Conservation of chromatin states and their association with transcription factors in land plants
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (Review Commons)
Abstract
The complexity of varied modifications of chromatin composition is integrated in archetypal combinations called chromatin states that predict the local potential for transcription. The degree of conservation of chromatin states has not been established amongst plants, and how they interact with transcription factors is unknown. Here we identify and characterize chromatin states in the flowering plant Arabidopsis thaliana and the bryophyte Marchantia polymorpha , showing a large degree of functional conservation over more than 450 million years of land plant evolution. We used this new resource of conserved plant chromatin states to understand the influence of chromatin states on gene regulation. We established the preferential association of chromatin states with binding sites and activity of transcription factors. These associations define three main groups of transcription factors that bind upstream of the transcription start site, at the +1 nucleosome or further downstream of the transcription start site and broadly associate with distinct biological functions. The association with the +1 nucleosome defines a list of candidate pioneer factors we know little about in plants, compared to their important roles in animal stem cells and early development.
Article activity feed
-
Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
The authors do not wish to provide a response at this time.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
Summary:
The manuscript by Shukla et al described the "chromatin states" in the bryophyte Marchantia polymorpha and compared it with that in Arabidopsis thaliana. They described the generally common features of chromatin states between these evolutionally distant plant species, but they also find some differences. The authors also studied the connection between chromatin states and TF bindings, mostly in Arabidopsis due to the scarcity of the TF binding data in Marchantia. Their analyses lead to interesting finding that specific transcription families tend to associate with specific chromatin state, which tend to associate with …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
Summary:
The manuscript by Shukla et al described the "chromatin states" in the bryophyte Marchantia polymorpha and compared it with that in Arabidopsis thaliana. They described the generally common features of chromatin states between these evolutionally distant plant species, but they also find some differences. The authors also studied the connection between chromatin states and TF bindings, mostly in Arabidopsis due to the scarcity of the TF binding data in Marchantia. Their analyses lead to interesting finding that specific transcription families tend to associate with specific chromatin state, which tend to associate with specific genomic regions such as promoter, TSS, gene body, and fucultative heterochromatin. Overall, the authors provide novel piece of information regarding the evolutional conservation of chromatin states and the relationship between chromatin states and TFs.
Major comments:
In the end of the abstract they state "The association with the +1 nucleosome defines a list of candidate pioneer factors we know little about in plants", which is one of their major points. This is based on the results Fig4F and 4G, described in P27 L16-17. Question is, is cluster 1 TFs really associated with the +1 nucleosome? From Fig. 1C, +1 nucleosome is characterized mostly by E1 state and also by E2, F3, F4. However, from Fig. 4F, cluster 1 TFs are not associated with E1/E2 and association is not particularly strong for F3/F4. Indeeed association with E1/E2 is much conspicuous for cluster 4 TFs. Therefore, authors should reconsider this point and consider rephrasing or showing further results of analyses.
P17 last line to P18, they state "The facultative heterochromatin states were primarily associated with the intergenic states I1 to I3, based on their enrichment in H3K27me3 and H2AK121ub, low accessibility, and low gene expression". I'm not sure about this statement. How can they say "primarily associated" from the data they cite? As far as the PTMs and variants patterns, I1 to I3 and facultative heterochromatin look different. The authors should explain more or rephrase.
P20 L15, the authors state "Contrary to Arabidopsis, the promoters of Marchantia defined by the region just upstream of the TSS showed enrichment of H2AUb and the elongation mark H3K36me3, along with other euchromatic marks. " I have a concern that the TSS annotation could be inaccurate in Marchantia compared to more rigorously tested annotation of Arabidopsis thaliana, so that the relationship between TSS and histone PTMs could be different between species. The authors should make sure this is not the case.
P21 last line to P22, they analyzed only H3K27me3 and H2Aub in the mutants of E(z) (Fig. 2E) and states that "we analyzed chromatin landscape in the Marchantia...". Is analyzing two histone marks enough to say "chromatin landscape"? In addition, they state "These findings suggest a strong independence of the two Polycomb repressive pathways in Marchantia. " However, they did not analyzed the effect of loss of PRC1 on H3K27me3; the opposite way. Actually, in Arabidopsis loss of PRC1 causes loss of H2Aub AND H3K27me3 (Zhou et al (2017) Genome Biol: DOI 10.1186/s13059-017-1197-z).
Related to the above comments, they states "To further compare the regulation by PRC2 in both species,". However, they did not describe the knowledge about regulation by PRC2 in Arabidopsis. They should consider describing.
P25 L14: "With this method to estimate TF activity, the scores of TF occupancy and activity converged. To look at different patterns of chromatin preferences among TFs, we kept ChIP-seq and DAP-seq data for ~300 TFs in Arabidopsis (after filtering out TFs with low scores of occupancy and activity)." This part is a little hard to follow. Perhaps better to explain in more detail.
In discussion section P30 L19-21: "This could be due to open chromatin, which is associated with highly expressed genes and permissive for TF binding, generating highly occupied target regions (HOT) with redundant or passive activity (19)." This part needs further explanation; espetially for the latter part, It's not clar what the authors claim.
Minor comments:
P17 L21: H2bUb should be H2Bub.
Legend of Fig. 4D: later should be latter.
Legend of Fig. 4G and H: "clusters defined in figure-H" should be "defined in Fig. 4F"?
Referee cross-commenting
Reviewer #1 raises thorough and important points that should be addressed before the manuscript is published. Particularly about the comparison of chromatin states between Arabidopsis and Marchantia, as this paper will make foundation for further research in the future and serve as a resource for community, the authors should thoroughly look into the points raised by reviewer #1 including annotation of transcriptional units.
Significance
Strength and limitation: Strength of this paper is the insights into chromatin-based transcriptional regulation by defining chromatin states using combination of many epigenome data and compare it with TF biding data. Limitation is lack of experimental support for their interesting claims by perturbing histone PTMs, for example. Also, a limitation is that comparing only two species can tell subjective "similar" or "different" between species.
Advance comparing past literature: One clear advance is studying chromatin states in a plant other than Arabidopsis thaliana. Another one is revealing that TFs can be classified into a number of groups according to the relationships with chromatin-based transcription regulation. However, experimental tests for these are awaited.
Audience: Epigenetics, chromatin, and transcription researchers, plant biologists interested in transcriptional regulation.
My expertise: Epigenome, genetics, histone PTMs, plants
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
Summary:
The authors characterize chromatin states in the flowering plant Arabidopsis thaliana and the bryophyte Marchantia polymorpha. Here, they draw from ChIP-seq data that was previously published, and from data generated as part of this study, in particular for Marchantia H2.A variants (H2A.X.1, H2A.X.1, H2A.Z, H2A.M.2). The authors compute chromatin states, which enables a comparison over more than 450 million years of land plant evolution. While comparisons of plant chromatin to other species highlighted conservation as well as differences, this study targets a knowledge gap of evaluating chromatin conservation during land …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
Summary:
The authors characterize chromatin states in the flowering plant Arabidopsis thaliana and the bryophyte Marchantia polymorpha. Here, they draw from ChIP-seq data that was previously published, and from data generated as part of this study, in particular for Marchantia H2.A variants (H2A.X.1, H2A.X.1, H2A.Z, H2A.M.2). The authors compute chromatin states, which enables a comparison over more than 450 million years of land plant evolution. While comparisons of plant chromatin to other species highlighted conservation as well as differences, this study targets a knowledge gap of evaluating chromatin conservation during land plant evolution. The authors investigate a connection between Transcription Factors binding sites and chromatin states. They propose a list of candidate pioneer factors associating with the +1 nucleosome.
Major comments:
For the Association of chromatin states with expression, the authors use the TAIR10 annotation for extracting TSSs and promoter sequences. When investigated, a comparison of data resolving TSS with this annotation (or Araport11) shows a pretty poor overlap between the TSS based on Tair10/Araport11 and experimentally derived TSSs. This information was captured in Arabidopsis genome annotation files where the experimental TSS matches the genome annotation. What is the advantage of using an annotation with the inaccurate TSSs in TAIR10? It seems to confound the study.
The TSS annotation in Marchantia polymorpha (Tak1 v7.1) may also match poorly to the experimentally derived TSS. I suggest that the authors generate data to detect TSS in their tissue of choice and compare the positions to the genome annotation they use (f.x. PMID: 38831668).
I am not convinced that it is a wise choice to utilize fewer ChIP-seq data in Marchantia than Arabidopsis. Can the missing Marchantia ChIP-seq experiments not be performed and included to complete the comparison?
P. 26 onwards, the authors investigate different TF clusters and their association with chromatin states. They state "cluster 1 TFs primarily associated with the first nucleosome downstream of the TSS". However, if the gene is not really expressed in these "leave" tissues, then how can the authors be sure that the same TSS position would be used in "flower" tissue? It could be an artifact of a genome annotation file that misses flower-tissue TSS data. It is not an obvious to conclusion to name these factors "pioneer TFs". Experiments testing this are missing as far as I can gather.
Minor comments:
Can the authors add files ( e.g. .bed) with their segmented chromatin states as part of their GEO submission? That could improve the impact and make the findings more accessible.
Can the authors rule out issues with the Marchantia annotation, for example missing read-through transcription or alternative isoforms, that would essentially have the effect that the genomic segmentation they use contains elongating upstream transcripts in from of promoter TSS? This could be an alternative explanation for the enrichment of H2AUb/H3K36me3 just upstream of the TSSs as they describe on p.21. If it can´t be ruled, the limitations from genome annotations, and examples offering improvements could be highlighted in the discussion. This may also be supported by the long persistence of E4 after the TTS p.23.
P.23 - This further suggests that in Marchantia, the orientation of genes defines
distinct chromatin environment in their vicinity, through mechanisms yet to be uncovered. Does this correlate with the distance of the closest (annotated) transcript pairs?
The E1 state highlighted on p.24 and in Fig.3A/d is not annotated in Fig.3A/D. It is also not clear in the legends which number it is.
P.30 - The marks H3K4me1 and H3K36me3 reflecting transcriptional elongation and confined to the gene bodies in Arabidopsis, extend beyond the TTS in Marchantia, suggesting that signals for transcriptional termination differ between flowering plants and bryophytes. There are multiple alternative explanations. Likely a combination of missing transcripts in their genome annotation (e.g. lncRNAs), annotation errors (e.g. wrong ends) and the segmentation of these regions (e.g. the transcripts are closer than in Arabidopsis). The discussion could extended significantly to address these issues and include the efforts to improve the genome annotations.
Referee cross-commenting
Reviewer #2 raises fair and valuable questions.
Significance
Significance: The authors corroborate prior chromatin state analyses in Arabidopsis and provide a chromatin state analysis for Marchantia. These data represent a resource that will be used and appreciated by the plant and ChromEvoDevo communities. The quality of the analyses are high and the description is transparent. I am not aware of a similar study comparing bryophytes and a land plant, so this study addresses a gap in knowledge.
General assessment: The quality of the manuscript is high. The analyses are described well, and in sufficient detail to be understood. The effort going into documentation is high, I rate the study as reproducible. The linked github deposition looks good. The data generated as part of this study is available in the linked GEO deposition. An experimental design of 2 biological repeats is used, which is OK, but the lower limit. The GEO-deposited .bw files should be of interest to the ChromEvoDevo community, and researchers interested in Marchantia epigenetics and gene expression. The manuscript is written clearly and to the point. The figures condense a lot of data and match the text. The figures are rather complex and not easily accessible to someone browsing through a journal issue. However, that is fine for these types of papers. The manuscript is strong on data analysis. Other approaches, for example mutants to validate their hypothesis, are not utilized. The calculation of chromatin states offers a way to condense complex information into simpler terms. Nevertheless, it re-organizes information that largely existed before. To me, the biggest value of this study appears to be to regard it as a resource that calculated the chromatin states in a comparable fashion between organisms.
Advance: The manuscript provides several advances. It provides new ChIP-seq data for Marchantia, it generates a chromatin state map for Marchantia, it compares Chromatin state maps between distant evolutionary time, and it generates a new hypothesis regarding pioneer TFs in plants. Some of the points described in the article hold true for even larger evolutionary distances, for example comparing plants to yeast and metazoans. The manuscript fills a knowledge gap and has offers a comparison via the computation of comparable chromatin states.
Audience: The audience will be colleagues interested in chromatin and epigenetics, the Marchantia and plant communities as well as researchers interested in EvoDevo of chromatin organization. Even though the study uses plant models, it is highly relevant for non-plant models.
-
