The clustered PHD fingers (KMT2A aa 1416-1640) are a chromatin-targeting module and the canonical t(4;11) junction removes them by ten residues
also called KMT2A PHD1-3 targeting · 2A13 clustered PHD · PHD fingers lost in the fusion
KMT2A has a cluster of small zinc-binding fingers that steer it toward already-active genes marked with a particular chemical tag. The commonest infant-leukemia fusion cuts the protein just ten residues before that cluster starts, so the fusion keeps the parts that get it into the nucleus and onto CG-rich DNA and loses the part that reads the activity mark.
The same thing, for a scientist
KMT2A carries a triplet of plant homeodomain fingers plus an extended PHD. Expressed as an isolated NLS-EYFP-tagged cluster in HeLa S3 and read out by greenCUT&RUN, the KMT2A cluster (designated 2A13, **aa 1416-1640** of UniProt Q03164) binds well-defined loci at H3K4me3- and H3-acetylation-rich **active promoters and enhancers**; the third PHD alone is aa 1565-1631. Strikingly, clusters from promoter-specific KMT2A/B and enhancer-specific KMT2C/D bind with little difference, and **fusing the KMT2A CXXC domain to the PHDs drastically shifts the preference toward promoters** - which the authors advance as the explanation for why full-length KMT2A/B are promoter-specific and KMT2C/D are not [E1]. The canonical t(4;11) chimera joins KMT2A residue 1406 to AFF1 [E2], so the entire cluster is lost, beginning ten residues past the junction. The fusion therefore retains the CXXC address [E3] with **no PHD partner for it to act on**, which is the reverse of the arrangement the 2023 paper shows produces promoter specificity. This is a structural prediction with a clean falsifier and it has not been tested.
No three-dimensional model is available for this protein, so the map above is the whole picture. That is itself informative: AlphaFold does not model chains this long, and the parts that have been crystallised are held under the entries for the individual domains.
The arithmetic, stated exactly
| element | KMT2A residues | in KMT2A::AFF1? |
|---|---|---|
| AT-hook region / nuclear localisation determinant | N-terminal, well upstream | retained |
| CXXC | ~1147-1195 | retained |
| canonical t(4;11) junction | 1406 | - |
| clustered PHD (2A13) | 1416-1640 | lost |
| PHD3 alone (2A3) | 1565-1631 | lost |
The 1416 figure is the start of the construct the 2023 authors expressed, taken verbatim from their methods, and is a construct boundary rather than a structural domain boundary; the true first zinc-coordinating residue may sit a little later. The junction residue 1406 is this graph's canonical value for the commonest breakpoint and varies with breakpoint. The correct statement is therefore the cluster lies immediately downstream of the junction and is entirely absent from the fusion, not that exactly ten residues separate them in every patient.
The prediction the CXXC-PHD result makes about the fusion
The 2023 result is that CXXC plus PHDs is what makes a KMT2 protein promoter-specific; PHDs alone do not discriminate promoters from enhancers, and adding CXXC to them does 1. The fusion has CXXC and no PHDs. Two readings follow and they are distinguishable:
- If CXXC is sufficient for the promoter preference, the fusion should be at least as promoter-biased as wild-type KMT2A, and the PHD loss changes little about where it lands.
- If the promoter preference is a joint property, the fusion should be less discriminating - broader, flatter, spread further into gene bodies and enhancers.
The second reading is consistent with the broad-domain occupancy pattern described for KMT2A fusions elsewhere in this graph, but that pattern is usually attributed to the partner's elongation activity rather than to a lost reader. The experiment that separates them is to re-supply the PHD cluster in cis to the fusion - the same design this graph already holds for MLLT10's PZP module - and ask whether occupancy narrows.
What the 2023 system can and cannot support
It is an isolated-domain, tagged, overexpressed, heterologous system: lentiviral NLS-EYFP-tagged domain clusters in HeLa S3, with greenCUT&RUN. Three specific limits.
- The NLS comes from the tag, so nothing in this system reports on nuclear import.
- An isolated cluster has no constraint from the rest of the protein, and the authors' own overlap analysis with full-length KMT2A peaks is partial rather than complete.
- The imaging control is informative and cuts against punctate models: the nuclear EYFP signal was diffuse, which the authors use to argue against aggregation artefacts, and endogenous full-length KMT2D also showed no punctate pattern 1. A diffuse nuclear signal from a chromatin-binding module is a reminder that binding thousands of sites looks diffuse at the diffraction limit, and that a punctate image is therefore not the default expectation for a chromatin reader.
Point 3 is the part most relevant to this beat: it is a published, internally-controlled observation that a KMT2-family chromatin-targeting module does not look punctate.
| uniprot | Q03164 |
|---|---|
| retained in fusion | false |
| canonical junction kmt2a residue | 1406 |
| system of evidence | isolated NLS-EYFP-tagged domain clusters, HeLa S3, greenCUT&RUN |
| imaging control | nuclear EYFP signal diffuse, not punctate |
domains
[
{
"name": "PHD1-3 cluster (2A13)",
"start": 1416,
"end": 1640
},
{
"name": "PHD3 (2A3)",
"start": 1565,
"end": 1631
}
]- If the PHD cluster restrains KMT2A's spread, a therapeutic that restored PHD function in cis would be re-installing a domain that also carries KMT2A's degradation signal, and could destabilise wild-type KMT2A as a side effect.
- Targeting the retained CXXC to narrow the fusion's genomic address would act identically on wild-type KMT2A and KMT2B, both of which use CXXC for promoter specificity.
- Does re-supplying the clustered PHD region in cis to KMT2A::AFF1 narrow its occupancy toward promoters, as the CXXC-plus-PHD result predicts?
- Is the broad-domain occupancy of KMT2A fusions partly a consequence of losing a reader rather than only of gaining an elongation module?
- Where exactly does the first zinc-coordinating residue of PHD1 sit, so that the gap between junction and cluster can be stated structurally rather than from a construct boundary?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Clustered PHD domains of KMT2A-D localise to H3K4me3 and H3 acetylation-rich active promoters and enhancers with little difference between promoter-specific KMT2A/B and enhancer-specific KMT2C/D clusters; fusion of the KMT2A CXXC domain to the PHDs drastically enhances preference for promoters over enhancers; the KMT2A cluster construct 2A13 spans aa 1416-1640 and 2A3 spans aa 1565-1631 of UniProt Q03164; and the nuclear EYFP signal of the expressed clusters was diffuse rather than punctate.
Clustered PHD domains in KMT2/MLL proteins are attracted by H3K4me3 and H3 acetylation-rich active promoters and enhancers · Stroynowska-Czerwinska AM, et al. · Cellular and molecular life sciences : CMLS 2023What this evidence cannot show
Read from Europe PMC full-text XML for PMC9813062. Isolated domain clusters, NLS-EYFP-tagged, lentivirally expressed in HeLa S3, read out by greenCUT&RUN - not full-length protein, not a haematopoietic cell, and not a leukaemia. The residue ranges are construct boundaries taken from the methods, not structurally determined domain limits. The supplied NLS means the system cannot report on nuclear import.
Evidence that argues against this (1)
- The authors' own Jaccard and Venn comparisons show the isolated clusters overlap full-length KMT2A, WDR5 and RbBP5 targets only partially, so an isolated cluster's genome-wide distribution is not a faithful proxy for the same domains inside the full-length protein. Stroynowska-Czerwinska AM et al. 2023
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The t(4;11)(q21;q23) translocation generates an MLL-AF4 fusion product in which the KMT2A contribution ends in the region upstream of the PHD cluster.
Acute mixed-lineage leukemia t(4;11)(q21;q23) generates an MLL-AF4 fusion product · Domer PH, et al. · Proceedings of the National Academy of Sciences of the United States of America 1993What this evidence cannot show
The founding report establishes the fusion product. The specific junction residue 1406 is this graph's canonical value for the commonest breakpoint and is not universal; breakpoint heterogeneity shifts how much KMT2A is retained, although no reported KMT2A breakpoint retains the PHD cluster intact.
Evidence that argues against this (1)
- Reported KMT2A breakpoints extend into intron 11 in some patients, which truncates PHD1 of the reciprocal product - so the amount of PHD sequence at stake is breakpoint-dependent and the 'entirely absent' statement should be checked per breakpoint. Butler LH et al. 1997
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The fusion retains the CXXC domain, whose CpG-binding function is essential for target recognition, transactivation and myeloid transformation by an MLL oncoprotein.
Binding to nonmethylated CpG DNA is essential for target recognition, transactivation, and myeloid transformation by an MLL oncoprotein · Ayton PM, et al. · Molecular and cellular biology 2004What this evidence cannot show
Identifiers taken from a live PubMed esummary/efetch run on 2026-09-20, abstract-level reading. The experiment is an MLL oncoprotein in a myeloid transformation assay, not KMT2A::AFF1 in human leukaemia, and this graph's CXXC entry already records that CpG-binding mutants lose transformation without changing subnuclear position.
Evidence that argues against this (1)
- Because CpG mutants lose transformation without moving in the nucleus, CXXC is a chromatin address rather than a compartment address, so the CXXC-plus-PHD promoter-specificity argument is about genomic targeting and may have no subnuclear correlate at all. Yano T et al. 1997
structural coordinates for the lost cassette
the PHD cluster carries CyP33's only route to a subnuclear compartment and is lost at the breakpoint
CXXC plus PHDs is what makes a KMT2 protein promoter-specific; the fusion keeps CXXC and loses the PHDs
- supports →Lost from the KMT2A half: PHD3 and the bromodomain are one cassette, not two modules (aa 1566-1765)
adds the measured chromatin-targeting function of the cluster that the subtraction entry records as lost
- related to →Lost from the KMT2A half: PHD2 (aa 1479-1533), an E3 ubiquitin ligase that limits KMT2A's own activity
the same lost cluster carries both the reader function and the E3 activity
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
the canonical junction at KMT2A 1406 removes the cluster beginning at 1416
- part of →KMT2A / MLL1 — the gene that breaks
clustered PHD fingers of KMT2A, aa 1416-1640