Oncofusion

start / graph / subtraction-mllt1-yeats-retained

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q11b-subtraction

MLLT1 (ENL) refutes the YEATS parallel: the reader is retained in 84% of patients and is required for leukemogenesis

also called ENL YEATS retained · MLL-ENL YEATS · refutation of the YEATS subtraction parallel

We expected the two commonest partners to work the same way — keep the useful bit, throw away the controlling reader. For ENL the opposite is true: the fusion keeps the reader in most patients, and deleting it makes the leukemia weaker, not stronger.

The same thing, for a scientist

The dispatch premise for this beat held that MLLT1 and MLLT3 both discard most or all of the YEATS reader. That is correct for MLLT3 and wrong for MLLT1. Approximately 84.1% of MLL-ENL patients have breakpoints at or before ENL residue 5, so the YEATS domain (residues 1-138), the central IDR and the AHD are all included in the fusion. Hu et al. (Leukemia 2022) deleted the YEATS domain from MLL-ENL and found impaired fusion-protein chromatin binding, reduced Eya1 and Meis1 expression, decreased stem-cell frequency and longer leukemia latency; YEATS-containing MLL-ENL cells are also more sensitive to the YEATS inhibitor SGC-iMLLT. For this partner YEATS is an addition, and a druggable one.

Go deeper
The premise, and why it fails here

The reasoning that produced this beat was: PZP is a targeting reader discarded by MLLT10; YEATS is also a targeting reader; MLLT1 and MLLT3 discard YEATS; therefore the two commonest lymphoblastic and myeloid partners are subtraction-driven and the retained-half framing collapses for most of the disease.

Step three is false for MLLT1. The YEATS domain is retained in 84.1% of MLL-ENL patients 1, because roughly 84% of breakpoints fall at or before ENL amino acid 5 2 — that is, essentially the whole ENL protein rides along in the chimera. This graph's own fusion record already said so (bp-kmt2a-intron-11-mllt1-e2 lists 'most of the ENL YEATS domain (exon-2 junction)' under domains_retained_3); the dispatch brief contradicted the graph, and the graph was right.

Retained, and load-bearing — the opposite of a deleted brake1 deeper

Retention alone would only make MLLT1 a non-example. The functional data make it an active counter-example.

Hu et al. tested the YEATS domain inside MLL-ENL itself, which is exactly the in-cis design missing for every other partner in this beat, run in the deletion direction rather than the add-back direction. Deleting YEATS from MLL-ENL impaired fusion-protein binding and decreased expression of pro-leukemic genes such as Eya1 and Meis1; point mutations that specifically disrupt YEATS binding to acetylated histones decreased stem-cell frequency and increased MLL-ENL-mediated leukemia latency; and YEATS-containing MLL-ENL leukemic cells showed increased sensitivity to the YEATS inhibitor SGC-iMLLT relative to control AML cells 1.

So for ENL the reader is not a constraint the fusion escapes. It is a capability the fusion exploits — a textbook addition, and one with an inhibitor already in hand.

What is actually discarded, and does it matter?

At the commonest (exon-2) junction the fusion loses roughly the first four or five residues of ENL. UniProt annotates the YEATS domain as beginning at residue 1, so this is a marginal N-terminal truncation of the fold rather than removal of a module; the acylated-histone-binding regions annotated at residues 56-58 and 78-80 are untouched. There is no published evidence that losing ENL residues 1-5 does anything, and none should be expected.

A minority architecture is more interesting. Downstream ENL junctions at exons 6-7 do exclude YEATS entirely (bp-kmt2a-intron-11-mllt1-e6-e7 in this graph), producing a fusion whose partner contribution is the AHD alone — architecturally equivalent to MLL-AF9. Those patients are the natural experiment: if YEATS retention is load-bearing, exon-6/7 MLL-ENL should behave like MLL-AF9 and should not respond to SGC-iMLLT. Nobody has looked.

The comparative question this opens

ENL and AF9 are close paralogues with homologous YEATS domains and homologous AHDs, fused to the same 5' partner, and their breakpoints sit at opposite ends of the protein: 84% before ENL residue 5, 92% at or after AF9 residue 376 12. That is a strikingly clean divergence for two proteins this similar, and the field states it as an unanswered question rather than a solved one 2.

Two classes of explanation compete, and they have different consequences for this project:

  • Functional selection. ENL YEATS helps the fusion and AF9 YEATS hinders it, so patients are selected accordingly. If so, MLLT3 really is a subtraction partner and the class is larger than one.
  • Structural accident. Where the breaks fall is set by chromatin accessibility, repeat content and recombination machinery at MLLT1 versus MLLT3 introns, and the functional consequence is downstream noise. If so, MLLT3 is not a subtraction partner and the class stays at one.

This graph has the material to attack the second explanation directly: breakpoint-kmt2a-partner-exon-preference and genome-alu-repeat-landscape-kmt2a-bcr record partner-side breakpoint architecture. Note also that the corresponding wild-type dependency runs the same direction as the retention: ENL is required for AML maintenance and AF9 is not 3.

Facts and figures
partnergene-mllt1
fusionfusion-kmt2a-mllt1
premise testedMLLT1 and MLLT3 both retain the AHD and lose most or all of the YEATS reader, making them subtraction partners like MLLT10
premise verdictrefuted for MLLT1, upheld as a sequence fact for MLLT3
discarded regionENL residues 1-4/5 only (marginal N-terminal truncation of the YEATS fold) at the commonest exon-2 junction
minority architectureENL exon 6-7 junctions exclude YEATS entirely, giving an MLL-AF9-like AHD-only fusion; frequency unquantified in this graph and clinical behaviour untested
genuinely discarded at observed breakpointsfalse
loss shown to contribute to transformationfalse
retention shown to contribute to transformationtrue
verdictaddition
druggable consequenceSGC-iMLLT (YEATS inhibitor) shows increased activity against YEATS-containing MLL-ENL cells, so for this partner the retained-half framing not only holds but names a partner-matched drug
discarded residues
{
  "start": 1,
  "end": 5,
  "uniprot": "Q03111",
  "numbering": "Q03111-1 (559 aa)"
}
retained modules
[
  "YEATS domain aa 1-138 (retained in 84.1% of patients)",
  "central IDR aa 147-484",
  "AHD"
]
breakpoint position in partner
{
  "residue_at_or_before": 5,
  "patient_fraction_pct": 84.1,
  "source": "Hu et al. 2022 (PMID 36435883); Hu & Muntean 2023 (PMID 37295550)"
}
How this could bite us back
  • A YEATS inhibitor deployed on genotype alone, without checking the ENL breakpoint, would be given to roughly one in six t(11;19) patients whose fusion does not contain the target at all.
  • Because YEATS-deleted MLL-ENL still transforms with longer latency, YEATS inhibition is unlikely to be curative as a single agent even in the retaining majority; expect delay, then escape.
Still open
  1. What fraction of MLL-ENL patients carry the exon-6/7 architecture that excludes YEATS, and do those patients differ in lineage, age or outcome? They are the human control group for the whole YEATS argument and nobody has assembled them.
  2. Would exon-6/7 MLL-ENL leukemias be insensitive to SGC-iMLLT? If YEATS retention drives the drug sensitivity reported in E1, that prediction is sharp, cheap and falsifiable.
  3. Is the ENL-versus-AF9 breakpoint divergence selected for function or imposed by the local recombination landscape? Comparing MLLT1 and MLLT3 intronic repeat content, Topo-II sites and accessibility would separate the two without any wet-lab work.
  4. Should MLL-ENL patients be stratified by YEATS retention in trials of YEATS-domain inhibitors? At 84%/16% the minority arm is large enough to matter and small enough to be invisible in a pooled analysis.
Every source behind this entry (3)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The YEATS domain is retained in 84.1% of MLL-ENL patients and is crucial for MLL-ENL-mediated leukemogenesis in mouse models: deletion impairs fusion-protein binding and decreases expression of pro-leukemic genes such as Eya1 and Meis1, point mutations disrupting YEATS binding to acetylated histones decrease stem-cell frequency and increase leukemia latency, and YEATS-containing MLL-ENL cells are more sensitive to the YEATS inhibitor SGC-iMLLT than control AML cells.

    The ENL YEATS epigenetic reader domain critically links MLL-ENL to leukemic stem cell frequency in t(11;19) Leukemia · Hu H, Saha N, Yang Y, Ahmad E, Lachowski L, Shrestha U, Premkumar V, Ropa J, Chen L, Teahan B, Grigsby S, Marschalek R, Nikolovska-Coleska Z, Muntean AG · Leukemia 2022
    PMID 36435883DOI
    What this evidence cannot show

    Mouse models plus cell lines; the 84.1% retention figure comes from the authors' survey of reported patient breakpoints, not from a prospectively typed cohort. Effects on latency and stem-cell frequency are quantitative rather than all-or-none — YEATS-deleted MLL-ENL still transforms, it is just weaker. Verified against PubMed on 2026-09-18.

    Evidence that argues against this (1)
    • Because YEATS-deleted MLL-ENL still causes leukemia with longer latency, the domain is a strong modifier rather than an absolute requirement. A partner-matched therapeutic argument built on it should be framed as modulating leukemic stem-cell frequency, not as removing an obligate dependency. Hu H et al. 2022
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.review

    Approximately 84% of MLL-ENL leukemia patients harbour breakpoints at or before amino acid 5 of ENL, giving fusions that include the YEATS domain, the IDR and the AHD, whereas approximately 92% of MLL-AF9 patients harbour breakpoints at or after AF9 amino acid 376 and exclude the YEATS domain while retaining the AHD; the authors pose the differential retention as an open question.

    PMID 37295550DOI
    What this evidence cannot show

    A review by the same group as E1, so the two sources are not independent. Used as a map of the comparative breakpoint statistics. Full text fetched and the quoted sentences read verbatim on 2026-09-18; PMID and DOI verified against PubMed the same day.

    Evidence that argues against this (1)
    • Both percentages come from one group's survey of reported breakpoints and neither has been recomputed on a systematically typed series, although the series exists. Meyer et al. determined genomic breakpoints in 2,345 patients, identified 135 distinct MLL rearrangements with 94 partners characterised at the molecular level, showed the MLL-side breakpoint shifts with age and lineage, and found that nine fusions carry more than 90% of cases - so per-partner denominators are small for everything outside those nine. The comparative statistic that carries this entry's headline is the one number in it that has never been independently produced, and the dataset capable of producing it is public. Meyer C et al. 2018
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Wild-type ENL, but not its paralogue AF9, is required for disease maintenance in acute myeloid leukaemia; ENL binds acetylated H3, colocalises with H3K27ac and H3K9ac at active promoters, and disrupting YEATS-acetyl binding reduces Pol II recruitment to ENL target genes.

    ENL links histone acetylation to oncogenic gene expression in acute myeloid leukaemia · Wan L, Wen H, Li Y, Lyu J, Xi Y, Hoshii T, Joseph JK, Wang X, Loh YE, Erb MA, Souza AL, Bradner JE, Shen L, Li W, Li H, Allis CD, Armstrong SA, Shi X · Nature 2017
    PMID 28241141DOI
    What this evidence cannot show

    Concerns wild-type ENL, not MLL-ENL. Cited here for the paralogue asymmetry, which runs in the same direction as the breakpoint asymmetry and is therefore suggestive but not decisive. Verified against PubMed on 2026-09-18.

    Evidence that argues against this (1)
    • A maintenance difference between the two paralogues in AML lines is not a difference in reading capacity, which is the property this entry transfers from it. Li et al. showed the AF9 YEATS domain is itself a competent acyl-lysine reader, with selectively HIGHER affinity for crotonyllysine than for acetyllysine through an extended aromatic sandwiching cage, co-localising with crotonylated histone H3 in cells and positively regulating gene expression in a YEATS-dependent manner. Wild-type AF9 being dispensable where wild-type ENL is not is then most simply about abundance, partner occupancy or target-gene set in the particular lines screened, not about AF9's YEATS being a weak module - and neither result speaks to the question this beat actually asks, which is whether AF9 YEATS would constrain the fusion if restored in cis. Li Y et al. 2016
Connected to