Immunoprecipitate the KMT2A fusion and map its phosphosites - the experiment this whole field has never done
also called fusion IP phospho-MS · oncoprotein phosphosite map
Pull the actual cancer fusion protein out of leukaemia cells, break it into pieces and ask a mass spectrometer where the phosphates are. Nobody has done this, so every claim about which switches the fusion keeps is currently an argument from where the chromosome broke rather than a measurement.
The same thing, for a scientist
A single, standard-platform experiment that converts every retention inference in this beat and the previous one into a measurement. Immunoprecipitate the fusion protein from KMT2A-rearranged cell lines using an N-terminal KMT2A antibody or an endogenously tagged allele, digest, enrich phosphopeptides by TiO2 or IMAC, and sequence by LC-MS/MS. The pulldown solves the attribution problem that makes patient-cohort phosphoproteomics unable to assign a retained-region site to the fusion: every peptide recovered is from the immunoprecipitated protein. Run untreated and after a genotoxic stimulus, with an isogenic KMT2A-wild-type line as the control.
Design
Lines. SEM (KMT2A::AFF1), THP-1 and MOLM-13 (KMT2A::MLLT3), KOPN-8 (KMT2A::MLLT1 - check karyotype before use). A KMT2A-wild-type AML line as the negative control for the N-terminal antibody.
Arms. Untreated; ionising radiation 10 Gy at 1 h; an anthracycline at a clinically relevant exposure at 1 h and 6 h. The 10 Gy / 1 h arm matches the condition under which AF9 S395 was mapped, which makes a negative result interpretable.
Pulldown. N-terminal KMT2A antibody, which captures the fusion and full-length KMT2A but not the wild-type partner - so any partner-derived peptide recovered is fusion-derived by construction. An endogenously tagged fusion allele is cleaner where one exists.
Readout. Phosphopeptide enrichment, LC-MS/MS, site localisation scoring reported per site. Targeted PRM for the specific residues at stake: AF9 S395, ENL S394/S398, ELL T551/S561/S589 and S308/S309, AF10 S658/S689, and the previous beat's KMT2A S136/S142.
Controls. The wild-type partner immunoprecipitated in parallel from the same lysate, which gives the direct fusion-versus-wild-type occupancy comparison nobody has. A phosphatase-treated aliquot as the technical negative.
What makes a negative interpretable. If wild-type AF9 in the same lysate shows phospho-S395 and the fusion does not, that is a real negative and it falsifies the retention hypothesis for that site. Without the parallel wild-type arm, a negative is uninterpretable.
| design | Immunoprecipitate the KMT2A fusion with an N-terminal KMT2A antibody or a tagged allele from SEM, THP-1, MOLM-13 and KOPN-8; digest; enrich phosphopeptides by TiO2/IMAC; LC-MS/MS with targeted PRM for the named residues. Arms: untreated, IR 10 Gy 1 h, anthracycline 1 h and 6 h. Parallel wild-type partner immunoprecipitation from the same lysate as the comparison arm. |
|---|---|
| readout | Per-site phosphorylation on the fusion protein with localisation scores, and fusion-versus-wild-type occupancy for every retained site |
| controls | KMT2A-wild-type AML line; phosphatase-treated aliquot; parallel wild-type partner pulldown; IgG |
| cost tier | med |
| timeline | 3-4 months including antibody validation |
| what a negative means | Interpretable only because of the parallel wild-type arm: wild-type positive plus fusion negative falsifies retention for that site |
targets
[ "MLLT3 S395", "MLLT1 S394/S398 (inferred)", "ELL T551/S561/S589", "ELL S308/S309", "MLLT10 S658/S689", "KMT2A S136/S142 (previous beat)" ]
- A negative result here would undercut the therapeutic rationale of this entire beat and of the previous beat's central CK2 claim. That is the point of running it first rather than last, and it is the reason the parallel wild-type arm is not optional.
- Is the fusion abundant enough for phosphopeptide enrichment after immunoprecipitation from a practical number of cells? This is the main feasibility risk and a pilot with an abundant known site would settle it.
- Does an N-terminal KMT2A antibody with adequate immunoprecipitation efficiency exist, or is an endogenous tag required?
- Would a junction-spanning phosphopeptide be detectable at all, given that the junction sequence is patient-specific?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
No study has mapped a phosphorylation site on a KMT2A fusion protein: PubMed returns zero records for KMT2A/MLL fusion terms combined with phosphosite, phosphorylation-site or phospho-specific-antibody terms.
PubMed esearch, three stated queries run 2026-09-19; counts 0, 0 and 1, the single record being a patient-cohort study · NCBI E-utilities 2026What this evidence cannot show
A counted absence over [tiab] queries; work reported only in a supplementary table or a figure panel would be invisible to it.
Evidence that argues against this (1)
- An integrative phosphoproteomic study of 74 KMT2A-rearranged AML patients does report phosphosites on MLLT1, MLLT3, MLLT10 and ELL, so the reagents and the platform plainly exist; what is missing is the immunoprecipitation step that would attribute them to the fusion. Casado P et al. 2023
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The specific sites the experiment would test include AF9 S395 (DNA-PKcs), ELL T551/S561/S589 (ATM) and ENL's conserved SQ sites, all of which lie on retained partner sequence and none of which has been measured on a fusion protein.
Post-translational modification-dependent oligomerization switch in regulation of global transcription and DNA damage repair during genotoxic stress. · Talukdar P, et al. · Nat Commun 2024What this evidence cannot show
The site list is assembled from three papers, all on wild-type proteins in non-haematopoietic cells. Whether each site is phosphorylated in a leukaemic cell at all is itself untested.
Evidence that argues against this (1)
- The same considerations that make this experiment worth doing make a negative likely for some sites: a kinase docking interaction may live on the partner's discarded N-terminus even where the acceptor residue is retained, in which case the fusion would carry the residue and never be phosphorylated at it. Pal S et al. 2022
measures whether the fusion carries the sites at all
the experiment that closes the absence
- ← depends onKnock S395A into the MLLT3 half of a KMT2A::MLLT3 model and ask whether it becomes chemotherapy-resistant
needs the fusion phospho-blot to confirm the perturbation
- ← related toQ25 enumeration: what was searched, found, rejected and could not be read across the five 3' partners
the experiment the enumeration points to first