The KMT2A::MLLT1 chimeric protein: exactly what survives the join on each side, in residue numbers
also called MLL-ENL fusion protein · MLL::ENL · KMT2A-MLLT1 chimera
The fusion protein is the front third of KMT2A welded onto most of ENL, and the weld point decides which working parts each parent contributes. KMT2A gives up every one of its own chromatin-writing and chromatin-reading modules and keeps only its targeting end; ENL gives up almost nothing.
The same thing, for a scientist
In the commonest architecture - a KMT2A intron 11 break joined to MLLT1 exon 2 - the chimera carries the KMT2A N-terminus through the CXXC zinc finger and the whole of MLLT1 from the YEATS domain onward. Reading residue coordinates back from UniProt on 2026-09-19: Q03164 (KMT2A_HUMAN, primary gene name KMT2A, 3969 aa) annotates the menin-binding motif at 6-25, integrase-binding motifs IBM1 123-134 and IBM2 147-152, three A.T hooks at 169-180, 217-227 and 301-309, and the CXXC-type zinc finger at 1147-1195; all of these lie 5' of the junction [E1]. Everything from PHD1 (1431-1482) onward is lost, including PHD2 (1479-1533), PHD3 (1566-1627), the bromodomain (1635-1765), the pre-PHD C2HC finger (1870-1910), PHD4 (1931-1978), FYRN (2018-2074), the 9aaTAD (2847-2855), FYRC (3666-3747), the WDR5-interaction WIN motif (3762-3767), the SET domain (3829-3945) and post-SET (3953-3969) [E1]. On the 3' side Q03111 (ENL_HUMAN, primary gene name MLLT1, 559 aa) annotates a YEATS domain at 1-138 with acylated-histone-binding contact regions at 56-58 and 78-80, and disorder over 147-484 [E2]. The functional C-terminus that every published MLL-ENL construct retains was mapped by deletion mutagenesis to the last 90 residues, i.e. approximately 470-559 [E3]. The asymmetry is the point: the chimera is a KMT2A targeting module with no KMT2A catalysis, fused to an essentially intact ENL.
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.
Why the residue numbers were re-fetched rather than reused
Both accessions were resolved by fetching the UniProt REST record and reading back the primary gene name and the sequence length before any coordinate was copied out. gene_exact: in the UniProt query API matches synonyms, so a search that looks exact is not one, and this project has twice recorded a coordinate from an entry that turned out to be a different protein or a different isoform. Q03164 reads back as KMT2A_HUMAN, primary gene name KMT2A, synonyms ALL1/CXXC7/HRX/HTRX/MLL/MLL1/TRX1, length 3969. Q03111 reads back as ENL_HUMAN, primary gene name MLLT1, synonyms ENL/LTG19/YEATS1, length 559, and the entry lists no alternative isoforms, so the isoform trap that produced the O00255-1/-2 error in cycle 14 does not arise here 12.
What the KMT2A side keeps
Menin-binding motif (6-25), the two LEDGF/PSIP1 integrase-binding motifs (123-134 and 147-152), the three A.T hooks (169-180, 217-227, 301-309) and the CXXC zinc finger (1147-1195) 1. That is the entire targeting apparatus - menin docking, LEDGF docking, minor-groove DNA binding and unmethylated-CpG binding - and nothing else. The CXXC finger ends at 1195 and PHD1 begins at 1431, so an intron 11 junction falls in the 235-residue stretch between them and cleanly separates targeting from reading.
What the KMT2A side loses, and why the list matters for the reciprocal question
Every module from PHD1 to post-SET 1. Two of them are load-bearing for the reciprocal arm of this beat. PHD1-3 mediate CYP33/PPIE binding and PHD1 self-dimerisation, and the recombinome reports that breaks upstream of KMT2A exon 11 leave PHD1 structurally intact on the reciprocal product while breaks within or downstream of exon 11 appear not to 4. Since MLLT1 is one of only two partners that prefer intron 11 constitutively 4, the reciprocal products that do exist in this partner are disproportionately the PHD1-impaired kind. The SET domain (3829-3945) and the WIN motif (3762-3767) go to the reciprocal allele too, which is why a reciprocal-positive patient carries an H3K4 methyltransferase module that a reciprocal-negative patient does not.
What the MLLT1 side keeps
In an exon-2 junction, essentially all of ENL: the YEATS acyl-lysine reader (1-138, with histone-contact regions at 56-58 and 78-80), the long disordered spacer (147-484) and the C-terminal AHD hub 2. The AHD has no experimental fold boundary in UniProt; this graph carries 485-559 as a predicted boundary from the end of the annotated disorder, and the independent functional boundary is Rubnitz's deletion mapping of Enl's transactivation to its C-terminal 90 amino acids, i.e. approximately 470-559, a region the paper notes is highly conserved with AF9 and retained in all HRX-ENL and HRX-AF9 fusion proteins 3. The two boundaries are 15 residues apart and neither is a crystallographic edge; treat the N-terminal edge as soft.
The junction is not fixed, and the point for this beat is that it varies with reciprocal status
MLLT1 junctions at exons 2, 4, 6 and 7 have all been reported, all in frame 5. A spliced fusion - the reciprocal-negative class - necessarily joins the last KMT2A exon to MLLT1 exon 2, because the break is upstream of MLLT1 exon 1 and exon 2 is the first exon carrying a bona fide splice acceptor 4. A balanced fusion may join exon 2, 4, 6 or 7. So the reciprocal-negative arm is architecturally homogeneous at the protein level and the reciprocal-positive arm is not, which is a confound running in a direction nobody has measured. See concept-the-two-mllt1-arms-do-not-carry-identical-direct-products.
| uniprot | Q03164 |
|---|---|
| gene symbol | KMT2A |
| length AA | 3969 |
| uniprot partner | Q03111 |
| gene symbol partner | MLLT1 |
| length AA partner | 559 |
| accession resolution | Both accessions fetched from https://rest.uniprot.org/uniprotkb/<acc>.json on 2026-09-19; primary gene name and sequence length read back before any coordinate was copied. Q03164 -> KMT2A_HUMAN / KMT2A / 3969 aa. Q03111 -> ENL_HUMAN / MLLT1 / 559 aa, no alternative isoforms annotated. |
| junction note | Commonest architecture: KMT2A intron 11 break spliced to MLLT1 exon 2. The junction falls between the CXXC finger (ends 1195) and PHD1 (starts 1431). |
| PDB note | No structure of any KMT2A::MLLT1 fusion protein exists. Component structures are catalogued in fusion-kmt2a-mllt1 and gene-mllt1 and are not duplicated here. |
domains — the numbers behind the chart
[
{
"name": "MBM1",
"start": 6,
"end": 25,
"note": "UniProt Q03164 annotates a single 'Menin-binding motif (MBM)' at 6-25; this graph's gene-kmt2a entry splits the same region into MBM1 4-15 and MBM2 23-40. Retained in the fusion."
},
{
"name": "AT-hook 1",
"start": 169,
"end": 180,
"note": "retained"
},
{
"name": "AT-hook 2",
"start": 217,
"end": 227,
"note": "retained"
},
{
"name": "AT-hook 3",
"start": 301,
"end": 309,
"note": "retained"
},
{
"name": "CXXC",
"start": 1147,
"end": 1195,
"note": "retained; last KMT2A module before the junction. UniProt gives 1147-1195; this graph's gene-kmt2a entry gives 1147-1242."
},
{
"name": "PHD1",
"start": 1431,
"end": 1482,
"note": "LOST to the reciprocal allele"
},
{
"name": "PHD2",
"start": 1479,
"end": 1533,
"note": "LOST"
},
{
"name": "PHD3",
"start": 1566,
"end": 1627,
"note": "LOST"
},
{
"name": "Bromodomain",
"start": 1635,
"end": 1765,
"note": "LOST"
},
{
"name": "pre-PHD C2HC zinc finger",
"start": 1870,
"end": 1910,
"note": "LOST"
},
{
"name": "PHD4",
"start": 1931,
"end": 1978,
"note": "LOST"
},
{
"name": "FYRN",
"start": 2018,
"end": 2074,
"note": "LOST"
},
{
"name": "TAD",
"start": 2847,
"end": 2855,
"note": "UniProt annotates a 9aaTAD at 2847-2855 inside the broader TAD this graph's gene-kmt2a entry gives as 2829-2883. LOST"
},
{
"name": "FYRC",
"start": 3666,
"end": 3747,
"note": "LOST"
},
{
"name": "WIN",
"start": 3762,
"end": 3767,
"note": "UniProt gives 3762-3767; gene-kmt2a gives 3762-3773. LOST"
},
{
"name": "SET",
"start": 3829,
"end": 3945,
"note": "LOST"
}
]domains partner — the numbers behind the chart
[
{
"name": "YEATS domain (acyl-lysine reader)",
"start": 1,
"end": 138,
"note": "UniProt Q03111 boundary. Retained in an MLLT1 exon-2 junction; lost in exon 6/7 junctions."
},
{
"name": "ENL central disordered region",
"start": 147,
"end": 484,
"note": "UniProt disorder annotation. Retained."
},
{
"name": "ANC1 homology domain (AHD)",
"start": 485,
"end": 559,
"note": "PREDICTED boundary (end of annotated disorder to the C-terminus); no ENL AHD structure exists. Rubnitz's functional boundary is the C-terminal 90 aa, approximately 470-559. Retained in every reported junction."
}
]PDB IDS
[]
- The fusion's only ligandable partner-derived pocket is the ENL YEATS aromatic cage, and it is present in exon-2 junctions and absent in exon-6/7 junctions. A YEATS-directed agent would therefore be active in some KMT2A::MLLT1 patients and inert in others, with no assay in routine diagnostic use that distinguishes them.
- Every KMT2A module that could be targeted directly - SET, WIN, the bromodomain, PHD1-3 - sits on the reciprocal allele, not on the fusion. A drug designed against them hits wild-type KMT2A and, in the reciprocal-positive half of this population, the reciprocal product as well.
- If the YEATS reader is only partially retained because exon 1 is lost, a YEATS-pocket inhibitor validated on wild-type ENL may have a different affinity for the fusion-borne domain than for the one it was optimised against.
- Where exactly does the chimera switch from KMT2A to MLLT1 in residue numbers? Every published description is at exon resolution; no paper gives the fusion protein's junction residue for this partner, so the 3D domain ruler for this fusion is drawn from two parent proteins and an inferred seam.
- Is the ENL YEATS fold actually intact in the chimera? An exon-2 junction removes the exon-1-encoded residues and the YEATS domain begins at residue 1, so the fusion may carry a truncated reader rather than an intact one. Nobody has expressed the fusion and tested acyl-lysine binding on it.
- Does the AHD begin at 470, 485 or 500? The published boundaries differ by about 30 residues, and the question matters for any attempt to design against the ENL AHD:AFF interface in the fusion specifically.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt Q03164 is KMT2A_HUMAN, primary gene name KMT2A (synonyms ALL1, CXXC7, HRX, HTRX, MLL, MLL1, TRX1), 3969 aa, and annotates a menin-binding motif at 6-25, IBM1 at 123-134, IBM2 at 147-152, A.T hooks at 169-180, 217-227 and 301-309, a CXXC-type zinc finger at 1147-1195, PHD-type fingers 1-4 at 1431-1482, 1479-1533, 1566-1627 and 1931-1978, a bromodomain at 1635-1765, a C2HC pre-PHD finger at 1870-1910, FYRN at 2018-2074, a 9aaTAD at 2847-2855, FYRC at 3666-3747, the WIN motif at 3762-3767, the SET domain at 3829-3945 and post-SET at 3953-3969.
UniProtKB entry Q03164 (KMT2A_HUMAN), Histone-lysine N-methyltransferase 2A · UniProt Consortium · UniProtKB/Swiss-Prot 2026What this evidence cannot show
Fetched 2026-09-19 from the REST API; the primary gene name and length were read back before any coordinate was used. UniProt domain boundaries are curated predictions except where a structure covers them; the SET and bromodomain edges are structure-backed, the motif edges are not. UniProt names one menin-binding motif at 6-25 where this graph's gene-kmt2a entry splits it into MBM1 4-15 and MBM2 23-40; the discrepancy is a naming convention, not a disagreement about the sequence.
Evidence that argues against this (1)
- Domain coordinates are not breakpoints. The exact residue at which any individual patient's chimera switches from KMT2A to MLLT1 is set by the genomic break and the splice, and is not published per patient for this partner, so 'retained' and 'lost' here describe the commonest architecture rather than any one person's protein. Meyer C et al. 2023
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt Q03111 is ENL_HUMAN, primary gene name MLLT1 (synonyms ENL, LTG19, YEATS1), 559 aa, with a YEATS domain at 1-138, acylated-histone-binding regions at 56-58 and 78-80, a disordered region at 147-484, subcellular location Nucleus, and no annotated alternative isoforms.
UniProtKB entry Q03111 (ENL_HUMAN), Protein ENL · UniProt Consortium · UniProtKB/Swiss-Prot 2026What this evidence cannot show
Fetched 2026-09-19; primary gene name and length read back. UniProt annotates no ANC1 homology domain for ENL, so the AHD boundary used elsewhere in this graph (485-559) is inferred from the end of the disorder annotation and is explicitly marked predicted. The 'Nucleus' location is a curated summary statement, not a subnuclear measurement.
Evidence that argues against this (1)
- A domain annotated on the wild-type protein is not evidence that the same fold exists in the chimera. No structure of any KMT2A::MLLT1 fusion protein has been solved, and no experiment has shown that the ENL YEATS fold is intact within the fusion rather than merely present in sequence. Erb MA et al. 2017
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Deletion mutagenesis localised the minimal portion of Enl required for transcriptional activation to its C-terminal 90 amino acids, a region highly conserved between Enl and the t(9;11) partner Af-9 and retained in all Hrx-Enl and Hrx-Af9 fusion proteins; Enl is a nuclear protein that activates transcription from synthetic reporters in lymphoid cells, myeloid cells and yeast.
ENL, the gene fused with HRX in t(11;19) leukemias, encodes a nuclear protein with transcriptional activation potential in lymphoid and myeloid cells. · Rubnitz JE, Morrissey J, Savage PA, Cleary ML · Blood 1994What this evidence cannot show
Read as abstract only. PubMed returns no DOI for this 1994 Blood record and Europe PMC returns no PMCID, so the deletion series itself was not inspected. 'C-terminal 90 amino acids' converts to 470-559 only if the construct numbering matches the 559-aa UniProt canonical, which the abstract does not state. A reporter assay measures the transactivation potential of an isolated fragment, not the behaviour of that fragment inside the chimera on chromatin.
Evidence that argues against this (1)
- A later study found that MLL-ENL transformation requires the ENL transactivation activity AND the HRX DNA-binding motifs together, so neither element is sufficient and the C-terminal 90 residues are necessary rather than the whole explanation. Slany RK et al. 1998
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Read in full text (PMC10169636): MLLT1 and SEPTIN6 are the fusion partners with a preferential KMT2A intron 11 break while other partners tend to intron 9 or upstream; breakpoints upstream of exon 11 retain PHD domain I structurally intact for the reciprocal fusion protein while breaks within or downstream of exon 11 seem to result in a different folding of PHD domain I, impairing CYP33 binding and PHD1 homodimerisation; and in spliced fusions the most 3' exon of the 5'-KMT2A usually splices to exon 2 of the partner gene because that is the first exon with a bona fide splice acceptor site.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, et al. · Leukemia 2023What this evidence cannot show
Read in full text via Europe PMC fullTextXML (body element present, 98,196 characters of flattened text). Single diagnostic-centre referral series of 3,401 patients, 2003-2022; partner denominators reflect referral patterns. The PHD1-folding consequence is the authors' summary of two prior studies, not a result measured in this paper.
Evidence that argues against this (1)
- The same paper states plainly that whether these breakpoint-position findings have an impact on clinical outcome is yet unclear, so the structural consequence has no demonstrated clinical correlate. Meyer C et al. 2023
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
MLL/ENL fusion transcripts use ENL breakpoints at exons 2, 4, 6 or 7, all in frame, and two of these were novel at the time of report.
Analysis of acute leukemias with MLL/ENL fusion transcripts: identification of two novel breakpoints in ENL. · Fu JF, Liang DC, Shih LY · Am J Clin Pathol 2007What this evidence cannot show
Identifiers and author list copied from the PubMed esummary record fetched 2026-09-19; abstract-level reading only. n = 15 MLL/ENL cases, so per-junction frequencies in an unselected population are not established by this series.
Evidence that argues against this (1)
- A therapy-related AML line was reported with an MLL-ENL transcript its authors call a novel isoform, showing the junction catalogue is not closed. Ninomiya M et al. 2006
residue-level domain map of the chimera named by this fusion entry
- part of →KMT2A / MLL1 — the gene that breaks
5' contribution: MBM, IBM1/2, AT-hooks, CXXC (UniProt Q03164)
3' contribution: YEATS 1-138, disordered 147-484, AHD (UniProt Q03111)
menin-binding motif at KMT2A 6-25 is retained in every KMT2A::MLLT1 architecture
LEDGF integrase-binding motifs IBM1 123-134 and IBM2 147-152 retained
supplies the residue coordinates from which the reciprocal product's contents are derived
junction heterogeneity of the chimera
residue-level assignment of the AHD in the architecture entry's 3' contribution
same 5-prime half, opposite half of the paralogous partner