Oncofusion

start / graph / subtraction-kmt2a-tad-cbp-contact

Mechanismstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q-subtraction-kmt2a

Lost from the KMT2A half: the transactivation domain (aa 2829-2883) and its CBP contact

also called MLL TAD · MLL activation domain · MLL-CBP KIX · 9aaTAD 2847-2855

KMT2A has its own 'switch genes on' patch that works by grabbing a shared coactivator called CBP. The fusion loses that patch and borrows an entirely different switching mechanism from its partner instead.

The same thing, for a scientist

MLL amino acids 2829-2883 form a potent transcriptional activation domain that binds the KIX (CREB-binding) domain of CBP at CBP residues 581-687; five hydrophobic residues, plausibly forming the face of an amphipathic helix, are critical, and the MLL activation domain facilitates CBP binding to phosphorylated CREB rather than competing with it [E1]. UniProt annotates a 9aaTAD motif at 2847-2855 inside that window [E2]. The domain is ~1,400 residues C-terminal to the breakpoint cluster region and is lost from every **der(11)** KMT2A::partner fusion - though the reciprocal der(4) AFF1::KMT2A product retains it [E5]. This was in fact the founding observation of the whole KMT2A-side subtraction idea: Zeleznik-Le et al. showed in 1994, before any partner was mechanistically understood, that 11q23 translocations separate an activation domain 3' of the breakpoint from a repression domain and a DNA-binding region 5' of it [E3]. The modern reading is that the fusion does not lose transcriptional activation - it **swaps activation mechanisms**, from a CBP-recruiting acidic TAD to a partner-supplied elongation module. **Added cycle 31:** two searches for a functional test of TAD loss - an add-back of the MLL TAD into a fusion, or a separation-of-function comparison against the partner's elongation module - returned nothing, so the 'swap, not a deficit' classification remains an inference rather than a measurement, and the entry says so.

Go deeper
Coordinates, confirmed from the primary source

The brief listed 'CBP at the TAD ~2829-2883' as an unverified lead. It is verified. Ernst et al. state it in the abstract of the primary paper: 'In vitro binding experiments refined the minimal region of interaction to amino acids 2829 to 2883 of MLL, a potent transcriptional activation domain, and amino acids 581 to 687 of CBP (the CREB-binding or KIX domain)' 1.

UniProt does not annotate a domain across that whole span, but does annotate a 9aaTAD motif at 2847-2855 2 - a nine-residue transactivation motif sitting inside the Ernst window. The two are consistent: 2829-2883 is the minimal binding region, 2847-2855 the curated core motif.

What the 1994 paper actually established, and why it is still the frame1 deeper

Zeleznik-Le, Harden and Rowley mapped MLL's regulatory geography with GAL4 fusion reporter assays and found 3:

  • a strong activation domain telomeric (3') to the breakpoint region, activating 18-fold to >200-fold depending on promoter and cell line;
  • a repression domain centromeric (5') to the breakpoint region, repressing ~4-fold;
  • and, separately, that the bacterially expressed MLL AT-hook region binds cruciform DNA, recognising structure rather than sequence.

Their conclusion - 'loss of an activation domain with retention of a repression domain and a DNA-binding domain on the der(11) chromosome could alter the expression of downstream target genes' - is the original statement of this beat's thesis, thirty-two years old. The graph should credit it as such rather than presenting KMT2A-side subtraction as a new idea.

What has changed since is the answer's direction. In 1994 the natural reading was that the fusion would be a repressor, having kept the repression domain and lost the activator. What actually happens is that the partner supplies a different activating machine (elongation, via SEC/DOT1L), so the fusion activates its targets by a mechanism KMT2A never used.

A swap, not a subtraction

This is the entry in the beat where 'lost' is least informative on its own. Losing the CBP-binding TAD would matter enormously if the fusion still needed to activate transcription the way KMT2A does. It does not. The replacement is not equivalent - a CBP-recruiting acidic TAD acetylates local chromatin, whereas a SEC/DOT1L module releases paused polymerase - but both end in activation of the same target loci.

The practical consequence: you cannot infer function loss from domain loss when a functional replacement arrives in the same molecule. The TAD is lost and its job is done by something else. Contrast PHD3, where the lost module's job (restraint) has no replacement anywhere in the fusion.

Cycle 31 counter-evidence pass: what was looked for and not found, plus one scope fix

Searched, not found. The swapped classification says the fusion replaces a CBP-recruiting acidic TAD with a partner-supplied elongation module, and therefore loses activation capability but not activation. Nobody has tested that. There is no published add-back of KMT2A residues 2829-2883 into a fusion, and no comparison of a KMT2A-N::MLL-TAD chimera against a real fusion at shared target genes. The classification is an inference from two separately established facts, and it should be read as one. This is a confirmed absence, and the searches are recorded so that a future reader can distinguish it from an unasked question.

One scope fix. Every 'lost' statement in this entry is about der(11). The reciprocal der(4) AFF1::KMT2A product retains the KMT2A C-terminus, including this TAD and the SET domain, and is itself taspase1-cleaved and FYRN/FYRC-stabilised 5. In a t(4;11) patient expressing both derivatives, the TAD is present in the cell and present in a fusion protein - just not in the one this beat is usually about.

One reframing that costs nothing and may matter. In wild-type MLL the TAD does not sit on the same polypeptide as the AT-hooks and CXXC after processing: taspase1 cleavage puts the DNA-binding N-terminus on N320 and the TAD and SET on C180, and the two are held together only by the FYRN/FYRC clamp 4. So 'wild-type KMT2A activates its targets through a CBP-recruiting TAD' is already a claim about a two-fragment complex whose assembly is regulated, and the fusion's departure from it is not only the absence of the TAD but the absence of the assembly step. Whether that distinction has any consequence is exactly what the untried add-back would test - and it predicts that inserting the TAD in cis into a fusion would not reconstitute wild-type-like activation, because the regulated assembly is what is missing.

Facts and figures
moduleTransactivation domain (CBP KIX-binding)
uniprotQ03164
start2829
end2883
fate in fusionlost
partner lostCREBBP (CBP), KIX domain aa 581-687
demonstrated functional consequence of lossnone directly measured, and specifically searched for in cycle 31 with nothing found. The 'swap' reading is an inference from two separately established facts, not a measurement.
loss classlost from der(11); inferred to be functionally replaced in cis by a partner-supplied elongation module - a swap on inference, never tested
scopelost from der(11) KMT2A::partner; retained by the reciprocal der(4) AFF1::KMT2A product (PMID 21233834)
core motif
{
  "name": "9aaTAD",
  "start": 2847,
  "end": 2855,
  "source": "UniProt Q03164"
}
position confirmed by
[
  "primary source abstract, residue range stated verbatim (PMID 11259575)",
  "UniProt 9aaTAD motif inside the same window"
]
How this could bite us back
  • Any strategy aimed at the KMT2A TAD-CBP interface would act only on wild-type KMT2A, since the fusion does not contain the TAD; CBP/KIX ligands are also broadly active across many transcription factors and would not be selective for this disease.
  • If the partner's elongation module is a functional replacement for the lost TAD, then blocking elongation (CDK9, DOT1L, BET) removes the fusion's only activation mechanism - which is the rationale for those programmes, but also means resistance need only restore activation by any route, including re-acquiring a conventional TAD-like contact.
Still open
  1. Is the CBP-recruiting TAD functionally equivalent to the partner's elongation module at shared target genes, or does the swap change WHICH targets are activated? A direct comparison - fusion versus an artificial KMT2A-N fused to the MLL TAD itself - has never been reported and would separate 'activation' from 'this particular activation'.
  2. CREBBP is itself mutated in relapsed paediatric ALL. Does CREBBP status modify KMT2A-rearranged disease, given that the fusion no longer binds CBP through this route while the wild-type allele still does?
  3. The 1994 prediction was that the der(11) product would be a repressor, because it keeps the repression domain and loses the activator. The graph records that the retained repression domain is required for transformation. Is the fusion's targeted repression (e.g. of BIM via PRC1) the surviving half of that 1994 prediction?
  4. Would inserting KMT2A 2829-2883 in cis into KMT2A::AFF1 reconstitute anything? In wild-type MLL the TAD and the DNA-binding N-terminus are on opposite sides of a taspase1 cleavage and are joined only by the FYRN/FYRC clamp, so a TAD placed directly next to the AT-hooks is not the wild-type arrangement and a null result would not mean the TAD is dispensable. That confound should be designed for, not discovered.
Every source behind this entry (5)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In vitro binding experiments refine the minimal MLL-CBP interaction region to amino acids 2829-2883 of MLL, a potent transcriptional activation domain, and amino acids 581-687 of CBP (the KIX domain); five hydrophobic MLL residues potentially forming the face of an amphipathic helix are critical; MLL transactivation depends on CBP; and the MLL activation domain facilitates binding of CBP to phosphorylated CREB rather than competing for it.

    MLL and CREB bind cooperatively to the nuclear coactivator CREB-binding protein · Ernst P, Wang J, Huang M, Goodman RH, Korsmeyer SJ · Molecular and Cellular Biology 21(7):2249-2258 2001
    PMID 11259575DOI
    What this evidence cannot show

    ABSTRACT ONLY - the residue range 2829-2883 is stated verbatim in the abstract, so the coordinate is confirmed, but the underlying binding data were not inspected. Yeast two-hybrid discovery plus in-vitro binding and reporter transactivation with isolated fragments; no chromatin, no haematopoietic model, and no fusion protein is studied. E1A inhibition of CBP is an indirect argument for CBP dependence.

    Evidence that argues against this (1)
    • An in vitro minimal-region mapping sets an upper bound on what is needed to bind, not what operates in a nucleus. CBP's KIX domain is engaged by many activators including CREB - the same paper's other subject - so occupancy in cells is a competition problem this measurement does not address. Molecular and Cellular Biology 2001
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt Q03164 annotates a 9aaTAD motif at residues 2847-2855, within the Ernst et al. minimal CBP-binding region, and annotates taspase1 cleavage sites at 2666/2667 and 2718/2719 N-terminal to it.

    What this evidence cannot show

    Accession verified by reading back the primary gene name (KMT2A) and length (3,969 aa). UniProt does not annotate a 'transactivation domain' feature spanning 2829-2883; only the 9aaTAD motif inside it. The 9aaTAD is a predicted sequence motif class, so it corroborates rather than independently establishes the activation function.

    Evidence that argues against this (1)
    • A UniProt feature table is a curated model rather than a measurement. Profile-based domain calls for KMT2A differ between UniProt, Pfam, SMART, InterPro and CATH, and the sibling atlas beat was corrected last cycle for quoting one source's range as though it were observed. Quoting coordinates without the spread states a choice as a fact. Molecular and Cellular Biology 2001
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    GAL4-fusion reporter mapping places a strong transcriptional activation domain telomeric (3') to the MLL breakpoint region, activating 18-fold to more than 200-fold depending on promoter and cell line, and a repression domain centromeric (5') to it repressing about 4-fold; in 11q23 translocations the activation domain is lost while the repression domain and the AT-hook DNA-binding region are retained on the der(11) chromosome.

    PMID 7938000DOI
    What this evidence cannot show

    ABSTRACT ONLY. 1994 transient-transfection GAL4 reporter assays in HeLa and other lines; a heterologous-DNA-binding-domain reporter measures the intrinsic activity of a fragment, not what that fragment does inside full-length MLL on native chromatin. The activation fold-range (18 to >200) is itself evidence of strong context dependence. No fusion protein was tested.

    Evidence that argues against this (1)
    • Zeleznik-Le et al.'s geography is the frame this entry inherits - activation domain 3' of the break, repression domain 5' of it - and it is the frame the cycle-31 correction pass found to be aimed at the wrong faculty. The KMT2A-side losses with measured functional consequences are not activation or repression modules at all; they are three degradation determinants (PHD2 E3 ligase, the taspase1 sites, the FYRN degron), each of which makes MLL1 more abundant and more chromatin-bound when disabled. A 1994 GAL4-reporter map of activation and repression could not have seen that, and this entry should not be read as establishing that activation/repression is the axis on which the KMT2A half is subtracted. Zhao Z et al. 2019
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Taspase1 cleaves MLL1 between the breakpoint cluster region and the transactivation domain, and cleavage primes MLL1 for proteasomal degradation; the transactivation domain and the SET domain both lie on the C180 fragment, so in wild-type MLL the TAD is carried on the shorter-lived, separately regulated half of the protein.

    Regulation of MLL/COMPASS stability through its proteolytic cleavage by taspase1 as a possible approach for clinical therapy of leukemia. · Zhao Z, Wang L, Volk AG, Birch NW, Stoltz KL, Bartom ET, Marshall SA, Rendleman EJ, Nestler CM, Shilati J, Schiltz GE, Crispino JD, Shilatifard A · Genes & Development 33(1-2):61-74 2019
    PMID 30573454DOI
    What this evidence cannot show

    FULL TEXT read via PMC6317322. The fragment topology is stated in that paper's introduction and is not itself a new result; the degradation direction is its finding. Cited here to make a point this entry did not make: the TAD is not simply a module the fusion lacks, it is a module that in the wild-type protein is physically segregated from the DNA-binding N-terminus by a regulated cleavage event, so 'KMT2A activates through a CBP-recruiting TAD' is already a statement about a two-fragment complex, not about one polypeptide.

    Evidence that argues against this (1)
    • Cleavage priming MLL1 for degradation is established with taspase1 knockout and CKII inhibition, neither of which is specific: taspase1 has substrates besides MLL and CX-4945 reaches a very large CKII substrate set. The link from cleavage position to the TAD's fate therefore rests on two pleiotropic perturbations. Genes & Development 2019
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    The reciprocal AFF1::KMT2A product of t(4;11) retains the KMT2A C-terminus including the transactivation domain and SET, is activated by taspase1 cleavage into AF4-MLL-N and MLL-C, and is stabilised against SIAH1/2-mediated polyubiquitinylation by FYRN/FYRC heterodimerisation.

    The heterodimerization domains of MLL-FYRN and FYRC--are potential target structures in t(4;11) leukemia · Pless B, Oehm C, Knauer S, Stauber RH, Dingermann T, Marschalek R · Leukemia 25(4):663-670 2011
    PMID 21233834DOI
    What this evidence cannot show

    ABSTRACT ONLY. Interaction mapping in bacteria and a HeLa biosensor assay, not a leukemia model. Cited to correct a scope ambiguity: the TAD is lost from der(11) KMT2A::AFF1, not from t(4;11) as a whole.

    Evidence that argues against this (1)
    • The reciprocal retaining the TAD and SET is true where a reciprocal exists and is expressed, which is roughly half of t(4;11) patients with an intact der(4). A 'restoration' argument built on it applies to a subset that this entry does not quantify. Leukemia 2011
Connected to