The Paf1 complex positively regulates enhancer activity in mouse embryonic stem cells

Using ChIP-seq and functional genomic analyses, the study shows that the Paf1 complex occupies transcriptional enhancers and positively regulates their activity.

Furthermore, both referee #1 and #2 find that the conclusion that NELF1 is displaced by PAF1C is not sufficiently supported by the data provided and this statement should be revised to accurately reflect the experimental evidence (ref#1-point 1, ref#2-point 5). Please also revise the manuscript text to reflect the concerns, as well as adding the requested additional information on experimental procedures, data analysis and statistics for the following points: ref#1 -points 2, 3, 6, 7, 8, ref#2points 2, 4, 5, 10 and ref#3 points 2, 3, 4. Moreover, the data presentation should be further expanded and a metagene plot (ref#1-point 1) and heatmaps (ref#2-point 6, 7, 9) provided as indicated. In addition, please also ensure that the current study is carefully discussed in the context of previous work (ref#2 -point 4, ref#-point 1).
If you find that you will be able to adequately address these issues, we would be happy to consider the study further for publication in Life Science Alliance. Therefore, we would now like to invite you to prepare and submit a revised version.
We realize that lab work worldwide is currently affected by the COVID-19/SARS-CoV-2 pandemic and that an experimental revision may be delayed or temporarily impossible. The typical timeframe for revisions is three months, but we can extend the revision time when needed, and we have extended our 'scooping protection policy' to cover the period required for a full revision. However, it is nonetheless important to clarify any questions and concerns at this stage and we encourage you to discuss a revision plan and any potential issues you may foresee as soon as possible.
To upload the revised version of your manuscript, please log in to your account: https://lsa.msubmit.net/cgi-bin/main.plex You will be guided to complete the submission of your revised manuscript and to fill in all necessary information. Please get in touch in case you do not know or remember your login name.
While you are revising your manuscript, please also attend to the below editorial points to help expedite the publication of your manuscript. Please direct any editorial questions to the journal office.
Please note that papers are generally considered through only one revision cycle, so strong support from the referees on the revised version is needed for acceptance.
When submitting the revision, please include a letter addressing the reviewers' comments point by point.
We hope that the comments below will prove constructive as your work progresses.
Thank you for this interesting contribution to Life Science Alliance. We are looking forward to receiving your revised manuscript. --An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--High-resolution figure, supplementary figure and video files uploaded as individual files: See our detailed guidelines for preparing your production-ready images, http://www.life-sciencealliance.org/authors --Summary blurb (enter in submission system): A short text summarizing in a single sentence the study (max. 200 characters including spaces). This text is used in conjunction with the titles of papers, hence should be informative and complementary to the title and running title. It should describe the context and significance of the findings for a general readership; it should be written in the present tense and refer to the work in the third person. Author names should not be mentioned.

B. MANUSCRIPT ORGANIZATION AND FORMATTING:
Full guidelines are available on our Instructions for Authors page, http://www.life-sciencealliance.org/authors We encourage our authors to provide original source data, particularly uncropped/-processed electrophoretic blots and spreadsheets for the main figures of the manuscript. If you would like to add source data, we would welcome one PDF/Excel-file per figure for this information. These files will be linked online as supplementary "Source Data" files. ***IMPORTANT: It is Life Science Alliance policy that if requested, original data images must be made available. Failure to provide original images upon request will result in unavoidable delays in publication. Please ensure that you have access to all original microscopy and blot data images before submitting your revision.*** Reviewer #1 (Comments to the Authors (Required)): Ding et al. examined the genome-wide occupancy of Ctr9, a subunit of the elongation factor Paf1C, and NELFA, a subunit of the negative elongation factor NELF, in mouse embryonic stem cells. To this end they tagged endogenous Ctr9 and NELFA proteins with GFP utilizing genome editing with CRISPR/Cas9. The authors show that Ctr9 is associated with protein coding genes and the occupancy correlates with expression levels. They next examined Ctr9 occupancy at enhancers and super enhancers, and show that Ctr9-occupancy is linked to enhancer activity in a plasmid reporter assay. The authors knock down Ctr9 with siRNA to show that it is needed for enhancer activity. The knockdown reduced the activity of Ctr9-bound enhancers in the reporter assay and was further supported by two examined endogenous enhancers, where Ctr9 knockdown reduced RNAPII-Ser2P occupancy at Oct4-and Tbx3-enhancers and transcription at the corresponding genes. The authors indicate that Ctr9-occuapncy could be a marker of active enhancers and substantiate their claim, by performing Ctr9 ChIP-seq for NIH3T3 cell line and the enhancer reporter assay with a subset of enhancers in the same cell line. Based on these results the authors propose that Paf1C occupancy could be utilized to classify active enhancers and that Paf1C is required for enhancer function. More specifically, that Paf1C activity at pluripotency enhancers is required for maintenance of stem cell self-renewal.
The authors make several statements based on weak or lacking evidence. In particular, in the abstract the authors claim that Ctr9 regulates gene expression by regulating the activity of enhancers, specifically pluripotency enhancers crucial for maintenance of mESC self-renewal. This is unsubstantiated and no evidence for specificity is shown (for example evidence that other genes are not affected). Also, the authors claim that they have shown "that the Paf1C positively regulates RNAPII pause-release at both protein-coding genes and at enhancers in mouse embryonic stem cells (mESC)", but provide no experimental quantification or analyses of pause-release. The Ctr9 ChIP-seq data analyses appear to be oversimplified and overinterpreted. Furthermore, the authors provide insufficient and questionable evidence to support their claim that Paf1C displaces NELF for active transcription of protein coding genes. Additionally, the authors tend to show one example and make global or wide-sweeping statements. Lastly, the methods section is lacking critical information and there is no indication that the ChIP-seq experiments were performed with replicates. Therefore I cannot recommend this study for publication. I hope the comments below help the authors to obtain an improved manuscript.
1. "Paf1C displaces NELF for active transcription of protein coding genes" -the statement is based on overinterpretation of ChIP-seq beyond the resolution of the method. 1E is a Gaussian kernel density distribution of ChIP-seq peaks. This is misleading, firstly, the resolution of ChIP-seq is approximately the size of the DNA fragments (200-500 bp), second, the peaks can be broad or narrow peaks -the midpoints do not represent the actual occupancy. A metagene profile for the actual coverage should be shown instead.
2. Figure 3B. It is unclear how was this distance measured -midpoint to midpoint or end to start of the peak range? In the heatmap there appears to be more Ctr9 bound in the window shown in the heatmaps in 3A for SE that for TE and also more overlapping intensities, which contradicts the information shown in 3B.
3. "In addition to protein coding genes, we found that RNAPII Ser5p and Ser2p were also significantly enriched on SEs, with 83% and 63% occupancy, respectively, suggesting that a majority of SEs are transcriptional enhancers (Dataset EV4, Fig 4A, 4B)." How did the authors get these numbers, also no data to support this in the referred figure panels. SEs are generally considered to be transcriptional enhancers.
5. "Strikingly, SEs with a strong decrease of RNAPII Ser2p after Ctr9 knockdown lost their enhancer activity dramatically, and this effect correlated with a strong decrease in transcription of their associated target genes, such as Oct4 and Tbx3 (Fig 5C-5F; Fig EV5A-5D), suggesting that Paf1C regulates gene expression by modulating enhancer activity." For the authors to make this claim, they would need to show that this is specific and not a global gene expression down-regulation upon Ctr9 knockdown and that this is not only valid for Oct4 and Tbx3 SE.
6. "... we have performed a panel of experiments to determine DNA-binding of relevant factors, including NELFA, RNAPII Ser5p, RNAPII Ser2P and Ctr9 in mESCs." The authors performed ChIPseq, this experiment does not prove direct binding to DNA. 7. "Our analyses are consistent with a dynamic transition of NELF, RNAPII Ser5p and Paf1C within a 150-bp region downstream of the TSS, indicating a coordinated action of events of these factors for active transcription in mESCs, where NELF is substituted by Paf1C to release RNAPII from promoter proximal pausing." The supposed distances between NELFA and Ctr9 peaks is 354 and 465 for TEs and SEs, respectively. This contradicts the concept that this transition would occur within a 150-bp window. 8. Incorrect statements: "...exchange of RNAPII-Ser5p for Paf1C...", "NELFA is an initiation factor"; "... Paf1C binding can be used in combination with histone marks to predict active enhancers in primary cells." -NIH3T3 are not primary cells.
Reviewer #2 (Comments to the Authors (Required)): In this manuscript, the authors investigate the connection between the PAF complex and enhancer function in mouse ES cells. Using ChIP-seq experiments on PAF subunit Ctr9 and reporter assays, the authors make several important discoveries. First, they show that Ctr9 occupies many but not all enhancers in a cell-type specific manner. Second, they show that Ctr9 occupancy correlates with transcription at the enhancer and the presence of eRNAs. Third, they show that Ctr9-occupied enhancers, but not those lacking Ctr9, activate transcription in their reporter assay. Further, depletion of Ctr9 reduces transactivation by these enhancers and Pol II Ser2p levels at the enhancer, suggesting PAF controls transcription at the enhancer. Finally, they provide evidence that Ctr9 occupancy, together with H3K27ac, is a predictor of enhancer activity. What is left unanswered by this work is what determines whether an enhancer is occupied by Ctr9 (and hence active) versus unoccupied (and hence inactive). Nonetheless, the contributions will be of interest to many in the enhancer field, as there is considerable debate about the role of PAF in enhancer function and Pol II pausing. However, in its current form, there are weaknesses in experimental design, data analysis and presentation that need to be addressed.
Major comments: 1. The paper requires a more rigorous description of reproducibility and controls. For the GFPtagging of Ctr9 and NELFA, how many independently tagged mESC and NIH3T3 cell lines were generated and tested? Given the possibility of CRISPR-generated indirect and off-target effects, a test of more than one independently derived line would substantiate the claims made in the paper regarding Ctr9 and NELF occupancy. Does the GFP tag on either protein affect occupancy? The authors should consider confirming their results with unaltered mESCs and antibodies against endogenous Ctr9 and NELF to show that the GFP tags are not having an influence and that CRISPR off-target effects are not impacting the data. 2. Regarding the ChIP-seq experiments, the methods do not describe a spike-in control or replicate numbers. The methods for normalizing ChIP-seq signals and data reproducibility (number of replicates, correlations between replicates) need to be stated. 3. The methods do not adequately describe how the GFP-tagged cell lines were confirmed or characterized, other than showing that the tagged proteins were expressed and nuclear. 4. Figure 1A,B. These data are consistent with substantial existing literature showing a strong correlation between gene expression (and Pol II occupancy) and PAF enrichment. Although the data were presented, the authors did not address the strong Ctr9 ChIP-seq signal downstream of the TTS. This should be discussed in the context of work by Yang et al. (PLOS Genetics 2016) who noted this downstream enrichment in mouse myoblasts and showed an effect of PAF on polyA site usage. 5. Figure 1D,E,F. The authors report the separated ChIP peaks for NELFA, Pol II Ser5p, and Ctr9. The authors arrive at a conclusion of factor exchange that is well supported by earlier studies. The data are consistent with recent structures from the Cramer lab, which showed that binding of NELF and PAF to Pol II is mutually exclusive. The Cramer lab also reported in 2010 that PAF is enriched downstream of Ser5p. While the authors' ChIP-seq results provide nice support for these results in mESCs, they do not address mechanism and are largely confirmatory. The title of this figure, "Paf1C displaces the NELF complex during promoter proximal pause release of target genes", is overstated. In addition, from the box plot in Figure 1F, it seems a significant number of genes show an overlap between the position of Ctr9 occupancy and the position of NELFA. An explanation for this observation is needed. 6. Figure 2. The occupancy data support previous work from the Shilatifard group, using HCT116 cells, showing PAF localization at enhancers. However, the advance here is the focus on ESCs, pluripotency genes and super enhancers. The demonstration that some super enhancers are unoccupied by Ctr9 and these are deficient in enhancer activity in a luciferase assay is an important result, as it counters current literature from the Shilatifard lab arguing that PAF negatively regulates enhancer function. My main suggestion to improve this figure and the paper in general is to provide genome-wide heatmaps. The paper is heavily reliant on metaplots and browser tracks, either of which can provide a skewed view of the data. What is the occupancy pattern of Ctr9, H3K27ac, and H3K4me1 at all proposed enhancers? 7. Figure 3. The data show distinct occupancy patterns for NELFA and Ctr9 at enhancers (both super and typical). Why are so few examples shown? The authors mention looking at 166 super enhancers but fewer appear to be shown in the heatmap. Based on the heatmap, the super enhancer in Panel D seems like an exception with respect to the patterns of NELFA and Ctr9 occupancy, making it unclear why it is emphasized or how frequent this pattern was observed. 8. Figure EV3. A loading control needs to be added to the western blot. 9. Figure 4 makes the point that Ctr9 occupancy and transcription of enhancers, as measured by Pol II phosphorylation or eRNA accumulation (from published GRO-seq data), are positively correlated. This is another example of where heatmaps would substantiate the more limited information provided by browser tracks and Venn diagrams. Also, given the binding of PAF to Pol II, the result of correlated PAF occupancy and active transcription is not especially surprising, though showing this result at enhancers is a notable advance. 10. Figure 5. The data show that Ctr9-occupied enhancers are more active than non-occupied enhancers in a reporter assay and enhancer activity is partially Ctr9-dependent. The data also show that depletion of Ctr9 reduces Pol II Ser2p, a measure of transcription at the enhancer. These are among the most significant advances in the paper. However, the authors need to be more accurate in the description of the data. Transactivation by the Ctr9-occupied enhancers is reduced 2-fold or more for only 4/10 of the tested enhancers upon Ctr9 depletion. The authors should soften their conclusions concerning the strength of the effects. In addition, for the reduction in Ser2p levels, the authors should show the generality of the data and not rely solely on browser tracks. Finally, while the reduction in genic transcription upon Ctr9 depletion correlates with enhancer inactivation, it is also possible that loss of PAF is affecting transcription through the gene more directly. This should be considered. 11. Figure 6 tests the hypothesis that H3K27ac and Ctr9 occupancy, together, predict enhancer activity better than H3K27ac. This is potentially an important discovery; however, the experiment is not acceptable as shown. There is no evidence of repetition or statistical significance.
Minor comments: 1. Last sentence on p. 6: "Consistent with its role as an elongation factor, Ctr9 binding exhibited cell-type specificity". The two thoughts in this sentence are not logically connected. Consider modifying to state: "Consistent with its role as an elongation factor associated with active gene transcription, ....". 2. Figure EV3. The term "LAP-tagging" is confusing. 3. The Chen et al reference is incomplete. 4. Sentence on p. 10 requires clarification. "Interestingly, the overlap of Ctr9 occupancy with TEs carrying H3K27ac modification (8%) was significantly lower than that in SEs (68%), suggesting that Ctr9 predominantly associates with active enhancers (Fig 2C, 2D, Dataset EV3)." How do the authors arrive at the main conclusion (ie. Ctr9 predominantly associates with active enhancers) from the lower occupancy overlap between Ctr9 and H3K27ac at TEs compared to SEs? 5. P. 12. The phrase "suggesting that a majority of SEs are transcriptional enhancers" is confusing. Why are they called super enhancers if they are not transcriptional enhancers? The sentence needs to be written more clearly. 6. Figure 6 title. Rephrase to avoid stating Ctr9 DNA-binding. The factor precipitated need not bind DNA directly.
Reviewer #3 (Comments to the Authors (Required)): In the study "The Paf1 complex positively regulates enhancer activity in mouse embryonic stem cells", Ding and colleagues reported functions of Paf1C in regulating enhancer activity, in particular in super enhancers. These results suggest the potential power of Paf1c in mECS self-renewal. This study is important to enlarge our knowledge on enhancer activity. This reviewer has several comments which may be addressed.
1. Previous studies (e.g., Hou et, al., PNAS 2019, Chen et.al., Genes & Development 2009) reported the importance of Paf1C in modulating Pol II elongation. This review suggests that comparison and discussion about previous studies is needed. 2. This review didn't observe characterization of super enhancer in each cell line. Could author describe the method they used in super enhancer definition? Some cases, e.g., figure 2e, figure 3cd, are unique enhancers in the region, which look like typical enhancer. 3. Author claimed that H3K27ac plus Ctr9 binding could improve prediction of activated enhancer. This reviewer agrees with this point by some way. While as author introduce in the manuscript, transcript enhancer (also known as enhancer RNA) could be a golden standard for active enhancer. Could author consider this important signal in definition of active enhancer? 4. H3K27ac is the marker for both active enhancer and promoter. If the author focus on enhancer only, other histone markers, e.g., H3K4me1, p300, are needed.

Reviewer #1 (Comments to the Authors (Required)):
Ding et al. examined the genome-wide occupancy of Ctr9, a subunit of the elongation factor Paf1C, and NELFA, a subunit of the negative elongation factor NELF, in mouse embryonic stem cells. To this end they tagged endogenous Ctr9 and NELFA proteins with GFP utilizing genome editing with CRISPR/Cas9. The authors show that Ctr9 is associated with protein coding genes and the occupancy correlates with expression levels. They next examined Ctr9 occupancy at enhancers and super enhancers, and show that Ctr9-occupancy is linked to enhancer activity in a plasmid reporter assay. The authors knock down Ctr9 with siRNA to show that it is needed for enhancer activity. The knockdown reduced the activity of Ctr9-bound enhancers in the reporter assay and was further supported by two examined endogenous enhancers, where Ctr9 knockdown reduced RNAPII-Ser2P occupancy at Oct4-and Tbx3-enhancers and transcription at the corresponding genes. The authors indicate that Ctr9-occuapncy could be a marker of active enhancers and substantiate their claim, by performing Ctr9 ChIP-seq for NIH3T3 cell line and the enhancer reporter assay with a subset of enhancers in the same cell line. Based on these results the authors propose that Paf1C occupancy could be utilized to classify active enhancers and that Paf1C is required for enhancer function. More specifically, that Paf1C activity at pluripotency enhancers is required for maintenance of stem cell self-renewal. Response: We thank reviewer#1 for the constructive criticism and suggestions that we have addressed as indicated for each point.
The authors make several statements based on weak or lacking evidence. In particular, in the abstract the authors claim that Ctr9 regulates gene expression by regulating the activity of enhancers, specifically pluripotency enhancers crucial for maintenance of mESC self-renewal. This is unsubstantiated and no evidence for specificity is shown (for example evidence that other genes are not affected). Also, the authors claim that they have shown "that the Paf1C positively regulates RNAPII pause-release at both protein-coding genes and at enhancers in mouse embryonic stem cells (mESC)", but provide no experimental quantification or analyses of pause-release Response: We have further strengthened the claim that Ctr9 regulates gene expression by regulating the activity of enhancers. We had already shown specificity for Ctr9 bound and non-bound enhancers in Fig. 5A and B. We now add an example of a gene not affected (e.g. Zfp638, Fig. EV5E-EV5H). Furthermore, we have added a global analysis of transcriptional changes of genes regulated by ESC SEs, TEs and other genes (Fig. 5C, 5D). These data substantiate the notion that Ctr9 regulates gene expression by regulating the activity of enhancers.
5. "Strikingly, SEs with a strong decrease of RNAPII Ser2p after Ctr9 knockdown lost their enhancer activity dramatically, and this effect correlated with a strong decrease in transcription of their associated target genes, such as Oct4 and Tbx3 (Fig 5C-5F; Fig EV5A-5D), suggesting that Paf1C regulates gene expression by modulating enhancer activity." For the authors to make this claim, they would need to show that this is specific and not a global gene expression down-regulation upon Ctr9 knockdown and that this is not only valid for Oct4 and Tbx3 SE.
Response: To address this point, we have analyzed the expression of super enhancer and typical enhancer target genes after Ctr9 knockdown. The expression of the majority of SEs target genes decreased after Ctr9 knockdown. In contrast, the expression of genes that are not associated with enhancers did not show this pattern. The new analysis is shown in the revised manuscript ( Figure  5C, 5D). 7. "Our analyses are consistent with a dynamic transition of NELF, RNAPII Ser5p and Paf1C within a 150-bp region downstream of the TSS, indicating a coordinated action of events of these factors for active transcription in mESCs, where NELF is substituted by Paf1C to release RNAPII from promoter proximal pausing." The supposed distances between NELFA and Ctr9 peaks is 354 and 465 for TEs and SEs, respectively. This contradicts the concept that this transition would occur within a 150-bp window.
Response: As mentioned in the text already, different mechanisms may account for the different distances between NEFLA and Ctr9 on enhancers and on protein coding genes. The reviewer might have missed this (page 11, bottom of original submission).

1: eRNA transcription is modulated by a panel of specific transcription factors. For instance, Mll3
and Mll4 are essential for RNAP II accumulation at enhancers to activate transcription. Loss of Mll3/4 from enhancers decrease the binding rate of RNAP II and thus eRNA production (Dorighi et al., 2017). Another study has suggested that condensin could modulate the binding of co-repressors or co-activators (like p300 and RIP140) by the recruitment of an E3 ubiquitin ligase to regulate eRNA transcription (Li et al., 2015).

2: Some eRNAs are bidirectionally transcribed, in contrast to the mostly unidirectional transcription of mRNA (Mikhaylichenko et al., 2018).
Therefore, we stand to our statement that the larger distance suggests that enhancer (e)RNA transcription may be orchestrated in a different way than protein coding genes. 8. Incorrect statements: "...exchange of RNAPII-Ser5p for Paf1C...", "NELFA is an initiation factor"; "... Paf1C binding can be used in combination with histone marks to predict active enhancers in primary cells." -NIH3T3 are not primary cells.

Reviewer #2 (Comments to the Authors (Required)):
In this manuscript, the authors investigate the connection between the PAF complex and enhancer function in mouse ES cells. Using ChIP-seq experiments on PAF subunit Ctr9 and reporter assays, the authors make several important discoveries. First, they show that Ctr9 occupies many but not all enhancers in a cell-type specific manner. Second, they show that Ctr9 occupancy correlates with transcription at the enhancer and the presence of eRNAs. Third, they show that Ctr9-occupied enhancers, but not those lacking Ctr9, activate transcription in their reporter assay. Further, depletion of Ctr9 reduces transactivation by these enhancers and Pol II Ser2p levels at the enhancer, suggesting PAF controls transcription at the enhancer. Finally, they provide evidence that Ctr9 occupancy, together with H3K27ac, is a predictor of enhancer activity. What is left unanswered by this work is what determines whether an enhancer is occupied by Ctr9 (and hence active) versus unoccupied (and hence inactive). Nonetheless, the contributions will be of interest to many in the enhancer field, as there is considerable debate about the role of PAF in enhancer function and Pol II pausing. However, in its current form, there are weaknesses in experimental design, data analysis and presentation that need to be addressed.

Response: We appreciate the constructive evaluation of our work and the thoughtful comments by reviewer #2. We have addressed his/her points as outlined below:
Major comments: 1. The paper requires a more rigorous description of reproducibility and controls. For the GFP-tagging of Ctr9 and NELFA, how many independently tagged mESC and NIH3T3 cell lines were generated and tested? Given the possibility of CRISPR-generated indirect and off-target effects, a test of more than one independently derived line would substantiate the claims made in the paper regarding Ctr9 and NELF occupancy.

Response: To generate GFP-tagged Ctr9 and NELFA cells we used a pool of GFP expressing cells, instead of a single cell line to perform ChIP experiments for Ctr9 and NELF. The pooling approach eliminates clonal variation which is frequently observed for single cell lines. We have added a statement in the methods section to emphasize this point (page 19).
To validate GFP tagging, we sequenced the flanking regions of GFP, and confirmed correct integration of GFP into the mESC genome as designed. GFP tagging was further confirmed on protein level by Western blot hybridization. GFP fusion proteins of expected molecular weight were detected by the anti-GFP antibody. When treating the cells with esiRNAs targeting Ctr9 or NELFA ( Figure EV1, EV3), specific reduction of GFP-tagged proteins was revealed, confirming correct and specific GFP tagging.
Does the GFP tag on either protein affect occupancy? The authors should consider confirming their results with unaltered mESCs and antibodies against endogenous Ctr9 and NELF to show that the GFP tags are not having an influence and that CRISPR off-target effects are not impacting the data.
Response: We thank reviewer #2 for this comment. To address this point, we have tried commercial antibodies targeting Ctr9 and NELFA, but the results were unfortunately not satisfactory, probably because the antibodies were not of sufficient quality. We would like to point out that the GFP antibody we used is of very high quality and that we have confirmed the GFP-tagging approach in numerous publications, many of which at very high throughput (e.g.  , 2008). We are therefore confident that the presented data is of high quality.
2. Regarding the ChIP-seq experiments, the methods do not describe a spike-in control or replicate numbers. The methods for normalizing ChIP-seq signals and data reproducibility (number of replicates, correlations between replicates) need to be stated. peak calling software macs2 (ver. 2.2.6). The method for normalizing ChIP-seq signals is now shown in the revised manuscript (page 20).

Response: The ChIP-Seq experiments were not done in replicates and we did not use any spike-in controls. However, we again refer to the validated quality of the GFP-tagging approach (see previous point). The number of sequenced fragments aligning to each base was normalized to the total number of reads in a sample. This normalization step was performed automatically by the
3. The methods do not adequately describe how the GFP-tagged cell lines were confirmed or characterized, other than showing that the tagged proteins were expressed and nuclear.
Response: Correct GFP-tagging was confirmed by sequencing of the GFP-tagging region from genomic DNA. In addition, the specific depletion of tagged protein is now shown by using specific esiRNAs to deplete the protein (Figure EV1, EV3). Figure 1A,B. These data are consistent with substantial existing literature showing a strong correlation between gene expression (and Pol II occupancy) and PAF enrichment. Although the data were presented, the authors did not address the strong Ctr9 ChIP-seq signal downstream of the TTS. This should be discussed in the context of work by Yang et al. (PLOS Genetics 2016) who noted this downstream enrichment in mouse myoblasts and showed an effect of PAF on polyA site usage.

4.
Response: We thank the reviewer for this suggestion. We have extended this aspect in the revised manuscript (page 6) and we reference the suggested manuscript. Figure 1D,E,F. The authors report the separated ChIP peaks for NELFA, Pol II Ser5p, and Ctr9. The authors arrive at a conclusion of factor exchange that is well supported by earlier studies. The data are consistent with recent structures from the Cramer lab, which showed that binding of NELF and PAF to Pol II is mutually exclusive. The Cramer lab also reported in 2010 that PAF is enriched downstream of Ser5p. While the authors' ChIP-seq results provide nice support for these results in mESCs, they do not address mechanism and are largely confirmatory. The title of this figure, "Paf1C displaces the NELF complex during promoter proximal pause release of target genes", is overstated. In addition, from the box plot in Figure 1F, it seems a significant number of genes show an overlap between the position of Ctr9 occupancy and the position of NELFA. An explanation for this observation is needed.

5.
Response: We would like to thank the reviewer for the thoughtful comments. We have rephrased the title of this figure. Due to a broader range of Ctr9 occupancy, a precise estimation of peak summit positions is quite challenging. We would attribute this large spread of distances to a poor accuracy of peak summit positioning.
6. Figure 2. The occupancy data support previous work from the Shilatifard group,using HCT116 cells, showing PAF localization at enhancers. However, the advance here is the focus on ESCs, pluripotency genes and super enhancers. The demonstration that some super enhancers are unoccupied by Ctr9 and these are deficient in enhancer activity in a luciferase assay is an important result, as it counters current literature from the Shilatifard lab arguing that PAF negatively regulates enhancer function. My main suggestion to improve this figure and the paper in general is to provide genome-wide heatmaps. The paper is heavily reliant on metaplots and browser tracks, either of which can provide a skewed view of the data. What is the occupancy pattern of Ctr9, H3K27ac, and H3K4me1 at all proposed enhancers?
Response: We have re-analyzed Ctr9, H3K27ac, and H3K4me1 at all studied enhancers, and prepared a heatmap. This analysis is now shown in the revised manuscript ( Figure 3A,Dataset EV3).
7. Figure 3. The data show distinct occupancy patterns for NELFA and Ctr9 at enhancers (both super and typical). Why are so few examples shown? The authors mention looking at 166 super enhancers but fewer appear to be shown in the heatmap. Based on the heatmap, the super enhancer in Panel D seems like an exception with respect to the patterns of NELFA and Ctr9 occupancy, making it unclear why it is emphasized or how frequent this pattern was observed.
Response: We thank the reviewer for the suggestion. We have increased the number of enhancers that we show in Figure 3A. To emphasize the relationships between the binding patterns, we decided to show enhancers at which we identified strong binding sites for both NELFA and Ctr9.
8. Figure EV3. A loading control needs to be added to the western blot.
Response: We have repeated the Western blot hybridization in NELFA-GFP cell line. Specific depletion of NELFA-GFP fusion protein was detected after RNAi. Tubulin was used as loading control. The Western blot hybridization is now shown in the modified version of the manuscript ( Figure EV3B). Figure 4 makes the point that Ctr9 occupancy and transcription of enhancers, as measured by Pol II phosphorylation or eRNA accumulation (from published GRO-seq data), are positively correlated. This is another example of where heatmaps would substantiate the more limited information provided by browser tracks and Venn diagrams. Also, given the binding of PAF to Pol II, the result of correlated PAF occupancy and active transcription is not especially surprising, though showing this result at enhancers is a notable advance.

9.
Response: We thanks review #2 for this suggestion. We have re-analyzed RNAPII Ser2P, Ser5P, NELFA, Ctr9 H3K27ac and H3K4me1 occupancy at super enhancers and typical enhancers. The heatmap and analysis are now shown in the revised manuscript ( Figure 3A and Dataset EV4).
10. Figure 5. The data show that Ctr9-occupied enhancers are more active than non-occupied enhancers in a reporter assay and enhancer activity is partially Ctr9-dependent. The data also show that depletion of Ctr9 reduces Pol II Ser2p, a measure of transcription at the enhancer. These are among the most significant advances in the paper. However, the authors need to be more accurate in the description of the data. Transactivation by the Ctr9-occupied enhancers is reduced 2-fold or more for only 4/10 of the tested enhancers upon Ctr9 depletion. The authors should soften their conclusions concerning the strength of the effects. In addition, for the reduction in Ser2p levels, the authors should show the generality of the data and not rely solely on browser tracks. Finally, while the reduction in genic transcription upon Ctr9 depletion correlates with enhancer inactivation, it is also possible that loss of PAF is affecting transcription through the gene more directly. This should be considered.
Response: We have rephrased and soften our conclusion in the revised manuscript (page 13). Furthermore, we checked Ser2p levels on SEs and gene body after Ctr9 knockdown, and observed a predominant decrease of Ser2p on enhancers, which suggests Paf1C regulate transcription through modulation of enhancer activity ( Figure 5C, 5D, 5E). Genome browser tracks of Ser2p over SEs and gene body are shown in the modified manuscript ( Figure 5G, Figure EV5B). Figure 6 tests the hypothesis that H3K27ac and Ctr9 occupancy, together, predict enhancer activity better than H3K27ac. This is potentially an important discovery; however, the experiment is not acceptable as shown. There is no evidence of repetition or statistical significance. 1st Revision -Editorial Decision Thank you for submitting your revised manuscript entitled "The Paf1 complex positively regulates enhancer activity in mouse embryonic stem cells". We would be happy to publish your paper in Life Science Alliance pending final revisions necessary to meet the reviewer 2's concerns and our formatting guidelines.

11.
Since repeating the ChIP-Seq data was not made a requirement in the first round of review, we will editorially overrule that request for publication in LSA. However, we do suggest you clarify that the ChIP-Seq data provided are from a single run -both in the manuscript text and in the figure legend. All other concerns raised by Reviewer 2 should be addressed in the revision.
Along with the Reviewer 2's points 2-6 and the points listed at the end of this email, please also attend to the following: -please consult our Manuscript Preparation Guidelines https://www.life-sciencealliance.org/manuscript-prep and put your manuscript sections in the correct order -please use the [10 author names, et al.] format in your references (i.e. limit the author names to the first 10) -please add the supplementary figure legends to the main manuscript text -LSA allows supplementary figures, but not EV Figures; please update your callouts for the Supplementary Figures in the manuscript Fig EV1A = Fig S1A) -please add a callout in your main manuscript text for Figure S1B -please add a scale bar for Figure S1B -please rename the datasets as supplementary tables -both in their titles and in their callouts in the manuscript text -please rename the 'Experimental Procedures' section as 'Materials and Methods' -please provide source data (original unprocessed gels) for Figure S3B If you are planning a press release on your work, please inform us immediately to allow informing our production team and scheduling a release date.
To upload the final version of your manuscript, please log in to your account: https://lsa.msubmit.net/cgi-bin/main.plex You will be guided to complete the submission of your revised manuscript and to fill in all necessary information. Please get in touch in case you do not know or remember your login name.
To avoid unnecessary delays in the acceptance and publication of your paper, please read the following information carefully.

A. FINAL FILES:
These items are required for acceptance.
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--High-resolution figure, supplementary figure and video files uploaded as individual files: See our detailed guidelines for preparing your production-ready images, https://www.life-sciencealliance.org/authors --Summary blurb (enter in submission system): A short text summarizing in a single sentence the study (max. 200 characters including spaces). This text is used in conjunction with the titles of papers, hence should be informative and complementary to the title. It should describe the context and significance of the findings for a general readership; it should be written in the present tense and refer to the work in the third person. Author names should not be mentioned.

B. MANUSCRIPT ORGANIZATION AND FORMATTING:
Full guidelines are available on our Instructions for Authors page, https://www.life-sciencealliance.org/authors We encourage our authors to provide original source data, particularly uncropped/-processed electrophoretic blots and spreadsheets for the main figures of the manuscript. If you would like to add source data, we would welcome one PDF/Excel-file per figure for this information. These files will be linked online as supplementary "Source Data" files. **Submission of a paper that does not conform to Life Science Alliance guidelines will delay the acceptance of your manuscript.** **It is Life Science Alliance policy that if requested, original data images must be made available to the editors. Failure to provide original images upon request will result in unavoidable delays in publication. Please ensure that you have access to all original data images prior to final submission.** **The license to publish form must be signed before your manuscript can be sent to production. A link to the electronic license to publish form will be sent to the corresponding author only. Please take a moment to check your funder requirements.** **Reviews, decision letters, and point-by-point responses associated with peer-review at Life Science Alliance will be published online, alongside the manuscript. If you do want to opt out of having the reviewer reports and your point-by-point responses displayed, please let us know immediately.** Thank you for your attention to these final processing requirements. Please revise and format the manuscript and upload materials within 7 days.
Thank you for this interesting contribution, we look forward to publishing your paper in Life Science