MLLT6 / AF17 — domain architecture, and where the leucine zipper actually is
also called AF17 domain map · AF17 · MLLT6
A map of the AF17 protein showing which parts survive into the leukaemia fusion and which are cut away. The self-sticking coil that survives sits at residues 719 to 753; the three zinc-finger reading modules that are lost sit at the very front of the protein.
The same thing, for a scientist
Coordinate-level domain map for MLLT6/AF17 (UniProt P55198, AF17_HUMAN, 1093 aa), written because `gene-mllt6` carries the DotCom and dependency content but no `data.domains` array, so the 3D domain viewer could not draw the MLLT6 lane of KMT2A::MLLT6 and `check-joins` reported it as a partner with no coordinates. UniProt annotates a **PHD-type zinc finger at aa 5-57**, a **C2HC pre-PHD-type zinc finger at aa 62-95** and a **second PHD-type zinc finger at aa 118-181** — the cysteine-rich, Br140/peregrin-homologous region described in the founding cloning paper — plus a **coiled coil at aa 719-753** and three disordered stretches (aa 185-500, 775-871, 1060-1093) [E1]. The coiled coil at 719-753 is, on the arithmetic of position, the leucine-zipper dimerisation motif that Prasad et al. reported as lying **3' of the fusion point** while the N-terminal cysteine-rich domain lies 5' of it [E2]. This entry asserts that correspondence as an inference and labels it as one: the 1994 paper gives no residue range, so the identification rests on there being exactly one annotated coiled coil in the protein and on it being C-terminal to a breakpoint that removes the zinc fingers. `gene-mllt6` is not duplicated here — this entry carries coordinates and nothing else.
This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.
I have never read one of these pictures before
A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.
Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.
Anything in yellow sticks is a bound small molecule, which is usually the drug.
The map, and which side of the breakpoint each element falls
| element | residues | fate in KMT2A::MLLT6 |
|---|---|---|
| PHD-type zinc finger 1 | 5-57 | lost (5' of the fusion point) |
| C2HC pre-PHD-type zinc finger | 62-95 | lost |
| PHD-type zinc finger 2 | 118-181 | lost |
| disordered | 185-500 | lost |
| coiled coil (leucine-zipper dimerisation motif) | 719-753 | retained (3' of the fusion point) |
| disordered | 775-871 | retained |
| disordered | 1060-1093 | retained |
All coordinates are from UniProt P55198 1. The retained/lost assignment follows the founding paper's explicit statement that the leucine-zipper dimerisation motif is located 3' of the fusion point and the cysteine-rich N-terminal domain 5' of it 2.
Two things this map makes visible that the prose does not. First, the fusion discards all three of AF17's zinc fingers — the PHD and pre-PHD modules that are the protein's chromatin-reading apparatus — and keeps a 35-residue coil plus a great deal of disorder. Second, the retained portion is roughly the C-terminal 400 residues of a 1093-residue protein, so the KMT2A::MLLT6 oncoprotein carries far less of its partner than KMT2A::MLLT11 (complete 90-residue ORF) or KMT2A::EPS15 (almost all of 896 residues) do of theirs.
The inference, stated so it can be challenged
Prasad et al. report "a leucine-zipper dimerization motif located 3' of the fusion point" and give no residue numbers 2. UniProt annotates exactly one coiled coil in P55198, at aa 719-753 1. The identification of the two is an inference from three facts: there is only one annotated coiled coil; it is C-terminal to the region the same paper places 5' of the breakpoint; and a leucine zipper is a coiled coil.
It could be wrong in two ways. A leucine zipper that UniProt does not annotate as a coiled coil would not appear in the feature list at all, and 1990s motif calls and 2020s structural annotations do not always agree. And no breakpoint residue is published for this fusion in any source read here, so "3' of the fusion point" bounds the zipper's position only relative to an unlocated junction.
Nobody has measured whether this coil dimerises. It has been a sequence prediction since 1994, and the graph's treatment of MLLT6 as a bridge case between the forced-oligomerisation and elongation-hijack models rests on it. That is recorded in fusion-kmt2a-mllt6 and is repeated here because a domain map is exactly the kind of artefact that makes an unverified motif look settled.
| uniprot | P55198 |
|---|---|
| uniprot verified | P55198 = AF17_HUMAN, PRIMARY gene name MLLT6, synonym AF17, 1093 aa, Homo sapiens. Fetched from the UniProt REST API on 2026-09-19 and the primary gene name and length read back from the record; the length agrees independently with the 1093 residues reported in PMID 8058765. |
| gene symbol | MLLT6 |
| length AA | 1093 |
| alphafold ID | AF-P55198-F1 |
| PDB note | No structure of MLLT6 or of any of its domains exists, and none is cross-referenced from UniProt P55198. |
| domain coordinate caveat | All four coordinates are UniProt sequence annotations; there is no structure of any part of this protein. The identification of the UniProt coiled coil at aa 719-753 with the 'leucine-zipper dimerization motif' of the 1994 cloning paper is an INFERENCE — that paper gives no residue range — resting on there being exactly one annotated coiled coil and on it being C-terminal to the zinc fingers the same paper places 5' of the fusion point. Disordered regions (aa 185-500, 775-871, 1060-1093) are deliberately omitted from the domains array: they are real annotations but not domains, and including them would make the viewer draw the majority of the protein as a feature. |
| relationship to other entries | gene-mllt6 (interactome3) holds the DotCom membership, the dependency data and the interaction-evidence class for this protein and is NOT duplicated here. This entry exists solely to supply coordinates, in the same pattern as protein-eps15-domains, because check-joins reported MLLT6 as a fusion partner with no domain coordinates once fusion-kmt2a-mllt6 was written. |
PDB IDS
[]
domains — the numbers behind the chart
[
{
"name": "PHD-type zinc finger 1",
"start": 5,
"end": 57
},
{
"name": "C2HC pre-PHD-type zinc finger",
"start": 62,
"end": 95
},
{
"name": "PHD-type zinc finger 2",
"start": 118,
"end": 181
},
{
"name": "Leucine-zipper dimerisation motif (coiled coil)",
"start": 719,
"end": 753
}
]fate in fusion
{
"retained": [
"Leucine-zipper dimerisation motif (coiled coil), aa 719-753"
],
"lost": [
"PHD-type zinc finger 1, aa 5-57",
"C2HC pre-PHD-type zinc finger, aa 62-95",
"PHD-type zinc finger 2, aa 118-181"
],
"basis": "the founding paper's statement that the leucine zipper lies 3' of the fusion point and the cysteine-rich N-terminal domain 5' of it; no breakpoint residue is published"
}- A coordinate table makes a 1994 motif prediction look like a measured fact. The leucine zipper in this map has never been shown to dimerise, and the graph's classification of MLLT6 as a bridge between two mechanistic models rests entirely on it.
- The fusion discards all three zinc fingers, which are AF17's chromatin-reading modules. Any argument that KMT2A::MLLT6 works through DotCom has to explain how a protein stripped of its PHD fingers still joins or functions within that complex — and DotCom composition was determined on the intact wild-type protein.
- Where exactly is the KMT2A::MLLT6 breakpoint in MLLT6 residues? Every domain-fate statement for this fusion is currently relative ('3' of the fusion point') rather than absolute, and no source read here gives a residue or a partner exon.
- Does the coiled coil at aa 719-753 dimerise? It has been a prediction since 1994 and the graph's bridge-case argument for MLLT6 depends on it.
- Are the 1994 'three zinc fingers' the same objects as UniProt's two PHD fingers plus a C2HC pre-PHD module, or did the 1994 call include something the current annotation splits differently? The two descriptions do not map one to one.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Human AF17 is UniProt P55198 (AF17_HUMAN), primary gene name MLLT6, synonym AF17, 1093 aa, Homo sapiens, with a PHD-type zinc finger at aa 5-57, a C2HC pre-PHD-type zinc finger at aa 62-95, a second PHD-type zinc finger at aa 118-181, a coiled coil at aa 719-753, and disordered regions at aa 185-500, 775-871 and 1060-1093; no PDB structure is cross-referenced.
UniProtKB P55198 (AF17_HUMAN) Protein AF-17, Homo sapiens, 1093 aa · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Canonical isoform. Resolved by fetching the accession and reading back the PRIMARY gene name (MLLT6) and length (1093 aa); the length agrees independently with the 1093 residues reported in the 1994 cloning paper, which is the cross-check the project's accession rule asks for. There is no structure of any part of this protein, so every coordinate is a sequence-based annotation.
Evidence that argues against this (1)
- The founding cloning paper describes the N-terminal cysteine-rich region as a domain that 'can be arranged in three zinc fingers' with homology to Br140/peregrin, which UniProt resolves into two PHD-type fingers plus a C2HC pre-PHD module at slightly different boundaries — so the 1994 description and the current annotation are not a one-to-one mapping, and a figure drawn from either alone will disagree with the other. Prasad R et al. 1994
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The AF17 gene encodes a protein of 1093 amino acids containing a leucine-zipper dimerisation motif located 3' of the fusion point and a cysteine-rich domain at the N terminus that can be arranged in three zinc fingers with homology to a domain within Br140 (peregrin); AF17 also contains stretches of amino acids previously associated with transcriptional repression or activation.
Leucine-zipper dimerization motif encoded by the AF17 gene fused to ALL-1 (MLL) in acute leukemia. · Prasad R, Leshkowitz D, Gu Y, Alder H, Nakamura T, Saito H, et al. · Proc Natl Acad Sci U S A 1994What this evidence cannot show
A 1994 cloning and sequence-analysis paper read in abstract. NO residue range is given for the leucine zipper and no breakpoint residue is given for the fusion, so 'located 3' of the fusion point' is a relative statement. The zipper is a motif PREDICTION; no dimerisation was measured then and none has been found since in the searches for this beat. The identification of this motif with the UniProt coiled coil at aa 719-753 is an inference made by this entry and labelled as such.
Evidence that argues against this (1)
- The mechanistic model this same paper proposed — that ALL-1 rearrangements cause LOSS of function, with the partner acting by repressing the truncated ALL-1 protein — is not the model the field holds, so the paper's framing of what the zipper is for should not be inherited uncritically along with its structural observations. Martin ME et al. 2003
- related to →MLLT6 (AF17) — the fourth KMT2A fusion partner inside DotCom, and the only one no cell needs
coordinate map for the same protein; gene-mllt6 holds the DotCom and dependency content and is not duplicated
leucine zipper (aa 719-753) retained; all three zinc fingers (aa 5-57, 62-95, 118-181) lost