Catalytic-dead UCH test: does KMT2A::USP2 need its deubiquitinase activity to transform?
also called USP2 C276A fusion test · UCH rescue falsification experiment
Build the leukaemia fusion protein twice, once normal and once with its enzyme switched off by a single amino-acid change, and see whether the broken version still turns blood progenitors into a cancer-like state. If it does, the leading explanation for this whole class of fusions is wrong.
The same thing, for a scientist
A three-arm murine progenitor colony-replating and half-life experiment that directly falsifies hyp-uch-domain-rescues-phd-mediated-mll-degradation. Arm 1: wild-type KMT2A(minor-BCR)::USP2. Arm 2: the identical construct carrying USP2 C276A, the catalytic nucleophile of the UCH domain [E1]. Arm 3: the identical KMT2A portion fused to an inert C-terminus of matched length. Readouts are (i) serial replating and Hoxa9/Meis1 expression, (ii) fusion-protein half-life by cycloheximide chase, (iii) K48-linked ubiquitin chains on immunoprecipitated fusion protein. Positive controls: MLL-AF9 (transforms, PHD-free) and MLL-AF9 with two PHD fingers added back (does not transform) [E2]. The design is deliberately cheap and conventional because the field currently has **no model of this fusion at all** - no cell line, no mouse, no xenograft - so the first useful thing to build is the simplest one.
Design2 deeper
Construct boundaries. The KMT2A portion runs from the start codon through the minor BCR, i.e. retaining AT-hooks, SNL/MBM, CXXC, PHD1-3, bromodomain and ePHD4, truncated at the exon 21-23 region where BCR2 breakpoints preferentially fall 3. The USP2 portion is the conserved UCH domain, residues 267-599 on UniProt O75604 numbering, the fragment the founding report states is retained 31.
Arms. (1) wild-type; (2) C276A catalytic-dead; (3) KMT2A portion + an inert, size-matched C-terminal tag (e.g. a catalytically inactive unrelated hydrolase fold) to control for mass and folding rather than a bare truncation.
Cells. Murine lineage-negative bone-marrow progenitors, retroviral transduction, serial methylcellulose replating - the same assay in which the PHD block was originally demonstrated, so the comparison is like-for-like 2.
Readouts.
- Colonies at rounds 1-4; Hoxa9 and Meis1 by qPCR.
- Fusion-protein half-life by cycloheximide chase at 0/2/4/8 h, with and without a proteasome inhibitor.
- K48-linked polyubiquitin on immunoprecipitated fusion protein.
- Ub-VS or Ub-AMC activity assay on immunoprecipitated fusion to confirm the wild-type arm actually has catalytic activity, which nobody has ever shown.
Controls. MLL-AF9 (positive, PHD-free), MLL-AF9+2PHD (negative) 2, empty vector. Every arm matched for transcript level by qPCR so that a protein-level difference is unambiguously post-transcriptional.
Why C276A and not a deletion
Deleting the UCH domain removes mass, folding and any protein-protein interaction surface at the same time as the catalysis, so a negative result would be uninterpretable. C276A removes only the nucleophile 1. If the domain contributes by oligomerisation or by binding something, C276A keeps that and the arm still transforms - which is itself the informative answer, and is why this design distinguishes the catalytic version of the hypothesis from the structural one.
What each outcome means
| result | reading |
|---|---|
| WT transforms, C276A does not, WT has longer half-life | hypothesis supported; USP2 catalysis is a drug target for this genotype |
| Both transform equally | catalysis is dispensable; the UCH domain acts structurally, or the PHD block is not degradation-based. Falsifies the hypothesis as written. |
| Neither transforms | the construct or the assay is wrong, or minor-BCR fusions need a co-operating lesion the assay does not supply. Re-derive before concluding anything |
| WT transforms and has no half-life advantage | transformation is real but the degradation explanation is not |
Honest limitations, stated before the experiment rather than after
- Murine progenitors are not the human infant B-cell progenitor in which this disease actually arises, and 16 of 17 KMT2A::USP2 patients had B-lineage or B/myeloid disease. A myeloid colony assay may be the wrong readout for a B-lineage fusion, and a negative result in arm 1 would be ambiguous for that reason alone. A parallel human CD34+ or iPSC-derived B-progenitor arm would be better and is roughly ten times the cost.
- The exact patient junction is unpublished, so the construct is a reconstruction from the stated interval and the stated UCH retention, not a patient's actual fusion. If the real junction includes more or less of USP2, the construct is wrong in a way this experiment cannot detect.
- Half-life in an overexpression system is not half-life at endogenous stoichiometry.
| design | Three-arm retroviral transduction of murine lineage-negative bone-marrow progenitors with (1) KMT2A(minor-BCR)::USP2-UCH wild-type, (2) the same carrying USP2 C276A, (3) the same KMT2A portion fused to an inert size-matched C-terminus. Controls: MLL-AF9, MLL-AF9+2PHD, empty vector. All arms transcript-matched by qPCR. |
|---|---|
| readout | Serial methylcellulose replating (rounds 1-4); Hoxa9 and Meis1 qPCR; fusion-protein half-life by cycloheximide chase +/- proteasome inhibitor; K48-linked polyubiquitin on immunoprecipitated fusion; Ub-AMC/Ub-VS activity assay on the immunoprecipitated wild-type fusion to demonstrate catalytic competence, which has never been shown. |
| controls | MLL-AF9 positive control and MLL-AF9-plus-two-PHD-fingers negative control, the exact pair in which the PHD block was established; empty vector; transcript-level matching across arms so any protein-level difference is post-transcriptional; a substrate-trap control to exclude C276A gaining a dominant trapping phenotype. |
| cost tier | low |
| timeline | Roughly 4-6 months: 6-8 weeks cloning and virus, 6 weeks replating, 4 weeks biochemistry, remainder for the parallel human CD34+ arm if funded. |
| decision rule | If arms 1 and 2 replate equivalently across four rounds with equivalent Hoxa9/Meis1 induction, hyp-uch-domain-rescues-phd-mediated-mll-degradation is falsified in its catalytic form and should be withdrawn to the structural form or abandoned. |
| what it does not test | Whether KMT2A::USP2 is menin-dependent; whether the fusion localises to chromatin; whether a B-lineage progenitor behaves like a myeloid one. Those need separate experiments and the menin one is the higher-value question. |
- A negative result in arm 1 - the wild-type fusion failing to transform - is the most likely single outcome and the least informative, because the assay was calibrated on a myeloid fusion and this is a B-lineage disease. Budget for the human arm from the start rather than treating it as a follow-up.
- If the experiment succeeds and catalysis is required, the immediate translational move is a USP2 inhibitor, and USP2's normal role in stabilising MDM2 means such an agent perturbs the p53 axis in every tissue. The therapeutic window would need establishing before this becomes a programme.
- The experiment produces a transforming KMT2A::USP2 construct if it works, which is the first model of this fusion. That is a genuine dual-use consideration for any laboratory generating it.
- Should the first arm instead be a human CD34+ or iPSC-derived pro-B system, given that 16 of 17 patients had B-lineage or B/myeloid disease and the murine myeloid assay may simply be the wrong tissue?
- Can the DCAL group supply the actual patient junction sequences? Every construct in this design is a reconstruction until they do.
- Should a KMT2A::AFDN minor-BCR construct be run as a fourth arm? AFDN supplies no UCH domain and is the second commonest BCR2 partner, so it is the natural internal control for whether the UCH story explains BCR2 or only USP2.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The catalytic nucleophile of the USP2 UCH domain is Cys276 and the proton acceptor is His557, within a USP domain spanning residues 267-599 of the 605-aa canonical protein, which makes C276A the minimal catalytic-dead mutation.
UniProtKB entry O75604 (UBP2_HUMAN), Ubiquitin carboxyl-terminal hydrolase 2, active sites and domain boundaries · UniProt Consortium; InterPro (EMBL-EBI) · UniProtKB / InterPro 2026What this evidence cannot show
Active-site positions are curated annotation for the canonical 605-aa isoform, largely assigned by family alignment rather than measured on human USP2 itself. If the fusion transcript corresponds to a shorter isoform with different numbering, the residue number shifts and the construct must be renumbered against the actual transcript. Accession resolved by reading back the primary gene name and length from the fetched record.
Evidence that argues against this (1)
- Cysteine-protease DUBs occasionally retain partial activity or acquire non-catalytic functions when the nucleophile is mutated, and a C-to-A mutant can act as a substrate trap that binds ubiquitinated proteins without releasing them. A trapping phenotype would look like a gain of function, not a loss, and must be excluded before a negative arm is read as loss of catalysis. Stevenson LF et al. 2007
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The murine lineage-negative progenitor retroviral serial-replating assay with MLL-AF9 and MLL-AF9-plus-PHD-fingers is an established system in which PHD retention abolishes immortalisation, and therefore supplies the positive and negative controls this design needs.
The PHD fingers of MLL block MLL fusion protein-mediated transformation. · Muntean AG, Giannola D, Udager AM, Hess JL · Blood 2008What this evidence cannot show
The assay measures immortalisation in methylcellulose, not leukaemia in an animal, and it was run with MLL-AF9 rather than with any minor-BCR fusion. Its calibration for a B-lineage-biased fusion is unknown. Abstract and central claim read in this session.
Evidence that argues against this (1)
- Thirty human KMT2A::USP2 patients establish that PHD retention does not prevent leukaemia in people, so a murine assay in which PHD retention abolishes immortalisation is measuring something the human disease evidently overcomes. If arm 1 fails to transform, the most likely explanation is assay-species mismatch rather than a true negative. Meyer C et al. 2023
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The construct boundaries are defined by the minor BCR (KMT2A intron 19/20 to exon 24, preference introns 21-23) and by the founding report's statement that the conserved UCH domain is what is fused to the extended 5'-MLL portion.
Human MLL/KMT2A gene exhibits a second breakpoint cluster region for recurrent MLL-USP2 fusions. · Meyer C, Lopes BA, Caye-Eude A, Cavé H, Arfeuille C, et al., Marschalek R · Leukemia 2019What this evidence cannot show
No per-patient exon-level junction is published for any KMT2A::USP2 case, so the construct is a reconstruction. The 5' boundary is also disputed between this paper (intron 19) and the 2023 recombinome (intron 20). Building two KMT2A truncations, one at each candidate boundary, costs almost nothing and removes the ambiguity. Full text read (PMC6756029).
Evidence that argues against this (1)
- The DCAL group holds per-patient breakpoint sequences that neither publication deposits, so a reconstruction from interval boundaries may differ from every real patient's fusion. Requesting the actual junctions before building anything would be cheaper than discovering the mismatch afterwards. Meyer C et al. 2023
- tests →The partner's deubiquitinase domain is what lets a PHD-retaining KMT2A fusion survive and transform
C276A catalytic-dead arm: equivalent transformation falsifies the catalytic form of the hypothesis
asks whether the fusion's UCH domain has catalytic activity at all, which has never been shown