S-sulfhydration of MEK1 leads to PARP-1 activation and DNA damage repair

Transcription

1 Manuscript EMBOR S-sulfhydration of MEK1 leads to PARP-1 activation and DNA damage repair Kexin Zhao, YoungJun Ju, Shuangshuang Li, Zaid Al Taany, Rui Wang and Guangdong Yang Corresponding author: Guangdong Yang, Lakehead University Review timeline: Submission date: 08 November 2013 Editorial Decision: 09 December 2013 Correspondence: 11 December 2013 Correspondence: 12 December 2013 Revision received: 18 February 2014 Editorial Decision: 13 March 2014 Accepted: 31 March 2014 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.) Editor: Nonia Pariente 1st Editorial Decision 09 December 2013 Thank you for your submission to EMBO reports. We have now received reports from the three referees that were asked to evaluate your study, which can be found at the end of this . As you will see, although especially referees 1 and 2 referees find the topic of interest, they all consider the study preliminary for publication here at this stage. All of them request a number of experiments that would strengthen the conclusiveness of your study and point out technical concerns with the data. Clearly, further experimentation would be necessary to provide convincing support for your model of the role of MEK1 sulfhydration in PARP1 activity. Given that the referees provide constructive suggestions on how to make the work more conclusive, I would like to give you the opportunity to revise your manuscript. Please note that for further consideration in EMBO reports, it would be essential to demonstrate that endogenous H2S levels are responsive to DNA damage and could lead to S-sulfhydration by CES in this setting, as well as demonstrate that NaHS-dependent PARP1 activation is dependent on MEK1 and ERK1/2 activity through knock-down studies. In addition, the key findings would need to be validated in at least another cell line and S-sulfhydration of MEK1 demonstrated directly. Finally, the various technical concerns pointed out by the three referees would also need to be addressed. Please note, however, European Molecular Biology Organization 1

2 that the elucidation of the molecular mechanism of ERK-dependent PARP1 activation, requested by referee 3, would be beyond the scope of a short report and does not need to be pursued. Please note that it is EMBO reports policy to undergo one round of revision only and thus, acceptance of your study will depend on the outcome of the next, final round of peer-review. We also welcome the submission of cover suggestions or motifs that might be used by our Graphics Illustrator in designing a cover. I look forward to seeing a revised form of your manuscript when it is ready. In the meantime, do not hesitate to get in touch with me if I can be of any assistance. REFEREE REPORTS: Referee #1: The authors, authorities on H2S signaling, report that sulfhydration of key targets underlies notable aspects of DNA repair. This reflects a substantial advance in the H2S field. There are a few modest comments: 1- In Figure S2E, the authors show increased levels of CSE mrna after MMS treatment in HUVECs. What about the protein levels? In addition, is there an increase in H2S production after MMS treatment? 2- In Figure 4F, the authors show that the levels of PAR are decreased when cysteine 341 is mutated. However it is not clear whether this experiment was conducted under basal conditions. Have the authors tested the sulfhydration of cys-341 when cells are treated with MMS? 3- In Figure 1C, 1D, the authors show convincingly, a decrease of PAR levels in kidney and liver of CSE knockout mice. It would be informative if the authors could provide the relative quantitation of PAR levels in wild-type and CSE knockouts. 4- In page 7 and page 10 there are some typos p7 line 22 "...(jjhg)..." similarly, p10 line 4, "...(hhdhaf)...". In addition, the word catalysis in page 10 line 4 should be change to catalytic. 5- The authors could also expand the discussion on the role of CSE in DNA damage repair and its relevance to disease states. 6- The authors may wish to rephrase the title as the sentence used is not accurate. Rather than "Mediation of MEK-1 Sulfhydration... damage repair" the authors could consider "MEK-1 S- sulfhydration by hydrogen sulfide mediates PARP-1 activation in DNA damage repair". Referee #2: The paper by Zhao et al describes a potential role of MEK1 S-sulfhydration in PARP1 activation and DNA repair. These findings are of interest however the authors would need to provide more convincing experiments to support the suggested model: It would be vital to see whether knocking down the MEK1 and ERK1/2 prevents the NaHSdependent PARP1 activation and leads to a diminished DNA repair. Figure 1B,C. Is PARG activity comparable between wild type and deficient extracts. The paper would have been more convincing if the NaHS-dependent PARP1 activation was demonstrated in at least one more cell line. European Molecular Biology Organization 2

3 C341G mutation could have affected the protein folding, so detecting directly S-sulfhydration of this residue in MEK1 by mass spectrometry would be essential. Suppl. Figure 2E. MMS-induced CSE mrna expression observed here complicates the model and it should be discussed. Is this something that is observed in other cell lines? The presented data seem to suggest that the treatment with NaHS activates PARP1 activity stronger and faster than the MMS treatment (Suppl. Figure 2A). Is this something that the authors consistently observe? This figure would benefit from including the control with the NaHS treatment only? Figure 1E. Are low PAR levels specific for heart tissue or it is something seen in many tissues? The westerns blots are lacking size markers. Some of the loading controls in western blot experiments are heavily overloaded which doesn't allow for a proper judgement of the data. Suppl. Figure 3B is missing a GAPDH positive control. Suppl. Figure 5B. Why a direct detection of S-sulfhydrated MEK1 couldn't be shown? Referee #3: In this manuscript, the authors report that hydrogen sulfide (H2S) facilitates DNA damage repair via MEK1 S-sulfhydration and PARP1 activation. H2S is a gastransmitter and plays important roles in a number of biological processes. Here, the authors demonstrate that H2S activates the enzymatic activity of PARP1 and promotes cell viability in response to DNA damage. During this process, the activation of PARP1 is regulated by MEK1 S-sulfhydration. Although the concept of the study is novel, this study is too preliminary to warrant publication. The molecular mechanism of ERKdependent PARP1 activation in the context of DNA damage is unclear. Thus, the major conclusions in this manuscript are questionable. Specific points are listed below: 1. Endogenous H2S should be measured in response to DNA damage. The regulation of endogenous H2S and S-sulfhydration by CES should be carefully studied. 2. PAR is a mixture of heterogeneous chains. It is very strange to see a neat band of PAR in SDS- PAGE presented in Fig 1A, B, 3A, B, E, and S3A. Without a molecular marker, it is hard to determine what has been shown in these figures. 3. Huge amount of PAR in both kidney and MEFs has been detected without genotoxic stress. Usually, the half-life of PAR is very short in vivo because PARG, the PAR hydrolase, is very potent in vivo. The authors need to provide a convincing explanation. 4. Page 4, line 2 "in which MEK/ERK pathway plays a critical role in initiating PARP automodification [12, 13]". Reference 12 does not support this conclusion. Correspondence - editor 11 December 2013 I have had some further communication with the referees and I think that two comments regarding points 2 and 3 raised by referee #3 could be helpful for you to address them, which I mentioned would be a requirement during revision. Please find them below. P.D.: Please note that all other points indicated in my decision letter need also be addressed European Molecular Biology Organization 3

4 "Comments to referee #3 points 2 and 3 Comment 2 - the authors surely have to show the size markers and more of the gel area, although a 'neat' PARP band might be OK. Comment 3 - a very significant PARP activation observed here quite likely happened only after the cells were open in the extraction buffer (not in vivo). To exclude this possibility, the authors should repeat the western blot using RIPA buffer supplemented with PARP inhibitors (that would suppress the artificial PARP activation during cell extract preparation)." Correspondence - author 12 December 2013 Thanks for all your s and kind suggestions. My colleagues and I will carefully perform additional experiments and try to address all the comments raised by the 3 reviewers. We will submit it back to EMBO reports within 3 months. 1st Revision - authors' response 18 February 2014 Responses to Comments from Editor Please note that for further consideration in EMBO reports, it would be essential to demonstrate that endogenous H 2 S levels are responsive to DNA damage and could lead to S-sulfhydration by CSE in this setting, as well as demonstrate that NaHS-dependent PARP1 activation is dependent on MEK1 and ERK1/2 activity through knock-down studies. Response: In the last several months, we performed additional studies to address all these comments. The data demonstrated that: 1) MMS (a well known DNA damage inducer) stimulates CSE expression and H 2 S production in HUVECs (Supplemental Figure 4C, 4D and 4E); 2) MMS induces MEK1 S-sulfhydration in HUVECs (Figure 4E); 3) CSE overexpression increases MEK1 S- sulfhydration in HEK-293 cells (Supplemental Figure 8C); 4) Transfection of HUVECs with MEK1-siRNA or ERK1/2-siRNA significantly decrease the expression of MEK1 or ERK1/2, prevent NaHS-induced PARP1 activation and attenuate NaHS-protected DNA repair (Supplemental Figure 7C, 7D and 7E). All these data have been incorporated into the revised manuscript. In addition, the key findings would need to be validated in at least another cell line and S- sulfhydration of MEK1 demonstrated directly. Response: By using another cell line, human fibroblasts (ATCC, CCL-75 TM ), we confirmed that NaHS stimulates PARP1 activity and attenuates MMS-induced DNA damage (Supplemental Figure 5A and 5B). We also provided evidence showing NaHS enhances MEK1 S- sulfhydration in human fibroblasts (Supplemental Figure 8B). We previously showed that knockdown of CSE by sirna significantly stimulates cellular senescence in human fibroblasts (Yang et al. Antioxid. Redox. Signal. 2013, 18: ). Finally, the various technical concerns pointed out by the three referees would also need to be addressed. Response: All the comments from the three referees were addressed point by point. Please note, however, that the elucidation of the molecular mechanism of ERK-dependent PARP1 activation, requested by referee 3, would be beyond the scope of a short report and does not need to be pursued Response: Thanks for the encouragement, and we have tried our best to address the comments from referee 3. European Molecular Biology Organization 4

5 Responses to Comments from Reviewer 1 The authors, authorities on H 2 S signaling, report that sulfhydration of key targets underlies notable aspects of DNA repair. This reflects a substantial advance in the H 2 S field. There are a few modest comments: Response: We thank the reviewer for appraising our work and raising many important concerns. 1- In Figure S2E, the authors show increased levels of CSE mrna after MMS treatment in HUVECs. What about the protein levels? In addition, is there an increase in H 2 S production after MMS treatment? Response: We performed additional experiments to explore the effects of MMS on CSE protein expression and H 2 S production in HUVECs. Similar to the change of CSE mrna level, incubation of HUVECs with 1 mm MMS significantly induced CSE protein level and H 2 S production rate in HUVECs (Supplemental Figure 4D and 4E), suggesting that MMS may activate CSE at the transcription level. 2- In Figure 4F, the authors show that the levels of PAR are decreased when cysteine 341 is mutated. However it is not clear whether this experiment was conducted under basal conditions. Have the authors tested the sulfhydration of cys-341 when cells are treated with MMS? Response: In the original Figure 4F, HEK293 cells were first transfected with His6-MEK1 or His6-MEK1 cysteine residue mutants for 48 hours, the cells were then incubated with NaHS (10 µm) for 2 hours. To follow the suggestion, we now performed new experiments to test the sulfhydration of cys-341 in the presence of NaHS with or without MMS. The results showed that NaHS has no effect on MEK1 S-sulfhydration when cysteine 314 is mutated even in the presence of MMS (Figure 4E). 3- In Figure 1C, 1D, the authors show convincingly, a decrease of PAR levels in kidney and liver of CSE knockout mice. It would be informative if the authors could provide the relative quantitation of PAR levels in wild-type and CSE knockouts. Response: Thanks for this suggestion. The relative quantitation of PAR level in kidney andliver tissues was provided in the revised manuscript (Figure 1B and 1C). 4- In page 7 and page 10 there are some typos p7 line 22 "...(jjhg)..." similarly, p10 line 4, "...(hhdhaf)...". In addition, the word catalysis in page 10 line 4 should be change to catalytic. Response: We are sorry for the overlook, and these have been corrected. 5- The authors could also expand the discussion on the role of CSE in DNA damage repair and its relevance to disease states. Response: To address this comment, we added new discussion on the role of CSE in DNA damage repair as well as its relevance to various diseases in the revised manuscript (last paragraph). 6- The authors may wish to rephrase the title as the sentence used is not accurate. Rather than "Mediation of MEK-1 Sulfhydration... damage repair" the authors could consider "MEK-1 S- sulfhydration by hydrogen sulfide mediates PARP-1 activation in DNA damage repair". Response: As suggested, the title for this revised manuscript was changed to MEK-1 S- sulfhydration by hydrogen sulfide mediates PARP-1 activation in DNA damage repair. European Molecular Biology Organization 5

6 Responses to Comments from Reviewer 2 The paper by Zhao et al describes a potential role of MEK1 S-sulfhydration in PARP1 activation and DNA repair. These findings are of interest however the authors would need to provide more convincing experiments to support the suggested model: It would be vital to see whether knocking down the MEK1 and ERK1/2 prevents the NaHSdependent PARP1 activation and leads to a diminished DNA repair. Response: We thank the reviewer for appraising our work and raising many important concerns. In the last several months, we performed new experiments to investigate the change of PARP1 activation and DNA damage repair after knockdown of MEK1 and ERK1/2 by sirnas. The results showed that transfection of HUVECs with MEK1-siRNA or ERK1/2-siRNA significantly decrease the expression of MEK1 or ERK1/2, prevent NaHS-induced PARP1 activation and attenuate NaHS-protected DNA repair (Supplemental Figure 7C, 7D and 7E). All these new data have been incorporated into the revised manuscript. Figure 1B,C. Is PARG activity comparable between wild type and deficient extracts. Response: To address this comment, we measured PARG activity in liver, kidney, and heart tissues from both wildtype and CSE-KO mice. We demonstrated that deficiency of CSE does not affect PARG activity in all these three tissues, suggesting H 2 S-stimulated PAR is not due to altered PARG activity (Supplemental Figure 1C). The paper would have been more convincing if the NaHS-dependent PARP1 activation was demonstrated in at least one more cell line. Response: Thanks again for this suggestion. By using another cell line, human fibroblasts (ATCC, CCL-75 TM ), we confirmed that NaHS stimulates PARP1 activity and attenuates MMSinduced DNA damage (Supplemental Figure 5A and 5B). We also found that NaHS strengthens MEK1 S-sulfhydration in this cell line (Supplemental Figure 8B). We previously showed that knockdown of CSE by sirna significantly stimulates cellular senescence in human fibroblasts (Yang et al. Antioxid. Redox. Signal. 2013, 18: ). C341G mutation could have affected the protein folding, so detecting directly S-sulfhydration of this residue in MEK1 by mass spectrometry would be essential. Response: We fully agree that mass spectrometry is a key technology for detecting proteins with post-translational modification, which can directly identify the given amino acid residue with chemical modification. LC-MS/MS is required to detect S-sulfhydrated cysteine residues. Unfortunately we do not have LC-MS/MS in our laboratory and all other laboratories in our University. We have contacted several commercial companies who can provide LC-MS/MS service. Because these companies never provide this kind of service to detect protein S- sulfhydration. We are now working with one of the companies (Southern Alberta Mass Spectrometry Centre, Canada) to develop and optimize the protocol for detecting protein S- sulhydration, and we are just afraid that we could not provide this MS data in so a short time. To further address this comment, we did more studies to investigate the possibility of other cysteine residues to be S-sulfhydrated. There are six cysteine residues in MEK1 protein, including cysteine 121, 142, 207, 277, 341, and 376. In the original manuscript, we have mutated four of these six cysteine residues (121, 207, 277 and 341), showing only cysteine 341 is responsible for S- sulfhydration. Now we further mutated another two cysteine residues (142 and 376) and performed biotin switch assay to detect MEK1 S-sulfhydration. The results showed that mutation of cysteine 142 or 376 does not affect MEK1 S-sulfhyration (Figure 4D). Suppl. Figure 2E. MMS-induced CSE mrna expression observed here complicates the model and it should be discussed. Is this something that is observed in other cell lines? Response: MMS-induced CSE expression probably represents one compensative response of the cells to protect DNA damage under stress condition. We further investigated the effect of MMS on CSE expression in human fibroblast cells. Interestingly, we also observed that incubation of human fibroblast cells with 1 mm MMS for 2 hour significantly induces CSE protein expression (Supplemental Figure 5C). The presented data seem to suggest that the treatment with NaHS activates PARP1 activity stronger and faster than the MMS treatment (Suppl. Figure 2A). Is this something that the authors European Molecular Biology Organization 6

7 consistently observe? This figure would benefit from including the control with the NaHS treatment only? Response: Thanks for this suggestion. We designed new studies to include the control with NaHS treatment and statistically analyze the change of PARP1 activity by MMS and NaHS. We observed that NaHS consistently activates PARP1 activity as early as at 5 minutes, and the supplement of MMS further strengthens NaHS-stimulated PARP-1 activity. However, MMS alone only activates PARP1 activity at 30 minutes (Supplemental Figure 4A). Figure 1E. Are low PAR levels specific for heart tissue or it is something seen in many tissues? Response: To address this comment, we further measured PARP1 activity in skeletal muscle and spleen tissues from both wildtype and CSE-KO mice. Compared with liver and kidney tissues, CSE expression in skeletal muscle and spleen tissues was pretty weaker (Ishii I et al., Biochem J. 2004, 381: ). Similar to heart tissues, PARP1 activity in both skeletal muscle and spleen tissues from both genotypes was hard to detect, suggesting the level of PAR is correlated with the expression level of CSE (Supplemental Figure 1A and 1B). The westerns blots are lacking size markers. Response: In the revised manuscript, the molecular marker for all the western blotting images was provided. Some of the loading controls in western blot experiments are heavily overloaded which doesn't allow for a proper judgement of the data. Response: We have replaced the heavily loaded controls with clear images. Suppl. Figure 3B is missing a GAPDH positive control. Response: As suggested, a GAPDH positive control for S-sulfhydration was provided (Supplemental Figure 6B). Suppl. Figure 5B. Why a direct detection of S-sulfhydrated MEK1 couldn't be shown? Response: We really appreciate this comment. Now we ordered a MEK1-specific antibody (Cell Signalling Technology), which does not cross-react with MEK2 and other MAP kinases. The results clearly showed that MEK1 can be basically S-sulfhydrated and the supplement of exogenous NaHS further strengthens MEK1 S-sulfhydration in HUVECs (Figure 4B). Responses to Comments from Reviewer 3 In this manuscript, the authors report that hydrogen sulfide (H 2 S) facilitates DNA damage repair via MEK1 S-sulfhydration and PARP1 activation. H 2 S is a gastransmitter and plays important roles in a number of biological processes. Here, the authors demonstrate that H 2 S activates the enzymatic activity of PARP1 and promotes cell viability in response to DNA damage. During this process, the activation of PARP1 is regulated by MEK1 S-sulfhydration. Although the concept of the study is novel, this study is too preliminary to warrant publication. The molecular mechanism of ERKdependent PARP1 activation in the context of DNA damage is unclear. Thus, the major conclusions in this manuscript are questionable. Specific points are listed below: Response: We first thank the reviewer raising many important issues. In the last several months, we did more experiments to address all these critical comments. 1. Endogenous H 2 S should be measured in response to DNA damage. The regulation of endogenous H 2 S and S-sulfhydration by CES should be carefully studied. Response: Our new data showed that H 2 S production is increased when HUVECs are exposed to MMS, a DNA damage inducer, suggesting one compensative response of the cells to protect from DNA damage under stress condition (Supplemental Figure 4C, 4D and 4E). It is still not clear how MMS induces CSE expression and H 2 S production in our system. We provided two lines of evidence demonstrating the S-sulfhydration regulation of MEK1 by endogenous H 2 S through CSE. By using kidney tissues from both wild-type and CSE knockout mice, we observed that wildtype mouse kidney displays more S-sulfhydrated MEK1/2 in European Molecular Biology Organization 7

8 comparison to that from CSE-KO mice (Supplemental Figure 8A). H 2 S production is reduced by 80% in kidney from CSE-KO mice when compared with that from wildtype mice (Bos EM et al. J Am Soc Nephrol 2013, 24: ). We further transfected HEK293 cells with both CSE and MEK1 genes, and found that CSE overexpression induces more MEK1/2 S-sulfhydartion (Supplemental Figure 8C). 2. PAR is a mixture of heterogeneous chains. It is very strange to see a neat band of PAR in SDS- PAGE presented in Fig 1A, B, 3A, B, E, and S3A. Without a molecular marker, it is hard to determine what has been shown in these figures. The authors surely have to show the size markers and more of the gel area, although a 'neat' PARP band might be OK. Response: In the revised manuscript, we have added molecular marker for all western blotting images. In our system, we consistently detected a clear band around 200 kda in HUVECs and human fibroblasts by using an anti-par mouse antibody from Trevigen (#4335-MC-100-AC), which is in accordance with the PAR antibody data sheet. However it is not the case in mouse tissues, where we detected a smear band for PAR, suggesting the different sensitivity of this antibody to detect PAR between cultured cells and tissues, at least in our cases. 3. Huge amount of PAR in both kidney and MEFs has been detected without genotoxic stress. Usually, the half-life of PAR is very short in vivo because PARG, the PAR hydrolase, is very potent in vivo. The authors need to provide a convincing explanation. A very significant PARP activation observed here quite likely happened only after the cells were open in the extraction buffer (not in vivo). To exclude this possibility, the authors should repeat the western blot using RIPA buffer supplemented with PARP inhibitors (that would suppress the artificial PARP activation during cell extract preparation)." Response: To address this comment, we performed two different studies to explore the change of PAR. Firstly, we measured PARG activity in liver, kidney, and heart tissues from both wildtype and CSE-KO mice. The results showed PAR level is positively correlated with PARG activity, and deficiency of CSE does not affect PARG activity in all these three tissues, suggesting H 2 S-stimulated PAR is not due to altered PARG activity (Supplemental Figure 1C). It is not clear why the tissue has higher PAR level even with higher PARG activity. Secondly, as suggested, we repeated western blotting using RIPA buffer with PARP inhibitor DPQ (10 µm) in the presence of protease and phosphatase inhibitor cocktail to detect PAR change. The result showed that the presence of DPQ in the lysis buffer does not affect PAR level in kidney tissues and MEFs from wildtype mice, suggesting higher PARP activation here is unlikely due to the operating procedure. In another way, we treated WT-MEFs with DPQ (10 µm) for 2 hours in vitro and then the cells were collected for lysis and run for a regular western blotting to check the change of PAR. We did observe that DPQ inhibits PAR when compared with the control without DPQ, suggesting that DPQ may inhibit PARP activity in a more complicated condition (Supplemental Figure 2A and 2B). In addition, there are several other papers showing higher basal PARP1 activity with clear and strong PAR band in mammalian cells and animals tissues (Eustermann et al. Nat Struct Mol Biol. 2010, 17: ; Rodríguez-Vargas et al. Cell Res. 2012, 22: ). 4. Page 4, line 2 "in which MEK/ERK pathway plays a critical role in initiating PARP automodification [12, 13]". Reference 12 does not support this conclusion. Response: We are sorry for the overlook. Now we replaced reference 12 with another two related references to support our findings (Cohen-Armon et al. Mol Cell. 2007, 25: ; Kauppinen et al. Proc Natl Acad Sci U S A. 2006, 103: ). 2nd Editorial Decision 13 March 2014 Thank you for your patience while we have reviewed your revised manuscript. Referees 1 and 3 have assessed your revised manuscript and support its publication with no further comments. I am therefore writing with an 'accept in principle' decision, which means that I will be happy to accept your manuscript for publication once a few minor issues regarding journal formatting have been addressed, as follows. European Molecular Biology Organization 8

9 - We can accommodate up to five main figures, and as you have quite a number of supplementary figures, I think it would be good to move one to the main text. Perhaps the most appropriate would be the scheme in supplementary figure 9 (which has a typo). Please provide it in ppt, tif or eps format. - We now encourage the publication of original source data -particularly for electrophoretic gels and blots, but also for graphs- with the aim of making primary data more accessible and transparent to the reader. If you agree, you would need to provide one PDF file per figure that contains the original, uncropped and unprocessed scans of all or key gels used in the figures and an Excel sheet or similar with the data behind the graphs. The files should be labeled with the appropriate figure/panel number, and the gels should have molecular weight markers; further annotation could be useful but is not essential. The source files will be published online with the article as supplementary "Source Data" files and should be uploaded when you submit your final version. If you have any questions regarding this please contact me. Once the above-mentioned issues have been attended to, you will receive an official decision letter from the journal accepting your manuscript for publication in the next available issue of EMBO reports. This letter will also include details of the further steps you need to take for the prompt inclusion of your manuscript in our next available issue. Thank you very much for your contribution to EMBO reports. 2nd Revision - authors' response 25 March 2014 Responses to Comments from Editor We can accommodate up to five main figures, and as you have quite a number of supplementary figures, I think it would be good to move one to the main text. Perhaps the most appropriate would be the scheme in supplementary figure 9 (which has a typo). Please provide it in ppt, tif or eps format. Response: Original supplementary Figure 9 now is moved to the main text as Figure 5. We also provide a tif file for this new Figure 5. We now encourage the publication of original source data -particularly for electrophoretic gels and blots, but also for graphs- with the aim of making primary data more accessible and transparent to the reader. If you agree, you would need to provide one PDF file per figure that contains the original, uncropped and unprocessed scans of all or key gels used in the figures and an Excel sheet or similar with the data behind the graphs. The files should be labeled with the appropriate figure/panel number, and the gels should have molecular weight markers; further annotation could be useful but is not essential. The source files will be published online with the article as supplementary "Source Data" files and should be uploaded when you submit your final version. If you have any questions regarding this please contact me. Response: We now include one separate PDF file per figure for the key figures (Figure 1 to 4 in the main text), which contain the original scans of gels. The files are labelled with the appropriate figure/panel number, and the gels are with molecular weight markers. In our laboratory, we often cut the western blotting membranes to develop with different antibodies with different molecular size. After getting the gel image, the researchers usually crop gels and make them to be smaller to post on the notebook with clear label. From now on, we encourage all researchers not to crop the membranes and gels for all western blotting experiments. European Molecular Biology Organization 9

10 3rd Editorial Decision 31 March 2014 I am very pleased to accept your manuscript for publication in the next available issue of EMBO reports. Thank you for your contribution to our journal. As part of the EMBO publication's Transparent Editorial Process, EMBO reports publishes online a Review Process File to accompany accepted manuscripts. As you are aware, this File will be published in conjunction with your paper and will include the referee reports, your point-by-point response and all pertinent correspondence relating to the manuscript. If you do NOT want this File to be published, please inform the editorial office within 2 days, if you have not done so already, otherwise the File will be published by default [contact: emboreports@embo.org]. If you do opt out, the Review Process File link will point to the following statement: "No Review Process File is available with this article, as the authors have chosen not to make the review process public in this case." Thank you again for your contribution to EMBO reports and congratulations on a successful publication. Please consider us again in the future for your most exciting work. European Molecular Biology Organization 10