Oncofusion

start / graph / ana-kmt2a-14-3-3-proximity-two-studies

Analysismoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q20-proximity-interactome

KMT2A sits in the 14-3-3 proximity neighbourhood, and the YWHAG link reproduces across two independent studies

also called KMT2A 14-3-3 · YWHAG KMT2A proximity · SFN KMT2A

A 2023 study tagged the 14-3-3 proteins - general-purpose molecular minders that hold other proteins in particular shapes - and found KMT2A among their neighbours. A different 2024 study, different laboratory, found the same thing for one of them.

The same thing, for a scientist

A proximity survey assigning 14-3-3 proteins a central chaperone-like role in human cells recovers **KMT2A as prey of six family members: YWHAB, YWHAE, YWHAG, YWHAH, YWHAQ and SFN** (SFN twice), with YWHAG's row carrying an elevated IntAct MI-score of 0.41 [E1]. Independently, in the K562 50-RBP map, **YWHAG is one of the 25 baits that recovers KMT2A** [E2]. That is the **only cross-study, cross-laboratory reproduction of a KMT2A proximity relation in the entire curated record**, and it is the strongest single piece of proximity evidence this beat found for any KMT2A-side claim. 14-3-3 binding is phospho-dependent, which makes it a testable and mechanistically specific proposal rather than a compartment coincidence.

Go deeper
Why this one is different from the rest of the beat

Everything else in this folder is a single observation from a single deposition. This is two: a dedicated 14-3-3 proximity study in one laboratory 1 and an RBP-focused BioID2 map in another 2, with different bait constructs, different experimental purposes and no shared authors, both placing KMT2A next to YWHAG.

Six of the seven human 14-3-3 isoforms recover KMT2A in the dedicated study 1. Family-wide recovery is what a genuine 14-3-3 client looks like, because the isoforms share a binding groove and recognise the same phosphopeptide motifs.

What it would mean if true, and how to break it

14-3-3 proteins bind phosphorylated motifs and act as conformational clamps and localisation switches. A 14-3-3 interaction with KMT2A would make KMT2A's behaviour phosphorylation-gated, which is a class of regulation this graph barely represents.

The question that matters here is whether KMT2A::AFF1 retains it. That depends entirely on where the motif is. If the 14-3-3 motif lies N-terminal to the breakpoint cluster region it is retained by every fusion; if it lies in the PHD-BRD-SET region it is discarded, like HCF-1, CyP33, CBP and MOF. This beat cannot tell, because proximity labelling has no positional resolution and no 14-3-3 motif has been mapped on KMT2A in any publication found here.

The experiment is small: a phosphosite-directed 14-3-3 pulldown from a t(4;11) line, or a scan of KMT2A for RSXpSXP-class motifs against published KMT2A phosphoproteomics, followed by mutation of the best candidate.

The deflating reading

14-3-3 proteins are among the most abundant proteins in a human cell and the 2023 study's own headline is that they are central and chaperone-like, i.e. near everything. A protein appearing in six 14-3-3 neighbourhoods may be reporting that fact rather than a specific client relationship 1. YWHAG's reproduction in the second study is real, but YWHAG is also a bait with 760 preys - the fourth largest in that dataset - so its list is broad by construction 2.

Facts and figures
scriptnone (database query)
statusrun
kmt2a roleprey in every row
result summarySix of seven 14-3-3 isoforms recover KMT2A in a dedicated 2023 proximity study, and YWHAG reproduces in an unrelated 2024 BioID2 map from another laboratory. It is the only cross-study reproduction of a KMT2A proximity relation in the curated record. No 14-3-3 motif has been mapped on KMT2A, so retention by the fusion is undetermined.
inputs
[
  "IntAct PSICQUIC identifier:Q03164 AND detmethod:\"MI:1314\"",
  "IntAct PSICQUIC pubid:39251607"
]
isoforms recovering kmt2a
[
  "YWHAB",
  "YWHAE",
  "YWHAG",
  "YWHAH",
  "YWHAQ",
  "SFN"
]
cross study reproduced
[
  "YWHAG"
]
publications
[
  "36931259",
  "39251607"
]
How this could bite us back
  • 14-3-3 proteins are near a large fraction of the proteome by the cited study's own account, so a client claim needs a mapped motif before it is a claim at all.
  • If KMT2A is phospho-gated by 14-3-3 and the fusion retains the motif, a kinase inhibitor already in the clinic could modulate the fusion non-specifically - which is a confounder for interpreting combination trials, not an opportunity.
Still open
  1. Where on KMT2A is the 14-3-3 motif, and is it N-terminal to the breakpoint cluster region - that is, retained by KMT2A::AFF1?
  2. Does a 14-3-3 pulldown from a t(4;11) line recover the fusion, wild-type KMT2A, both, or neither?
  3. Is the interaction phospho-dependent, and if so which kinase writes the site? That kinase becomes a candidate node this graph does not have.
  4. Do any of the six isoforms recover AFF1? In this beat's data they do not, which would place the contact on the KMT2A side of the fusion.
Every source behind this entry (3)
  1. E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In a proximity-labelling study assigning 14-3-3 proteins a central chaperone-like role in human cells, KMT2A is recovered as prey by six family members - YWHAB, YWHAE, YWHAG, YWHAH, YWHAQ and SFN (two rows) - with the YWHAG row carrying an IntAct MI-score of 0.41 against the 0.27 default for single-method proximity records.

    PMID 36931259DOI
    What this evidence cannot show

    Abstract read; full text not read. Counts taken from the IntAct deposition rather than from the paper's figures, so the study's own confidence thresholds are not reflected here. 14-3-3 proteins are extremely abundant and the study's own claim is that they are near a very large fraction of the proteome, which weakens any single client assignment. KMT2A is the prey in every row; no 14-3-3 motif on KMT2A is mapped.

    Evidence that argues against this (2)
    • In the largest human BioID survey (192 baits, HEK293) a prey's bait profile is dominated by subcellular compartment rather than binding, so an abundant chromatin protein recovered by many nuclear baits is the expected background of the method. Go CD et al. 2021
    • The founding BioID report describes the reagent as labelling proteins near the bait without binding it; promiscuous labelling of the vicinity is designed behaviour. Roux KJ et al. 2012
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    Independently, in the K562 50-RBP BioID2 map from a different laboratory, YWHAG is one of the 25 baits whose neighbourhood contains KMT2A; YWHAG's own prey list in that dataset numbers 760.

    What this evidence cannot show

    Retrieved from IntAct 2026-09-19. A 760-prey bait list is the fourth largest in the dataset, so YWHAG reaching KMT2A is not by itself discriminating; the value is in the agreement between two independent experiments, not in either alone.

    Evidence that argues against this (1)
    • Reproduction between two studies that both used biotin-ligase proximity labelling in human cell lines reproduces the method's biases as well as its signal. Two BioID experiments agreeing is weaker than a BioID result agreeing with an orthogonal assay, and no orthogonal assay of a KMT2A-14-3-3 interaction exists. Lambert JP et al. 2015
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Most famous KMT2A partners bind the portion of KMT2A discarded by the fusion - HCF-1/HCF-2 at aa 1799-1802, CyP33 at PHD3, CBP at the transactivation domain, MOF at the C-terminus - so whether the fusion retains any newly proposed KMT2A contact depends on where the motif lies relative to the breakpoint cluster region.

    PMID 15199122DOI
    What this evidence cannot show

    The cited paper establishes that oncogenic MLL forms retain menin but not other identified complex components; the residue-level partner map is this graph's own synthesis and is recorded in the KMT2A-side subtraction beat, not in this citation.

    Evidence that argues against this (1)
Connected to