Hypothesis: what the KMT2A half loses is control of its own abundance, not control of its activity
also called the fusion is what is left after the disposal instructions are cut off · degradation-determinant subtraction · stability-first reading of KMT2A-side loss
Everything the broken half of KMT2A throws away that anyone has actually tested turns out to be a piece that tells the cell when to destroy the protein. So the leukemic protein's main advantage may not be that it does something new - it may simply be that it never goes away.
The same thing, for a scientist
Three independent determinants of KMT2A turnover lie C-terminal to the breakpoint cluster region and are therefore absent from every der(11) KMT2A::partner fusion: the **PHD2 CDC34-dependent E3 ligase** (1479-1533), whose mutation stabilises MLL1 and raises its transactivation and recruitment [E1]; the **taspase1 cleavage sites** (2666/2667, 2718/2719), whose loss removes a cleavage event that primes MLL1 for proteasomal degradation and whose blockade raises MLL1 chromatin occupancy genome-wide [E2]; and the **FYRN domain** (2018-2074), the feature targeted by the mechanism that destroys a free, unpartnered MLL N-terminal fragment [E3]. Every one of these three perturbations, done in wild-type MLL1, makes the protein more abundant, longer-lived and more chromatin-bound. The hypothesis is that this is the KMT2A half's principal subtraction: **the fusion is a KMT2A N-terminus with its disposal instructions deleted**, and its defining property is persistence rather than novel activity. This reframes the beat, which has been organised around restraint-of-activity (the PHD3/CyP33 repressor switch), and it is falsifiable by a single measurement nobody has published.
Why this pattern was not visible before
Each of the three determinants was discovered by a different group asking a different question, none of them about subtraction, and each entered this graph as a separate entry:
- PHD2 entered as an E3 ligase - a curiosity of MLL1 biochemistry 1.
- The taspase1 sites entered as 'proteolytic maturation', framed by the 2003 model in which cleavage plus reassociation confers stability. The 2018 result that cleavage instead primes MLL1 for degradation 2 reverses that sign, and the reversal is what makes the three results line up.
- FYRN entered as half of a structural clamp. That it is also the degron on an unpartnered N-terminal fragment 3 is stated in a 2011 paper about wild-type MLL subunit turnover, in a sentence nobody in this field has had a reason to connect to fusion biology.
Read together they say one thing. And the fusion is, topologically, exactly the object that degron exists to dispose of: an MLL N-terminal fragment with no C-terminal partner - except that the breakpoint falls ~600 residues short of the degron.
What this predicts that the restraint-of-activity reading does not
- P1. In a KMT2A-rearranged cell, the fusion protein has a longer half-life than the wild-type KMT2A protein in the same cell. Everyone in this field asserts this; nobody has published the head-to-head chase. Under the restraint-of-activity reading there is no reason for it to be true.
- P2. Re-installing FYRN in cis into a fusion destabilises it, and any suppression of transformation tracks the lost protein rather than any regained regulation. The restraint reading predicts nothing from a FYRN insert, since FYRN binds no regulator.
- P3. Fusion abundance, not fusion activity, is the variable that most strongly predicts target-gene output - so a degron-driven titration of fusion protein should produce a steeper dose-response at HOXA9/MEIS1 than an equivalent titration of a partner-side activation function.
- P4. Interventions that raise wild-type KMT2A abundance are therapeutic by competition, without any agent binding the fusion. This one is already observed: CKII inhibition stabilises wild-type MLL1, displaces MLL-AFF1/AFF4/ELL2 from chromatin in SEM cells, and prolongs survival in MLL-AF9 mice 2.
- P5. Because the fusion has thrown away three disposal routes, a heterologous degrader should face less endogenous competition and produce a cleaner, more complete clearance of the fusion than of wild-type KMT2A - which is the opposite of the usual assumption that an oncoprotein is hard to degrade.
What this hypothesis does not claim
It does not claim the fusion's partner contributes nothing - the SEC/DOT1L machinery is the best-established thing in this disease and is not in question here. It does not claim the PHD3/CyP33 switch is unreal; that axis has a genuine in-cis add-back result behind it, unreplicated but real.
What it claims is narrower and about ranking: among the things the KMT2A half loses, the modules with measured consequences are degradation determinants, they all point the same way, and the beat's framing should follow the evidence rather than the 1994 activation/repression-domain geography.
It is also worth stating the weakest link plainly. All three determinants were measured on wild-type MLL1, not on a fusion. Three wild-type results pointing one way is a pattern that deserves a direct test, not a conclusion.
| statement | The modules the KMT2A half loses that have measured functional consequences are all determinants of KMT2A's own turnover - the PHD2 E3 ligase, the taspase1 cleavage sites and the FYRN degron - and all point the same way. The fusion's principal KMT2A-side subtraction is therefore loss of abundance control, not loss of activity control, and its defining property is persistence. |
|---|---|
| origin | data-pattern |
| prior probability | 0.55 |
predictions
[ "The fusion protein has a longer half-life than wild-type KMT2A protein measured in the same KMT2A-rearranged cell", "Re-installing FYRN (aa 2018-2074) in cis into KMT2A::AFF1 lowers fusion protein abundance, and any suppression of serial replating scales with the protein lost rather than with any regained regulatory contact", "Degron-titrated fusion abundance produces a steeper HOXA9/MEIS1 dose-response than an equivalent titration of a partner-side activation function", "Raising wild-type KMT2A abundance displaces the fusion from chromatin and is therapeutic without any agent binding the fusion - already observed for CX-4945 in SEM cells and MLL-AF9 mice", "A heterologous degrader clears the fusion more completely than wild-type KMT2A, because the fusion has discarded three endogenous disposal routes and therefore competes less for the same machinery" ]
experiments
[ "exp-hbm-reinsertion-kmt2a-aff1", "exp-wildtype-allele-dosage-vs-fusion-occupancy" ]
analyses
[ "ana-kmt2a-lost-module-ledger", "ana-fusion-halflife-degron-vs-menin" ]
falsifiers
[ "A head-to-head cycloheximide chase or degron washout in one KMT2A-rearranged line shows the fusion is NOT longer-lived than wild-type KMT2A. The hypothesis's central premise is then false and the three wild-type results do not transfer.", "A FYRN insert in cis does not destabilise the fusion. FYRN would then be a clamp component whose degron behaviour is specific to the cleaved wild-type fragment, and the 'the fusion is a free MLL(N) that escaped its degron' framing collapses.", "Takeda et al. turn out to be right that taspase1 cleavage activates rather than destabilises MLL1. One of the three pillars inverts, and with only two determinants left pointing one way the pattern is no longer distinguishable from coincidence.", "Fusion abundance is shown to be buffered - e.g. transcription or translation of the fusion allele compensates - so that removing degradation routes does not raise steady-state fusion protein. The mechanism would then be real but inconsequential.", "An in-cis add-back panel shows that suppression of transformation by PHD3 does NOT track fusion protein abundance, while an abundance-matched control arm has no effect. Regulation would then be doing the work after all, and the ranking this hypothesis proposes would be wrong." ]
degradation determinants lost — the numbers behind the chart
[
{
"name": "PHD2 (CDC34-dependent E3 ligase)",
"start": 1479,
"end": 1533,
"effect_of_loss": "stabilisation, increased transactivation, increased recruitment",
"source": "PMID 23129768"
},
{
"name": "FYRN (degron on a free MLL N-terminal fragment)",
"start": 2018,
"end": 2074,
"effect_of_loss": "escape from the mechanism that destroys unpartnered MLL(N)",
"source": "PMID 21670200"
},
{
"name": "Taspase1 cleavage site 1",
"start": 2666,
"end": 2667,
"effect_of_loss": "escape from cleavage-primed proteasomal degradation",
"source": "PMID 30573454"
},
{
"name": "Taspase1 cleavage site 2",
"start": 2718,
"end": 2719,
"effect_of_loss": "escape from cleavage-primed proteasomal degradation",
"source": "PMID 30573454"
}
]prior revision 2026 09 19
{
"from": 0.4,
"to": 0.55,
"reason": "Bursen 2004 (PMID 15221006) demonstrates degradation escape as the oncogenic mechanism for AF4-MLL in this exact translocation, via SIAH1/SIAH2. The mechanism class is established in t(4;11) rather than hypothetical, which raises the prior that the same logic applies to the der(11) product. Not raised further because the paper is about the reciprocal fusion, not the one this hypothesis concerns.",
"found_by": "scripts/check-absence-vs-library.py - the paper was already cited by 4 other entries"
}- If persistence is the fusion's defining property, then any therapy that lowers fusion abundance transiently will be defeated by re-accumulation, and dosing schedules matter more than potency - which is a different development problem from the one the menin-inhibitor programme has been solving.
- A degrader strategy that exploits the fusion's missing disposal routes is attractive, but the same routes are missing from the reciprocal product and from nothing else, so selectivity against wild-type KMT2A would have to come entirely from the induced-proximity ligand, not from the biology.
- Raising wild-type KMT2A to compete the fusion off chromatin has a hard ceiling and an obvious escape: silence or lose the wild-type allele. This graph already records wild-type KMT2A silencing in a related context.
- Is the fusion actually the longer-lived species? The entire hypothesis rests on a measurement that takes a week and has never been published head-to-head.
- Which ubiquitin ligase acts on FYRN? Yokoyama et al. show the domain is targeted but do not name the machinery, and naming it would turn a descriptive degron into a druggable node - and would say immediately whether the fusion could be made a substrate again.
- If the fusion has discarded three disposal routes, why does menin-inhibitor-driven eviction from chromatin lead to fusion turnover at all? Either a fourth route exists on the retained N-terminus, or eviction works by something other than degradation, and the graph's degron/half-life analysis should distinguish them.
- Does partner identity set fusion half-life, as Hsieh predicted in 2003? With three KMT2A-side determinants gone, the partner is the only remaining source of stability determinants, so the prediction is sharper now than when it was made.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The second PHD finger (PHD2) of MLL1 is an E3 ubiquitin ligase in the presence of the E2-conjugating enzyme CDC34; this activity is conserved in MLL4's second PHD finger but not in MLL2 or MLL3; and mutation of PHD2 leads to MLL1 stabilisation, increased transactivation ability and increased MLL1 recruitment to target gene loci, indicating that PHD2 negatively regulates MLL1.
A subset of mixed lineage leukemia proteins has plant homeodomain (PHD)-mediated E3 ligase activity · Wang J, Muntean AG, Wu L, Hess JL · Journal of Biological Chemistry 287(52):43410-43416 2012What this evidence cannot show
ABSTRACT ONLY. HEK293 biochemistry with mutant constructs; no haematopoietic model and no fusion protein.
Evidence that argues against this (1)
- The hypothesis needs PHD2 to be an abundance controller specifically, and the cited result does not isolate that: PHD2 mutation raises MLL1 levels AND its transactivation AND its recruitment together. Conservation in MLL4's second PHD finger but not others is interesting and does not separate the two control axes either. Journal of Biological Chemistry 2012
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Taspase1-dependent cleavage of MLL1 primes it for proteasomal degradation; taspase1 loss leaves MLL1 uncleaved, longer-lived under cycloheximide chase and markedly increased in genome-wide chromatin occupancy, with localisation, bulk H3K4 methylation and MLL/COMPASS assembly unchanged; KMT2A breakpoints lie N-terminal to the cleavage sites so the chimeras are resistant to this route; and stabilising wild-type MLL1 with the CKII inhibitor CX-4945 displaces MLL-AFF1, AFF4 and ELL2 from chromatin in SEM cells and prolongs survival in an MLL-AF9 mouse model.
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 2019What this evidence cannot show
FULL TEXT read via PMC6317322. Mechanism in 293T/MCF7/HCT116; leukemia arms in SEM, MV4-11, RS4;11 and an MLL-AF9 transplant. The claim that chimeras are more stable than wild-type MLL1 is asserted in discussion, not measured. Takeda et al. 2006 (PMID 16951254) read the same perturbation in the opposite direction - as loss of activation - and that conflict is unresolved; if Takeda is right, this hypothesis loses its central pillar.
Evidence that argues against this (1)
- Takeda et al. 2006 concluded that uncleaved precursor MLL has reduced H3K4 methyltransferase activity with decreased H3K4me3 at Cyclin E1/E2 and that non-cleavable MEFs proliferate poorly - i.e. cleavage activates rather than destabilises. A third study cited by Zhao et al. found non-cleavable knock-in mice with no apparent defect. Takeda S et al. 2006
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The stability of MLL(N) depends on intramolecular interaction with MLL(C) through PHD1, PHD4 and the FYRN domain, and free MLL(N) is destroyed by a mechanism that targets the FYRN domain, while free MLL(C) is exported to the cytoplasm and degraded by the proteasome.
Proteolytically cleaved MLL subunits are susceptible to distinct degradation pathways · Yokoyama A, Ficara F, Murphy MJ, Meisel C, Naresh A, Kitabayashi I, Cleary ML · Journal of Cell Science 124(Pt 13):2208-2219 2011What this evidence cannot show
ABSTRACT ONLY. An MLL(C)-null mouse plus construct biochemistry; the ubiquitin machinery acting on FYRN is not identified, so 'degron' is a functional description rather than a mapped degradation signal.
Evidence that argues against this (1)
- Intramolecular MLL(N)-MLL(C) stabilisation cannot operate in a fusion, which has no MLL(C). So this supports the claim that the fusion escapes an abundance control only by assuming the fusion's N-terminal portion behaves like free MLL(N) - and fused to a large partner it may not. Journal of Cell Science 2011
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
FYRN and FYRC are two halves of a single alpha+beta fold rather than two independent domains, with the MLL arrangement carrying a large insertion between two strands of one beta-sheet.
The structure of the FYR domain of transforming growth factor beta regulator 1 · Garcia-Alai MM, Allen MD, Joerger AC, Bycroft M · Protein Science 19(7):1432-1438 2010What this evidence cannot show
ABSTRACT ONLY, solved on TBRG1 with the MLL statement an explicit homology extrapolation. Cited for why an unpartnered FYRN would be recognised as defective: it is half a beta-sheet.
Evidence that argues against this (1)
- A TBRG1 structure again, and the hypothesis leans on FYRN/FYRC being one module to argue that splitting them removes a control. That architecture is inferred across proteins, and the MLL arrangement's large insertion is a known difference. Protein Science 2010
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The wild-type Mll allele is required for MLL-AF9-induced leukemogenesis and for maintenance of MLL-AF9-transformed cells.
MLL-AF9-induced leukemogenesis requires coexpression of the wild-type Mll allele · Thiel AT, Blessington P, Zou T, Feather D, Wu X, Yan J, Zhang H, Liu Z, Ernst P, Koretzky GA, Hua X · Cancer Cell 17(2):148-159 2010What this evidence cannot show
ABSTRACT ONLY; mouse genetics in one fusion. Cited as the tension this hypothesis must live with: if wild-type MLL1 is required, then P4's competition mechanism has a floor below which removing wild-type protein helps the leukemia rather than hurting it, and the dose-response is non-monotonic.
Evidence that argues against this (1)
- This one-line statement of the wild-type requirement is the fifth appearance of the same mouse result in this beat. Used here it also runs against the hypothesis rather than for it: if the leukaemia needs the wild-type allele, the cell retains whatever control that allele exerts, and the fusion's escape is partial by construction. Cancer Cell 2010
- E_bursen_siah_precedentstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The mechanism this hypothesis proposes is already demonstrated in t(4;11) - on the other derivative. Bursen 2004 shows growth-transforming potential of AF4 wild-type and the AF4.MLL fusion, and states the underlying oncogenic mechanism involves the E3 ubiquitin ligases SIAH1 and SIAH2, the N-terminal portion of AF4, and PROTECTION OF THE AF4.MLL FUSION PROTEIN AGAINST PROTEASOMAL DEGRADATION. Escape from degradation control is therefore an established oncogenic mechanism in this exact translocation, not a novel proposal - which raises the prior on the same mechanism operating on the der(11) side.
Interaction of AF4 wild-type and AF4.MLL fusion protein with SIAH proteins: indication for t(4;11) pathobiology? · Bursen A, Moritz S, Gaussmann A, Dingermann T, Marschalek R · Oncogene 23(37):6237-49 2004What this evidence cannot show
Read as abstract only, re-fetched 2026-09-19. This is the RECIPROCAL fusion, AF4-MLL on der(4), not the KMT2A::AFF1 der(11) product this hypothesis is about - a different protein with a different N-terminus and different ligases. It establishes the mechanism CLASS in t(4;11) and says nothing directly about KMT2A half-life. A 2004 paper using overexpression and transformation assays, from the same laboratory this project has flagged as the single source behind much of its descriptive layer.
Evidence that argues against this (1)
- The precedent cuts against the hypothesis as well as for it. If degradation escape is already the documented oncogenic mechanism for AF4-MLL, then the interesting question is not whether abundance control matters in t(4;11) - that is answered - but whether the der(11) product needs it at all, given that the reciprocal may already supply it. This project also records that AF4-MLL alone induced pro-B ALL in mice while MLL-AF4 alone induced none over 13 months (PMID 20194896), which would make the reciprocal the relevant molecule and this hypothesis aimed at the wrong half. Bursen A et al. 2004
the corrected rows all point one way: the losses with measured consequences are degradation determinants
- related to →KMT2A / MLL1 — the gene that breaks
[describes] three degradation determinants C-terminal to the breakpoint are absent from every der(11) fusion
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
[describes] predicts the fusion is the longer-lived species in its own cell
[supported_by] FYRN is the degron on a free MLL N-terminal fragment
[supported_by] cleavage primes MLL1 for proteasomal degradation; the fusion has neither site
[supported_by] PHD2 mutation stabilises MLL1 and raises transactivation and recruitment
[tested_by] the head-to-head half-life measurement that would falsify or confirm it
- tests →Experiment: put four residues back - the minimal add-back control the PHD3 result has never had
[tested_by] the add-back panel, with abundance promoted to primary readout and a FYRN arm added
[tested_by] maps the competition arm of the hypothesis
[contradicts] reframes the lost set from restraint-of-activity to restraint-of-abundance