With no eraser, how fast the histone itself is replaced decides how much H3K79 methylation there is - and that sets the clock a DOT1L drug has to beat
also called H3K79 methylation turnover · replication-coupled dilution of H3K79me · nucleosome turnover sets H3K79 valency
Because nothing removes the H3K79 tag, the only way to lose it is to swap out the histone carrying it - either when the cell copies its DNA or as histones are routinely exchanged. That means how much tag a gene has depends on how fast its histones are replaced, and it means a drug blocking the enzyme has to stay on board through several histone turnovers before the tag disappears.
The same thing, for a scientist
H3K79 methylation is the textbook case of a mark with a single writer and **no known demethylase**, which makes it the cleanest system in which to ask how a histone modification is lost. Chory et al. used chemical-induced proximity to show that the **valency** of H3K79 methylation - whether a nucleosome carries me1, me2 or me3 - is determined by **nucleosome turnover rates**, that propagation of the mark is predicted by turnover simulations across the genome, and that turnover accounts for the asymmetric distribution of H3K79me toward the transcriptional unit [E1]. This is a third and distinct model of state determination, alongside enzyme-flux redundancy [E2] and state-specific zoning: here the enzyme's distributive kinetics and the nucleosome's residence time jointly set the distribution. A mass-spectrometric measurement in synchronised human cells complicates it usefully - **old histones continue to acquire K79 mono- and dimethylation at the same rate as newly synthesised ones**, implying positional scrambling through the cell cycle [E3]. For a KMT2A-fusion graph the operational consequence is a number nobody has measured in this disease: how many cell divisions of continuous DOT1L inhibition are required to clear H3K79me2 from *HOXA9*. That number is the dosing schedule, and pinometostat was infused continuously for exactly this reason while never reaching its maximum tolerated dose [E4].
Three models of the same distribution, and what distinguishes them
| model | what sets me1:me2:me3 | testable prediction |
|---|---|---|
| flux / redundancy 2 | total enzyme-time per nucleosome | the three states move together when DOT1L activity changes |
| nucleosome turnover 1 | how long a nucleosome survives to be re-methylated | state ratio tracks turnover rate, measurable independently of DOT1L |
| state-specific zones | dedicated regulation per state | loci hold their state ratio when flux changes |
The turnover model is the most attractive for this disease because it is the only one that makes a prediction about the drug: a locus with fast nucleosome turnover loses the mark fast under inhibition and one with slow turnover keeps it. If leukaemia-defining broad H3K79me2 domains are slow-turnover chromatin, they are the last to clear - which would be the mechanism by which the very loci you most want to affect are the ones the drug reaches last.
Nobody has measured nucleosome turnover at KMT2A-fusion target genes and related it to the rate of mark loss under a DOT1L inhibitor. That is the experiment exp-h3k79-state-resolved-icechip-dose-series proposes in its arm (b).
Why the clock is a clinical variable and was treated as a pharmacology variable
If the mark can only be lost through histone replacement, the time to clear it is a property of the cell, not of the drug. Three things follow that this graph should hold together:
- The trial schedule was reasoned correctly and tested incompletely. Continuous IV infusion is the right shape of answer to a slow mark; a maximum tolerated dose that was never reached means the depth of that continuous exposure was never established 4.
- Quiescent cells are the worst case. A leukaemia-initiating cell that divides rarely replaces its histones rarely and therefore retains H3K79 methylation longest - so the compartment responsible for relapse is the compartment a dilution-limited drug reaches last. This is the sharpest bite-back in the whole beat.
- Recovery is on the same clock. With no eraser, normal marrow that has lost the mark also cannot re-set it quickly, so both the therapeutic effect and the toxicity have long time constants and neither is well described by plasma pharmacokinetics.
The scrambling result 3 argues the clock is not a clean exponential: if old histones keep gaining methyl groups at the same rate as new ones, then the mark does not simply halve per division.
Correction: quiescence removes one hand of the clock, not the clock
This entry argued that clearance of a mark with no eraser tracks nucleosome replacement, and it drew the clinical consequence that a cell that divides rarely replaces its histones rarely. That is too strong, and the measurement exists.
Postmitotic adult mouse cardiomyocytes replace histones with a half-life of about two weeks (PMID 31104571, Replication-Independent Histone Turnover Underlines the Epigenetic Homeostasis in Adult Heart, Circ Res 2019; record fetched and title, journal, year and DOI read back). Replication-independent turnover is a real and substantial route in a cell that never divides.
So the argument survives with its sign intact and its magnitude softened. A quiescent cell clears H3K79 more slowly than a cycling one because it loses the replication-coupled hand; it does not stop clearing. The therapeutic worry is a rate, not a wall.
The caveat that must travel with the number: H2B-GFP pulse-chase measures H2A-H2B exchange, which is faster than H3-H4 exchange, so two weeks is an upper bound on the speed of H3.3 replacement and the true H3 clock in a quiescent cell is slower than this. And it is cardiomyocytes, not a relapse-initiating lymphoblast.
What is genuinely unmeasured is supply rather than division: a quiescent cell has no S-phase histone biosynthetic burst, and nobody has measured the soluble H3-H4 pool in one. If the limiting factor is new histone availability rather than eviction rate, the clock could be much slower than the cardiomyocyte number implies.
And the writer is itself a histone chaperone
Yeast Dot1p has intrinsic histone chaperone activity and drives histone exchange methylation-independently (PMID 29339748, Dot1 regulates nucleosome dynamics by its inherent histone chaperone activity in yeast, Nat Commun 2018; fetched and read back). The abstract is explicit that a methylation-independent function of Dot1p is required for suppressing cryptic transcription.
If that transfers to human DOT1L, it generates a falsifiable prediction this graph did not have: a DOT1L DEGRADER should clear H3K79me2 more slowly than a SAM-competitive inhibitor. A catalytic inhibitor leaves the protein in place to keep driving exchange; a degrader removes both the writer and the exchange activity, and slower exchange means slower dilution of the existing mark.
It is yeast. Yeast Dot1p lacks a SET domain and does not methylate free histones, and the human enzyme has not been tested for chaperone activity. The prediction is worth stating precisely because it is cheap to falsify and because the field is actively building DOT1L degraders.
| mechanism class | mark kinetics / loss mechanism |
|---|---|
| writer | gene-dot1l |
| eraser | none known |
| unmeasured number | cell divisions of continuous DOT1L inhibition required to clear H3K79me2 from HOXA9 in a KMT2A-rearranged line |
| clinical consequence | the clearance clock, not plasma exposure, sets the minimum treatment duration; pinometostat used continuous IV infusion and never reached its maximum tolerated dose |
mark
[ "mark-h3k79me1", "mark-h3k79me2", "mark-h3k79me3" ]
loss mechanisms
[
{
"mechanism": "replication-coupled dilution of the parental histone pool",
"rate_determined_by": "cell division rate"
},
{
"mechanism": "replication-independent nucleosome turnover/exchange",
"rate_determined_by": "locus-specific turnover rate",
"source_pmid": "30472189"
}
]state determination models — the numbers behind the chart
[
{
"model": "nucleosome turnover sets valency",
"pmid": "30472189",
"title": "Nucleosome Turnover Regulates Histone Methylation Patterns over the Genome.",
"authors": "Chory EJ, Calarco JP, Hathaway NA, et al.",
"venue": "Mol Cell",
"year": 2019,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"model": "enzyme flux with distributive kinetics",
"pmid": "18511943",
"title": "Nonprocessive methylation by Dot1 leads to functional redundancy of histone H3K79 methylation states.",
"authors": "Frederiks F, Tzouros M, Oudgenoeg G, et al.",
"venue": "Nat Struct Mol Biol",
"year": 2008,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"model": "state-specific zones",
"pmid": "41925229",
"title": "Trans-histone crosstalk establishes distinct H3K79 methylation zones with differential transcriptional functions.",
"authors": "Park NH, Kim HR, Kim HY, et al.",
"venue": "Nucleic Acids Res",
"year": 2026,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
}
]- A dilution-limited mechanism is weakest against non-dividing cells, so the quiescent leukaemia-initiating population that drives relapse is the one a DOT1L inhibitor clears last - the opposite of the selectivity a stem-cell-directed therapy needs.
- With no eraser, normal haematopoietic recovery after the drug is stopped is also slow, so toxicity persists on the same long time constant as efficacy and neither tracks plasma levels.
- If nucleosome turnover rate rather than enzyme activity sets the mark, then anything that slows turnover - quiescence, chromatin compaction, differentiation arrest - is a drug-resistance mechanism that involves no change to DOT1L at all.
- How many divisions of saturating DOT1L inhibition are needed to clear H3K79me2 at HOXA9 in a KMT2A-rearranged line, measured by calibrated ChIP? Nobody has published this number and it is the dosing schedule.
- Are the broad H3K79me2 domains over HOXA in leukaemia lines slow-turnover chromatin? If so the loci the drug most needs to reach are the last to clear.
- Do quiescent KMT2A-rearranged leukaemia-initiating cells retain H3K79me2 through a course of DOT1L inhibition that clears it from cycling blasts?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
H3K79 methylation is produced by a single methyltransferase, Dot1l, with no known demethylase; using chemical-induced proximity, the valency of H3K79 methylation (mono-, di- and tri-) is determined by nucleosome turnover rates, propagation of the mark across the genome is predicted by nucleosome turnover simulations, and turnover accounts for the asymmetric distribution of H3K79me toward the transcriptional unit. A meta-analysis of other conserved histone modifications shows turnover models predict both valency and chromosomal propagation of methylation marks.
Nucleosome Turnover Regulates Histone Methylation Patterns over the Genome. · Chory EJ, Calarco JP, Hathaway NA, et al. · Mol Cell 2019What this evidence cannot show
Abstract read from the eutils record; full text not read. PubMed types this record as a Meta-Analysis as well as a research article, and the cross-species part (worms, flies, mice) is indeed a meta-analysis of published data rather than new measurement. The core H3K79 experiments use chemical-induced proximity to force a methyltransferase to a locus, which is an artificial recruitment regime, and the propagation conclusions rest on simulation fitted to those data. Not a leukaemic system and no KMT2A fusion.
Evidence that argues against this (1)
- A turnover-only model cannot explain a locus where the enzyme is tethered by an oncogenic fusion at high local concentration: there, re-methylation after histone exchange is near-immediate and the state distribution should decouple from turnover rate. Graded weakening of the DOT1L-AF9 interaction changes the mark at target genes without changing nucleosome turnover, which is the observation a pure turnover model does not accommodate. Kuntimaddi A et al. 2015
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Dot1 introduces multiple methyl groups by a non-processive mechanism, so the H3K79 methylation states cannot be generated independently; yeast Dot1-dependent silencing relied on global H3K79 methylation levels rather than on one state, and multiple methylation is expected to limit the possibilities for regulation by putative demethylases or binding proteins.
Nonprocessive methylation by Dot1 leads to functional redundancy of histone H3K79 methylation states. · Frederiks F, Tzouros M, Oudgenoeg G, et al. · Nat Struct Mol Biol 2008What this evidence cannot show
Abstract read from the eutils record; full text not read. Yeast Dot1 and yeast silencing; the redundancy conclusion is drawn from one phenotype in one organism. Cited here to contrast the enzyme-flux model with the turnover model, not to settle between them.
Evidence that argues against this (1)
- The turnover model explains the same state distribution without invoking flux at all, and explains something the flux model does not - the asymmetric distribution of H3K79me toward the transcriptional unit. Two mechanisms that predict the same observable are not both needed, and the field has not chosen. Chory EJ et al. 2019
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In synchronised human cells labelled with stable isotopes on both histones and their methyl groups, and resolved by H3 isoform, pre-existing histones continue to be K79 mono- and di-methylated at a rate equal to that of newly synthesised histones, implying some degree of positional scrambling of K79 methylation through the cell cycle; H3.3K79me1 levels are higher than H3.2/H3.1 but the rate of establishing K79 methylation is the same for all three isoforms.
Kinetics of re-establishing H3K79 methylation marks in global human chromatin. · Sweet SMM, Li M, Thomas PM, et al. · J Biol Chem 2010What this evidence cannot show
Abstract read from the eutils record; full text not read. Bulk chromatin in a synchronised human cell line; no locus resolution and no trimethylation kinetics in the abstract. 'Scrambling' is an interpretation of equal labelling rates rather than an observed redistribution. Cell-cycle synchronisation itself perturbs chromatin metabolism.
Evidence that argues against this (1)
- Equal labelling rates on old and new histones are exactly what nucleosome turnover predicts without any positional scrambling: if nucleosomes are exchanged several times per cell cycle, an 'old' histone in bulk chromatin is not a histone that has stayed at one position, and the measurement cannot distinguish redistribution of methyl groups from redistribution of histones. Chory EJ et al. 2019
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Pinometostat was administered by continuous intravenous infusion, reduced H3K79 methylation in patients and produced 2 complete remissions in 51 adults with acute leukaemia.
The DOT1L inhibitor pinometostat reduces H3K79 methylation and has modest clinical activity in adult acute leukemia. · Stein EM, Garcia-Manero G, Rizzieri DA, et al. · Blood 2018What this evidence cannot show
Abstract read from the eutils record; full text not read. The maximum tolerated dose was not reached. The claim in this entry that the continuous schedule was chosen because of the mark's slow kinetics is a reading of the design, not a statement made in the abstract.
Evidence that argues against this (1)
- A slow-clearance argument predicts that longer exposure helps, but modelled acquired resistance to DOT1L inhibition in KMT2A-rearranged ALL shows the leukaemia adapts transcriptionally under sustained exposure, so extending exposure is not monotonically beneficial and may select the adaptation that defeats it. Schneider P et al. 2023
- E-heart-turnoverestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.journal-article
Replication-independent histone turnover in postmitotic adult mouse cardiomyocytes, measured by H2B-GFP pulse-chase, gives a histone half-life of about two weeks, so histone replacement continues substantially in a cell that never divides.
Replication-Independent Histone Turnover Underlines the Epigenetic Homeostasis in Adult Heart · Li Y, ..., He A · Circulation Research 2019What this evidence cannot show
Record fetched 2026-09-20 and title, journal, year and DOI read back; **the full text was not read** and the half-life is taken from the abstract's reported result. Established as a result within the reported system. **H2B-GFP measures H2A-H2B exchange, which is faster than H3-H4**, so this is an upper bound on H3.3 replacement speed. Cardiomyocytes are not lymphoblasts and nothing here is leukaemia.
Evidence that argues against this (1)
- This softens a claim in this entry rather than overturning it, and the direction still favours the original worry: a quiescent cell clears more slowly than a cycling one. What remains genuinely unmeasured is the soluble H3-H4 pool in a quiescent cell - if supply rather than eviction is limiting, the clock could be far slower than two weeks. Lee S et al. 2018
- E-dot1-chaperoneestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.journal-article
Yeast Dot1p displays histone chaperone activity and regulates nucleosome dynamics via histone exchange, and a methylation-independent function of Dot1p is required for suppressing cryptic transcription within transcribed regions.
Dot1 regulates nucleosome dynamics by its inherent histone chaperone activity in yeast · Lee S, ..., Lee D · Nature Communications 2018What this evidence cannot show
Fetched 2026-09-20, title and journal read back; abstract only. Established as a result in **yeast**. Yeast Dot1p lacks a SET domain and does not methylate free histones, so it is not simply a smaller DOT1L, and **no chaperone activity has been demonstrated for the human enzyme**. The degrader-versus-inhibitor prediction that follows is an inference beyond these numbers and is graded as such wherever it is stated.
Evidence that argues against this (1)
- If human DOT1L has no chaperone activity, the prediction inverts to a null result that would look like a failed experiment rather than a informative one - which is an argument for testing it as a stated prediction rather than discovering it inside a combination study. Li Y et al. 2019
replication-independent arm of the histone replacement that clears a mark with no eraser
- ← supportsHistone H3.3 - one protein, two genes, and the reason this graph should not file it as gene-h3f3a
the replacement histone whose exchange rate sets the clearance clock
replication-coupled arm: per-division dilution of parental histones
supplies the machinery and qualifies the quiescence claim
nucleosome turnover rate sets methylation valency in the absence of an eraser
independent 2019 statement that H3K79me has no known demethylase
- related to →Pinometostat (EPZ-5676) — the DOT1L inhibitor, and the most instructive failure in this field
the clearance clock, not plasma exposure, sets the minimum treatment duration
- related to →DOT1L methylates H3K79 non-processively - so the three states are one distribution, not three signals
rival model of what sets the me1:me2:me3 distribution
arm (b): clearance half-life against nucleosome turnover, per locus
identifies what the field's belief in a long half-life actually rests on: the absence of a known demethylase, cited to a review, not a measured rate
a 4-24 h window under a menin inhibitor is shorter than one division of these cells, so a share change at the fusion's sites inside it is not replicative dilution; but a within-sample instrument cannot see uniform loss, so dilution after 24 h is untouched
a 52-91x fall over nine days is the quantity a dilution-plus-turnover model has to account for
4-6 divisions by day 9 predicts 16-64x by passive dilution; measured 52-91x
bounds that entry's unmeasured_number with a dated, controlled query