H3K79me3 - the third methyl group, the one that needs the most enzyme flux and the one human DOT1L makes least of
also called H3K79 trimethylation · H3K79me3 · histone H3 lysine 79 trimethylation
H3K79me3 is the fully methylated version of the same histone position, and it is the hardest state for the enzyme to reach because each methyl group has to be added in a separate binding event. In human cells it is the rarest of the three states, which makes it the first thing an inhibitor removes and a poor thing to measure a drug against.
The same thing, for a scientist
H3K79me3 is the terminal state of a **distributive** reaction: Dot1 enzymes across yeast, trypanosome and human share a distributive (non-processive) mechanism but have highly diverged catalytic properties and generate very different steady-state methylation patterns [E1]. Because each successive methyl group requires a fresh binding event, the trimethyl state sits at the far end of a flux-dependent distribution and is therefore the most sensitive state to any reduction in enzyme activity, recruitment or SAM availability. Three consequences follow for this project. **First**, H3K79me3 is the earliest casualty of partial DOT1L inhibition, so a pharmacodynamic assay reading me3 will report target engagement at doses that leave most H3K79 methylation in place. **Second**, its dependence on upstream H2B ubiquitination is the strongest of the three: loss of Rad6 causes the complete loss of the yeast H3K79me3 zone, converting it into an H3K79me1-enriched region [E2]. **Third**, its transcriptional polarity is genuinely disputed - in human CD4+ T cells H3K79me3 correlates with repression [E3], while in yeast Dot1-activated genes are the ones enriched for H3K79me3 [E2]. The graph should not assert a polarity for this state.
Why distributive kinetics make me3 the fragile state
A processive methyltransferase holds its substrate and adds methyl groups in succession, so the three states are not independent readouts of activity. A distributive enzyme releases the nucleosome between additions, so reaching me3 requires three productive encounters and the steady-state ratio of me1:me2:me3 is set by the rate constants and by how much enzyme-time each nucleosome gets 1.
That makes me3 the state most strongly amplified by enzyme flux, and the state most strongly suppressed when flux falls. It also means the three states cannot be treated as three independent biomarkers of the same process: they are three points on one distribution.
The experimental signature of this is visible in the KMT2A literature. Graded weakening of the DOT1L-AF9 interaction produced differential loss of H3K79me2 and me3 at MLL-AF9 target genes 4 - differential, not proportional, which is what a distributive mechanism predicts.
The polarity problem, stated rather than resolved
Two well-executed genome-wide studies disagree about what H3K79me3 means transcriptionally. Barski et al. place trimethylation of H3K79 with H3K27me3 and H3K9me3 as repression-linked in human CD4+ T cells 3. Park et al. find in yeast that Dot1-activated genes are enriched for H3K79me3 while Dot1-repressed genes carry H3K79me1 2.
These are different organisms and different logics - a correlation with expression versus an enrichment among genes whose expression changes when the writer is deleted - and neither is obviously wrong. The honest position for this graph is that H3K79me3 has no established transcriptional polarity in human haematopoietic cells, and any entry that uses 'H3K79 methylation is an activating mark' as a premise is using a statement established for me1 and me2, not for me3.
There is a reagent problem underneath both. Antibodies raised against one methylation state cross-react with adjacent states, and since the states co-occur on the same residue in a distributive distribution, cross-reactivity is not a small correction here. Calibrated internal-standard ChIP (ICeChIP) exists and has been applied to H3K79me2 in this disease; it has not been applied to the three states side by side.
| mark | H3K79me3 |
|---|---|
| histone | H3 |
| residue | K79 (K76 in trypanosomes) |
| writer | gene-dot1l |
| deposition mechanism | distributive - each methyl addition requires a separate productive binding event |
| h2bub dependence | strongest of the three states; Rad6 loss abolishes the yeast H3K79me3 zone entirely |
| transcriptional polarity | DISPUTED - repression-linked in human CD4+ T cells (PMID 17512414), enriched at Dot1-activated genes in yeast (PMID 41925229) |
| sensitivity to partial inhibition | highest of the three states, by the distributive argument; makes it a poor pharmacodynamic endpoint for depth of target engagement |
| erasers known | none established; KDM2B has been reported as a di- and tri-methyl H3K79 demethylase and is contested - see mechanism-h3k79-demethylation-contested |
| measured in kmt2a | ChIP-seq alongside H3K79me2 in MLL-AF9 recruitment-mutant series (PMID 25921540) and at KEE enhancers (PMID 31243293) |
- Choosing the most flux-sensitive state as a pharmacodynamic endpoint makes a drug look target-engaged at exposures that barely dent the mark the disease runs on - the precise way a dose-finding study can stop escalating for the wrong reason.
- If the three states are three points on one distribution rather than three signals, then an intervention that shifts the distribution without removing methylation may leave the oncogenic programme intact while every biomarker moves.
- What is the absolute me1:me2:me3 stoichiometry at KMT2A-fusion target loci in patient blasts, measured with calibrated internal-standard ChIP rather than with state-specific antibodies alone? No published dataset reports all three states quantitatively side by side in this disease.
- Does H3K79me3 have any transcriptional polarity in human haematopoietic cells, or is the human correlation with repression an artefact of one cell type and 2007 reagents?
- If me3 is the flux-limited state, does it fall first under a DOT1L inhibitor in patient marrow - and did the pinometostat pharmacodynamic assay, which read 'H3K79 methylation', in fact read the state that moves at the lowest exposure?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Dot1 enzymes establish the H3K79 methylation pattern via a distributive mechanism; the yeast, Trypanosoma brucei (TbDot1A/TbDot1B, on H3K76) and human hDOT1L enzymes generate very divergent methylation patterns, and computational simulation of their output when expressed in yeast at varying levels shows they have highly distinct catalytic properties while sharing the distributive mechanism.
Dot1 histone methyltransferases share a distributive mechanism but have highly diverged catalytic properties. · Stulemeijer IJ, De Vos D, van Harten K, et al. · Sci Rep 2015What this evidence cannot show
Abstract read from the eutils record; full text not read. The comparison is made by expressing heterologous Dot1 enzymes in budding yeast and fitting computational models to the resulting bulk methylation distributions, so human DOT1L's catalytic properties are inferred in a yeast chromatin context rather than measured on human chromatin. The rate constants are model parameters, not directly measured. Nothing is measured in a leukaemic cell.
Evidence that argues against this (1)
- Distributivity is established on bulk chromatin and says nothing about a locus where the enzyme is held in place by a fusion protein. A tethered DOT1L at an MLL-fusion target gets a far higher local enzyme-time than the genome average, and the resulting local state distribution cannot be predicted from bulk kinetics - which is the whole point of the graded-recruitment result in this disease. Kuntimaddi A et al. 2015
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In yeast, loss of Rad6 leads to the complete loss of the H3K79me3 zone, converting it into an H3K79me1-enriched region while simultaneously diminishing the H3K79me1 zone; Dot1-activated genes are enriched for H3K79me3 whereas a majority of Dot1-repressed genes are associated with H3K79me1; and Dot1 occupancy does not always correlate with H3K79me3 level, since translation-related genes show high Dot1 occupancy but are depleted of H3K79me3.
Trans-histone crosstalk establishes distinct H3K79 methylation zones with differential transcriptional functions. · Park NH, Kim HR, Kim HY, et al. · Nucleic Acids Res 2026What this evidence cannot show
Abstract read from the eutils record; full text not read. Saccharomyces cerevisiae; Rad6-Bre1 is the yeast H2B E3 machinery and the human equivalent is RNF20/RNF40 with a disputed E2. Published 2026 and not independently replicated. State-specific antibodies carry cross-reactivity risk that the abstract does not address with a specificity control, and the zone model depends entirely on that discrimination.
Evidence that argues against this (1)
- The older kinetic work argues the states are functionally redundant rather than zoned: because Dot1 methylates non-processively the states cannot be generated independently, yeast Dot1-dependent silencing tracked GLOBAL H3K79 methylation rather than any one state, and the authors explicitly expected multiple methylation to 'limit the possibilities for regulation by putative demethylases or binding proteins'. Frederiks F et al. 2008
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In high-resolution genome-wide maps of 20 histone methylations in the human genome, trimethylation of H3K79 is linked to repression, alongside H3K27me3 and H3K9me3, while monomethylation of H3K79 is linked to gene activation.
High-resolution profiling of histone methylations in the human genome. · Barski A, Cuddapah S, Cui K, et al. · Cell 2007What this evidence cannot show
Abstract read from the eutils record; full text not read. Resting human CD4+ T cells only. A correlation between ChIP-seq occupancy and expression with no perturbation, at 2007 sequencing depth and with 2007 antibodies. The assignment of H3K79me3 to repression is one line in a 20-mark survey and was not the paper's focus.
Evidence that argues against this (1)
- The opposite association is reported in yeast with a perturbative readout - genes ACTIVATED by Dot1 are the ones enriched for H3K79me3. Whether the disagreement is species, assay or antibody is unresolved, and this graph should therefore not assert a transcriptional polarity for H3K79me3. Park NH et al. 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Graded reduction of the DOT1L-MLL-AF9 interaction by structure-guided point mutation produces differential loss of H3K79me2 and H3K79me3 at MLL-AF9 target genes, and the degree of recruitment tracks transformation potential.
Degree of recruitment of DOT1L to MLL-AF9 defines level of H3K79 Di- and tri-methylation on target genes and transformation potential. · Kuntimaddi A, Achille NJ, Thorpe J, et al. · Cell Rep 2015What this evidence cannot show
Abstract read from the eutils record; full text not read. The word 'differential' is the abstract's; the direction and magnitude of the me2-versus-me3 difference is not stated there and this entry does not assert one. MLL-AF9 and the AF9 interface only. ChIP-seq with state-specific antibodies inherits the cross-reactivity caveat that applies to all three states.
Evidence that argues against this (1)
- Mutating the recruitment interface leaves catalytically competent DOT1L in the cell and only relocates it, which is a different perturbation from inhibiting or degrading it. DOT1L has functions that are independent of H3K79 methylation, so a recruitment-mutant series measures the consequences of mislocalised enzyme rather than of absent mark. Cao K et al. 2020
terminal state of a distributive methylation reaction
three points on one flux-set distribution, not three independent marks
strongest H2Bub dependence of the three states; Rad6 loss abolishes the me3 zone
state-resolved granularity for the step-4 mark entry
distributive kinetics make me3 the flux-limited state
KDM2B claim is for the di- and tri-methyl states
raises global H3K79me2 and me3