Oncofusion

start / graph / gene-kat8

Proteinstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q5d-interactome3

KAT8 (MOF) — a genuine KMT2A partner that binds the half of KMT2A the fusion discards

also called KAT8 · MOF · MYST1 · hMOF · Histone acetyltransferase KAT8

MOF is the enzyme that puts a specific acetyl tag on histones, and it forms a stable complex with normal KMT2A so the two marks land together. The contact is on the far end of KMT2A, which the leukemia fusion cuts off, so MOF stays behind with the healthy protein.

The same thing, for a scientist

KAT8/MOF (Q9H7Z6, 458 aa) forms a stable, reciprocally validated complex with MLL1/KMT2A; the interaction sites were mapped to the MLL1 **C-terminus** and the MOF zinc finger. The complex has both MLL1 H3K4 mono/di/trimethyltransferase activity and MOF H4K16 acetyltransferase activity, and both are required for optimal activation of an endogenous MLL1 target (HOXA9). Because the mapped contact is C-terminal — the SET-domain end of KMT2A — it is on the fragment removed by every leukemia breakpoint, 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 no KMT2A-selectivity (difference +0.004). This entry exists because KAT8 is routinely drawn inside 'the MLL complex' in review figures without the C-terminal qualifier, and the graph already carries the related acetyltransferase KAT7/HBO1, so the omission looked like an oversight rather than a decision.

What it looks like
loading AF-Q9H7Z6-F1…
AlphaFold AF-Q9H7Z6-F1 ↗
colourshape

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

very high (90+)confident (70–90)low (50–70)very low (<50)
I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper
What was actually shown

Method: immunoaffinity purification from a human cell line stably expressing epitope-tagged WDR5, with the MLL1-MOF interaction then confirmed by reciprocal immunoprecipitation, cosedimentation and cotransfection, and the contact mapped to the MLL1 C-terminal region and the MOF zinc finger 1. Functionally, both enzymatic activities were required for optimal activation on a chromatin template in vitro and at endogenous HOXA9 in vivo 1.

That is strong biochemistry by any standard — three orthogonal association assays plus a mapped interface plus a functional requirement.

Shown for: wild-type MLL1/KMT2A only. The paper does not test any fusion.

Which part of KMT2A: the C-terminus. Every clinical KMT2A breakpoint falls in the exon 8-13 cluster region (approximately aa 1100-1400), which leaves the SET-domain end on the other side of the break. KAT8 is not part of the KMT2A::AFF1 interactome.

Why the negative still matters

It matters in two ways, both about how the graph reads.

First, KAT8 is one of the most cited members of 'the MLL complex' and its exclusion from the fusion is not obvious from the way the complex is usually drawn. A reader who sees H3K4me and H4K16ac drawn as a coordinated pair on MLL1 targets may carry that coordination over to fusion targets, where the acetyltransferase half has no anchor.

Second, KAT8 and KAT7 (HBO1, already in the graph) are both MYST-family acetyltransferases with KMT2A-adjacent biology, and they behave completely differently in the dependency data. KAT8 is depleted in 99.5% of cell lines with mean effect -1.45 — as pan-essential as it gets — with KMT2A-selectivity of +0.004, which is zero with the sign pointing the wrong way 2. Whatever the case for KAT7, it does not transfer to KAT8.

Facts and figures
uniprotQ9H7Z6
gene symbolKAT8
length AA458
protein nameHistone acetyltransferase KAT8
alphafold IDAF-Q9H7Z6-F1
accession resolutionUniProt query gene_exact:KAT8 AND organism_id:9606 AND reviewed:true; primary gene name read back as KAT8 and length 458 confirmed
enzymatic activityH4K16 acetyltransferase (MYST family)
interaction evidence classaffinity purification, reciprocal co-IP, cosedimentation, cotransfection, domain mapping
shown for fusionNO - KAT8 has never been localised in a KMT2A-rearranged cell, and no measurement in this disease separates the NSL-bound from the MSL-bound pool
druggabilityNot a target for this disease: pan-essential, contact on the discarded fragment, zero selectivity.
subnuclear localizationcomplex-dependent chromatin address: NSL-KAT8 at proximal promoters and transcription start sites (human, measured); MSL-KAT8 at gene bodies (by similarity from mouse). No nuclear-body observation exists.
contact on kmt2a
{
  "kmt2a_region": "C-terminal region of MLL1/KMT2A (residue boundaries not given in the source)",
  "kat8_region": "MOF zinc finger",
  "retained_in_fusion": false,
  "reason": "the mapped contact is C-terminal to the KMT2A breakpoint cluster region (exons 8-13, approx. aa 1100-1400)"
}
essentiality
{
  "depmap_common_essential": true,
  "chronos_mean": -1.4515,
  "n_lines": 1208,
  "fraction_lines_depleted": 0.995,
  "chronos_mean_kmt2a_rearranged": -1.4888,
  "chronos_mean_other_leukemia": -1.4929,
  "kmt2a_selectivity_difference": 0.0041,
  "n_kmt2a_rearranged_lines": 11,
  "n_other_leukemia_lines": 43,
  "benchmark_men1_difference": -0.8656
}
h3k79 zone link
{
  "source_pmid": "41925229",
  "organism": "yeast",
  "finding": "H4K16ac loss disrupts the H3K79me1 zone, weakly; H2Bub loss abolishes the me3 zone",
  "relevance": "connects KAT8's mark to the DOT1L axis, but via the state this disease does not read out"
}
localization
{
  "subcellular": "nucleus (chromatin-associated) AND mitochondrion. The mitochondrial pool is experimentally coded (UniProt Q9H7Z6, ECO:0000269|PubMed:27768893): MOF and a subset of NSL partners reside in mitochondria, MOF binds mtDNA in a KANSL3-dependent manner, and a mitochondrially targeted MOF - but not a catalytically dead mutant - rescues respiratory and mtDNA transcription defects. Nuclear entry uses two importin-alpha-1/KPNA2 NLSs resolved by crystallography (PMID 28991411).",
  "subnuclear": "complex-dependent chromatin address rather than a nuclear body: NSL-KAT8 associates with the proximal promoter and transcription start site (measured in human, PMID 33657400), MSL-KAT8 with gene bodies (BY SIMILARITY, ECO:0000250 from mouse Q9D1P2). No nuclear-compartment observation exists.",
  "how_shown": "fractionation with complex-partner co-fractionation, mtDNA binding, compartment-targeted rescue with a catalytic-dead control, conditional knockout mouse, X-ray crystallography of the NLS/importin-alpha1 complex, complex-resolved acetylation profiling",
  "systems": "mouse cardiomyocytes and conditional knockout, cultured aerobically respiring cells, human cell lines; nothing leukaemic",
  "complex_dependence": "NSL-KAT8 catalyses H4K5ac/H4K8ac and is essential; MSL-KAT8 catalyses bulk H4K16ac and is dispensable for proliferation - so the famous mark is made by the dispensable complex",
  "node": "localization-kat8-mof-has-a-mitochondrial-pool"
}
cell type
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "identifier_note": "the alias MOF is unusable as a search token: (KAT8 OR MOF OR MYST1 OR hMOF)[tiab] returns 28,480 records, almost all metal-organic frameworks. The on-topic set is reached only by pairing the gene with H4K16 or H4K16ac. Brief alias-precision rule, worked example.",
  "expressed_in": {
    "class": 1,
    "summary": "ubiquitous"
  },
  "required_in": {
    "class": 2,
    "floor_compressed": true,
    "pct_all_lines_depleted": 99.5,
    "all_1258_lines_mean_gene_effect": -1.453,
    "kmt2a_rearranged_lines_mean_gene_effect": -1.489,
    "myeloid_kmt2ar_mean": -1.516,
    "blineage_kmt2ar_mean": -1.441,
    "verdict": "PAN-ESSENTIAL and the strongest floor in the worklist: depleted below -0.5 in 99.5% of 1,258 lines. KAT8 is the brief's own worked example of floor compression and it behaves exactly as described - the KMT2A-rearranged arm is -0.036 relative to all lines, which is not a small selectivity, it is an unmeasurable one.",
    "the_cell-type_answer_that_a_conditional_knockout_gives_and_DepMap_cannot": "the histone acetyltransferase activity of MOF is required for ADULT but NOT for early FETAL haematopoiesis in mice (PMID 27827827). That is a genuine cell-type-and-developmental-stage answer, it is invisible to any cancer dependency screen, and it is the pattern this beat recommends for all six floor-compressed genes."
  },
  "context": "wild-type protein in normal murine haematopoiesis for the conditional-knockout arm; no fusion-context requirement measurement",
  "graph_links": [
    "cell-type-hsc-bone-marrow",
    "cell-type-fetal-liver-hsc"
  ]
}
cell state
{
  "maintains": {
    "class": 3,
    "states": [
      "cell-state-quiescence"
    ],
    "claim": "In mice, DIFFERENTIAL H4K16 acetylation levels - the mark KAT8/MOF writes - set the balance between QUIESCENCE and activation in haematopoietic stem cells. This is the only explicit quiescence result for any gene in this worklist.",
    "pmids": [
      "34362741",
      "27827827"
    ],
    "system_named_in_the_claim": "murine haematopoietic stem cells in normal haematopoiesis, not leukaemia and not human (brief rule 73)"
  },
  "bounded_query": {
    "date": "2026-09-20",
    "tool": "NCBI esearch db=pubmed",
    "query": "(KAT8[tiab] OR \"MOF\"[tiab] OR \"MYST1\"[tiab] OR \"hMOF\"[tiab]) AND <cell-state clause> AND <leukaemia clause>",
    "count": 6,
    "on_topic_after_title_check": 1,
    "controls": {
      "gene_clause_alone": 28480,
      "state_clause_alone": 96743
    },
    "alias_contamination": "three of the six hits are metal-organic-framework nanomaterial papers that matched on MOF. The bare count of 28,480 is meaningless and a reader who saw only it would conclude KAT8 was one of the best-studied genes in the set.",
    "rescue_query_that_worked": "(KAT8 OR hMOF OR MYST1 OR Mof)[tiab] AND (H4K16 OR H4K16ac)[tiab] AND (hematopoie* OR haematopoie* OR quiescen*)[tiab] -> 5 records, of which 34362741 and 27827827 are the two that matter"
  }
}
How this could bite us back
  • KAT8 is depleted in 99.5% of DepMap lines; targeting it would be essential-function toxicity with no fusion engagement.
  • If H4K16ac turns out to matter at fusion-occupied loci through a KAT8-independent route, any KAT8-directed strategy would fail for a reason that has nothing to do with dose.
Still open
  1. What are the residue boundaries of the MLL1 C-terminal MOF contact? Without them the fusion-exclusion argument rests on the word 'C-terminal'.
  2. H4K16ac is a coordinated mark at wild-type MLL1 targets. Is it still present at KMT2A::AFF1-occupied loci, and if so which acetyltransferase puts it there?
  3. KAT7/HBO1 and KAT8/MOF are both MYST enzymes adjacent to this biology with opposite dependency profiles. Is the KAT7 dependency in KMT2A-rearranged cells a MYST-family effect or specific to HBO1?
Every source behind this entry (3)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    A stable complex containing MLL1 and MOF was immunoaffinity purified; the interaction was confirmed by reciprocal immunoprecipitation, cosedimentation and cotransfection, and interaction sites were mapped to the MLL1 C-terminal and MOF zinc finger domains; both the MLL1 H3K4 methyltransferase and the MOF H4K16 acetyltransferase activities are required for optimal transcriptional activation on a chromatin template and at endogenous HOXA9.

    Physical association and coordinate function of the H3 K4 methyltransferase MLL1 and the H4 K16 acetyltransferase MOF. · Dou Y, Milne TA, Tackett AJ, Smith ER, Fukuda A, Wysocka J, Allis CD, Chait BT, Hess JL, Roeder RG · Cell 2005
    PMID 15960975DOI
    What this evidence cannot show

    Abstract read in full; not full text. Work is in HeLa-derived cells on wild-type MLL1; no fusion protein is tested anywhere in the paper. The purification was via tagged WDR5, so the complex recovered is a WDR5-containing one, which is a selection the reader should keep in view. 'C-terminal' is the resolution the abstract gives — no residue numbers, so the fusion-exclusion inference is at domain rather than residue precision.

    Evidence that argues against this (1)
    • The fusion-exclusion argument here is weaker than it is for HCFC1, because the abstract gives no residue boundaries for the MLL1 C-terminal contact. If any part of that contact extended N-terminally past aa 1400 the conclusion would change. No paper testing MOF against a KMT2A fusion protein was retrieved, so this remains an inference from domain position rather than a measured negative. Oncofusion Knowledge Graph project et al. 2026
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

    KAT8 has DepMap mean gene effect -1.4515 across 1,208 lines with 99.5% of lines depleted, and a KMT2A-rearranged minus other-leukemia difference of +0.004 (11 vs 43 lines), against -0.866 for MEN1 in the same lines.

    DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API, with KMT2A-rearrangement status from Cellosaurus; per-gene recomputation for this beat · Broad Institute DepMap; Cellosaurus; Open Targets Platform; analysis run by this project · Database and derived analysis 2026
    What this evidence cannot show

    DepMap release version not stated by the API. Eleven KMT2A-rearranged lines. A pan-essential gene has little dependency variance, so a near-zero difference is the expected result whether or not there is fusion biology.

    Evidence that argues against this (1)
    • Pan-essentiality is the finding here and it argues against the target rather than for it. A gene depleted in more than 90% of all cell lines is required by the patient's normal marrow as much as by the leukaemia, so its appearance in a KMT2A complex is a statement about the complex and not a therapeutic opening. A difference of +0.004 at 99.5% depletion is floor-compressed and points, if anywhere, the wrong way. Oncofusion knowledge graph (internal analysis) 2026
  3. E_h4k16ac_h3k79me1_zoneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    KAT8's mark connects to the DOT1L axis this project cares about. In yeast, H3K79 methylation is organised into three state-specific zones, and loss of H4K16 acetylation disrupts the H3K79me1 zone. The dominant input is H2B ubiquitination - Rad6 loss causes COMPLETE loss of the H3K79me3 zone, converting it to an H3K79me1-enriched region - with H4K16ac a secondary contributor to the me1 zone specifically. The same work finds Dot1 occupancy does not track H3K79me3 level, and that Dot1-activated genes carry me3 while most Dot1-repressed genes carry me1.

    Trans-histone crosstalk establishes distinct H3K79 methylation zones with differential transcriptional functions · Park NH, Kim HR, Kim HY, Lee SY, Cho YJ, Roe JS, Kim T · Nucleic Acids Research 54(6):gkag291 2026
    PMID 41925229DOI
    What this evidence cannot show

    YEAST, not human: Saccharomyces ChIP-seq, where the H4K16 acetyltransferase is Sas2 and the E3 is Bre1, not KAT8/MOF and RNF20/RNF40. The effect is also explicitly hedged by the authors - 'loss of H4K16 acetylation ALSO SIMILARLY DISRUPTED the H3K79me1 zone, ALBEIT WEAKLY' - and it touches the me1 zone, while KMT2A-rearranged leukaemia is read out in me2/me3. Read as a mechanistic connection worth knowing, not as a KMT2A result. Surfaced by the histone-marks beat, which reported it rather than editing this entry; the PMID was re-fetched and the abstract read here before adding.

    Evidence that argues against this (1)
    • This weakens as much as it adds. If H4K16ac contributes only weakly and only to the me1 zone, then KAT8's route into the DOT1L axis is the least relevant of the three states for this disease - H2Bub dominates, and the states that matter clinically are me2/me3. Taking this as support for KAT8 as a KMT2A target would over-read a hedged yeast result, and the graph's separate DepMap lookup does not show the H2Bub writers themselves ranking as selectively required. Park NH et al. 2026
Connected to