually cause mTORC1-dependent transcriptional regulation of V-ATPase genes via TFEB. Therefore, in addition to TSC depletion, the authors need to address whether signals capable of activating mTORC1 such as from growth factors, amino acids and/or glucose are indeed able to trigger changes in phosphorylation and nuclear localization of TFEB and the subsequent expression of V-ATPase. These mTORC1-specific changes should be verified with raptor knockdown and/or rapamycin treatment. The reviewer brings up an important point. TFEB is indeed regulated in wild-type cells, but we find that TFEB regulation is rather complex, and despite that the interventions proposed have predictable effects on mTORC1, consistent with the observation that TFEB is extensively phosphorylated (more than a dozen phosphorylation sites have been identified in phosphoproteomic studies [Dephoure et al, PNAS 2008; Mayya et al, Sci Sign 2009]), we find that TFEB regulation is dynamic and rather complex and most certainly involving pathways other than mTORC1. Furthermore, the complexity is such that we believe that we were able to identify mTORC1 as a critical regulator of TFEB nuclear localization because we had an extremely tight system in which mTORC1 activity could be manipulated very specifically by removing TSC1/TSC2 or treating with rapamycin. This is illustrated in Fig. 9, which on the other hand shows that TFEB is extensively regulated in wild-type cells. 2. In figure 6c, the authors should provide Western data as well in order to show that shTFEB actually induces a decrease in V-ATPase protein expression. The data has been added. The figure is now the new Fig. 5B. 3. Since the authors did not observe any inhibitory effect of V-ATPase knockdown on mTORC1 activity, it is more likely that V-ATPase activity may not be involved in mTORC1 regulation, and that the inhibition of mTORC1 by bafilomycin A1 treatment may be caused by off-target or indirect effects. If the authors insist on leaving this data in the manuscript, they must complete a more thorough analysis, the authors should complete AA stimulation or GF stimulation experiments, in addition to starving and stimulating with both to get an idea if Bafilomycin is inhibiting the AA or GF signal to mTORC1. They should also provide how many times they completed the bafilomycin experiments? Show that knockdown of V-ATPase and other genes linked to lysosome biogenesis identified as regulated by mTORC1, also inhibit mTORC1 activation. Additionally, what is the concentration of bafilomycin and how long is it pre-treated? This information is not provided. Is the inhibitory effect of bafilomycin


17


also reproducible when another V-ATPase-specific inhibitor such as conconamycin A is used? Without substantial characterization, the link between bafilomycin and mTORC1 activation should be removed. As requested by the reviewer, we have taken the bafilomycin timecourse experiments out of the paper. Having said that, the bafilomycin experiments were highly reproducible and more than 10 experiments had been performed, showing that preincubation with bafilomycin (1 mM) consistently blocked the activation of mTORC1 by amino acids. There is no doubt of this result. However, the interpretation is confounded because, in keeping with the referee’s query it seems that bafilomycin activates AMPK, which would similarly result in the inhibition of mTORC1. We were not able to obtain independent evidence that the effects of bafilomycin were mediated through its effects on V-ATPases and thus, the data, which in the previous version was placed in the supplement, has been removed altogether.

3rd Editorial Decision

03 June 2011

Thank you for sending us your (re-)revised manuscript. Let me first of all apologise for the slow rereview process. However, as you will see, this case is still complicated. In the meantime our original referees 1 and 2 have seen the manuscript again. Referee 3 was not available at the time of resubmission. As you will see, referee 1 is still not convinced that you are making a sufficiently strong case for a causal line of events that link mTORC1-dependent regulation of the TFEB phosphorylation status and TFEB-dependent regulation of V-ATPase gene transcription. Overall, referee 1 therefore does not support publication of the paper here. Referee 2 is clearly more positive (similar as referee 3 on the previous version) and supports publication here. Still, after consulting with him/her on the arguments put forward by referee 1, it becomes clear that he/she agrees with referee 1 that the causality of the line of events you are putting forward is not demonstrated by the data, even though the data are certainly consistent with your model and highly suggestive. Having looked at the manuscript again myself in depth, I can see both points of view as well, and I have now also had a chance to discuss the case in depth with our Chief Editor. On balance, we have come to the conclusion that the manuscript will ultimately be publishable here. Still, given the very recent paper by the Ballabio lab that links ERK2 signalling to TFEB localisation and activity in response to nutrient levels (PMID: 21617040), we think that one experiment to see whether ERK signalling is activated in the context of your experiments and whether ERK signalling could act in the (direct) phosphorylation of TFEB downstream of mTORC1 in your experiments would be rather straight forward and could address the remaining concerns at least to some extent. We would therefore like to ask you to do such an experiment and include it into the paper. In addition, you may wish to take the referees' concerns into account and to discuss these further in the manuscript and/or response. We will then take the final decision at the editorial level. I am sorry that we have to be that insistent at this late stage. Still, given the reservations put forward by both referees, we feel that this is the most constructive way forward in this case. We are looking forward to receiving the final version of this manuscript. Yours sincerely, Editor The EMBO Journal


18


-----------------------------------------------REFEREE COMMENTS Referee #1: In this second revised version the authors have addressed some concerns, particularly the ones related to the use of bafilomycin. However, I still have major concerns regarding the phosphorylation status of TFEB and its regulation by mTOR. Again, the authors assess this regulation mainly in conditions in which TSC2 is missing. In control HeLA cells Fig 8D, there is no differences -at least in the blot presented- of TFEB in the nucleus, it is therefore difficult to attribute changes of gene expression due to this transcription factor. In Fig. 9 MEFs and HeLa cells are used but the shift is assessed in nutrient conditions that correlate with mTOR activity, but never used mTOR inhibitors or Raptor knock down. In addition, they would need to analyze the cytoplasm and nuclear fractions. In my view, the authors need to clearly demonstrate this crucial point. I have seen now this manuscript three times, while it has certainly improved it is still for me hard to believe that the mechanism of TFEB phosphorylation that is proposed can account for the gene expression that depends on mTOR activity. For a very mechanistic type of journal such as EMBO J I would expect that the authors would need to critically address this issue. Referee #2: The authors have addressed my concerns. They have also added new data to support regulation of TFEB phosphorylation by nutrient signals to correlate this phosphorylation with mTORC1 regulation.

2nd Revision

22 June 2011

Please find attached a revised version of our manuscript “Regulation of TFEB and V-ATPases by mTORC1”, EMBOJ-2010-76845R. I appreciate your looking carefully at our manuscript. We have now not only performed the experiment that you requested, but also added a whole new figure that substantially adds to the manuscript. As you mentioned, Ballabio’s group recently reported in Science that TFEB nuclear localization is regulated in response to starvation and they propose that TFEB is driven into the nucleus in a manner which involves ERK2-mediated phosphorylation of S142. As you probably observed however, there is no evidence in their paper that this site is actually phosphorylated in response to starvation in cells. We have nonetheless addressed the role of ERK2 on mTORC1-dependent TFEB nuclear localization as requested. These experiments are in Fig. 8E. As we show, mutation of S142 to A has no effect on mTORC1-mediated TFEB nuclear localization. Furthermore, we have added a full new figure with mutagenesis studies that not only shows that the site proposed to be involved in ERK2mediated regulation is dispensable for mTORC1-dependent TFEB nuclear localization, but more importantly, defines a C-terminal serine-rich motif that is required for TFEB regulation by mTORC1. Mutagenesis studies of the 5 serine residues in this motif show that these residues are necessary for mTORC1-dependent nuclear localization of TFEB. Furthermore, mutation to phosphomimetic residues is sufficient to drive TFEB to the nucleus and obviates the need for mTORC1 activation. These data strongly suggest that TFEB is driven to the nucleus through phosphorylation at one (or several) serine residues in this motif we have identified. These data, not only address your request, but adds very substantially to the manuscript. It has been a long process and I believe that for different reasons pretty much every round of review took 30-40 days, and despite that this manuscript is the first one to show that (1) endogeneous TFEB is regulated at the subcellular localization level (the Science paper deals exclusively with overexpressed protein), and (2) that TFEB is regulated by mTORC1, it would have been nice if its publication had predated the Science paper.


19


We have added a complete new figure, with the corresponding amount of text. We were already close to the character limit, but I have managed by rearranging and deleting some text to stay below the limit. As you state in your letter, I hope that the decision will be made at the editorial level and that there won’t be more delays. I think this is a very exciting manuscript and I hope to hear from you very soon of the manuscript’s acceptance. Rebuttal Referee 1 In this second revised version the authors have addressed some concerns, particularly the ones related to the use of bafilomycin. However, I still have major concerns regarding the phosphorylation status of TFEB and its regulation by mTOR. Again, the authors assess this regulation mainly in conditions in which TSC2 is missing. In control HeLA cells Fig 8D, there is no differences -at least in the blot presented- of TFEB in the nucleus, it is therefore difficult to attribute changes of gene expression due to this transcription factor. As indicated in the text, all the experiments in this figure are performed under serum starvation conditions where mTORC1 is largely inactive. Thus, we would not expect to see TFEB in the nucleus. In Fig. 9 MEFs and HeLa cells are used but the shift is assessed in nutrient conditions that correlate with mTOR activity, but never used mTOR inhibitors or Raptor knock down. In addition, they would need to analyze the cytoplasm and nuclear fractions. In my view, the authors need to clearly demonstrate this crucial point. As we clearly state in the manuscript, and as illustrated by the effects of withdrawal of glucose, amino acids and serum as well as following DTT, the regulation of TFEB is dynamic and complex and we do not believe that can be accounted for exclusively by changes in mTORC1 activity. In addition, TFEB has been shown to be phosphorylated in over 10 sites and its regulation is likely to be very complex. In fact, as mentioned in the discussion, we believe that the discovery that TFEB is regulated by mTORC1 was only possible because of a very tight experimental system in which mTORC1 activity could be precisely manipulated genetically and pharmacologically.

4th Editorial Decision

24 June 2011

Thank you for sending us your re-revised manuscript. I have now had a chance to look at the manuscript again, and you will be pleased to learn that you have now addressed all criticisms in a satisfactory manner. The paper will now be publishable in The EMBO Journal and you will receive a formal acceptance letter shortly. Thank you very much again for considering our journal for publication of your work. Yours sincerely, Editor The EMBO Journal


20