Oncofusion

start / graph / gene-h2az1

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q37-histone-chaperones

H2AZ1 - the one histone variant with a direct KMT2A-fusion paper: MLL-AF10 and MLL-ENL recruit Tip60 to Hoxa9 to acetylate H2A.Z

also called H2AZ1 · H2AFZ · H2A.Z · H2A.Z-1 · H2AZ · P0C0S5

H2A.Z is a swap-in version of a different core histone that marks genes as ready to be switched on. Unlike the H3 variant this beat mostly deals with, somebody has already shown that MLL fusion proteins act on it directly at the HOXA9 gene.

The same thing, for a scientist

H2A.Z (H2AZ1, P0C0S5, 128 aa) is the H2A variant enriched at promoters and regulatory elements. **It is the only histone variant in this beat with a published, direct connection to a KMT2A fusion:** Tip60 is required for HoxA cluster and Meis1 expression and for the development of MLL-fusion leukaemia; **Tip60 is recruited by MLL-AF10 and MLL-ENL to the Hoxa9 locus where it acetylates H2A.Z, promoting Hoxa9 expression**, and conditional Tip60 deletion prevents the development of MLL-AF10 and MLL-ENL leukaemia [E1]. Its exchange machinery is unusually well defined. **ANP32E removes H2A.Z from chromatin** [E2]; **YL1/VPS72 is the deposition chaperone**, delivering H2A.Z-H2B to the SRCAP and p400/TIP60 remodellers [E3]. ANP32E is itself a member of the p400/TIP60 complex [E2] - the same complex Tip60 acts in. **And the dependency data run the opposite way to the literature.** H2AZ1 is more required in KMT2A-rearranged lines (difference -0.184, pan-essentiality -0.770), while every component of its exchange machinery is **less** required: ANP32E +0.093, SRCAP +0.118, EP400 +0.121, VPS72 -0.051 [E4]. The variant matters more; the machines that move it matter less.

Go deeper
The pattern nobody in this graph has stated

In the 29-gene scan, the three most positive differences - the genes LEAST required in KMT2A-rearranged lines relative to other leukaemia lines - are SSRP1 (+0.265), EP400 (+0.121) and SRCAP (+0.118), with SUPT16H (+0.117) and ANP32E (+0.093) immediately behind 4.

That is FACT (SSRP1, SUPT16H) and the H2A.Z exchange machinery (EP400, SRCAP, ANP32E), and this graph already carries protein-fact-complex, gene-ssrp1 and gene-supt16h as part of the fusion's proposed elongation mechanism. The screen says KMT2A-rearranged lines need FACT less than other leukaemia lines do.

Three readings, and this beat cannot choose between them:

  1. Noise. Eleven lines, no null, and a positive tail is as likely to be noise as a negative one.
  2. Lineage. The index arm is myeloid-enriched; FACT and H2A.Z machinery requirements may differ by lineage. The split does not resolve it (SSRP1 +0.206 myeloid, +0.354 lymphoid - positive in both).
  3. Real. A fusion that supplies its own elongation drive may need less of the general elongation and exchange apparatus.

Reading 3 would be an important negative for this graph's own mechanism and it is recorded here so it is not lost. It is a hypothesis, not a finding, and the honest prior is reading 1.

The chromatin address, and the fusions are MLL-AF10 and MLL-ENL (Q87, 2026-09-21)

Subcellular. Nucleus and chromosome (UniProt P0C0S5 - both terms UNCODED, no citation); HPA v25.1 IF nucleoplasm, reliability Supported. For a histone variant this is the uninformative half of the question.

Subnuclear = which chromatin. Promoters and enhancers in murine ESCs, where H2A.Z knockdown reduces MLL-complex binding to active genes and PRC2 binding to repressed ones (PMID 23260488); the centromere and pericentric heterochromatin (PMID 17194760); and, in leukaemia, the Hoxa9 locus.

Shown for fusion: YES, in MLL-AF10 and MLL-ENL. Tip60 recruited by those two fusions acetylates H2A.Z at Hoxa9 (PMID 33967269, mouse AML). Not KMT2A::AFF1, not B-lineage, and one locus rather than a map.

The unresolved direction. H2A.Z is the substrate in the leukaemia work and upstream of MLL-complex recruitment in the ESC work. Two live causal orderings, two systems, neither testing the other.

Facts and figures
uniprotP0C0S5
gene symbolH2AZ1
retired symbolH2AFZ
length AA128
paralogueH2AZ2
removal chaperonegene-anp32e
deposition chaperoneVPS72/YL1
kmt2a fusion linkTip60 recruited by MLL-AF10 and MLL-ENL to Hoxa9; acetylates H2A.Z (PMID 33967269)
subnuclear localizationCHROMATIN - specifically promoters and enhancers (PMID 23260488, murine ESCs) and the centromere/pericentric heterochromatin (PMID 17194760), and in a KMT2A-fusion context the Hoxa9 locus, where Tip60 recruited by MLL-AF10 or MLL-ENL acetylates H2A.Z (PMID 33967269). UniProt's 'Nucleus' and 'Chromosome' terms for P0C0S5 are both UNCODED and carry no citation; the nuclear call is corroborated independently by HPA v25.1 IF at reliability Supported.
cell stateSELF-RENEWAL VERSUS DIFFERENTIATION, at the chromatin-accessibility layer. H2A.Z is required for both efficient self-renewal AND efficient differentiation of murine ESCs, and it gates the access of both activating (MLL) and repressing (PRC2) complexes to chromatin - so it is a permissiveness setting rather than a switch pointing one way (PMID 23260488). In leukaemia the corresponding state statement is fusion-driven HoxA/Meis1 programme maintenance (PMID 33967269).
shown for fusionYES, and the fusions are MLL-AF10 and MLL-ENL - not the index fusion. Tip60 is recruited by MLL-AF10 and MLL-ENL to the Hoxa9 locus where it acetylates H2A.Z, promoting Hoxa9 expression; conditional Tip60 deletion impairs MLL-fusion leukaemia development (PMID 33967269, mouse AML). This is a chromatin-ADDRESS result in a named fusion, which is rarer and more valuable than a compartment call - but it is NOT KMT2A::AFF1 and it is NOT B-lineage, and it is an acetylation result at one locus rather than a genome-wide map of where H2A.Z sits in a fusion cell.
depmap
index_mean-1.166index_mean: -1.1658other_leukaemia_mean-0.982other_leukaemia_mean: -0.9816pan_essentiality-0.770pan_essentiality: -0.7702lymphoid_difference-0.204lymphoid_difference: -0.2037difference-0.184difference: -0.1842myeloid_difference-0.177myeloid_difference: -0.177n_index11.0n_index: 11
deposition remodellers
[
  "gene-srcap",
  "gene-ep400"
]
depmap — the numbers behind the chart
{
  "index_mean": -1.1658,
  "other_leukaemia_mean": -0.9816,
  "difference": -0.1842,
  "pan_essentiality": -0.7702,
  "myeloid_difference": -0.177,
  "lymphoid_difference": -0.2037,
  "n_index": 11
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -1.1658,
  "chronos_mean_other_leukaemia": -0.9816,
  "difference": -0.1842,
  "pan_essentiality_all_lines": -0.7702,
  "selectivity_rank": 302,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
localization
{
  "subcellular": "NUCLEUS and CHROMOSOME. UniProt P0C0S5 lists exactly those two terms and BOTH ARE UNCODED - no ECO:0000269, no ECO:0000305, no citation - so this is a database statement rather than a measurement (rule 91). HPA v25.1 IF, reliability Supported, gives a single main location, Nucleoplasm, which is the independent antibody corroboration the UniProt record lacks. For a replication-independent histone variant the compartment question is close to trivial; the address that matters is the chromatin one below.",
  "subnuclear": "CHROMATIN, at three resolvable addresses. (1) PROMOTERS AND ENHANCERS: H2A.Z is highly enriched at both in murine ESCs, its deposition produces an abnormal nucleosome with decreased occupancy and increased accessibility, and H2A.Z knockdown DECREASES binding of MLL complexes to active genes and of PRC2 to repressed genes (PMID 23260488) - so H2A.Z is upstream of this graph's central complex at the chromatin level. (2) THE CENTROMERE AND PERICENTRIC HETEROCHROMATIN: H2A.Z is a structural component of the centromere, distributed nonuniformly through heterochromatin and interspersed with H3K4me2 nucleosomes in centric chromatin (PMID 17194760); (\"H2A.Z\"[tiab] OR H2AFZ[tiab] OR H2AZ1[tiab]) AND heterochromatin[tiab] returns 80 records (2026-09-21), a literature that was NOT title-checked and is quoted only as a size. (3) A NAMED LOCUS IN A NAMED FUSION: the Hoxa9 locus in MLL-AF10 and MLL-ENL AML (see shown_for_fusion).",
  "why_this_is_the_right_answer_shape": "For a core histone variant, 'nucleus' is not informative and a compartment name like 'speckle' would be wrong. The meaningful subnuclear answer for H2A.Z is WHICH chromatin, and that is what is recorded here.",
  "how_shown": "ChIP-seq, nucleosome-occupancy and accessibility assays plus knockdown in murine ESCs (PMID 23260488); chromatin-fibre extension and immunofluorescence at the centromere (PMID 17194760); ChIP at Hoxa9 with conditional deletion in mouse AML (PMID 33967269)",
  "node": "localization-chromatin-bound-versus-soluble-fraction-kmt2a"
}
cell type
{
  "haematopoietic_answer": "MOUSE AML BLASTS driven by MLL-AF10 or MLL-ENL. Tip60 is recruited by those two fusions to the Hoxa9 locus, where it acetylates H2A.Z and promotes Hoxa9 expression; Tip60 is required for HoxA-cluster and Meis1 expression and for the development of MLL-fusion leukaemia. This is the strongest haematopoietic chromatin-address statement in the tranche, and it names the fusions.",
  "the_other_system": "murine embryonic stem cells for the promoter/enhancer and MLL-complex-recruitment result (PMID 23260488) - not haematopoietic, and self-renewal in an ESC is not self-renewal in an LSC.",
  "expression_atlas": "HPA v25.1: RNA tissue specificity Low; IF reliability Supported (nucleoplasm)."
}
How this could bite us back
  • H2A.Z is an essential histone variant at promoters throughout the genome and is not buffered by its paralogue in these lines, so a direct block has no window.
  • Tip60 (KAT5) is a tumour suppressor in several contexts and is required for DNA damage signalling; inhibiting it in a child carries a genome-instability risk, and the EPC1/EPC2 work shows the same complex restrains MYC accumulation - losing that restraint is not obviously therapeutic.
  • If a KMT2A fusion recruits an acetyltransferase to H2A.Z at HOXA through its C-terminal partner, that is a route to HOXA activation that a menin inhibitor cannot touch, and it would be fusion-partner-specific - a resistance mechanism that differs between t(4;11) and t(9;11).
Still open
  1. Does KMT2A::AFF1 - the index fusion - recruit Tip60 to HOXA9 the way MLL-AF10 and MLL-ENL do? Both published fusions carry different C-terminal partners and neither is AFF1.
  2. Which H2A.Z residue does Tip60 acetylate at Hoxa9, and is the acetylated form the functional species? The abstract does not name it and this graph files marks individually.
  3. Why are FACT and the H2A.Z exchange machinery the LEAST required genes in the KMT2A-rearranged arm of this scan? A permutation null on the positive tail would say whether that is noise, and nobody has run one on a positive tail in this project.
  4. Is there a menin-inhibitor-independent H2A.Z acetylation route to HOXA activation? If Tip60 is recruited by the fusion's C-terminal partner rather than its N-terminus, menin inhibition would not displace it.
Every source behind this entry (7)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Tip60, a MYST-family histone acetyltransferase, is required for expression of the HoxA cluster and Meis1 and for the development of MLL-fusion leukaemia; Tip60 is recruited by MLL-AF10 and MLL-ENL fusions to the Hoxa9 locus where it acetylates H2A.Z, thereby promoting Hoxa9 gene expression; conditional deletion of Tip60 prevented the development of MLL-AF10 and MLL-ENL leukaemia.

    PMID 33967269DOI
    What this evidence cannot show

    Abstract read from efetch 2026-09-20; full text not read (no PMCID in the record). **Mouse models of MLL-AF10 and MLL-ENL acute MYELOID leukaemia.** The index disease of this graph is infant B-ALL with KMT2A::AFF1, which is neither of these fusions and not myeloid. The acetylation claim is site-specific on H2A.Z but the abstract does not name the residue. Graded strong as a generalisation from two fusions in one lineage to 'KMT2A fusions act on H2A.Z'; within the reported models it is established.

    Evidence that argues against this (1)
    • The same complex Tip60 works in - p400/TIP60 - is LESS required in KMT2A-rearranged lines in DepMap (EP400 +0.121), and EPC1/EPC2, two other EP400-complex components, were found to be required in MLL-mutated AML through MYC accumulation rather than through HOXA. Two published routes through the same complex disagree about which output matters. Huang X et al. 2014
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    ANP32E is a specific H2A.Z chaperone and a member of the p400/TIP60 histone-exchange complex; it interacts with a short region of the H2A.Z docking domain through an H2A.Z-interacting domain (ZID), and the 1.48 A crystal structure of the ANP32E-ZID/H2A.Z-H2B complex supports a mechanism of H2A.Z/H2B eviction from the nucleosome stabilised by extension of the H2A.Z C-terminal alpha-helix; in ANP32E-null cells H2A.Z shows genome-wide enrichment, redistribution and accumulation at enhancers and insulators.

    ANP32E is a histone chaperone that removes H2A.Z from chromatin. · Obri A, Ouararhni K, Papin C, et al. · Nature 2014
    PMID 24463511DOI
    What this evidence cannot show

    Abstract read from efetch 2026-09-20; full text not read (no PMCID in the record). Mouse and human cell systems with crystallography; established for the biochemistry and the knockout redistribution. Not leukaemia. **No residue spans stored** - this project requires them to be derived by alignment to a freshly fetched sequence, not read off the deposit.

    Evidence that argues against this (1)
    • A 2026 preprint reports cooperative AND opposing functions of ANP32E and VPS72 at gene promoters, so 'ANP32E removes, YL1 deposits' is an incomplete account of a two-way system. The preprint is not peer reviewed and is recorded as a bound on the current model rather than as a result. Murphy P et al. 2026
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    YL1 (VPS72) is a specific H2A.Z deposition chaperone; the 2.7 A crystal structure of the human YL1-H2A.Z-H2B complex shows YL1 binding triggers extension of the H2A.Z alphaC helix, similarly to ANP32E, with a more extensive interaction covering the extended acidic patch and the whole DNA-binding surface; substitution of only four H2A residues creates an H2A.Z-like interface recognised by YL1, and the structure illuminates H2A.Z transfer to the nucleosome by the SRCAP and p400/TIP60 remodelling complexes.

    Molecular basis and specificity of H2A.Z-H2B recognition and deposition by the histone chaperone YL1. · Latrick CM, Marek M, Ouararhni K, et al. · Nature Structural & Molecular Biology 2016
    PMID 26974126DOI
    What this evidence cannot show

    Abstract read from efetch 2026-09-20; full text not read (no PMCID in the record). Crystallography of a chaperone-histone complex; established for the interaction. The statement about SRCAP and p400/TIP60 is the authors' inference from the structure ('shed light on the mechanism'), not a measurement of transfer. Same laboratory as E2, so E2 and E3 are not independent observations. No leukaemia.

    Evidence that argues against this (1)
    • Both deposition remodellers named here - SRCAP and EP400 - are LESS required in KMT2A-rearranged lines than in other leukaemia lines in DepMap, which means the structurally defined deposition route has no positive dependency evidence in this disease. Oncofusion knowledge graph (internal analysis) 2026
  4. E4weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset

    In DepMap CRISPR gene effect over 11 KMT2A-rearranged and 43 other leukaemia lines, H2AZ1 is MORE required in the index arm (index -1.166, other -0.982, difference -0.184, pan-essentiality -0.770) while its exchange machinery is LESS required: ANP32E +0.093, SRCAP +0.118, EP400 +0.121, VPS72 -0.051, YEATS4 -0.052, DMAP1 -0.043. The three most positive differences among 29 genes scanned are SSRP1 (+0.265), EP400 (+0.121) and SRCAP (+0.118).

    What this evidence cannot show

    Run 2026-09-20. **Weak.** No permutation null; 29 genes scanned; 11 index lines, 43 comparators, no infant KMT2A::AFF1 line. H2AZ1's own pan-essentiality of -0.77 is past this project's -0.5 convention, so its difference is partly floor-compressed. **The positive tail is reported because a positive difference is as much a result as a negative one and this project has never recorded one**, not because it is established.

    Evidence that argues against this (1)
    • H2AZ2 exists as a paralogue (difference -0.039, pan-essentiality -0.057, i.e. required nowhere), so H2AZ1's apparent requirement is not buffered the way the ASF1 and UBN pairs are - which makes it more likely to be a genuine requirement and less likely to be a selective one, since an unbuffered essential gene is required everywhere. Oncofusion knowledge graph (internal analysis) 2026
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    H2A.Z is UniProt P0C0S5 (H2AZ_HUMAN), histone H2A.Z, 128 amino acids, primary gene symbol H2AZ1 with alias symbols H2AFZ and H2AZ; the paralogue H2AZ2 is a separate gene. As of 2026-09-20 '(H2AFZ OR H2AZ1 OR H2A.Z) AND (MLL OR KMT2A)' returns 3 PubMed records, of which one is the Tip60/MLL-AF10 paper, one an ES-cell H2A.Z study and one a C. elegans DCC-localisation paper.

    UniProtKB P0C0S5 (H2AZ_HUMAN), Histone H2A.Z, 128 aa · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
    What this evidence cannot show

    Fetched 2026-09-20 with `gene_primary` requested and read back per rule 62: exactly one reviewed human record with primary gene name H2AZ1. **The retired symbol H2AFZ was included in the literature query deliberately** - this project's documented tenfold under-count for the septins came from searching only current symbols, and here the retired symbol is the one most of the literature uses. All three returned titles were fetched and read.

    Evidence that argues against this (1)
    • Three records is a very small literature for a variant with this much mechanistic work behind it, which suggests the symbol-pair query is missing papers that frame the result at the level of Tip60, p400 or NuA4 rather than naming the histone - the same complex-name blind spot that hid the MLL/TAF-I result from every symbol query in this beat. Huang X et al. 2014
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In murine embryonic stem cells, H2A.Z is highly enriched at promoters and enhancers and its deposition produces an abnormal nucleosome with decreased occupancy and increased chromatin accessibility; knocking H2A.Z down compromises OCT4 binding, DECREASES binding of MLL complexes to active genes, and decreases PRC2 binding to repressed genes. H2A.Z's subnuclear address is therefore promoter and enhancer chromatin, and it sits upstream of MLL-complex recruitment there.

    PMID 23260488DOI
    What this evidence cannot show

    Murine EMBRYONIC STEM CELLS, not haematopoietic cells and not leukaemia; ESC self-renewal is not LSC self-renewal, and the MLL complexes assayed are wild-type, not a fusion. Knockdown of a core histone variant perturbs nucleosome composition globally, so a decrease in MLL-complex binding is consistent with a general loss of accessibility as well as with a specific requirement.

    Evidence that argues against this (1)
    • 'H2A.Z marks active promoters and enhancers' is the version of this fact that travels, and it is incomplete in a way that would mislead anyone designing a ChIP experiment: H2A.Z is also a structural component of the CENTROMERE, distributed nonuniformly through pericentric heterochromatin and interspersed with H3K4me2-containing nucleosomes in centric chromatin, contributing to the centromere's 3D organisation. So a genome-wide H2A.Z signal in a leukaemic cell contains a heterochromatic component that an enhancer-focused analysis would silently discard, and a change in bulk H2A.Z is not automatically a change at enhancers. Greaves IK et al. 2007
  7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a KMT2A-fusion context the chromatin address is a named locus: Tip60 is recruited by MLL-AF10 and MLL-ENL to the Hoxa9 locus, where it acetylates H2A.Z and thereby promotes Hoxa9 expression; Tip60 is required for HoxA cluster and Meis1 expression and for the development of MLL-fusion leukaemia in mouse.

    PMID 33967269DOI
    What this evidence cannot show

    MOUSE AML, and the fusions are MLL-AF10 and MLL-ENL - NOT the index fusion KMT2A::AFF1, and not B-lineage. The experiment measures an acetylation event at one locus, not where H2A.Z sits genome-wide in a fusion cell, and the genetics is on Tip60 rather than on H2A.Z, so H2A.Z's requirement is inferred from its being the substrate rather than demonstrated by removing it.

    Evidence that argues against this (1)
    • The direction of the arrow is worth arguing about before this is built on. Here H2A.Z is the SUBSTRATE that a fusion-recruited acetyltransferase modifies; in murine ESCs H2A.Z is UPSTREAM, and its knockdown reduces MLL-complex binding to active genes. Those are opposite causal orderings for the same pair of molecules, in two different systems, and neither experiment was designed to test the other's ordering. A proposal that treats H2A.Z as a downstream mark of fusion activity, or as an upstream licence for fusion binding, is picking one of two live models rather than reading a settled fact. Hu G et al. 2013
Connected to