Does the fusion need the PWWP arm? An artificial CXXC + H3K36me2-reader + AF4-interaction protein substitutes for a whole MLL fusion - but LEDGF itself is not required for the fusion to reach chromatin
also called LEDGF PWWP requirement · minimal MLL fusion · H3K36me2 targeting module
A stripped-down artificial protein built from a DNA-recognising piece, a histone-mark-reading piece and an elongation-recruiting piece works as well as a real leukaemia fusion. That says the histone-reading function matters. It does not say LEDGF has to be the protein doing it.
The same thing, for a scientist
The question the graph needs answered is whether a PWWP-directed agent would do anything to a KMT2A fusion, and the literature answers two adjacent questions instead. **Yes, an H3K36me2/3-reading module matters:** only two targeting modules are necessary and sufficient for MLL-fusion target recognition - one binding non-methylated CpG and one binding H3K36me2/3 - and an artificial protein of those two modules plus an AF4-family coactivator interaction domain can functionally substitute for an MLL fusion protein. **No, LEDGF is not required for the fusion's chromatin retention:** LEDGF knockdown increases fusion occupancy, and the fusion's own CXXC/RD1 domain alone suffices for recruitment. The reconciliation is that in the artificial construct the reader module is supplied **in cis**, fused to the CXXC; in the real fusion it is supplied **in trans** by menin-bridged LEDGF, and a factor supplied in trans through a micromolar interaction is much more dispensable than one welded to the same polypeptide. So the mark matters, the module matters, and whether the endogenous protein carrying it is load-bearing in an established leukaemia is exactly what nobody has measured.
The three results that have to sit together
| Result | What it shows | What it does not show |
|---|---|---|
| Artificial CXXC + H3K36me2/3-reader + AF4-interaction protein substitutes for an MLL fusion 1 | an H3K36me2/3 reader module is sufficient in cis | that endogenous LEDGF is necessary in trans |
| LEDGF knockdown increases fusion occupancy; CXXC/RD1 alone suffices for fusion recruitment 2 | LEDGF is not the fusion's address | that the fusion does not need LEDGF for anything |
| ASH1L writes the H3K36me2 that LEDGF reads; KDM2A overexpression lowers H3K36me2 and reduces LEDGF and wild-type MLL occupancy but not fusion occupancy 2 | the mark controls LEDGF and wild-type MLL, not the fusion | that the mark is irrelevant to the leukaemia - ASH1L loss does impair transformation |
The consistent thread is that the fusion's own targeting is CpG-driven, and the H3K36 arm belongs to LEDGF and wild-type MLL. Since wild-type MLL is itself required for MLL-fusion leukaemogenesis, the H3K36 arm still matters to the disease - just one step removed from the oncoprotein.
The uncomfortable consequence for a PWWP inhibitor
A PWWP occupancy inhibitor is a drug aimed at the LEDGF-H3K36me2/3 contact. On the evidence above, breaking that contact:
- does not evict the fusion (LEDGF loss raises fusion occupancy)
- does displace wild-type MLL (KDM2A overexpression lowering H3K36me2 reduced wild-type MLL and LEDGF occupancy)
- would remove whatever the fusion gets from LEDGF functionally rather than positionally - the AEP/SEC, BRD4 and P-TEFb recruitment described in
mechanism-ledgf-recruits-elongation-machinery-not-the-fusion
That is a coherent rationale but it is a different one from the rationale the field states, and its pharmacodynamic readout is transcription rather than fusion ChIP. It also means a PWWP agent's efficacy runs partly through inhibiting wild-type MLL, which is a dependency of the leukaemia but also of normal haematopoiesis.
And there is a specific warning: SETD2 loss of function, prevalent in MLL leukaemias, increases H3K36me2, LEDGF and MLL occupancy at MLL target genes 2. A SETD2-mutant patient therefore has more LEDGF on target genes, which could mean more dependence or simply a higher ligand burden for a competitive PWWP inhibitor to displace.
The experiment that is missing
A PWWP separation-of-function mutant of endogenous LEDGF, expressed in place of the endogenous protein in an established KMT2A-rearranged cell. The nearest existing experiment is the pME chimera's W21A mutation, which abolished Hoxa9 upregulation and transformation 5 - but that tests a PWWP fused in cis inside an engineered construct, not the endogenous protein acting in trans. The IBD arm has this experiment - IBM2 point mutants, F148A/F151A fusion mutants, F129A - and the PWWP arm has nothing equivalent in a leukaemic system. Rescue arms 6 and 7 of exp-psip1-inducible-knockdown-established-kmt2a-aff1-xenograft are designed to supply it.
Until that exists, the arm of LEDGF with chemical matter is the arm with no genetic validation in this disease, and that asymmetry should be stated in any target dossier.
| question | Does the KMT2A fusion require the LEDGF PWWP arm? |
|---|---|
| in cis module sufficient | true |
| endogenous ledgf required for fusion occupancy | false |
| mark written by | ASH1L (H3K36me2) |
| mark erased by | KDM2A |
| setd2 lof effect | increases H3K36me2, LEDGF and MLL occupancy at MLL target genes |
| missing experiment | PWWP separation-of-function mutant rescue in an established KMT2A-rearranged cell |
| nearest existing pwwp separation of function | pME chimera W21A (LEDGF PWWP fused to MLL-ENL), abolishes Hoxa9 upregulation and transformation - but the PWWP is supplied in cis |
- A PWWP inhibitor's efficacy would run partly through displacing wild-type MLL, which normal haematopoiesis also requires.
- Multivalent chromatin binding means a cage ligand may change residence time without changing measured occupancy, so a real effect and no effect look the same in the standard assay.
- SETD2 loss of function, common in this disease, increases LEDGF occupancy, so the competitive burden on a PWWP agent is highest in a recognisable patient subgroup.
- Is there any PWWP separation-of-function experiment in a KMT2A-rearranged system? The IBD arm has three; the PWWP arm has none, and that asymmetry is invisible in every dossier that treats the two arms as alternatives.
- Do SETD2-mutant KMT2A-rearranged patients have more LEDGF on target genes, and does that make them more or less sensitive to a PWWP-directed agent? A competitive inhibitor faces a higher ligand burden in exactly the patients with the most LEDGF.
- Would an ASH1L inhibitor be a better way to attack the same axis? It removes the mark rather than the reader, has published transformation data, and is not subject to the avidity problem.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Only two functional modules are necessary and sufficient for MLL-fusion target recognition - one binding non-methylated CpG and one binding H3K36me2/3 - and an artificial protein composed of those two targeting modules plus an interaction domain for AF4-family coactivators can functionally substitute for MLL fusion proteins.
MLL fusion proteins link transcriptional coactivators to previously active CpG-rich promoters. · Okuda H, Kawaguchi M, et al. · Nucleic Acids Res 2014What this evidence cannot show
An engineered-substitution experiment: it establishes sufficiency of the module combination, not necessity of the endogenous protein that normally supplies the reader. In the artificial construct the reader is fused in cis, which is a much stronger arrangement than a trans interaction through a micromolar interface. Open access in Europe PMC (PMC3985622).
Evidence that argues against this (1)
- In the endogenous setting, LEDGF knockdown increases rather than decreases fusion occupancy and the fusion's CXXC/RD1 domain alone suffices for recruitment, so the in-trans reader is evidently not required for targeting. Zhu L et al. 2016
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
ASH1L writes the H3K36me2 that LEDGF preferentially binds and facilitates recruitment of LEDGF and wild-type MLL to key leukaemia target genes; KDM2A overexpression reduces H3K36me2 and reduces occupancy of LEDGF and wild-type MLL but not of the fusion; and SETD2 loss of function, prevalent in MLL leukaemias, increases occupancy of H3K36me2, LEDGF and MLL at MLL target genes, so H3K36me3 is not essential for MLL occupancy or leukaemogenesis.
ASH1L Links Histone H3 Lysine 36 Dimethylation to MLL Leukemia. · Zhu L, Li Q, et al. · Cancer Discov 2016What this evidence cannot show
Murine transformed HSPCs; ChIP-qPCR at a small number of loci; the SETD2 statement is the authors' synthesis of their KDM2A and ASH1L data with the known mutation spectrum rather than a SETD2-null experiment. Read via efetch db=pmc after fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- ASH1L itself is a crucial regulator of MLL-dependent transcription and leukaemic transformation in the same paper, so the H3K36me2 axis clearly matters to the disease even though it does not control fusion occupancy - which means 'not required for fusion occupancy' must not be read as 'not required'. Zhu L et al. 2016
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The LEDGF PWWP-nucleosome contact is multivalent, engaging the methylated H3 tail and nucleosomal DNA across both gyres, which is why occupancy is dominated by avidity rather than by cage affinity alone.
Structure of H3K36-methylated nucleosome-PWWP complex reveals multivalent cross-gyre binding. · Wang H, Farnung L, et al. · Nat Struct Mol Biol 2020What this evidence cannot show
Structural; cited to explain why a cage-occupancy inhibitor faces an avidity problem. Cryo-EM at 3.2 A on a reconstituted methylated mononucleosome.
Evidence that argues against this (1)
- Multivalency also means a PWWP ligand need not fully displace LEDGF to change its residence time, so the avidity argument bounds the modality without eliminating it. Koutná E et al. 2023
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
LEDGF recognition of the nucleosome is multivalent with contributions from the PWWP cage, the acidic patch/DNA interface and DNA-binding motifs in the disordered linker, so the protein is a multivalent chromatin binder rather than a single-site reader.
Multivalency of nucleosome recognition by LEDGF. · Koutná E, Lux V, et al. · Nucleic Acids Res 2023What this evidence cannot show
Biophysical and structural work on reconstituted nucleosomes; cited for the multivalency argument.
Evidence that argues against this (1)
- If chromatin binding is distributed across several weak contacts, a drug occupying one of them may reduce residence time without producing a measurable occupancy change, so the standard ChIP readout could miss a real pharmacological effect. Wang H et al. 2020
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
A PWWP separation-of-function experiment does exist, in a chimera rather than in endogenous LEDGF: in the pME construct (LEDGF PWWP fused to MLL-ENL), mutation of the conserved PWWP tryptophan to alanine (W21A) abolished both Hoxa9 upregulation and myeloid transformation; pME also occupied the Hoxa9 locus in the absence of both menin and LEDGF, and induced transplantable acute leukaemia in vivo with latency and penetrance similar to ME.
Menin critically links MLL proteins with LEDGF on cancer-associated target genes. · Yokoyama A, Cleary ML · Cancer Cell 2008What this evidence cannot show
The PWWP is supplied in cis inside an engineered fusion, so W21A tests whether an H3K36me-reading module is required by that construct, not whether endogenous LEDGF's PWWP is required by a natural KMT2A fusion. Murine myeloid progenitors. Read via efetch db=pmc (29,353 characters, body element present) after Europe PMC fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- That pME occupies Hoxa9 in the absence of both menin and LEDGF is the same result as the later occupancy work: a construct carrying CXXC plus a reader module does not need the endogenous bridge, so the chimera cannot report on whether the natural fusion needs LEDGF in trans. Zhu L et al. 2016
whether the fusion requires this contact at all
the in-cis versus in-trans reconciliation
the arm with chemistry is the arm without genetic validation
- related to →KMT2A / MLL1 — the gene that breaks
CXXC-driven targeting is the fusion's own, LEDGF's H3K36 arm belongs to wild-type KMT2A
the PCE chimera bypass is the same module-sufficiency architecture