Oncofusion

start / graph / ptm-runx1-runx1t1-k43-acetylation-a-retained-site-the-fusion-requires

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.q165-the-retained-acetylation-site

RUNX1 K43 acetylation - a retained-portion site that AML1-ETO genuinely requires, in the 5' partner, with the 3' partner supplying the acetyltransferase in trans

also called AML1-ETO K43 · RUNX1 K43 p300 · Wang 2011 Science

In a different leukaemia fusion, p300 adds a chemical tag to a single lysine in the front half of the protein. Take that one lysine away and mice that would have died in five months survive past eleven. The back half of the fusion is what recruits the tagging enzyme - so the fusion builds the modification that it then depends on.

The same thing, for a scientist

p300 acetylates RUNX1 K43 (UniProt Q01196, 453 aa) within the retained 5' portion of AML1-ETO. Acetyl-K43 is detected in Kasumi-1, in two t(8;21) patient samples and in leukaemic mouse spleen. A K43R AE9a extends median survival from 145 to 328 days and abolishes self-renewal in human CD34+ cord blood without affecting the differentiation block or DNA binding. The p300 docking site is the NHR1 domain of ETO (aa 245-436) in the 3' partner, so the fusion assembles its own writer. This is the closest published analogue to what this beat is asking about KMT2A K1130 - and it flatly corrects a premise in the dispatch brief.

Go deeper
🔴 Two corrections to the brief's premise, both verified here

The dispatch reasoning assumed "p300-mediated acetylation of a specific lysine in the ETO/RUNX1T1 portion (I think K43)", citing PMID 21596992.

1. The residue is in RUNX1, the 5' partner, not in ETO. Verified against live UniProt in this beat 2:

accessionproteinresidue 24residue 43
Q01196RUNX1KK
Q06455RUNX1T1 / ETONV

Q06455 position 43 is valine. K43 cannot be an ETO residue, and the paper says so directly: "A-E, but not ETO, was acetylated by p300, localizing the potential acetylation sites to the AML1 portion of A-E (amino acids 1 to 177)" 1.

2. PMID 21596992 is a bacterial ribosome structure paper. Checked: "Structures of the bacterial ribosome in classical and hybrid states of tRNA binding", Science 2011. The correct PMID is 21764752 13. Two Science 2011 papers, adjacent-looking accession numbers, completely different content - a clean example of why a remembered PMID is not a citation.

The effect sizes, which are the reason this is the benchmark for a retained ACETYLATION site

Mouse AE9a retroviral transduction and transplant, median survival 1:

constructmedian survival
AE9a wild type145 d
AE9a K24R173 d (still leukaemogenic)
AE9a K43R328 d
AE9a K24R/K43Rnot reached

And the specificity of the defect matters as much as its size: K43R abolished self-renewal (CAFC in human CD34+ cord blood) and the self-renewal gene programme, but did NOT affect the differentiation block or DNA binding (EMSA unchanged, expression matched). One acetyl group separates two functions of one fusion protein.

p300 inhibitors reproduce it pharmacologically and selectively: Lys-CoA-Tat 36→54 d, C646 32→43 d in the AE9a model, with no or minimal effect in MLL-AF9 1. 🟠 That MLL-AF9 negative is directly relevant here and cuts against the beat's own direction: a p300-inhibitor experiment has already been run in a KMT2A fusion model and was largely inactive.

The structural lesson: the fusion builds the writer it depends on

Deleting NHR1 (ETO aa 245-436) abolishes both p300 binding and K43 acetylation 1. So the modification is not a property RUNX1 carries into the fusion - it is created by the fusion, because the 3' partner recruits the enzyme to a site in the 5' partner.

This is the mechanism that would make a KMT2A retained site fusion-specific, and it is exactly what nobody has looked for. KMT2A::AFF1 brings AFF1, ELL and P-TEFb to the retained KMT2A N-terminus; whether that changes which enzymes reach K1130/K1133 is unasked. A retained site can become a fusion-specific site if the partner changes its enzymatic environment - which is a far better argument for this beat than 'the site is retained'.

Facts and figures
fusionRUNX1::RUNX1T1 (AML1-ETO)
siteRUNX1 K43
uniprotQ01196
length AA453
partner side5' (RUNX1) - NOT the 3' ETO partner
secondary siteRUNX1 K24
modificationacetylation
writerEP300 / p300
writer docking siteETO NHR1 domain, aa 245-436, in the 3' partner - the fusion recruits its own writer in trans
fusion statusRETAINED (within RUNX1 aa 1-177, the retained 5' portion)
shown for fusionAML1-ETO and AE9a - acetyl-K43 detected in Kasumi-1, 2 t(8;21) patient samples and leukaemic mouse spleen; K43R tested in vivo
residue verified in this beattrue
runx1t1 residue 43V (disproves the 3'-partner assignment)
corrects briefthe dispatch brief placed K43 in RUNX1T1/ETO and cited PMID 21596992 (a bacterial ribosome paper); the site is in RUNX1 and the PMID is 21764752
measurements
median survival, AE9a wil…145.0median survival, AE9a wild type: 145median survival, AE9a K24R173.0median survival, AE9a K24R: 173median survival, AE9a K43R328.0median survival, AE9a K43R: 328
measurements — the numbers behind the chart
[
  {
    "quantity": "median survival, AE9a wild type",
    "subject": "mouse AE9a retroviral transduction/transplant",
    "value": 145,
    "units": "days",
    "n": null,
    "universe": "AE9a mouse model, Wang 2011 Science",
    "null": null,
    "note": "the comparator for every mutant below"
  },
  {
    "quantity": "median survival, AE9a K43R",
    "subject": "mouse AE9a retroviral transduction/transplant",
    "value": 328,
    "units": "days",
    "n": null,
    "universe": "AE9a mouse model, Wang 2011 Science",
    "null": 145,
    "note": "2.26x extension from a single K-to-R substitution; K24R alone gives 173 d and the double mutant does not reach median"
  },
  {
    "quantity": "median survival, AE9a K24R",
    "subject": "mouse AE9a retroviral transduction/transplant",
    "value": 173,
    "units": "days",
    "n": null,
    "universe": "AE9a mouse model, Wang 2011 Science",
    "null": 145,
    "note": "still leukaemogenic - the two lysines are not equivalent"
  }
]
How this could bite us back
  • The AE9a K43R survival result is in a retroviral overexpression mouse model of a truncated fusion variant. The gap between 'K43R extends survival in AE9a mice' and 'a p300 inhibitor helps a t(8;21) patient' is the same gap this beat would face, and the p300 inhibitors in that very paper bought 18 and 11 days.
  • This precedent makes retained-site work look more tractable than the KMT2A case is. AML1-ETO had a clean in vivo model, a single-copy dependency and an identified writer before the site work started. KMT2A K1130 has none of those.
Still open
  1. Does a KMT2A fusion partner change which enzymes reach the retained KMT2A N-terminus, as ETO's NHR1 does for RUNX1 K43? This is the mechanism that would make a retained site fusion-specific, and it is the strongest available argument for this beat - much stronger than retention alone. Nobody has asked it for KMT2A.
  2. Why did p300 inhibitors work in AE9a and not in MLL-AF9 in the same paper? If the answer is that MLL fusions do not depend on a p300-written acetylation, that is a published negative for the direction this beat is exploring, and it deserves to be read rather than worked around.
  3. Is RUNX1 K43 acetylation present in wild-type RUNX1 as well? If it is, the site is another shared-sequence modification and the AE9a phenotype must be explained by the fusion's altered enzymatic context rather than by the site's existence.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    p300 acetylates AML1-ETO at RUNX1 K43 (and K24); acetyl-K43 is present in Kasumi-1, two t(8;21) patient samples and leukaemic mouse spleen; AE9a K43R extends median survival from 145 to 328 days and abolishes self-renewal without affecting the differentiation block or DNA binding; the ETO NHR1 domain (aa 245-436) is required for p300 binding and for K43 acetylation.

    The leukemogenicity of AML1-ETO is dependent on site-specific lysine acetylation · Wang L, Gural A, Sun XJ, Zhao X, Perna F, Huang G, Hatlen MA, Vu L, Liu F, Xu H, Asai T, Xu H, Deblasio T, Menendez S, Voza F, Jiang Y, Cole PA, Zhang J, Melnick A, Roeder RG, Nimer SD · Science 333(6043):765-9 2011
    PMID 21764752DOI
    What this evidence cannot show

    The title and journal were verified against Europe PMC in this beat; the full text and the survival figures were read by a worker agent and are reported at second hand. AE9a is a truncated AML1-ETO variant, not full-length AML1-ETO, and the survival experiment is a mouse retroviral transduction/transplant model. The two patient samples are an n of 2 for the human relevance of the mark.

    Evidence that argues against this (1)
    • p300 inhibitors in the same paper had no or minimal effect in MLL-AF9 - a KMT2A fusion. That is a published negative for the acetyltransferase axis in the fusion family this graph is about, and it should be weighed before any p300-directed proposal here.
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt Q01196 (RUNX1, 453 aa) has lysine at positions 24 and 43; UniProt Q06455 (RUNX1T1/ETO, 604 aa) has asparagine at 24 and valine at 43, so K43 cannot be an ETO residue.

    What this evidence cannot show

    Residue identity verified in this beat against the live UniProt FASTA on 2026-09-24, independently of the worker agent that supplied the reference. Q01196 is the RUNX1b isoform numbering used by the Wang 2011 paper; RUNX1 has RUNX1a and RUNX1c isoforms with different N-termini, and a K43 in RUNX1c numbering would be a different residue. The match of BOTH candidate lysines (K24 and K43) to the canonical Q01196 sequence is what makes the isoform assignment secure here.

    Evidence that argues against this (1)
    • The disproof of the ETO assignment is only as good as the isoform assumption. RUNX1 has RUNX1a/b/c isoforms with different N-termini, so a 'K43' in RUNX1c numbering would be a different residue, and this check would not detect that. What protects it is that BOTH candidate lysines the paper names, K24 and K43, match Q01196 - a coincidence unlikely under the wrong isoform. Note also what the check does NOT establish: it rules out ETO as the host of residue 43, but a sequence lookup cannot confirm that the acetylated residue is the one the paper says it is.
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    PMID 21596992 is 'Structures of the bacterial ribosome in classical and hybrid states of tRNA binding', Science 2011 - not the AML1-ETO acetylation paper.

    What this evidence cannot show

    Recorded because the wrong PMID entered this beat through a dispatch brief and would have propagated into an entry. It resolves to a real, unrelated paper - the most dangerous kind of wrong citation, because it passes every existence check.

    Evidence that argues against this (1)
    • This is a negative identification and it supports less than it appears to. Showing that 21596992 is an unrelated paper does not show that 21764752 is the paper the brief meant - that rests on the title matching the described content. If the brief's author had a third paper in mind, this entry has substituted its own best guess for their intent. The substantive claims in this entry all rest on 21764752 as read, so the risk is confined to the provenance of the error rather than to the science.
Connected to