EP300 — the acetyltransferase whose paralogue split is the finding: p300 is the 108th most KMT2A-selective gene, CBP is the 16,950th

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.

• p300 acetylates hundreds of non-histone substrates including p53. On-target toxicity is the dominant risk for this target class and is not a formulation problem.
• Every clinical CBP/p300 inhibitor is dual, so the paralogue selectivity this entry argues for does not currently exist in any molecule, and building a strategy on it means building a molecule first.
• CREBBP is a tumour suppressor in lymphoid malignancy. Inhibiting it in a B-lineage leukaemia - which a dual agent necessarily does - could be actively harmful, and this project's disease is B-lineage.
• I-CBP112 induced differentiation in MLL-AF9 cells. In a KMT2A-rearranged patient, induced differentiation is the physiology behind differentiation syndrome, the dose-limiting toxicity of the menin inhibitors.
• The reported synergy with JQ1 combines two epigenetic agents with overlapping haematological toxicity, which is the combination most likely to fail on tolerability rather than on efficacy.

1. Is the EP300/CREBBP selectivity split real? Permutation testing, lineage conditioning and line normalisation - all of which this repository has scripts for - should be run before anything is built on it. This is the single highest-value cheap analysis this beat leaves open.
2. Would a p300-selective agent outperform a dual CBP/p300 agent in KMT2A-rearranged leukaemia? Every clinical compound is dual, and the dependency data predict the CBP arm contributes toxicity without selectivity.
3. Does EP300 contact KMT2A::AFF1, or is the requirement entirely downstream? A co-IP in a fusion-positive B-ALL line would settle the classification.
4. Does I-CBP112 or a clinical CBP/p300 inhibitor work in KMT2A::AFF1 infant B-ALL? Every published result is myeloid.
5. Bromodomain versus HAT inhibition give different phenotypes in principle. Which arm delivers the MLL-AF9 effect?

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   UniProtKB Q09472 (EP300_HUMAN, Histone acetyltransferase p300) is 2414 aa and annotates a KIX domain at aa 566-645, a bromodomain at aa 1049-1156, a CBP/p300-type HAT domain at aa 1287-1663, a TAZ-type 1 zinc finger at aa 331-417, a ZZ-type zinc finger at aa 1665-1713 and a TAZ-type 2 zinc finger at aa 1728-1809.

   UniProtKB Q09472 (EP300_HUMAN), Histone acetyltransferase p300 · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
   ▸What this evidence cannot show

   Record fetched 2026-09-19; primary gene name read back as EP300 and length 2414 confirmed from the same fetched record, and every coordinate copied from it. Domain boundaries are curator annotations and differ from crystallographic construct boundaries.

   ▸Evidence that argues against this (1)

   • A domain map of p300 conceals what makes it hard to drug: the HAT and bromodomain are embedded in one of the most intrinsically disordered large proteins in the proteome, and p300's substrate range extends far beyond histones to hundreds of non-histone proteins including p53. Reading these coordinates as a target map understates both the selectivity problem and the on-target toxicity. Giotopoulos G et al. 2016
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   In vitro and in vivo genetic ablation and knockdown experiments in murine models demonstrate a role for CBP and p300 in the induction and maintenance of AML; selective small-molecule inhibitors of their lysine acetyltransferase activity validate CBP/p300 as therapeutic targets across a wide range of human AML subtypes, producing growth retardation through transcriptional changes that induce apoptosis and cell-cycle arrest, and decreasing clonogenic growth of primary AML patient samples.

   The epigenetic regulators CBP and p300 facilitate leukemogenesis and represent therapeutic targets in acute myeloid leukemia · Giotopoulos G, Chan WI, Horton SJ, Ruau D, Gallipoli P, Fowler A, Crawley C, Papaemmanuil E, Campbell PJ, Gottgens B, Van Deursen JM, Cole PA, Huntly BJP · Oncogene 35(3):279-89 2016
   PMID 25893291DOI
   ▸What this evidence cannot show

   Abstract fetched and read on 2026-09-19; full text not read. CBP and p300 are treated together throughout, so this paper cannot separate the paralogues and provides no support for the p300-specific reading this entry proposes. 'A wide range of human AML subtypes' is not resolved to KMT2A-rearranged in the abstract, and no B-lineage model appears.

   ▸Evidence that argues against this (1)

   • The DepMap split directly contradicts the paper's framing of CBP and p300 as an interchangeable pair: in 1,208 cell lines EP300 is required (-0.359) and CREBBP is close to dispensable (-0.091), and in KMT2A-rearranged lines they move in opposite directions. Whatever the murine genetics showed, the two genes are not equivalent in human cells. Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   I-CBP112, a specific and potent acetyl-lysine competitive protein-protein interaction inhibitor targeting the CBP/p300 bromodomains, substantially impaired colony formation and induced cellular differentiation in human and mouse leukaemic cell lines without significant cytotoxicity; it significantly reduced the leukaemia-initiating potential of MLL-AF9-positive acute myeloid leukaemia cells in a dose-dependent manner in vitro and in vivo; and it increased the cytotoxic activity of the BET bromodomain inhibitor JQ1 and of doxorubicin. CBP/p300 are themselves involved in recurrent leukaemia-associated chromosomal translocations.

   Generation of a Selective Small Molecule Inhibitor of the CBP/p300 Bromodomain for Leukemia Therapy · Picaud S, Fedorov O, Thanasopoulou A, Leonards K, Jones K, Meier J, Olzscha H, Monteiro O, Martin S, Philpott M, Tumber A, Filippakopoulos P, Yapp C, Wells C, Che KH, Bannister A, Robson S, Kumar U, Parr N, Lee K, Lugo D, Jeffrey P, Taylor S, Vecellio ML, Bountra C, Brennan PE, O'Mahony A, Velichko S, Muller S, Hay D, Daniels DL, Urh M, La Thangue NB, Kouzarides T, Prinjha R, Schwaller J, Knapp S · Cancer Research 75(23):5106-5119 2015
   PMID 26552700DOI
   ▸What this evidence cannot show

   Abstract fetched and read on 2026-09-19; full text not read. I-CBP112 is a **dual CBP/p300** bromodomain inhibitor and cannot distinguish the paralogues, so it does not test the p300-specific hypothesis this entry raises. The fusion is MLL-AF9 (myeloid), not KMT2A::AFF1. 'Without significant cytotoxicity' is a statement about cell lines in culture and not a safety claim. I-CBP112 is a chemical probe, not a clinical candidate.

   ▸Evidence that argues against this (1)

   • The bromodomain is not the catalytic module, and a bromodomain inhibitor and a HAT inhibitor of the same protein can give different and even opposing phenotypes - the same genetic-versus-chemical direction problem this graph documented for USP22. Giotopoulos et al. validated CBP/p300 with **KAT** (catalytic) inhibitors while Picaud et al. used a **bromodomain** inhibitor, so the two papers do not straightforwardly corroborate each other. Giotopoulos G et al. 2016
4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

   In the project's genome-wide KMT2A-rearranged selectivity scan of 17,728 genes, EP300 ranks 108 (0.609th percentile) with a difference of -0.2837 (KMT2A-rearranged mean -0.9967 against other-leukaemia mean -0.7130) and a pan-essentiality of -0.3585 across 1,208 lines, while its paralogue CREBBP ranks 16,950 (95.611th percentile) with a difference of +0.1032 and a pan-essentiality of -0.0908. For comparison in the same scan DOT1L ranks 184 (-0.2310), AFF1 ranks 160 (-0.2472), KMT2A ranks 11 (-0.5411) and MEN1 ranks 1 (-0.8656).

   Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap Chronos via Open Targets) · Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026
   ▸What this evidence cannot show

   All six rows read out of analyses/depmap-kmt2a-selective-scan-full/results.json on 2026-09-19 in one pass. Eleven KMT2A-rearranged against 43 other leukaemia lines, AML-heavy, none infant-derived. Not permutation-tested for EP300 specifically, not lineage-conditioned, not line-normalised - all three of which this project has scripts for and none of which was run here.

   ▸Evidence that argues against this (1)

   • CREBBP is an established tumour suppressor in lymphoid malignancy, so its positive gene effect may reflect that rather than anything about KMT2A rearrangement, and the striking 'paralogue split' could be one real dependency next to one unrelated tumour-suppressor signal rather than a single interpretable contrast. The split is the observation; the interpretation offered in the prose is not established. Picaud S et al. 2015
5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

   PubMed returns 66 records for '(MLL OR KMT2A) AND (EP300 OR p300) AND leukemia' and 15 for '(p300 OR EP300 OR CBP) AND ("MLL-AF9" OR "MLL-AF4" OR "MLL fusion" OR "KMT2A-rearranged")' as of 2026-09-19, so unlike most entries in this beat EP300 has a real literature in this disease.

   PubMed E-utilities esearch record counts for EP300 and p300 against KMT2A, MLL and MLL-fusion terms, run for this beat · NCBI PubMed, queried 2026-09-19 2026
   ▸What this evidence cannot show

   'p300' as a search term also matches unrelated uses of the string in other contexts, so the 66-record count is an upper bound and was not title-checked record by record. Only two of the fifteen fusion-term records were read in full abstract for this entry.

   ▸Evidence that argues against this (1)

   • A count of 66 says the topic is studied, not that EP300's role in KMT2A-rearranged leukaemia specifically is settled. Only two of these records were read here, and neither of them separates EP300 from CREBBP - which is precisely the question this entry says matters most. Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026

• related to →CREBBP (CBP) — binds the KMT2A transactivation domain, which no fusion keeps

  paralogue split: EP300 rank 108 (-0.284) against CREBBP rank 16,950 (+0.103) in the same KMT2A selectivity scan, yet every clinical inhibitor is dual

• depends on →KMT2A::MLLT3 (MLL-AF9), t(9;11)(p21.3;q23.3)

  I-CBP112, a selective CBP/p300 bromodomain inhibitor, reduced MLL-AF9 leukaemia-initiating potential in vitro and in vivo and induced differentiation without significant cytotoxicity

• related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)

  NOT a demonstrated interactor: the evidence is genetic requirement and pharmacology, with no experiment placing p300 on a KMT2A fusion protein

• related to →MYB — rank 17 of 17,728 for KMT2A selectivity, and the only node in this beat where partial, transient suppression eradicated leukaemia without stopping normal blood production

  the p300 KIX domain (aa 566-645) is the classic MYB cofactor interface and the most obvious indirect route to an otherwise unligandable transcription factor

• ← related toTwo of the three top new dependencies are myeloid identity, not KMT2A

  EP300's dependency is weak in both halves once split, which bounds the EP300-versus-CREBBP paralogue claim.

• ← related toH3K27ac - named in 40 entries with no entry of its own, and the strongest KMT2A::AFF1 measurement of it is invisible to every abstract search

  EP300 writes this mark and is bound at the MLL-AF4 BCL-2 enhancer

• ← depends onAKT and ATM push ELL between two complexes: K355 acetylation sends it to SEC, ATM phosphorylation sends it to LEC

  p300 is the K355 writer

• ← bindsNUTM1 (NUT) - a testis-restricted, fully disordered protein that writes no mark and instead switches ON a writer, and the 3' partner of the KMT2A-wild-type infant B-ALL with the opposite prognosis

  NUT TAD1 (403-418) and TAD2 (419-470) each bind the p300 TAZ2 domain as an amphipathic helix and allosterically ACTIVATE the acetyltransferase; NMR 7XEZ/7XFG

• ← recruitsBRD9::NUTM1, t(5;15)(p15.33;q14) - one fusion is enough to make infant B-ALL in a mouse, and it is the one that behaves best in a child

  INFERRED, NOT SHOWN for this fusion: p300 recruitment and allosteric activation is established for BRD4-NUT via the retained NUTM1 TADs; the H3K27ac rise in NUTM1-r leukaemia is reported without naming the writer

• ← bindsATF4 / CREB2 (P18848) - the hub this graph inferred from its targets and never wrote, and its own transcript does not move under either menin inhibitor

  p300 stabilises ATF4 and is required for its recruitment into nuclear speckles, independently of p300 acetyltransferase activity (PMID 16219772)

• regulates →KMT2A / MLL1 — the gene that breaks

  p300 acetylates MLL1 (pan-acetyl readout, C-terminally biased); not shown at K1130/K1133 specifically

• ← depends onRUNX1 K43 acetylation - a retained-portion site that AML1-ETO genuinely requires, in the 5' partner, with the 3' partner supplying the acetyltransferase in trans

  p300 is the writer; the ETO NHR1 domain docks it in trans

• ← part ofEP300::ZNF384 - the commonest ZNF384 fusion, and it throws away the acetyltransferase

  EP300 exon 6 contributes about 450 N-terminal residues carrying TAZ1; the HAT and bromodomain are deleted

• ← refutesXPO1 / exportin-1 (CRM1) - the cell's main nuclear export receptor, and a chromatin-bound transcriptional scaffold

  XPO1 and p300 co-occupy HUNDREDS of chromatin loci and an XPO1 degrader phenocopies p300 inhibition (PMID 38917791) - so chromatin-bound XPO1 is a general p300-associated cofactor, not a HOX-restricted postal address, and finding it at HOXA in a HOX-driven line discriminates nothing

• ← refutesChromatin-bound XPO1 as a shared, druggable address for HOXA-driven leukaemia - and the objection that it may be a general chromatin cofactor rather than an address at all

  the hypothesis opens on 'a LOCUS-RESTRICTED pool'; an independent group with a degrader and an XPO1-C528S resistance control finds XPO1 across hundreds of p300-co-occupied loci, which removes the locus restriction the address metaphor needs

• ← refutesTwo pools of one protein: the chromatin-bound XPO1 that matters here has never been separated biochemically from the soluble export machine

  the missing chromatin annotation in UniProt is missing a LARGE and general pool rather than a rare specialised one - XPO1 co-occupies hundreds of loci with p300 and its removal detaches p300 genome-wide

• ← refutesThree measurements the export axis needs and nobody has published: XPO1 occupancy in KMT2A-r, H3K79 under an XPO1 inhibitor, and CRM1 occupancy under a menin inhibitor

  the counted absence is not 'nobody has mapped chromatin-bound XPO1' but 'nobody has done it in a KMT2A-rearranged cell': a genome-wide XPO1 occupancy map exists in a chemical-biology paper framed on p300 and is unreachable from any KMT2A-scoped query (brief rule 77)

• ← related toJMY - one AP-MS row with AFF4, in HCT116 only, from a paper whose whole argument is that interactomes differ between cell lines

  JMY's nuclear function is as a p53 co-activator acting through p300/EP300; recorded as an adjacency with no experiment joining it to AFF4

• ← related toCREBBP (CBP) — binds the KMT2A transactivation domain, which no fusion keeps

  paralogous acetyltransferases with the same readout; A-485 and the knockouts in PMID 38693103 cannot separate them

• ← inhibitsSET / TAF-I beta - a histone chaperone that binds the retained N-terminus of MLL, co-activates HOXA with it, and is itself a HOXA-activating fusion partner that recruits DOT1L

  the same INHAT activity inhibits p300 alongside CBP and PCAF