Dot1 regulates nucleosome dynamics by its inherent histone chaperone activity in yeast

Dot1 (disruptor of telomeric silencing-1, DOT1L in humans) is the only known enzyme responsible for histone H3 lysine 79 methylation (H3K79me) and is evolutionarily conserved in most eukaryotes. Yeast Dot1p lacks a SET domain and does not methylate free histones and thus may have different actions with respect to other histone methyltransferases. Here we show that Dot1p displays histone chaperone activity and regulates nucleosome dynamics via histone exchange in yeast. We show that a methylation-independent function of Dot1p is required for the cryptic transcription within transcribed regions seen following disruption of the Set2–Rpd3S pathway. Dot1p can assemble core histones to nucleosomes and facilitate ATP-dependent chromatin-remodeling activity through its nucleosome-binding domain, in vitro. Global analysis indicates that Dot1p appears to be particularly important for histone exchange and chromatin accessibility on the transcribed regions of long-length genes. Our findings collectively suggest that Dot1p-mediated histone chaperone activity controls nucleosome dynamics in transcribed regions.


1.
2. Figure 4: In the chromatin sliding assay, can you be certain that the ability of Dot1 to stimulate Chd1 isn't a general function of other nucleosome-binding enzymes? The 101-140 region of Dot1 is clearly necessary for its stimulatory function but is it sufficient? 3. "long gene" sequencing analysis: Have the necessary measures been taken to ensure that gene length bias hasn't been introduced into the analysis?
Reviewer #2 (Remarks to the Author): Soyun Lee et al have investigated the role of Dot1 in regulating nucleosome dynamics. They show that a cross talk between H4K16Ac and H3K79me and show that H4K16Ac regulates the Dot1p mediated distribution of H3K79 methlation on euchromatin. They show loss of Dot1 suppresses histone exchange and cryptic transcription of set2∆ cells in a methylation -independent manner. They show Dot1 has nucleosome assembly activity and enhances chromatin remodeler activities, including that of Asf1p. They demonstrate that Dot1 preferentially affects histone exchange in long genes. Lastly they show that Dot1 regulates chromatin accessibility, especially on transcribed regions of genes.
General comments: The manuscript presents a number of novel and important advances in our understanding of the mehtyltransferase, Dot1p. The manuscript is very well written and the conclusions drawn are sound and based on high quality data.
As a general comment it is difficult to see differences in the peak values in set2 and set2∆ dot1∆ double mutants in the way the data has been presented here. Perhaps a close-up of the peaks showing a modulated response in the set2 dot1 doubel might be more appropriate?
Minor comments: 201: We used (the) a differential expression program 209: Set2-dep(e)leted  References: Reference manager has failed to abbreviate and capitalize many journal names, eg Molecular cell should be Mol. Cell etc.
RPKM is not clearly defined.
The rationale for using MNase-seq and ATAC-seq, and what question being addressed here is not clearly explained. They should not assume that the reader is familiar with these techniques.
Reviewer #3 (Remarks to the Author): In this manuscript Lee et al report that S. cerevisiae Dot1 regulates nucleosome dynamics through a histone chaperone activity that is independent of its previously well-characterised H3K79me activity. This is an important finding given that Dot1 is evolutionarily conserved in humans and its dysregulation is associated with MLL-leukemia. It also raises the possibility that other histone methyltransferases may function in a similar manner. This is of interest because methylationindependent roles have been suggested for another KMTase (S. pombe Set1). The authors show that H4K16ac is important for Dot1 recruitment to chromatin, that loss of Dot1 suppresses cryptic transcription in set2Δ cells, that Dot1 promotes nucleosome assembly in vitro and that Dot1 promotes histone exchange independently of its KMTase activity. The conclusions of the study are generally supported by the data, however there are some points which need to be addressed.
1. The number of biological repeats used to generate the ChIP-seq, RNA-seq, MNase-seq and ATAC-seq datasets is not specified (as far as I can tell). It should be, because single experiments are not acceptable. This is particularly important as the impact of Dot1 deletion on cryptic transcript levels, histone occupancy and exchange are relatively modest. Some attempt to show the levels of variability in these experiments should also be included.
2. There is a 'disconnect' between the data showing that H4K16ac is important for Dot1 recruitment and the failure of sas2Δ to suppress the cryptic transcript phenotype of set2Δ. If H4K61ac is really critical for Dot1 recruitment then deletion of SET2 should suppress the cryptic transcription phenotype of set2Λ, but it does not. This should be addressed/discussed. The supplementary information also shows that H4 tail binding is not necessary to potentiate the action of ATP-dep remodelers.
3. It has previously been shown that loss of Asf1 suppress the increased histone exchange that occurs in set2Δ cells. The author's indicate that that loss of Dot1 does the same. Therefore for Fig  5 it would have been very useful to have also compared histone turnover in set2Δ and set2Δ dot1Δ cells using the FLAG/Myc approach (rather than just measuring H4K16ac as in Fig 3). 5. The materials and methods section lacks a description of the GST-pulldown experiment (Fig 4a).
Here the reference to histone octamers (line 302) is probably misleading as octamers are only stable in the presence of DNA. Lines 309 and 310 also suggest that both Asf1 and Nap1 were used as controls in this experiment. Only the data relating to Nap1 is shown.
6. Fig 2a. The authors indicate that "sas2D set2D cells displayed a slight decrease in growth compared to set2D" I don't think that this conclusion can be made on the basis of the data shown in this Fig.   7. In the discussion the authors suggest that Dot1 may participate in chromatin dynamics in subtelomeric regions. Why did they not analyse these regions in their genome-wide experiments?

Reviewer #1 (Remarks to the Author):
In this paper, Lee et al. describe an uncharacterized and non-enzymatic role of the histone H3 lysine-79 methyltransferase, Dot1. Although the methyltransferase activity of Dot1 has been well characterized for nearly two decades, this manuscript provides strong evidence for methyltransferase-independent functions of Dot1 including in vitro nucleosome assembly activity and stimulating ATP-dependent chromatin remodeling. The authors also show that H4K16ac is a key factor in the distribution of Dot1 mediated H3K79 methylation and that the loss of Dot1 can suppress cryptic transcription in set2∆ cells. The results presented in the paper are rigorous and shed new light on the role of Dot1 in transcriptional regulation. As non-enzymatic roles of several other methyltranferases are also now being elucidated, this work seems very timely. Thus, this paper will be of interest to the chromatin and transcription community. Before publication, however, we have several suggestions listed below that should be addressed: 1. Fig. 3b Our comment: Thank you for your comment. We now have included the graphs of H4ac and H3 for sas2∆ and sas2∆set2∆ in Supplementary Fig. 3. The level of H4ac which is an indicative of the histone exchange do not show a significant difference in sas2∆set2∆ compared to set2∆ (Supplementary Fig. 3b). The level of H3 in sas2∆ was lower than wildtype cells ( Supplementary Fig. 3c), and histone turnover measured by Flag-histone H3 in inducible strain displayed a slight decrease in sas2∆ (Supplementary Fig. 3d). These results support the approach of Dot1 by Sas2-mediated H4K16ac on transcribed region regulates histone exchange.
Although the unchanging level of histone exchange supports our previous observation of sas2∆set2∆ not suppressing cryptic transcription (Fig. 2a), it seems contradictory to the Sas2mediated H4K16ac playing a role in the recruitment of Dot1 to the transcribed regions implicated by the decrease in H3K79me3 in sas2∆ cells (Fig. 1a). This is may be due to the difference in the chromatin context in wild-type and set2∆ cells. In set2∆ cells, the chromatins are hyper-acetylated as the HDAC enzymes are unable to be recruited via Set2-Rpd3S pathway. The hyper-acetylation state increases chromatin dynamics and increases chromatin accessibility, such that distributive enzymes like Dot1 can access the chromatin more easily. Moreover, the increased cryptic transcription in set2∆ cells may result in the recruitment of other HATs to the transcribed regions of cryptic transcription. We have shown the involvement of other HATs such as NuA4 at the promoter region at the promoters (Fig.  1e) and the emergence of cryptic promoters in the transcribed region of the genes in set2∆ cells may therefore imply the involvement of other HAT in cryptic transcription situation. Taken together, the chromatin state in set2∆ cells are different from that of wild-type, such that Sas2-mediated H4K16ac may no longer necessary for the recruitment of Dot1. Fig. 4: In the chromatin sliding assay, can you be certain that the ability of Dot1 to stimulate Chd1 isn't a general function of other nucleosome-binding enzymes? The 101-140 region of Dot1 is clearly necessary for its stimulatory function but is it sufficient?

2.
Our comment: The chaperonic activity of Dot1 led to postulate that Dot1 may stimulate remodeling effect of remodelers, as histone chaperones like Asf1 has been shown to stimulate remodeling in vitro. As the reviewer suggested, it is possible that other chaperone proteins that have nucleosome binding activity stimulate chromatin remodeling of Chd1. Further characterization of other nucleosome binding proteins would be required to identify the role of nucleosome binding domain per se, in stimulation of Chd1.
As we mentioned in the line 256, it has been reported by Oh et al,. (2010) that the Dot1 domain 101-140 plays a crucial role in the nucleosome binding of Dot1. In this paper, the authors characterized the domains of Dot1 by performing EMSA experiment using several truncation mutants including the Dot1(101-140Δ) mutant. We tested all of these mutants for the stimulating role of Chd1 and only Dot1(101-140Δ) mutant showed defect in the stimulation.

"long gene" sequencing analysis: Have the necessary measures been taken to ensure that gene length bias hasn't been introduced into the analysis?
Our comment: Thank you for raising an important issue. The gene length bias is one of the common bias that occur during analysis of high-throughput sequencing to produce metazoan average plot. We eliminated gene length bias by RPKM analysis in Fig. 5d. Usage of RPKM is commonly used in RNA-seq analysis, and is also used in ChIP-seq to remove gene length bias, by adopting the equation RPKM = (number of reads mapped to a gene × 1E+09)/ (length of the gene × number of total mapped read counts in the experiment), which take gene length into consideration. We added the sentence "To eliminate analysis error due to variability of gene length and the effect of strong histone exchange near the promoters and 5'-ends of genes," in line 385-387 to clarify the analysis method. Also, we would like to emphasize that the heatmap in Fig. 5a was sorted by gene length to prevent gene-length bias.

Reviewer #2 (Remarks to the Author):
Soyun Lee  Our comment: Thank you for your suggestion. We agreed with the reviewer's concern and revised Fig. 2 and Fig. 3 to contain a close-up of the peaks of mRNA-seq and log 2 H3 data for set2∆ and set2∆dot1∆ mutants.  Our comment: We apologize for not making it clear enough. For Fig. 1, we used wzy42 strains which the expression of histone H3 and H4 are maintained via HHT-HHF2 plasmid. The histone mutant strains are not overexpressed strains but strains where HHT-HHF2 plasmid is replaced with plasmids of respective histone mutations via FOA selection. We added the details in the legends and in the Methods sections. Our comment: We included mRNA-seq results and the box-plot comparison of expression of single mutant sas2∆, dot1∆ in Supplementary Fig. S2.
References: Reference manager has failed to abbreviate and capitalize many journal names, eg Molecular cell should be Mol. Cell etc.
Our comment: Thank you for letting us know. There was an error on 'Term List' in the reference program. We updated the program and fixed the error.

RPKM is not clearly defined.
Our comment: We apologize for not elaborating the term. We elaborated the term RPKM in the Line 139. We also added the equation for calculating RPKM in the analysis part of the Materials and Methods section.
The rationale for using MNase-seq and ATAC-seq, and what question being addressed here is not clearly explained. They should not assume that the reader is familiar with these techniques.
Our comment: Thank you for the constructive comment. We added the following sentences in Line 429 to further explain the rationale for using MNase-seq and ATAC-seq: "If Dot1 regulates the dynamics of chromatin structure, changes in nucleosome positioning or chromatin accessibility may be observed in dot1∆ cells. We performed MNase-seq to assess the effect on nucleosome positing and nucleosome occupancy, and ATAC-seq, a technique that of similar rationale to DNase-seq, to assess the effect on chromatin accessibility."

In this manuscript Lee et al report that S. cerevisiae Dot1 regulates nucleosome dynamics through a histone chaperone activity that is independent of its previously well-characterised H3K79me activity. This is an important finding given that Dot1 is evolutionarily conserved in humans and its dysregulation is associated with MLL-leukemia. It also raises the possibility that other histone methyltransferases may function in a similar manner. This is of interest because methylation-independent roles have been suggested for another KMTase (S. pombe Set1). The authors show that H4K16ac is important for Dot1 recruitment to chromatin, that loss of Dot1 suppresses cryptic transcription in set2Δ cells, that Dot1 promotes nucleosome
assembly in vitro and that Dot1 promotes histone exchange independently of its KMTase activity. The conclusions of the study are generally supported by the data, however there are some points which need to be addressed.

The number of biological repeats used to generate the ChIP-seq, RNA-seq, MNase-seq and
ATAC-seq datasets is not specified (as far as I can tell). It should be, because single experiments are not acceptable. This is particularly important as the impact of Dot1 deletion on cryptic transcript levels, histone occupancy and exchange are relatively modest. Some attempt to show the levels of variability in these experiments should also be included.
Our comment: We apologize for not specifying the number of biological repeats in the legends. The sentence specifying the number of repeats was added in the figure legends of the appropriate figures in the revised manuscript. We also included the scatter plot and R 2 value to show the sample variation of the repeat samples in Supplementary Fig. 7-9.

There is a 'disconnect' between the data showing that H4K16ac is important for Dot1 recruitment and the failure of sas2Δ to suppress the cryptic transcript phenotype of set2Δ. If H4K16ac is really critical for Dot1 recruitment then deletion of SET2 should suppress the cryptic transcription phenotype of set2Λ, but it does not. This should be addressed/discussed. The supplementary information also shows that H4 tail binding is not necessary to potentiate the action of ATP-dep remodelers.
Our comment: We appreciate your comment for pointing out an important issue. The phenotype of sas2∆set2∆ not suppressing the cryptic transcript phenotype of set2∆ was the point that puzzled us well. Fortunately, we managed come to a conclusion after checking the H4ac level in sas2∆set2∆, which was the experiment that Reviewer 1 asked for. We have discussed in detail at the above section (Reviewer 1, comment 1) why sas2∆set2∆ would not suppress the cryptic phenotype of set2∆. Briefly, the chromatin context in set2∆ cells are different to wild-type that the set2∆ chromatin are in its hyper-acetylation status and may stimulate non-specific recruitment Dot1, and may not necessitate Sas2 for Dot1 recruitment.
This hyper-acetylation status in set2∆ also helps to explain why the H4 tail binding domain of Dot1 (EDVDE domain) is not necessary for the stimulation of ATP-dependent remodelers. The chromatin accessibility increases when chromatin is hyper-acetylated and nucleosome binding proteins such as Dot1 may easily access to the transcribed region and facilitate histone exchange. This indicates that the histone exchange activity of Dot1 is mediated by the nucleosome binding activity but not the binding of histone H4 tail. In fact, histone H4 tail binding seems dispensable for the histone exchange activity per se. In the revised manuscript, we discussed Supplementary Fig. 3 in detail to address this issue. Fig 3).

It has previously been shown that loss of Asf1 suppress the increased histone exchange that occurs in set2Δ cells. The author's indicate that that loss of Dot1 does the same. Therefore for Fig 5 it would have been very useful to have also compared histone turnover in set2Δ and set2Δ dot1Δ cells using the FLAG/Myc approach (rather than just measuring H4K16ac as in
Our comment: Thank you for your suggestion. As the reviewer suggested, we attempted FLAG/Myc approach in set2∆ and set2∆dot1∆ cells, and the result is presented as a box-plot in the figure below. The result of FLAG/Myc assay was a verification of our previous observation in H4ac: the turnover rate increased in set2∆, and the increase was suppressed in dot1∆set2∆. We fully agree that FLAG/Myc is an excellent approach to visualize the rate of histone exchange. However, we thought that H4ac data would represent more authentic phenotype than FLAG/Myc assay as FLAG/Myc assay includes media exchange from glucose to galactose while H4ac assay uses glucose media throughout. Moreover, since the data in Fig. 3 proceeds the cryptic transcription data in Fig. 2, we thought using the same glucose medium would be a better choice. We therefore decided to present the FLAG/Myc assay only as reviewer's inspection. Fig. legends Fig 4a? Our comment: We again apologize for the legends. We added details in the figure legends such that it would contain all of the experimental conditions, symbols and scale bar. The arrow head in Fig. 1A was initially deployed for emphasis purpose but was removed in the revised manuscript as it did not have scientific significance. The type of gel shown in Fig. 4A was coomassie blue-stained gel. The Materials and Methods section and legends were changed accordingly. (Fig 4a). Here the reference to histone octamers (line 302) is probably misleading as octamers are only stable in the presence of DNA. Lines 309 and 310 also suggest that both Asf1 and Nap1 were used as controls in this experiment. Only the data relating to Nap1 is shown.

The materials and methods section lacks a description of the GST-pulldown experiment
Our comment: We apologize for the errors. The word 'Asf1' was removed. We described the GST-pulldown experiment in detail by adding a separate section in Materials and Methods section. The histone octamer was purified from YS14 core histone expression vector and stored a high salt buffer. We agree with the reviewer's concern that the instability of octamers without presence of DNA as the octamers may not be stable under 150mM GST-pulldown experiment condition and disassemble into dimer forms. However, what we intended to test here was just the binding with the GST-proteins with octameric core-histone. We believe that the possible instability of the octamers would not affect our overall conclusion.
6. Fig 2a. The authors indicate that "sas2∆ set2∆ cells displayed a slight decrease in growth compared to set2∆" I don't think that this conclusion can be made on the basis of the data shown in this Fig. Our comment: We agreed to the reviewer's comment, and we revised the manuscript to describe that there was no change in the growth between set2∆ and set2∆sas2∆ cells. Also, we believe that the mRNA-seq is sufficient to demonstrate the change in cryptic transcription, and thus the spotting data in Fig. 2a was moved to Supplementary Fig. 2a. 7. In the discussion the authors suggest that Dot1 may participate in chromatin dynamics in subtelomeric regions. Why did they not analyse these regions in their genome-wide experiments?
Our comment: We removed sections regarding telomere silencing (Line 77-87, 498-505) as the sections were considered as deviations from our main findings. To answer the reviewer's question, it is difficult to examine a precise role of Dot1 in transcribed regions in subtelomeric regions, because the subtelomeric regions contain wide intergenic regions and little long-length genes, making it difficult to directly analyze of the genes in this region (see the figure below) Moreover, we believe the analysis of the genes in these regions is beyond the scope of this paper.
ChIP-seq data near TEL06R region of Chromosome VI. The figure represents the region from TEL06R to YFR050C (the distance is 21kb). The magenta box indicates the TEL06R region. The normalized ChIP-seq signal of H3K79me3 (blue), H3K79me1 (orange), and H4K16ac (purple) in wild-type cells is shown. Histone exchange in wild-type is represented as H4ac signal normalized by H3 (turquoise). In addition, red signal indicates histone turnover (red, log 2 (Flag/Myc)) from Flag/Myc assay in wild-type cells.
Minor Points 1. Line 32. Change "cryptic transcriptions" to "cryptic transcription" 2. Line 40 Change "The chromatin structure" to "Chromatin structure" Our comment: Thank you. We made the changes accordingly.

Lines 183-184. I don't understand the use of "stabilized" in this context. These lines should be reworded.
Our comment: We rephrased the Lines 183-184: "As transcription elongation is closely related to Set2-Rpd3S pathway which suppresses cryptic transcription, we examined the effects of Dot1p on cryptic transcription through Set2p."

Lines 472-47This sentences reads as if loss of Dot1 up-regulates cryptic transcription in set2Δ cells, when in fact the opposite is true.
Our comment: Thank you for pointing it out. We rephrased the sentence: "Second, the loss of DOT1 suppresses histone exchange on the transcribed regions of genes and down-regulates cryptic transcripts in set2Δ cells."

Lines 476. Asf1 is not an ATP-dep remodeler. I think the authors mean Chd1 here
Our comment: Thank you. We corrected it to Chd1.

Lines 500-501. The sense of this sentence is not clear, please reword.
Our comment: We removed Line 498-505 as we decided that the subject of telomere silencing deviates from our main story. 7. Line 207 "DOT1" should be italicized, 8. Line 223 replace "translation" with "transcription" Our comment: We made the changes accordingly.
Again, we thank the reviewers for their time and effort reviewing our manuscript. The comments were constructive and, in addressing them, we greatly improved our insight into the histone chaperone activity of Dot1. P. S. The following is a list of the major changes made to the revised manuscript: 1. To respond to a specific comment from Reviewer 3, we removed sections regarding telomere silencing (Line 77-87, 498-505) as the sections were considered as deviations from our main findings.
2. To respond to a general comment from Reviewer 2, we revised Fig. 2 and Fig. 3 to contain a close-up of the peaks of mRNA-seq and log 2 H3 data for set2∆ and set2∆dot1∆ mutants.
3. To respond to a specific comment from Reviewer 2, we moved originally submitted Fig. 2a to Supplementary Fig. 2a, as we agreed to the reviewer's comment that sas2∆set2∆ does not suppress the cryptic phenotype of set2∆ 4. To respond to a specific comment from Reviewer 2, we included mRNA-seq results and the box-plot comparison of the gene expression of sas2∆, dot1∆ single mutants in Supplementary Fig. 2.

5.
To respond to a specific comment from Reviewer 3, we performed FLAG/Myc experiment and showed the data as reviewer's inspection to verify our previous observation of the histone turnover suppression in dot1∆set2∆.
6. To respond to a specific comment from Reviewer 3, we expensively discussed the issue of sas2∆set2∆ not suppressing the cryptic phenotype of set2∆ by discussing the data in Supplementary Fig. 3.

7.
To respond to a specific comment from Reviewer 1, we included the graphs of H4ac sas2∆ and sas2∆set2∆ and the graph of H3 in sas2∆ in Supplementary Fig. 3 to show the effect of the respective mutant on the levels of H3 and on histone exchange. 8. To respond to a specific comment from Reviewer 2, we revised the Results section to include a detailed rationale of using MNase-seq and ATAC-seq (Line 429) 9. To respond to a specific foment from Reviewers 1 and 2, we revised the Results section to include the definition of RPKM (Line 139) and revised the Materials and Methods section to include the equation used for the calculation of RPKM.
10. To respond to a specific comment from Reviewer 3, we revised the manuscript to describe that there was no change in the growth between set2∆ and set2∆sas2∆ cells which we previously argued for a marginal change. We therefore moved the previous Fig. 2a to Supplementary Fig. 2a.
11. To respond to a specific comment from Reviewer 2, we revised the figure legends and the Materials and Methods section to include the details of the wzy42 strain system for analyzing histone mutant strains.
12. To respond to a specific comment from Reviewers 2 and 3, we revised the figure legends to include the details of the number of biological repeats, meaning of axis and colors for further clarity, and Materials and Methods section was revised accordingly.
13. To respond to a specific comment from Reviewer 3, we included scatter plots and R 2 values to show the sample variation of the repeat samples in newly added Supplementary Fig. 7-9.
14. To respond to a specific comment from Reviewer 3, we described the GST-pulldown experiment in detail by adding an independent section in Materials and Methods section.
15. To respond to a specific comment from Reviewer 3, we included the genome-wide H3K79me distribution and histone exchange rate as reviewer's inspection to show the difficulty of assessing the role of Dot1 in subtelomeric regions.
16. To respond to a specific comment from Reviewer 3, we revised the Results section to clarify the meaning of transcription elongation being stabilized by the Set2-Rpd3S pathway.
17. We removed FLO8 data from originally submitted Fig. 2b due to space restriction as it was a redundant example among the four genes (SPB4, STE11, PCA1, FLO8).
18. We revised the Abstract section to reduce its length to meet the requirement of maximum 150 words.
19. We trimmed the Introduction and Discussion sections to reduce the number of words.