Oncofusion

start / graph / mechanism-mll1-bookmarks-mitotic-chromatin-and-carries-menin-there

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.

MLL1 stays on condensed mitotic chromosomes and tethers menin there — a menin-dependent function absent from this graph until cycle 34

also called mitotic bookmarking · MLL mitotic chromatin · menin on mitotic chromosomes · transcriptional memory through mitosis

When a cell divides almost everything is stripped off the chromosomes, but MLL1 stays bound to the promoters of genes that were busy beforehand and holds menin there with it, so cells that lose this are slower to switch those genes back on. This project is entirely about the menin-MLL contact, and the 2009 paper describing this job it does was not cited anywhere here.

The same thing, for a scientist

**MLL associates with gene promoters packaged within condensed mitotic chromosomes**, which the authors report is unlike the other chromatin-modifying enzymes they examined. Genome-wide location analysis found a **globally rearranged** MLL occupancy pattern during mitosis, favouring genes highly transcribed during interphase; knockdown showed that MLL retention at promoters **accelerates transcriptional reactivation after mitotic exit**. **Three things about this matter here, and one is uncomfortable.** 1. **MLL tethers Menin, RbBP5 and ASH2L to its occupied sites during mitosis.** Menin is on mitotic chromatin, carried by MLL. This graph is about the menin-KMT2A interface and had no entry for that. 2. **MLL is dispensable for preserving H3K4 methylation** in this setting. The bookmarking function is separable from the catalytic output the field usually reasons about. 3. **It is wild-type MLL1 in HeLa cells, in 2009. Whether a KMT2A FUSION bookmarks mitotic chromatin is, as far as I can find, unmeasured** - and that is now a much sharper question than the one this graph was asking.

Go deeper
Why this should change how the graph reasons about menin inhibition

If menin rides mitotic chromatin as part of an MLL complex, menin-inhibitor pharmacology may be cell-cycle-dependent in a way nobody here has considered. A pocket blocker that displaces menin from KMT2A would, on this model, also strip it off mitotic chromosomes — and the consequence would land at mitotic exit rather than during interphase.

Three concrete consequences, none of which this graph currently carries:

  • Timing. Effects of a menin inhibitor might depend on where cells are in the cycle when the drug arrives, which bears on scheduling and on why in vitro assays with different treatment windows disagree.
  • A separable function to lose. Bookmarking is dispensable for H3K4 methylation here, so a menin inhibitor could impair bookmarking without the H3K4me3 readout moving — and H3K4me3 is exactly what the field measures. An unchanged H3K4me3 does not mean an unchanged complex.
  • A mechanism for transcriptional memory. This project already carries an entry on the two meanings of 'transcriptional memory' in this disease. Bookmarking is a third, literal, mitotic one, and it is the one with a direct menin connection.

None of that is established for the fusion and I am not asserting it. It is what follows if the wild-type observation transfers, and whether it transfers is the open question.

How this was found, and what it says about the graph

From the stale-correction queue, checking a different claim. cell-state-unsynchronisable-cell-cycle asserted that 'nobody has measured' whether fusion chromatin occupancy, complex composition or activity vary through the cycle. A bounded query for MLL and mitotic bookmarking returned 2 records against a control of 85 for "mitotic bookmarking" alone and 64 for KMT2A chromatin occupancy in leukaemia.

Both records were uncited in this graph, and the word 'bookmarking' appeared nowhere in it. A 2009 Molecular Cell paper putting menin on mitotic chromatin was missing from a knowledge graph whose subject is the menin-KMT2A contact.

The claim it was checked against survives in narrowed form — nobody has measured this for the fusion — but the unqualified version was wrong, and the gap it concealed is larger than the claim. This is the second time the stale queue has found not an overstatement but a missing body of work, after the RAS deep-sequencing paper.

Facts and figures
bounded query
{
  "target": "(KMT2A OR MLL1 OR mixed lineage leukemia) AND (mitotic bookmarking OR bookmarking)",
  "hits": 2,
  "both_uncited_in_graph_before": true,
  "controls": {
    "\"mitotic bookmarking\" alone": 85,
    "KMT2A/MLL chromatin occupancy in leukemia": 64
  },
  "run": "PubMed esearch 2026-09-20"
}
system
{
  "protein": "wild-type MLL1",
  "cells": "HeLa",
  "year": 2009,
  "fusion_tested": false
}
How this could bite us back
  • This is wild-type MLL1 in HeLa cells in 2009 and it is NOT evidence about a fusion, a leukaemia cell or a patient. The reason it is in the graph is that it was missing, not that it is established here - importing a wild-type-allele function into a fusion argument is a mistake this project has documented repeatedly and this entry is a candidate for it.
  • And do not use it to argue for cell-cycle-timed menin-inhibitor dosing. The chain from 'menin is present on mitotic chromatin' to 'dose at a particular phase' has at least three unmeasured steps in it, and the first one - does the fusion do this at all - has never been tested.
Still open
  1. Does a KMT2A FUSION bookmark mitotic chromatin? Nothing found. This is the sharp version of the question the cell-cycle entry was groping at, and it is answerable with one mitotic CUT&RUN in an index-fusion line.
  2. Does a menin inhibitor displace menin from mitotic chromatin, and if so does the phenotype appear at mitotic exit rather than during interphase? That would make drug effect cell-cycle-phase dependent.
  3. Has the 2009 mitotic-retention result been re-tested with the fixation controls that the bookmarking field adopted afterwards? If not, the whole entry rests on a method later shown to generate artefacts in related systems.
Every source behind this entry (2)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.journal-article

    In HeLa cells, wild-type MLL associates with gene promoters within condensed mitotic chromosomes unlike other chromatin-modifying enzymes examined, shows a globally rearranged genome-wide occupancy pattern during mitosis that favours genes highly transcribed during interphase, and its retention at promoters accelerates transcriptional reactivation following mitotic exit.

    PMID 20064463DOI
    What this evidence cannot show

    **Wild-type MLL1 in HeLa cells, published 2009 — not a KMT2A fusion, not a leukaemia cell, not the index disease** (rule 73). Occupancy is by genome-wide location analysis of the era, so resolution and antibody specificity are those of 2009 ChIP rather than modern CUT&RUN, and the 'unlike other enzymes examined' comparison covers the set the authors chose to examine rather than all chromatin modifiers. Read from the abstract; full text not obtained.

    Evidence that argues against this (1)
    • Mitotic retention is technically contested as a class. Formaldehyde fixation can produce artefactual retention signals on mitotic chromosomes, and several proteins reported as bookmarking factors in the 2000s were later shown to be fixation artefacts when re-examined with live imaging or alternative fixation. This result predates that methodological reckoning and has not, as far as this entry establishes, been re-tested with those controls. Bina M et al. 2013
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.journal-article

    MLL tethers menin, RbBP5 and ASH2L to its occupied sites on mitotic chromatin while being dispensable for preserving histone H3K4 methylation at those sites, so the mitotic bookmarking function is separable from the H3K4 methylation output usually used to read out this complex.

    PMID 20064463DOI
    What this evidence cannot show

    Same system and vintage as E1 — wild-type MLL1, HeLa, 2009. **Moderate rather than strong** because 'tethers' is a co-occupancy observation and the dispensability claim is a negative result at the sensitivity of the assays used. That menin is present at these sites does not establish that menin is required for bookmarking, which is the inference this graph would most want to draw and which the experiment as described does not support.

    Evidence that argues against this (1)
    • The graph's central hypothesis concerns the menin-KMT2A contact as a leukaemogenic determinant, and menin's presence on mitotic chromatin is a WILD-TYPE housekeeping observation that may have nothing to do with transformation. Treating it as relevant risks the error this project has documented repeatedly: importing a wild-type-allele function into an argument about a fusion. The honest status is an unexplored adjacency, not a mechanism of disease. Bach C et al. 2009
Connected to