The fusion may reach the H2B ubiquitin mark through MOF-MSL rather than through RNF20/RNF40 — the one configuration in which this layer could be selective
also called MSL2 route hypothesis · fusion-proximal H2B ubiquitination
The KMT2A fragment kept in the fusion is reported to contact MOF, and the MOF complex contains a second enzyme that puts a ubiquitin tag on H2B at a different spot — a tag that also switches on DOT1L, and that was shown to matter at HOXA9 and MEIS1. If the fusion is pulling that enzyme to its target genes, the ubiquitin layer would be fusion-specific after all.
The same thing, for a scientist
Everything else in this beat points to a general, transcription-coupled route to H2Bub (RNF20/RNF40-WAC via PAF1C), which is why the writers show no DepMap selectivity. There is one route that would not be general. MSL2, within the MOF-MSL complex, is an E3 ligase for **H2BK34**; H2BK34ub regulates H3K4 and H3K79 methylation in trans, and MSL1/2 activity is important for transcription activation at the **HOXA9 and MEIS1** loci [E1]. Structurally, H2BK34ub positions Dot1L for the same productive conformation as H2BK120ub [E2]. This graph separately records a KMT2A–MOF contact (`subtraction-kmt2a-mof-contact`) and carries MOF/KAT8 as a protein entry. The hypothesis is that the retained KMT2A fragment delivers MOF-MSL, and therefore MSL2's H2B ubiquitination activity, to fusion target genes. No experiment tests this. It is the single most consequential untested proposition this beat generated, because it is the only way the H2Bub layer becomes fusion-proximal rather than general.
Why this is worth stating despite thin support
The components are all present and independently established:
| Component | Status | Where |
|---|---|---|
| MSL2 is an H2BK34 E3 whose mark stimulates DOT1L | established in vitro and structurally, but the nucleosomal activity is contested (PMID 30930284) | 12 |
| MSL1/2 activity matters at HOXA9 and MEIS1 | reported | 1 |
| KMT2A contacts MOF | recorded in this graph | subtraction-kmt2a-mof-contact |
| The fusion retains the KMT2A N-terminus | established | fusion-kmt2a-aff1 |
What is missing is the join: nobody has asked whether the fusion co-purifies with MSL2, or whether H2BK34ub is present at fusion target genes.
The strongest objection, stated up front. The graph's MOF entry is a subtraction entry — it exists in the context of what the fusion loses, and whether the MOF contact survives the breakpoint in KMT2A::AFF1 is exactly the question that entry examines. If the MOF contact maps to a discarded region, this hypothesis dies immediately and cheaply. Anyone pursuing it should read subtraction-kmt2a-mof-contact first; this Queen flagged that entry as the prerequisite rather than assuming its conclusion.
The second objection is quantitative: MSL2 is essentially dispensable in DepMap (mean -0.083, difference -0.015, rank 7,112) 3. A route that matters should leave some mark on a knockout screen. It does not.
The third objection, and the reason the prior is 0.15 rather than 0.4. Until this repair the low prior was asserted without a source. It has one: Krajewski & Vassiliev found purified MSL1/MSL2 ubiquitylate free histone substrates efficiently but modify nucleosomes very poorly, and inferred that efficient H2BK34 ubiquitylation requires the nucleosome to be structurally altered first (PMID 30930284, counter-evidence on 1). Recruitment is not sufficient if the enzyme cannot act on ordinary chromatin once it arrives. Added to that, H2BK34 also carries a conserved meiotic methyl mark in mouse testes (PMID 32933557), so even a positive K34 read-out needs the modification identified, not assumed.
| statement | The KMT2A fragment retained in the fusion recruits the MOF-MSL complex to fusion target genes, delivering MSL2's H2BK34 ubiquitin ligase activity and thereby licensing DOT1L by a route independent of RNF20/RNF40 and PAF1C - which would make the H2B ubiquitination layer fusion-proximal and potentially selective. |
|---|---|
| origin | literature-gap |
| prior probability | 0.15 |
| reason prior is low | MSL1/MSL2 modify nucleosomes very poorly in vitro (PMID 30930284); MSL2 and MSL1 show no KMT2A selectivity in DepMap; and the KMT2A-MOF contact is not confirmed to survive the breakpoint |
predictions
[ "KMT2A::AFF1 co-immunoprecipitates MSL2 and MSL1", "H2BK34ub is detectable by CUT&RUN at HOXA9 and MEIS1 in KMT2A-rearranged cells and falls on fusion degradation", "MSL2 loss lowers H3K79me2 at fusion target genes more than at matched non-target transcribed genes", "Menin inhibition reduces H2BK34ub at fusion targets" ]
experiments
[ "exp-msl2-fusion-coip-and-h2bk34ub-cutrun" ]
analyses
[ "ana-depmap-h2bub-machinery-kmt2a-selectivity" ]
falsifiers
[ "No MSL2 or MSL1 in a KMT2A::AFF1 pulldown from a KMT2A-rearranged B-ALL line", "No detectable H2BK34ub at fusion target genes", "The KMT2A-MOF contact maps to a region discarded at the breakpoint, so the fusion cannot carry it", "MSL2 depletion leaves H3K79me2 at fusion targets unchanged", "MSL1/MSL2 prove unable to ubiquitinate ordinary nucleosomal H2BK34 in a leukaemia chromatin extract, consistent with PMID 30930284" ]
- Even if the recruitment is real, MOF/KAT8 is one of the most essential genes in DepMap (-1.452), so a complex-level intervention is not available and a subunit-level one hits an enzyme that is dispensable.
- A fusion-proximal route to H2Bub would make the layer selective and would also make it redundant with menin inhibition, which already removes the fusion from chromatin - so the therapeutic gain might be zero even if the biology is right.
- Does the KMT2A-MOF contact survive the breakpoint in KMT2A::AFF1? This graph's own subtraction entry is the place to answer it and this beat did not resolve it.
- What fraction of H2B ubiquitin in a leukaemia cell is at K34? If it is a trace species the hypothesis is uninteresting even if the recruitment is real.
- Is there an H2BK34ub-specific antibody good enough for CUT&RUN? The experiment depends on it and this beat did not check.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
MSL2 in the MOF-MSL complex is a histone ubiquitin E3 ligase which, with MSL1, targets nucleosomal H2B at lysine 34; H2BK34ub directly regulates H3K4 and H3K79 methylation through trans-tail crosstalk in vitro and in cells; MSL1/2 activity is important for transcription activation at the HOXA9 and MEIS1 loci; and the MOF-MSL complex therefore carries two chromatin-modifying activities, H4K16 acetylation and H2BK34 ubiquitylation.
The RING finger protein MSL2 in the MOF complex is an E3 ubiquitin ligase for H2B K34 and is involved in crosstalk with H3 K4 and K79 methylation. · Wu L, Zee BM, Wang Y, Garcia BA, Dou Y · Mol Cell 2011What this evidence cannot show
Abstract read in full; full text not read. No KMT2A-rearranged system is named; the HOXA9/MEIS1 result is not a leukaemia result and the connection drawn here to the fusion programme is this project's inference.
Evidence that argues against this (1)
- This is the published reason the prior is 0.15 rather than higher, and it was previously asserted in prose without a citation. Krajewski & Vassiliev re-examined purified MSL1/MSL2 and found that while the enzyme ubiquitylates free histone substrates efficiently it modifies NUCLEOSOMES very poorly, implying that efficient H2BK34 ubiquitylation requires the nucleosome to be structurally altered first. If MSL1/2 cannot readily act on ordinary chromatin, a recruitment-based route to H2Bub at fusion target genes is unlikely to deliver a functionally meaningful amount of the mark even if the recruitment is real. Krajewski WA et al. 2019
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Cryo-EM shows H2BK34ub induces an almost identical Dot1L orientation and binding pattern on the nucleosome as H2BK120ub, positioning Dot1L for the productive conformation, though by nucleosome distortion rather than direct ubiquitin-enzyme contact.
H2B Lys34 Ubiquitination Induces Nucleosome Distortion to Stimulate Dot1L Activity. · Ai H, Sun M, Liu A, Sun Z, Liu T, Cao L, Liang L, Qu Q, Li Z, Deng Z, Tong Z, Chu G, Tian X, Deng H, Zhao S, et al. · Nat Chem Biol 2022What this evidence cannot show
Abstract read in full; full text not read. Reconstituted nucleosomes; no cellular or leukaemia data.
Evidence that argues against this (1)
- A structure showing K34-ubiquitin can license Dot1L does not establish that K34-ubiquitin exists at fusion target genes, and the residue is not dedicated to ubiquitin: H2BK34 is enriched for METHYLATION during meiosis in mouse testes (with the yeast counterpart H2BK37me). The hypothesis needs the mark detected and attributed, not just shown to be sufficient in a reconstituted system. Crespo M et al. 2020
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
MSL2 is essentially dispensable across DepMap (mean gene effect -0.083 across 1,208 lines) and shows no KMT2A selectivity (difference -0.015, rank 7,112 of 17,728); MSL1 is -0.365 with a difference of -0.073 (rank 2,078).
DepMap CRISPR (Chronos) genome-wide KMT2A-selectivity scan of 17,728 protein-coding genes, 11 Cellosaurus-annotated KMT2A-rearranged leukaemia lines vs 43 other leukaemia lines · Broad Institute DepMap; Cellosaurus; Open Targets Platform; analysis run by this project (analyses/depmap-kmt2a-selective-scan-full) · Database and derived analysis 2026What this evidence cannot show
A gene with near-zero pan-essentiality has little dynamic range, so this null is weak evidence against the hypothesis rather than strong evidence.
Evidence that argues against this (1)
- SCAN CONFOUND: this number comes from the project's genome-wide KMT2A-selectivity scan, whose index arm is MYELOID-ENRICHED (8 AML lines against 5 B-ALL). A gene whose requirement differs by lineage is mis-scored in BOTH directions by that design — CEBPA ranks 25 genome-wide with no dependency at all in KMT2A-rearranged B-ALL, and conversely a gene required in one lineage and not the other can average to apparent non-selectivity. See `ana-lineage-confound-myeloid-tf-dependencies`. A mean gene effect of -0.083 across 1,208 lines says MSL2 is dispensable in cancer generally, which is a much weaker statement than non-selectivity in KMT2A-rearranged disease specifically and does not test the hypothesis's claim about fusion TARGETS as opposed to cell survival. Oncofusion knowledge graph (internal analysis) 2026
- ← related toMSL2 — the second H2B ubiquitin ligase, sitting inside the MOF complex the KMT2A fragment contacts
Counter-evidence (Krajewski and Vassiliev 2019, PMID 30930284): this is the published reason the prior is 0.15, and it was previously asserted without a source - MSL1/MSL2 modify nucleosomes very poorly, so recruitment is not sufficient if the enzyme cannot act on ordinary chromatin once it arrives
- ← related toLost from the KMT2A half: the MOF/KAT8 contact - required in the cell, but not shown to matter to the fusion
Prerequisite and potential cheap kill: whether the KMT2A-MOF contact survives the KMT2A::AFF1 breakpoint is examined in this subtraction entry, and if the contact maps to a discarded region the hypothesis dies without any bench work
the hypothesised fusion-proximal route runs through the K34 mark
- ← testsDoes the fusion carry an H2B ubiquitin ligase? A co-IP and an H2BK34ub CUT&RUN in KMT2A::AFF1 cells
co-IP plus H2BK34ub CUT&RUN in KMT2A::AFF1 cells
Supplies the missing scaffold subunit for the hypothesis that the fusion reaches H2Bub through MOF-MSL rather than RNF20/RNF40.