P-body proteins regulate transcriptional rewiring to promote DNA replication stress resistance

mRNA-processing (P-) bodies are cytoplasmic granules that form in eukaryotic cells in response to numerous stresses to serve as sites of degradation and storage of mRNAs. Functional P-bodies are critical for the DNA replication stress response in yeast, yet the repertoire of P-body targets and the mechanisms by which P-bodies promote replication stress resistance are unknown. In this study we identify the complete complement of mRNA targets of P-bodies during replication stress induced by hydroxyurea treatment. The key P-body protein Lsm1 controls the abundance of HHT1, ACF4, ARL3, TMA16, RRS1 and YOX1 mRNAs to prevent their toxic accumulation during replication stress. Accumulation of YOX1 mRNA causes aberrant downregulation of a network of genes critical for DNA replication stress resistance and leads to toxic acetaldehyde accumulation. Our data reveal the scope and the targets of regulation by P-body proteins during the DNA replication stress response.

This manuscript examines the functions of P-bodies in regulating cellular responses to replication stress. P-bodies are sites of mRNA decapping to decrease mRNA abundance. Stresses including replication stress induce P-body formation and function and P-body proteins are important for cell viability in response to these stresses. The authors utilized RNA sequencing and genetics to identify mRNAs that may be regulated in a P-body dependent manner to yield resistance to replication stress. Their analysis identified Yox1, a transcriptional repressor as one of several candidates and they validated Yox1 mRNA regulation by P-body processing as important for replication stress responses. Furthermore, they identified two gene targets of Yox1 regulation that contribute to these responses.
Overall, I found the data in the manuscript to be compelling and the conclusions interesting. While it is not surprising that Yox1 is involved in controlling gene expression in response to replication stress, the control of Yox1 by P-body dependent processing is novel and interesting. The only thing that I would have liked the authors to do was to demonstrate unequivocally that ALD6 and ICS2 are really direct targets of Yox1 important for stress responses. For example, it would be helpful to demonstrate direct binding of Yox1 protein to the putative Yox1 binding sites in the promoters of these genes. Also, monitoring the effects of deletion of these binding sites on gene expression and cell viability would be useful.

Reviewer #2 (Remarks to the Author)
This manuscript addresses the role of mRNA decay factors in the control of gene expression during DNA replication stress. The main point of the work is that Lsm1 and Pat1 play roles in shaping the transcriptome during HU stress, and that by keeping specific mRNAs expressed at lower levels, they contribute to successful DNA damage response. This conclusion is supported by, i) RNA-Seq of Wt and lsm1∆ strains identifying mRNAs that are overexpressed either normally or in HU, ii) showing that deletion of 6 of these overexpressed mRNAs can partially suppress the HU sensitivity of lsm1∆ or pat1∆ strains, iii) focusing on YOX1, a transcriptional repressor, as key regulator whose over-expression in pat1∆ and lsm1∆ strains leads to HU sensitivity. They then go on and show the Yox1 mRNA can be localized to P-bodies, and that the Yox1 protein goes up a little bit in lsm1∆ with or without HU and shows increased nuclear localization. Finally, they show that overexpression of Yox1 via galactose induction, or deletion of the Lsm1 gene leads to changes in the expression of genes that can contribute to the sensitivity of the DNA damage response. However, taken together these observations are not striking because, a) broadly, a role for mRNA decay in regulating gene expression is well established, and b) an actual mechanistic role for any of the downstream targets in mediating HU-related stress (which might not be DNA replication stress-please see comment 5 below) is lacking in the manuscript. Hence, the insights on mRNA decay/P-bodies playing a role in gene expression during replicative stress not sufficient to warrant publication in Nature Communications, unless the specific targets and regulatory loops identified in this manuscript are important to the DNA-damage community. Specific comments: 1. Expression of several genes increased as well as decreased in the RNA-seq dataset as a function of HU stress in Wt vs lsm1Δ. One confounding issue with such analyses is that if some mRNA increase in abundance, they take up more sequence space, as a result some mRNAs get underrepresented and register as reduced levels. Is the subset of decreased mRNA in these datasets controlled for this eventuality? 2. An additional limitation of the RNA-seq data is the lack of target validation with an alternative technique such as northern blotting or RT-PCR. Such an additional validation can strengthen the argument for changes in expression of the identified target genes.
3. On that note, the authors model a role for P-bodies in regulating the level of Yox1 mRNA, yet the actual levels of total Yox1 mRNA in lsm1∆ and pat1Δ have not been measured. 4. Are these changes truly because of mRNA decay? Steady-state levels in gene expression don't often change much even with a change in decay rates. Do deletions in other mRNA decay genes yield the same changes in levels of targets identified? 5. Are the effects observed in gene expression, genetic suppression etc., due to HU-related stress or DNA replication stress, specifically? Is the expression of the identified targets affected in a similar manner upon exposure to other agents that lead to DNA replication stress? 6. One concern is whether the effects in gene expression observed in lsm1Δ (or pat1∆) are due to P-bodies. It is my understanding that lsm1Δ and pat1∆ strains do not prevent P-body assembly. While this does not impact the key observations that these proteins can affect gene expression during DNA stress, it does affect the suggested role for P-bodies per se. 7. Figure 4.-nRNA should be mRNA on Y-axis. 8) Why is td-tomato used as control in Figure 5a? Shouldn't the controls be done in +/-HU to show how that alteration affects the signal?
Reviewer #3 (Remarks to the Author) The manuscript Characterizes mRNAs associated with p-bodies caused by HU-induced replication stress. The transcriptional repressor Yox1 is identified as localizing in P-bodies, and accumulates in the nucleus of lsm1-mutant cells. In general, the manuscript makes great strides in characterizing the function of P-bodies. I think the manuscript is beautifully written and the figures look very clean and professional.
The subject matter seems to appeal to a specific readership because it combines DNA-replication, P-bodies, and hydroxyurea stress. Nevertheless, it is written in a very accessible way, and could appeal to a broad readership given that it may reveal some fundamental biology of P-bodies.
In summary, I would accept this manuscript. I point out a few places where the manuscript could be improved.
A few minor formatting issues such as two periods after the sentence: …exonuclease Xrn1, which together determine the decapping or degradation rate of mRNAs..
There is a super-script error in the sentence "Both RAD54 and RAD51 are induced during the DNA replication stress response or upon X-ray exposure 8; and our data,54. ".

Minor issues:
This particular sentence: "Amazingly, we could identify specific YOX1 targets whose de-repression is critical to avoid replication stress induced toxicity, ALD6 and ICS2." has two issues. First, "Amazingly" is a very strong word, and might be too strong for a publication. Would "Surprisingly" be more appropriate? Secondly, "replication stress induced" functions as an adjective, and should be hyphenated as "replication stress-induced toxicity".
Suggestions to the authors for improvement of readability: How are "fitness values" defined? In the methods I see that it is the ratio of colony size in HU vs no drug. Would it be more clear to briefly define fitness when first introduced in the Results on page 6, or at least mention that it is defined by colony size?
The section title "Suppressors of replication stress sensitivity of P-body mutants" seems grammatically strange to me.
Concerning the statement: "Alternatively, absence of LSM1 could stabilize transcriptional repressors, resulting indirectly in mRNA abundance decreases, as has been observed in cells lacking the 5'-3' RNA exonuclease Xrn1" This is an interesting hypothesis. I think it would strengthen this point to identify specific transcriptional repressors in the set of transcripts that increase in lsm1-mutants (other than Yox1) and mention here, although more detail is given in the discussion on this point. Along those lines, the statement in the manuscript about these differentially expressed genes does not state that the list of the 333 up-and 258 down-regulated genes in lsm1-mutants (in the absense of HU-stress) is also part of Table S2. As far as I can tell, Table S2 is only introduced in the ms in the context of HU-treatment. It would be helpful to reference this table when discussing these genes at the end of page 3.

Point-by-point responses:
Reviewer #1: This manuscript examines the functions of P-bodies in regulating cellular responses to replication stress. P-bodies are sites of mRNA decapping to decrease mRNA abundance. Stresses including replication stress induce P-body formation and function and P-body proteins are important for cell viability in response to these stresses. The authors utilized RNA sequencing and genetics to identify mRNAs that may be regulated in a P-body dependent manner to yield resistance to replication stress. Their analysis identified Yox1, a transcriptional repressor as one of several candidates and they validated Yox1 mRNA regulation by P-body processing as important for replication stress responses. Furthermore, they identified two gene targets of Yox1 regulation that contribute to these responses.
Overall, I found the data in the manuscript to be compelling and the conclusions interesting. While it is not surprising that Yox1 is involved in controlling gene expression in response to replication stress, the control of Yox1 by P-body dependent processing is novel and interesting. The only thing that I would have liked the authors to do was to demonstrate unequivocally that ALD6 and ICS2 are really direct targets of Yox1 important for stress responses. For example, it would be helpful to demonstrate direct binding of Yox1 protein to the putative Yox1 binding sites in the promoters of these genes. Also, monitoring the effects of deletion of these binding sites on gene expression and cell viability would be useful.
We agree that demonstrating a direct binding of Yox1 on ALD6 or ICS2 promoters would be of interest. However, it is a reasonable possibility that Yox1 could regulate ALD6 or ICS2 expression indirectly (for example by repressing ALD6 or ICS2 transcriptional activator) without changing the conclusions of our study. Yox1 or Mcm1 binding sites are present in multiple copies in both promoters, but Yox1 binding has not been detected in high throughput studies. Since direct binding by Yox1 is not an essential component of our model, we have modified the manuscript to clarify that Yox1 regulation could be indirect: We identified binding sites for both Yox1 and its co-repressor Mcm1 in the 1000-bp promoter regions of ALD6 and ICS2 using YeTFaSCo 44 (Table S9), although it is also possible that both are indirect targets. (p.11) Reviewer #2 Specific comments: 1. Expression of several genes increased as well as decreased in the RNA-seq dataset as a function of HU stress in Wt vs lsm1Δ. One confounding issue with such analyses is that if some mRNA increase in abundance, they take up more sequence space, as a result some mRNAs get underrepresented and register as reduced levels. Is the subset of decreased mRNA in these datasets controlled for this eventuality?
We agree that upregulation of hundreds of genes could take up more sequencing space and artificially decrease abundance of other transcripts. While this is a known limitation of normalization by RPKM, most current methods to identify differentially expressed genes from RNA-seq data (including Cuffdiff, which is what we used) apply more sophisticated normalization routines to overcome the limitation. It is true that different analysis methods can produce some differences in results, so we re-analyzed our RNA-seq data using two alternative RNA-Seq data analysis methods: EBSeq and edgeR. In particular, EBSeq uses a Bayesian statistics approach, which takes into account the compositional structure of RNA-seq data. edgeR uses a alternative normalization method as compared to cuffdiff (our initial analysis method). Using both methods, we were able to confirm our conclusions. The comparison of the three analysis methods has been added (Supplemental Table S3), and the text has been modified on p.3: Finally, to confirm that the differentially expressed genes that we identified were independent of the data analysis method used, we applied two different analyses of the RNA-Seq data to identify differentially expressed genes: EBSeq 23 and edgeR 24 . Between 34 and 79% of the genes identified in our initial analysis were also identified using EBSeq or edgeR, depending on the time point analyzed (Table S3). And on p.6: Independent reconstruction of the pat1∆ and lsm1∆ double mutants with each of the 11 genes resulted in validation of 6 putative target genes: ARL3, ACF4, HHT1, TMA1, RRS1 and YOX1. Increased mRNA abundance in HU for 5 of these transcripts was confirmed by two independent data analysis methods. ACF4 was confirmed by edgeR but not by EBSeq (Table S6).
We also note that the correlation between biological replicates in our RNA-Seq experiments was very high (R>0.92 for biological replicates, mentioned in the Methods section).
2. An additional limitation of the RNA-seq data is the lack of target validation with an alternative technique such as northern blotting or RT-PCR. Such an additional validation can strengthen the argument for changes in expression of the identified target genes.
We addressed the reviewer's concern by validating YOX1 up-regulation and ALD6 downregulation in lsm1∆ cells using qRT-PCR. The data are presented in Supplemental Figures S4  and S5. 3. On that note, the authors model a role for P-bodies in regulating the level of Yox1 mRNA, yet the actual levels of total Yox1 mRNA in lsm1∆ and pat1Δ have not been measured.
We have now measured YOX1 mRNA in both lsm1∆ and pat1∆ by qRT-PCR ( Figures S4 and  S5), in addition to the original measurement in lsm1∆ by RNA-seq. 4. Are these changes truly because of mRNA decay? Steady-state levels in gene expression don't often change much even with a change in decay rates. Do deletions in other mRNA decay genes yield the same changes in levels of targets identified?
We found that YOX1 mRNA increases in pat1∆ and xrn1∆ cells ( Figure S4 and text on p. 9 (xrn1∆:wildtype = 3.7 ± 1.8)). Pat1 is an mRNA decapping protein, and Xrn1 is the predominant 5' to 3' exoribonuclease, indicating a role for mRNA decay functions in the reduction in YOX1 mRNA abundance. 5. Are the effects observed in gene expression, genetic suppression etc., due to HU-related stress or DNA replication stress, specifically? Is the expression of the identified targets affected in a similar manner upon exposure to other agents that lead to DNA replication stress?
We tested whether the lsm1∆ differentially expressed genes in HU overlapped with genes whose expression is affected during DNA replication induced by treatment with MMS and found good overlap (as much as 53%, depending on the dataset) suggesting that the transcriptional program that we identified is likely a response to DNA replication stress in general and not only HUspecific. The text has been modified on p.4: The correlation with data obtained using a distinct replication stress agent, MMS, indicates that a substantial fraction of the transcriptional program that we identified is due to DNA replication stress (Fig. S2b,c). 6. One concern is whether the effects in gene expression observed in lsm1Δ (or pat1∆) are due to P-bodies. It is my understanding that lsm1Δ and pat1∆ strains do not prevent P-body assembly. While this does not impact the key observations that these proteins can affect gene expression during DNA stress, it does affect the suggested role for P-bodies per se.
Deletion of LSM1 induces Dcp1, Dcp2, Edc3, Xrn1 and Dhh1 foci formation (due to the accumulation of RNA in the cytoplasm) and reduces Pat1 foci formation (see Teixeira & Parker, 2007, Mol Biol Cell). Deletion of PAT1 prevents the formation of P-body granules for almost all core P-body proteins, including Lsm1 (see Teixeira & Parker, 2007, Mol Biol Cell). Given that YOX1 mRNA abundance increases in lsm1∆ and pat1∆ cells, we suggest that the regulation of Pbody formation is required for the regulation of YOX1 mRNA abundance.
This has been corrected.
8. Why is td-tomato used as control in Figure 5a? Shouldn't the controls be done in +/-HU to show how that alteration affects the signal?
Td-tomato was not used as a control in Fig. 5a. We used Hta2-mCherry as a nuclear marker to segment the nuclei in order to quantify nuclear and cytoplasmic Yox1-GFP. Both RFP and GFP channels are shown in both conditions (-/+ HU) in Fig. 5a.
Reviewer #3 (Remarks to the Author): The manuscript characterizes mRNAs associated with p-bodies caused by HU-induced replication stress. The transcriptional repressor Yox1 is identified as localizing in P-bodies, and accumulates in the nucleus of lsm1-mutant cells. In general, the manuscript makes great strides in characterizing the function of P-bodies. I think the manuscript is beautifully written and the figures look very clean and professional.
The subject matter seems to appeal to a specific readership because it combines DNAreplication, P-bodies, and hydroxyurea stress. Nevertheless, it is written in a very accessible way, and could appeal to a broad readership given that it may reveal some fundamental biology of Pbodies.
In summary, I would accept this manuscript. I point out a few places where the manuscript could be improved.
A few minor formatting issues such as two periods after the sentence: …exonuclease Xrn1, which together determine the decapping or degradation rate of mRNAs.. This has been corrected.
There is a super-script error in the sentence "Both RAD54 and RAD51 are induced during the DNA replication stress response or upon X-ray exposure 8; and our data,54. ". This has been corrected.

Minor issues:
This particular sentence: "Amazingly, we could identify specific YOX1 targets whose de-repression is critical to avoid replication stress induced toxicity, ALD6 and ICS2." has two issues. First, "Amazingly" is a very strong word, and might be too strong for a publication. Would "Surprisingly" be more appropriate? Secondly, "replication stress induced" functions as an adjective, and should be hyphenated as "replication stress-induced toxicity". The text has been corrected as suggested Suggestions to the authors for improvement of readability: How are "fitness values" defined? In the methods I see that it is the ratio of colony size in HU vs no drug. Would it be more clear to briefly define fitness when first introduced in the Results on page 6, or at least mention that it is defined by colony size?
This has been clarified in the main text. The manuscript now reads: We then assessed the fitness of every double mutant, by measuring and comparing colony size in the presence and absence of HU, in triplicate.
The section title "Suppressors of replication stress sensitivity of P-body mutants" seems grammatically strange to me.
The title has been changed as follows Suppressors of the replication stress sensitivity of P-body mutants Concerning the statement: "Alternatively, absence of LSM1 could stabilize transcriptional repressors, resulting indirectly in mRNA abundance decreases, as has been observed in cells lacking the 5'-3' RNA exonuclease Xrn1." This is an interesting hypothesis. I think it would strengthen this point to identify specific transcriptional repressors in the set of transcripts that increase in lsm1-mutants (other than Yox1) and mention here, although more detail is given in the discussion on this point.
We looked whether there were up-regulated repressors and whether their targets were enriched in the subset of lsm1∆ down-regulated genes at the same time points but did not find repressors that showed this pattern, with the exception of YOX1.
Along those lines, the statement in the manuscript about these differentially expressed genes does not state that the list of the 333 up-and 258 down-regulated genes in lsm1-mutants (in the absense of HU-stress) is also part of Table S2. As far as I can tell, Table S2 is only introduced in the ms in the context of HU-treatment. It would be helpful to reference this table when discussing these genes at the end of page 3.
We added an earlier reference to Table S2 as suggested.

Reviewers' Comments:
Reviewer #1: Remarks to the Author: If the authors were able to show that it was direct and map the responsible DNA element, those mechanistic insights would significantly strengthen the overall conclusions. However, I agree with the author's response that the model does not require direct binding and the results would still be of interest.
Reviewer #2: Remarks to the Author: The authors have presented several lines of evidence to solidify claims made in the initial version of this manuscript. Specifically, the additional data presented are sufficient to address technical concerns raised previously. There is one major issue that I still find problematic. Specifically, I disagree with the claim made by the authors regarding the role for "P-bodies" per se in regulating levels of transcripts, such as Yox1 in vivo. Measurable, yet insufficient effects of single gene deletions on abrogation of P-body assembly is well documented since the manuscript by Teixiera andParker 2007 (e.g., Buchan et al., 2008, JCB). Furthermore, the change in mRNA levels and P-body assembly are correlative, and not causative. As result, a model for P-bodies in DNA replication stress response by "rewiring" transcriptome is not supported by the data, and most likely is incorrect. Certainly, the proteins found in P-bodies can have an effect, but whether it is P-body assembly per se has not been demonstrated. I recommend one of two things: A) I suggest that the authors change the title and the tone of the manuscript such that the title and tone does not overstate the observations, which implicates Pbodies in regulating the mRNA, or B) Examine how edc3∆ or edc3∆ lsm4∆c strains, which have very strong effects on P-bodies (Decker et al., 2007, JCB), affect this process. If they also have a strong effect, then I would be more convinced P-bodies per se are involved in the response.