iiinova — where your digital self keeps evolving

The Lewis Lab @peterlewislab.bsky.social · 08/06/2026

Allosteric activation of PRC2 is coupled to adoption of the compact state. Our new bioRxiv manuscript shows that EZH2 S21-phos, adjacent to the SANT1 domain, restrains this conformational transition and limits activation. With @kasinath-lab.bsky.social & @garcialabms.bsky.social tinyurl.com/4369kmh2

021

The Lewis Lab @peterlewislab.bsky.social · 05/06/2026

How do cells tune PRC2 allostery? We find EZH2 Ser21-phos restrains PRC2 adoption of the compact active state. Loss of this restraint promotes H3K27me3, redistributes cPRC1, and impairs differentiation. New bioRxiv with @kasinath-lab.bsky.social & @garcialabms.bsky.social! tinyurl.com/4369kmh2

13312

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

These separation-of-function mutants show that H3.3 deposition can be genetically uncoupled from ERV silencing. This supports our earlier finding that H3.3 deposition at ERVs is dispensable for repression. Instead, repression depends on the C-terminal SIM and interaction with SUMOylated effectors.

010

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

Which DAXX functions are required for ERV silencing? DAXX mutants defective in ATRX binding or H3.3 deposition still restore ERV repression. By contrast, deletion of the C terminal SIM abolishes silencing. Thus, ERV repression requires the C terminal SIM, but not ATRX binding or H3.3 deposition.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

One SUMOylated protein that binds the DAXX C-terminal SIM is the ATPase MORC3. Removing the SIM abolishes the DAXX-MORC3 interaction and prevents MORC3 recruitment to ERVs. MORC3 knockout cells likewise lose H3.3 accumulation at these loci.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

We previously showed that the DAXX HBD alone is insufficient for H3.3 deposition. Deleting the final eight amino acids, which removes the C-terminal SIM, blocks H3.3 accumulation at ERVs without affecting DAXX localization.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

Previously, we demonstrated that the DAXX four-helix bundle mediates ATRX interaction. But point mutations in this domain, or deletion of it, do not disrupt DAXX binding to ERVs or H3.3 accumulation at these sites.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

The HBD basic patch is required for productive nucleosome assembly in vitro. In cells, basic patch mutants still localize to ERVs but do not restore H3.3 accumulation at those sites.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

We identified a conserved basic patch within the tower domain of the DAXX histone-binding domain. Using recombinant HBD-H3.3-H4 complexes, we found that mutation of this surface weakened interaction with DNA in vitro.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

In our study, we asked how DAXX promotes H3.3 deposition mechanistically. Genome-wide, DAXX is not widely distributed across repeats. It is concentrated at a restricted subset of evolutionarily young ERVK elements, where it restrains ERV expression and influences nearby genes.

110

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

A major challenge in interpreting H3.3 knockout phenotypes is that loss of H3.3 affects more than chromatin incorporation. It also causes a marked drop in DAXX abundance. Thus, ERV derepression in H3.3-null cells could reflect loss of H3.3 nucleosome assembly, loss of DAXX, or both.

100

The Lewis Lab @peterlewislab.bsky.social · 23/03/2026

New preprint from our lab! An H3.3 knockout does two things at once: it removes H3.3 from chromatin and destabilizes DAXX. We disentangle those functions and find that DAXX-mediated H3.3 deposition can be uncoupled from ERV silencing. www.biorxiv.org/content/10.6...

24820

The Lewis Lab @peterlewislab.bsky.social · 22/03/2026

The paper closes with a model: Gcn5-containing HAT complexes are targeted to active genes by the transcription machinery, providing a direct link between histone acetylation and gene activation. A fitting paper to revisit 30 years later, on what would have been Dave Allis's 75th birthday, March 22.

060

The Lewis Lab @peterlewislab.bsky.social · 22/03/2026

And then decisive experiment. Recombinant Gcn5p expressed in E. coli had HAT activity in the in-gel assay, whereas uninduced and vector-only controls lacked activity. With that, Gcn5p was assigned a direct biochemical activity, linking histone acetylation more directly to transcriptional activation.

120

The Lewis Lab @peterlewislab.bsky.social · 22/03/2026

The conceptual surprise was in the sequence. Tetrahymena p55 turned out to be highly homologous to yeast Gcn5p, a known transcriptional co-activator: ~40% identity, ~60% similarity, with a conserved bromodomain.

110

The Lewis Lab @peterlewislab.bsky.social · 22/03/2026

Then came the validation. Anti-peptide antibodies detected a single 55 kDa species, tracked with the active RP-HPLC fractions, and immunodepleted ~60% of HAT activity. p55 was not just associated with HAT A. It was the catalytic subunit.

110

The Lewis Lab @peterlewislab.bsky.social · 22/03/2026

Starting from the previously purified active p55 HAT A subunit, Brownell et al. sequenced six internal peptides, designed degenerate primers, and assembled the cDNA by PCR and RACE. The predicted protein sequence then recovered all six peptides from the purified enzyme.

120

The Lewis Lab @peterlewislab.bsky.social · 22/03/2026

Happy 30th birthday to this paper published March 22, 1996, one of the studies that moved chromatin biology from correlation toward mechanism. Brownell et al. linked a transcription-associated Tetrahymena HAT to yeast Gcn5p, providing a biochemical link for chromatin modification to gene activation.

24822

The Lewis Lab @peterlewislab.bsky.social · 07/01/2026

Nice structural and biochemical dissection of how ATRX engages nucleosomes and couples ATP hydrolysis to productive sliding via a distinctive mechanism. www.biorxiv.org/content/10.6...

092

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

VEFS: the NBD is gone. RNA binding is rare (small red bubble). Most PRC2 is RNA-free, catalytically active, and binds both CpG and non-CpG sites.  Result: diffuse H3K27me3 that dilutes PRC1 away from its normal targets, disrupting gene silencing. Read more in our preprint: tinyurl.com/4yrwftrn

110

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

ncMUT: PRC2 still binds RNA efficiently (large red bubble, large red arrows), so catalysis stays limited. But without accessory subunits, PRC2 targeting is unfocused. Result: most PRC2 remains in the RNA-bound state, with little productive methylation at CpG islands.

100

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

Wild-type SUZ12: much of PRC2 can interact with RNA (large red bubble), which blocks engagement with nucleosomes. Accessory subunits (PRC2.1/2.2) target CpG islands. Result: strong H3K27me3 at intended sites, little elsewhere. Arrow size = rate. Bubble size = how much PRC2 is in that state.

100

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

PRC2 silences genes by adding H3K27me3 at the right spots in the genome. Our model shows how the nucleic acid-binding domain (NBD) of SUZ12 keeps this activity under control, and what happens when you remove it.

1268

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

VEFS: the NBD is gone. RNA binding is rare (small red bubble).  Most PRC2 is RNA-free (large green bubble), catalytically active, and binds everywhere. Result: diffuse H3K27me3 that pulls PRC1 away from its normal targets, disrupting gene silencing. Read more in our preprint: tinyurl.com/4yrwftrn

000

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

ncMUT: PRC2 still binds RNA efficiently (large red bubble, large red arrows), so catalysis stays limited.  But without accessory subunits, PRC2 targeting is unfocused.  Result:most PRC2 remains in the RNA-bound state, with little productive methylation at CpG islands.

100

The Lewis Lab @peterlewislab.bsky.social · 13/08/2025

Wild-type SUZ12: much of PRC2 can interact with RNA (large red bubble), which blocks engagement with nucleosomes. Accessory subunits (PRC2.1/2.2) target CpG islands.  Result: strong H3K27me3 at intended sites, little elsewhere. Arrow size = rate. Bubble size = how much PRC2 is in that state.

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Our results support a model where SUZ12-NBD interactions with RNA act as a kinetic buffer, limiting PRC2 activity and ensuring specific targeting to CpG islands. Loss of the NBD (VEFS) disrupts this equilibrium, driving widespread, non-specific H3K27me3 deposition. 13/13

020

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

PRC2 complexes containing the SUZ12-NBD strongly bind diverse forms of nucleic acids, including structured RNA and double-stranded DNA, significantly inhibiting enzymatic activity. PRC2 lacking the NBD (VEFS) is resistant to nucleic acid-mediated inhibition. 12/13

130

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Domain analysis identifies a discrete nucleic acid-binding domain (NBD) within SUZ12 responsible for constraining PRC2 activity. CryoEM studies showed the NBD in close proximity to nucleic acids. Loss of this region (NEC, VEFS) causes global H3K27me3 hypermethylation. 11/13

132

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Widespread, diffuse H3K27me3 deposition induced by VEFS expression titrates canonical PRC1 away from key genomic targets. This dispersal of PRC1 disrupts Polycomb domains, resulting in transcriptional derepression and upregulation of genes normally silenced by PRC1. 10/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Expression of the VEFS truncation globally elevates H3K27me3 independently of non-core accessory subunits. This increase occurs across multiple cell types, including both K27M-mutant DMG cells and wild-type H3 cells, highlighting intrinsic SUZ12 regulatory mechanisms. 9/13

110

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

In vitro and in vivo data confirm PRC2 subcomplexes collectively support K27M-DMG proliferation. Disrupting PRC2 recruitment (ncMUT, VEFS) dramatically reduces tumor growth, and remarkably, VEFS expression alone leads to 100% survival in orthotopic xenograft models. 8/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

EZH2 inhibition (Tazemetostat) or disruption of PRC2 subcomplexes (shSUZ12, ncMUT, VEFS) derepresses CDKN2A, causing significant growth defects. Knockout of CDKN2A rescues DMG cells from these proliferation defects, highlighting its critical role downstream of PRC2. 7/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

RNA-seq data show that disrupting PRC2 subcomplexes (ncMUT, VEFS) in DMG cells significantly derepresses genes critical for neurodevelopment. PRC2.1/2.2 complexes thus maintain silencing of developmental regulators essential for tumor cell identity. 6/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

ChIP-seq analyses reveal PRC2.1 and PRC2.2 subcomplexes act collectively to recruit PRC2 to critical genomic loci in K27M DMGs. Mutants lacking these accessory subunits (ncMUT, VEFS) show severely impaired PRC2 localization at key target genes, including CDKN2A. 5/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

To investigate PRC2 targeting, we disrupted PRC2 subcomplexes using SUZ12 mutants. VEFS truncation removes all accessory subunits, while the non-core mutant (ncMUT) selectively disrupts PRC2.1 and PRC2.2 subcomplexes, preserving core PRC2 integrity. 4/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Recent studies show non-core (accessory) subunits are necessary for PRC2 targeting in mammalian cells. These subunits, interacting via SUZ12, guide PRC2 precisely to CpG islands, ensuring accurate H3K27me3 deposition and gene silencing. 3/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Diffuse midline gliomas (DMGs) carrying the H3 K27M inhibitor histone exhibit global H3K27me3 reduction but retain focal enrichment at PRC2-bound CpG islands. Understanding how this residual methylation is maintained and its functional role is a critical unresolved question. 2/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

We are excited to share our new preprint demonstrating that nucleic acid interactions with SUZ12 constrain PRC2 activity, establishing a kinetic buffer essential for targeted gene silencing and revealing vulnerabilities in diffuse midline gliomas. www.biorxiv.org/content/10.1...

14819

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Domain analysis identifies a discrete nucleic acid-binding domain (NBD) within the SUZ12 N-terminus responsible for constraining PRC2 activity. CryoEM studies previously showed the NBD in close proximity to nucleic acids. Loss of this region (NEC, VEFS) causes global H3K27me3 hypermethylation. 11/13

000

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

110

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Expression of the VEFS truncation globally elevates H3K27me3 independently of non-core accessory subunits. This substantial increase occurs across multiple cell types, including both K27M-mutant DMG cells and wild-type H3 cells, highlighting intrinsic SUZ12 regulatory mechanisms. 9/13

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

110

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

100

The Lewis Lab @peterlewislab.bsky.social · 24/07/2025

Diffuse midline gliomas (DMGs) carrying the H3 K27M inhibitor histone exhibit global loss of H3K27me3 but retain focal enrichment at PRC2-bound CpG islands. Understanding how this residual H3K27me3 is maintained, and its functional significance, may reveal critical therapeutic vulnerabilities. 2/13

100

The Lewis Lab @peterlewislab.bsky.social · 16/06/2025

It was a privilege to be part of the Allis lab. Dave fostered an environment of rigor, curiosity, and kindness- his impact lives on in the many he mentored. Grateful to reconnect at Rockefeller with this extraordinary group. Thank you to the organizers for an inspiring symposium.

0112