Oncofusion

start / graph / concept-epitope-tagged-knock-in-at-the-endogenous-fusion-allele

Explainermoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q27b-fusion-attribution

The experiment nobody has done: an epitope tag knocked into the endogenous KMT2A fusion allele in a patient-derived line

also called endogenous fusion tagging · HA knock-in KMT2A::AFF1 · HaloTag fusion allele · tag knock-in attribution

The cleanest way to study a fusion protein is to attach a small unique handle to it in the patient's own leukaemia cells, so that any measurement using that handle can only be seeing the fusion. Despite being routine technology, nobody has published this for any KMT2A fusion.

The same thing, for a scientist

Every attribution failure in this beat has the same shape - a reagent that binds sequence the fusion shares with a wild-type parent - and one construct removes all of them at once. A short epitope (HA, FLAG, V5), a HaloTag or a degron cassette knocked into the **endogenous** rearranged allele of a patient-derived line creates a species that is unique in the cell by definition. Anti-tag ChIP or CUT&RUN, anti-tag AP-MS, anti-tag immunofluorescence and anti-tag live imaging then all become fusion-specific without a single new antibody against the fusion itself. **It has not been done.** Two PubMed queries run 2026-09-20 bound the absence [E3]. Searching KMT2A::AFF1 under four spellings against knock-in, knockin, degron, dTAG and 'endogenously tagged' returns **one** record, and that record is the 2006 **murine Mll-AF4 fusion-gene knock-in** [E2] - a knock-in of the fusion, not of a tag onto it. Searching KMT2A or MLL against 'endogenously tagged', 'endogenous tagging' or dTAG returns **zero**. The strongest single piece of corroboration is not a search at all: the authors of the only MLL-fusion degron paper write in their own Limitations section that **'future studies should include endogenously degron-tagged MLL::AF9'** [E1] - i.e. the group best placed to have done it states in 2022 that it has not been done. **What it would settle, in one line each.** Composition: an anti-tag AP-MS gives the first fusion-specific interactome at endogenous stoichiometry. Localisation: live imaging gives the first attributable picture of where the fusion is. Occupancy: anti-tag CUT&RUN validates or breaks every MLL-N-based map in the literature. Turnover: a tag plus a degron gives both half-life and washout kinetics at the endogenous locus. PTMs: anti-tag IP concentrates the fusion for junction-peptide PRM. **Five claim classes, one cell line.**

Go deeper
Design constraints that are already known
  • Tag position must not be N-terminal. Olsen et al. found that an N-terminally HA-FKBP12-tagged MLL::AF9 did not reproduce the leukaemogenesis kinetics of the untagged construct, which they attribute to interference with chromatin localisation because the DNA-binding motifs and the menin-binding domain are N-terminal; only the C-terminal tag behaved like the parent 1. For KMT2A::AFF1 the C-terminus is AFF1-derived, which is where the tag should go - and that then has to be validated in its own right, because the AFF1 C-terminal region carries the AFF4, ENL and P-TEFb contacts.
  • The line must be patient-derived and the allele must be the rearranged one. SEM and RS4;11 are the obvious hosts. Targeting must be verified to be on the der(11) allele and not on the wild-type KMT2A allele, which is trivially checkable by PCR across the junction.
  • Heterozygosity is a feature, not a bug. The untagged wild-type KMT2A and AFF1 alleles remain, so the tagged line still contains the confounding species - which is exactly what makes it the right validation set for the antibody-based methods.
  • Chemical synchronisation does not work in these lines. Mimosine, double thymidine and nocodazole all failed in SEM, RS4;11, THP-1, NOMO-1 and MOLM-13; centrifugal elutriation is the published workaround 4. Any cell-cycle-resolved experiment on the tagged line has to plan for that.
Why the absence is surprising

CRISPR knock-in tagging of endogenous loci is a standard method and has been applied across cancer biology. The specific reason it may not have happened here is that the KMT2A fusion literature is unusually well served by the MLL-N minus MLL-C subtraction (ana-attribution-audit-chromatin-occupancy-kmt2a-aff1), which gives a workable qualitative answer for the readout the field cares about most. The subtraction is good enough for occupancy and useless for everything else, so the field has a partial solution to one of five problems and has not felt the pressure to solve the other four.

A second reason may be that the cell lines are hard: t(4;11) lines are slow-growing, poorly transfectable and, as above, cannot be chemically synchronised.

Facts and figures
reagent classepitope or HaloTag knocked into the endogenous rearranged allele
published instances in kmt2a leukaemia0
bounded absence queries
[
  {
    "query": "(\"MLL-AF4\"[tiab] OR \"MLL::AF4\"[tiab] OR \"KMT2A::AFF1\"[tiab] OR \"KMT2A-AFF1\"[tiab]) AND (knock-in[tiab] OR knockin[tiab] OR degron[tiab] OR dTAG[tiab] OR \"endogenously tagged\"[tiab])",
    "count": 1,
    "hit": "PMID 16551973, murine Mll-AF4 fusion-gene knock-in",
    "date": "2026-09-20",
    "parser_note": "both double-colon spellings were collapsed to the hyphenated form by the query parser"
  },
  {
    "query": "(KMT2A[tiab] OR MLL[tiab]) AND (\"endogenously tagged\"[tiab] OR \"endogenous tagging\"[tiab] OR dTAG[tiab])",
    "count": 0,
    "date": "2026-09-20"
  }
]
claim classes it would settle
[
  "complex composition",
  "subnuclear localisation",
  "chromatin occupancy",
  "protein half-life and turnover",
  "post-translational modifications"
]
design constraints
[
  "tag C-terminal or internal, never N-terminal",
  "target the der(11) allele, verified by PCR across the junction",
  "validate transformation and growth against the untagged parent",
  "plan cell-cycle work around elutriation, not chemical synchronisation"
]
How this could bite us back
  • A tagged line becomes the reference model for the whole field the moment it exists, so a tag that subtly attenuates the fusion would bias every downstream drug sensitivity measured in it towards apparent efficacy.
  • If the tag sits on the AFF1 C-terminus and perturbs P-TEFb recruitment, the line will under-report the elongation arm of the mechanism - which is precisely the arm that distinguishes KMT2A::AFF1 from a menin-interface-only model.
  • Knock-in clones are single-cell derived, so any phenotype measured in one clone confounds the tag with clonal drift; at least three independent clones and an untagged sibling clone are needed before a tagged-line result can be attributed to the tag rather than to the cloning.
Still open
  1. Does a C-terminal tag on the AFF1 portion of KMT2A::AFF1 perturb AFF4/ENL/P-TEFb assembly? The constraint that rules out an N-terminal tag was derived in MLL::AF9 and does not transfer, so tag position for this fusion is an open experimental question rather than a settled design choice.
  2. Would a tagged SEM line behave like the parent in a xenograft? A tag that is neutral in culture and attenuating in vivo would invalidate every in vivo conclusion drawn from it, and the transformation control used for MLL::AF9 was a mouse leukaemogenesis assay that has no direct equivalent for a human t(4;11) line.
Every source behind this entry (4)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The authors of the only published MLL-fusion degron system state in their Limitations section that future studies should include endogenously degron-tagged MLL::AF9, and report that only a C-terminal tag on MLL::AF9 reproduced the leukaemogenesis kinetics of the untagged construct while an N-terminal tag did not.

    PMID 35245435DOI
    What this evidence cannot show

    Read verbatim from the PMC body (PMC9044330, efetch db=pmc, <body> present). A statement of intent in a 2022 Limitations section is evidence that it had not been done by 2022 in that laboratory; it is not proof that no other group has since done it, which is why the PubMed bound in E3 is also given.

    Evidence that argues against this (1)
    • The tag-position constraint is derived from MLL::AF9, where the C-terminus is AF9-derived. It does not transfer automatically to KMT2A::AFF1, whose C-terminus is the AFF1 region carrying the AFF4/ENL/P-TEFb contacts - so a C-terminal tag could perturb complex assembly in the AFF1 fusion even though it is safe in the AF9 one. Olsen SN et al. 2022
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The one published KMT2A::AFF1 knock-in is a murine Mll-AF4 fusion-gene knock-in producing lymphoid and myeloid deregulation and haematologic malignancy - a knock-in of the fusion, not an epitope tag knocked onto an existing fusion allele.

    PMID 16551973DOI
    What this evidence cannot show

    Cited from the PubMed record for the existence and nature of the model; the full text was not retrieved and no quantity from it is asserted. It is a mouse model, so it does not address the patient-derived-line question at all; it is cited because it is the single record the tag-knock-in query returns and a reader would otherwise expect it to be the thing sought.

    Evidence that argues against this (1)
    • A murine fusion knock-in does give one attribution advantage the human lines lack: the fusion is expressed from the endogenous Mll locus at endogenous level, so its abundance is not a transfection artefact. It still shares every epitope with the wild-type mouse proteins, so it solves the stoichiometry problem and none of the reagent problem. Olsen SN et al. 2022
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Two bounded PubMed queries run 2026-09-20 return one record and zero records respectively for an epitope-tag or degron knock-in at a KMT2A fusion allele.

    What this evidence cannot show

    Bounded absence, and the bounds matter. PubMed's parser collapsed both double-colon spellings to the hyphenated form, so 'MLL::AF4' and 'MLL-AF4' were the same query and only two distinct fusion spellings were actually searched. The searches are title/abstract only, so a knock-in described only in a methods section is invisible to them. A tagged line distributed as a resource and never published would also be invisible. The absence should be read as 'not reported in a title or abstract', not as 'does not exist'.

    Evidence that argues against this (1)
    • A counted absence is only as good as the symbol set it was run over, and this one did not search partner-specific spellings (MLL-ENL, MLL-AF9, KMT2A::MLLT3) against tagging terms, nor the phrase 'knock-in tagging'. The Olsen Limitations statement is the stronger evidence here and the query is corroboration. Olsen SN et al. 2022
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    SEM, RS4;11, THP-1, NOMO-1 and MOLM-13 could not be synchronised with mimosine, double thymidine or nocodazole, and centrifugal elutriation was used instead.

    PMID 17908926DOI
    What this evidence cannot show

    Read verbatim from the PMC HTML page for PMC1993870 after fullTextXML returned HTTP 500 and efetch returned no <body>. A 2007 observation in one laboratory; other groups may have since synchronised these lines by other means. It is included as a practical design constraint, not as a law.

    Evidence that argues against this (1)
    • If the synchronisation failure reflects a technical rather than a biological problem, then a modern method - for example FUCCI reporters or Hoechst-based sorting - would sidestep it entirely, and the elutriation workaround would be unnecessary. Nobody has revisited it. Olsen SN et al. 2022
Connected to