H4K16ac - the one mark rule 80 actually rescued in this beat: elevated at MLL-AF9 targets in a supplementary figure, and the reason MLL1's own enzyme turns out to be dispensable

(MLL-rearranged OR KMT2A-rearranged OR MLL-AF9 OR MLL-AF4 OR KMT2A OR "MLL fusion")[tiab] AND (H4K16ac OR "H4K16 acetylation")[tiab] returns 3 records on 2026-09-20: PMIDs 33879367 (a crosstalk review), 28202522 (the MOF paper, 1) and 24813891 (MLL1 SET dispensability). H4K16ac[tiab] alone returns 364; the KMT2A term set alone returns 3,209. The instrument works and the operands are large.

PMID 25822366 is not in the three, and it contains the only locus-resolved H4K16ac observation in this disease: "We also note that H4K16ac (another known Sirt1 substrate) showed the same elevated pattern as H3K9ac (Supplementary Fig. 8)", and in the discussion, "it is also important to note that H4K16ac ... follows the same pattern as H3K9ac" 2. The mark appears twice in a 66,570-character full text and never in the abstract.

The prior ledger ana-histone-mark-audit-rejected-candidates recorded H4K16ac as 1 KMT2A-intersecting record and concluded the graph's eight existing mentions were "proportionate to a literature of one". The count is now three by a slightly wider name list, and the literature is four once the supplementary panel is included. The conclusion that followed from the count - that no entry was owed - does not survive, because the measurement the graph most wanted was the one the count could not see.

Acetylation at H4K16 is the only single histone modification shown to change the physical state of a chromatin fibre on its own. Using chemically defined nucleosomal arrays, H4K16ac inhibits formation of the compact 30-nm-like fibre and impedes cross-fibre interactions, and it also inhibits ACF-dependent chromatin assembly and remodelling 4. That matters for this graph specifically because the KMT2A fusions work by holding long transcription units open for elongation: a mark that mechanically resists compaction is a plausible partner for that, and it is one of the few places where a chromatin claim in this project could be tested biophysically rather than by ChIP.

It is also a trans-histone input to DOT1L, the enzyme at the centre of this disease. In yeast, loss of H4K16 acetylation disrupts the H3K79me1 zone, although weakly - much more weakly than loss of Rad6-dependent H2B ubiquitination, which abolishes the H3K79me3 zone entirely 3. So H4K16ac sits upstream of the mark this whole graph is organised around, on a pathway that is established in yeast and untested in a human leukaemia cell.

mark-h3k4me3 and this project's SET-domain discussions rest on KMT2A being an H3K4 methyltransferase whose catalytic half the fusion discards. PMID 24813891 tests that directly and finds the SET domain dispensable - for HSC maintenance and for MLL-AF9-driven leukaemogenesis alike 6.

What is not dispensable is the acetyltransferase MLL1 brings with it. Within 48 hours of Mll1 deletion in lineage-negative cells, H4K16Ac falls at MLL1 target promoters (Hoxa9, Evi1/Mecom, Prdm16), and so do MOF itself, BRD4, CDK9 and Ser2-phosphorylated elongating Pol II; H3K27me3 rises only later, consistent with being a consequence of transcriptional cessation rather than a cause. A SET-deleted mutant shows unchanged H4K16Ac, gene-specific and global. And the SIRT1 inhibitor Ex-527 restores both H4K16Ac and target-gene expression without touching H3K4me3 6.

Three things follow. (1) The functional histone output of the KMT2A locus at its own targets is H4K16 acetylation, delivered by a recruited enzyme. (2) The MOF-KMT2A contact maps to the KMT2A C-terminus - the fragment every leukaemia breakpoint removes (gene-kat8) - so the fusion loses the module that recruits this activity, while wild-type KMT2A on the second allele retains it. (3) This is a wild-type MLL1 experiment: the H4K16ac ChIP is in normal HSPCs, and the MLL-AF9 arm of the paper carries no H4K16ac measurement at all. The locus-resolved observation in a KMT2A-rearranged cell remains the supplementary panel in 2.

A chromatin-regulator RNAi screen in murine MLL-AF9 leukaemia cells found hairpins against Kat8/Mof among the most potent suppressors of growth, alongside Brd4. Conditional Mof deletion reduced tumour burden and prolonged survival; HAT-dead rescue constructs (CoA-binding-site deletion and the G327E point mutant) failed to rescue colony formation, so the enzymatic activity is required 1.

But the mechanism the paper lands on is DNA damage, not transcription at fusion targets. RNA-seq showed downregulation of DNA damage repair genes and a significant increase in gamma-H2AX nuclear foci in Mof-deficient MLL-AF9 cells, and Mof inactivation also suppressed a NUP98-HOXA9 model 1. So the H4K16ac requirement in this disease is currently a genome-maintenance requirement that happens to be measured in a KMT2A-rearranged cell, not a demonstration that the fusion uses the mark.

And this graph already records the wild-type qualifier. gene-kat8 establishes that the MOF-KMT2A contact maps to the KMT2A C-terminus, the fragment every leukaemia breakpoint removes - so KAT8 is a wild-type KMT2A partner and not a member of the KMT2A::AFF1 interactome. KAT8 is also DepMap common-essential (mean -1.45, 99.5% of 1,208 lines depleted) with a KMT2A-selectivity difference of +0.004, which this project's own floor-compression rule (QUEEN_BRIEF.md) marks as uninterpretable rather than small. The writer is essential everywhere; the mark's selectivity for this disease is unmeasured.

• No H4K16ac ChIP, CUT&RUN or CUT&Tag has been performed in any KMT2A::AFF1 cell, line or patient sample. All four KMT2A-r records are myeloid: MLL-AF9 (mouse and human) and a NUP98-HOXA9 comparator.
• No H4K16ac measurement exists under a menin inhibitor or a DOT1L inhibitor at locus resolution. The DOT1L-inhibitor link in 2 is inferred from H4K16ac following H3K9ac's pattern; the H4K16ac panel itself is a baseline comparison, not a drug time course.
• The eraser side is empty in this disease. SIRT1 is named as the H4K16ac deacetylase in 2 and this graph carries gene-sirt1, but no experiment has measured H4K16ac after SIRT1 activation or inhibition at fusion targets - even though the same paper does exactly that for H3K9ac with SRT1720.
• MG149 is not a selective tool. It is a MYST-family inhibitor, so its activity spans KAT5/TIP60, KAT6A/MOZ, KAT7/HBO1 and KAT8/MOF - four proteins this graph carries separately - and a growth effect under MG149 cannot be attributed to H4K16ac.

This entry was written without linking to subtraction-kmt2a-mof-contact, which already answers most of the question it proposes to test.

That entry records the MOF/KAT8 contact as sitting on the KMT2A C-terminus that every breakpoint discards, and — importantly — it has already been tested in the fusion context. Its conclusion: Mof is required for MLL-AF9 leukaemia, but the mechanism is global H4K16ac loss and DNA-damage-repair failure, not the fusion's target programme, and its loss_class is "dependency-of-the-cell-not-of-the-fusion".

That is substantially the prediction made here — that fusion loss should not move H4K16ac at fusion targets — reached by a different route and already graded.

What is NOT settled, and is the honest residue of the proposal:

• The existing result is MLL-AF9 and murine; nothing equivalent exists for KMT2A::AFF1 in B lineage.
• It is a whole-leukaemia dependency result (survival, colony formation), not a locus-resolved H4K16ac measurement at fusion-bound sites. The ChIP proposed here would still be new — it would show whether H4K16ac at fusion targets specifically is unaffected, which the dependency experiment cannot.
• No residue range exists for the contact. That entry deliberately records coordinates: not assigned because the primary source says only "MLL1 C-terminal" — so the claim that the contact lies beyond every breakpoint is qualitative, not coordinate-checked, and should not be stated as though it were mapped.

So the experiment is narrowed rather than cancelled, and any proposal of it should cite the existing fusion-context result rather than present the question as open.