RINGs, DUBs and Abnormal Brain Growth—Histone H2A Ubiquitination in Brain Development and Disease

Round 1

Reviewer 1 Report

Illingworth and group have written an excellent review covering histone ubiquitination and de-ubiquitination pathway in brain development and dysregulation of this axis in neurological disorders.  Overall, it is a very original, comprehensive review and useful to wider readers. 

Author Response

Thank you for taking the time to appraise our review and provide us with such positive feedback. 

Author Response File: Author Response.pdf

Reviewer 2 Report

This is a well written, highly informative review on the role of Polycomb repressive complexes 1 and 2 and H2AK119 ubiquitination on brain development and disease.

Towards improving the review, the authors should do the following.

1. The title should be revised to

RINGs, DUBs and Abnormal Brain Growth – Histone H2A Ubiquitination in Brain Development and Disease

Many other histones are ubiquitinated and this is not the focus of this study.

2. The authors should comment on the following

- As the nucleosome has two H2As, is it known whether ubiquitination event ubiquitinates both H2As in a nucleosome or just one? If this has not been tested, the authors should say so.

- H2A can also be multi-ubiquitinated. Which enzymes do this?

- the authors need to comment on bivalent histone modifications (e.g. PMID: 34930904, 28793256)

- has it been demonstrated that H3K4me3 and H3K27me3 are on the same histone tail of a bivalently marked nucleosome?

- Are these two histone PTMs in the same nucleosome; has this been tested?

- Is H2AK119Ub also a mark of bivalent marked regions?

Author Response

Thank you for taking the time to appraise our review and provide us with such positive feedback. Please find our specific responses to your suggestions below.

Major point 1 - The title should be revised to - RINGs, DUBs and Abnormal Brain Growth – Histone H2A Ubiquitination in Brain Development and Disease. Many other histones are ubiquitinated and this is not the focus of this study.

We agree and the title has been changed as suggested.

Major point 2 - The authors should comment on the following

- As the nucleosome has two H2As, is it known whether ubiquitination event ubiquitinates both H2As in a nucleosome or just one? If this has not been tested, the authors should say so.

- H2A can also be multi-ubiquitinated. Which enzymes do this?

- the authors need to comment on bivalent histone modifications (e.g. PMID: 34930904, 28793256)

- has it been demonstrated that H3K4me3 and H3K27me3 are on the same histone tail of a bivalently marked nucleosome?

- Are these two histone PTMs in the same nucleosome; has this been tested?

- Is H2AK119Ub also a mark of bivalent marked regions?

Bivalency is an interesting phenomenon and worthy of an entire review in its own right, however as it is established by PRC2 and members of the trithorax family of complexes it is somewhat independent of PRC1 and PR-DUB (and therefore H2AK119ub). Given the focus of this review, we feel that including a detailed commentary on bivalency would detract from the primary focus, however we have tried to address these points as fully as possible in an additional paragraph in the discussion section (Lines 544-558):

‘Protein ubiquitination is used as a control mechanism for a wide array of cellular processes, the mechanics of which relate, in part, to the linkage chemistry of the underlying ubiquitin conjugate [179]. With respect to PRC1 function, modification of H2A (and H2AZ/H2Av) is exclusively monomeric, however whilst K119 is the favoured target of RING1A/B (K120 in H2AZ), additional proximal lysines can also be modified, albeit at lower frequency (K118 in H2A; K121 and K125 in H2AZ) [8,9,81,180-184]. Structural analysis suggests that modification is restricted to monoubiquitination due to the con-strained conformation of the chromatin-E2:E3 interface and steric hindrance with the monomeric ubiquitin product [180,182]. Current experimental evidence does not however exclude the possibility that both H2A copies are simultaneously ubiquitylated within a single nucleosome. PRC1 and H2A/H2AZ monoubiquitination are thought to be primarily repressive, however the identification of dually modified H2AZ bearing both ubiquitylation and acetylation (acH2AZub1) provides an interesting parallel to bivalent nucleosomes (coexistence of H3K4me3 and H3K27me3 within a single nucleosome) [181,185,186].’

 

Author Response File: Author Response.pdf

Reviewer 3 Report

In this article, Doyle and colleagues perform an excellent review about the role of polycomb group complexes in neurodevelopmental disorders, with a focus on the importance of maintaining appropriate levels of H2AK119ub1 by canonical PRC1 and PR deubiquitinase. The first describe the composition and functions of several polycomb repressive complexes, including canonical and non-canonical PRC1 and PR-DUB. Then, after a short section about the role of polycomb complexes in neurogenesis, the focus on the alterations of different components of cPRC1, ncPRC1 and PR-DUB that lead to a variety of neurodevelopmental disorders. They end the manuscript with an interesting discussion about the mechanism of gene repression mediated by H2AK119ub1 and the importance of the balance of H2AK119ub1 levels to ensure proper gene regulation and avoid pathological situations.

            I find this review very interesting, well written and structured, and therefore an excellent resource for other researchers focused on gene regulation.

Author Response

Thank you for taking the time to appraise our review and provide us with such positive feedback. 

Author Response File: Author Response.pdf

Reviewer 4 Report

Doyle et al, present a well-written and insightful overview of the role of PRC1 and PR-DUB in neurodevelopment. The review details well the function and mechanisms of these complexes, and their partner complex PRC2, in healthy cells/tissue. There have been significant developments into our understanding of the mechanisms of these complexes in the last 2-4 years, therefore this review is timely. Furthermore, PRC1 and PR-DUB are often mutated in neurodevelopmental syndromes, such as Shukla-Vernon, AUTS2, Bohring Opitz, Bainbridge Ropers and Shasi-Pena syndromes. The aetiologies of such syndromes tend to overlap. This review links clinical phenotypes of patients with these syndromes to the aforementioned mechanisms and poses intriguing questions about future research on PRC1 and PR-DUB in the area of neurodevelopment and transcriptional regulation.

 

Major

 

·       Figure 1. While YY1 was initially reported to interact with BAP1, this interaction has not been observed/replicated in several mass spectrometry experiments from multiple groups (Campagne et al, 2019, Kloet et al, 2016, Hauri et al, 2016, Conway et al, 2021). On the other hand, KDM1B (LSD2) has been successfully replicated as a PR-DUB member and would be worthy of inclusion in figure 1 instead of YY1.

 

·       Figure 2. As the authors state in the text immediately below figure 2, there is little evidence that PRC1.2 or PRC1.4 contribute to H2AK119ub1 catalysis. Therefore, the arrow in figure 2 showing cPRC1 can catalyse H2AK119ub1 should be removed.

 

·       Line 160-164. It is not clear what role the authors are referring to when they say that cPRC1 can recruit Polycomb complexes to chromatin. cPRC1 is considered to be the last of the complexes to be recruited in the currently understood model, including in figure 2.

 

·       Line 188-190- While mice with hypomorphic RING1B may be viable during early development, these mice do have a perinatal lethal phenotype which is not insignificant and should be mentioned.

 

·       Line 382-383- A caveat should be added that this inhibition of RING1B by CK2 enzymes has only been shown in vitro to date as it has not been confirmed by other research groups. This is particularly important as several groups (such as Fursova et al 2019) have shown that PCGF3+5 are the most catalytically active forms of PRC1 in mESCs. Importantly, it has been shown by mass spec that CK2 interacts with PRC1 in the same cell context therefore the in vitro data do not fit well with mESC data.

 

Minor

 

·       Figure 2 could be improved by labelling the respective PRC2 subcomplexes shown as PRC2.1 and PRC2.2, and briefly explaining the subunit differences in the text or legend.

·       Mouse Embryonic Stem Cells are first mentioned on line 105, so the abbreviation should be defined here, not later on as is the case currently.

·       Line 365- BCORL1 is spelled incorrectly.

• Line 167- ‘RING and’ is missing from the full description of the RYBP acronym (RING and YY1 Binding protein).
• Line 178- The authors use the terminology PRC1.1 to refer to complexes containing PCGF1 and similar for the other sub-complexes. They should define that this is the sub-complexes naming system at its first use.
• Table 1: Bohring-Opitz is abbreviated as BOPS here instead of BOS which is the abbreviation used throughout the review. One should be used consistently throughout.
• Line 528: change vPRC1 to ncPRC1 to remain consistent with nomenclature in the rest of the text.

 

Author Response

Thank you for taking the time to appraise our review and provide us with such positive feedback. Please find our specific responses to your suggestions below.

Major point 1 - Figure 1. While YY1 was initially reported to interact with BAP1, this interaction has not been observed/replicated in several mass spectrometry experiments from multiple groups (Campagne et al, 2019, Kloet et al, 2016, Hauri et al, 2016, Conway et al, 2021). On the other hand, KDM1B (LSD2) has been successfully replicated as a PR-DUB member and would be worthy of inclusion in figure 1 instead of YY1.

This is an excellent point and YY1 has been removed from Figure 1 and the text. KDM1B was mentioned in the text originally but had been missed from the figure in error and has now been included in Figure 1. The legend  has been updated to reflect this change.

Major point 2 - Figure 2. As the authors state in the text immediately below figure 2, there is little evidence that PRC1.2 or PRC1.4 contribute to H2AK119ub1 catalysis. Therefore, the arrow in figure 2 showing cPRC1 can catalyse H2AK119ub1 should be removed.

We agree that canonical PRC1 contributes little to PRC1-mediated ubiquitination, however there remains low levels of H2AK119ub and some genomic sites have some limited ubiquitination, presumably dependent on PRC1.2 (Fursova 2019 – Supplemental figure 5). Also, it remains unclear if PRC1.4, in contexts where PCGF4 is more abundant (e.g. neural lineages), contributes more significantly to the H2AK119ub pool. As such, we have not removed the arrow entirely but made it substantially smaller and highlighted the reason for this in the figure legend. We have substantially reworked Figure 2 and its legend to address this and other comments (Lines 152-167).  

Major point 3 - Line 160-164. It is not clear what role the authors are referring to when they say that cPRC1 can recruit Polycomb complexes to chromatin. cPRC1 is considered to be the last of the complexes to be recruited in the currently understood model, including in figure 2.

We agree with the reviewer that this statement was unclear and misleading. As such it has been rephrased as follows (Lines 146-149): ‘Whilst the exact contribution of cPRC1 to developmental gene regulation remains unclear, repression is likely reliant on its ability to both stabilise PRC1 recruitment onto chromatin and to nucleate chromosomal interactions between target sites in the genome [37,42,54].’

Major point 4 - Line 188-190- While mice with hypomorphic RING1B may be viable during early development, these mice do have a perinatal lethal phenotype, which is not insignificant and should be mentioned.

We agree, and the text has been changed as follows to reflect this (Lines 191-194): ‘Hypomorphic RING1B mice (Ring1bI53A/I53A) die perinatally, however, as for Drosophila bearing a functionally equivalent mutation in the RING1B homolog Sce (SceI48A/I48A), they do not present with the earlier embryonic developmental deficits associated with RING1B deficiency [79,81].’

Major point 5 - Line 382-383- A caveat should be added that this inhibition of RING1B by CK2 enzymes has only been shown in vitro to date as it has not been confirmed by other research groups. This is particularly important as several groups (such as Fursova et al 2019) have shown that PCGF3+5 are the most catalytically active forms of PRC1 in mESCs. Importantly, it has been shown by mass spec that CK2 interacts with PRC1 in the same cell context therefore the in vitro data do not fit well with mESC data.

This is a very good point and has been addressed in the final section as follows (Lines 637-640):’ It should also be noted however, that inhibition of RING1B by CK2 has only been demonstrated in vitro, and interaction between PRC1.3/1.5 and CK2 in mESCs does not lead to reduced H2AK119ub deposition [27,54,129]. Further experiments are therefore required to resolve these seemingly discrepant observations.

And also

(Lines 398-401) Consistent with this observation, and contrary to the canonical view of PRC1 as a repressor, AUTS2 appears to have a role in transcriptional activation via recruitment of the co-factors casein kinase 2 (CK2) and P300 [129]

 

Minor points  

Figure 2 could be improved by labelling the respective PRC2 subcomplexes shown as PRC2.1 and PRC2.2, and briefly explaining the subunit differences in the text or legend.

We agree and have substantially reworked Figure 2 and its legend to address this and other comments.

Mouse Embryonic Stem Cells are first mentioned on line 105, so the abbreviation should be defined here, not later on as is the case currently.

Corrected (Line 106) – now defined at first use and not as it was in the original manuscript.

Line 365- BCORL1 is spelled incorrectly.

Now corrected (Line 384).

Line 167- ‘RING and’ is missing from the full description of the RYBP acronym (RING and YY1 Binding protein).

Now corrected (Line 68).

Line 178- The authors use the terminology PRC1.1 to refer to complexes containing PCGF1 and similar for the other sub-complexes. They should define that this is the sub-complexes naming system at its first use.

This is a very good point and has now been clarified by the addition / modification to the text as follows (Lines 93-97): “The resulting PRC1 sub-complexes are denoted as PRC1.1-6 to indicate the inclusion of PCGF1-6 respectively. Biochemical studies have identified additional PcG subunits which allow the classification of these PRC1s into cPRC1 (PRC1.2 and 1.4) and ncPRC1 (PRC1.1, 1.3, 1.5 and 1.6) forms with distinct chromatin targeting and regulatory functions [24,27,28]”

Table 1: Bohring-Opitz is abbreviated as BOPS here instead of BOS which is the abbreviation used throughout the review. One should be used consistently throughout.

Agree and now corrected both in the table and the main text (Lines 480 and 497).

Line 528: change vPRC1 to ncPRC1 to remain consistent with nomenclature in the rest of the text.

Now corrected.

 

Author Response File: Author Response.pdf