DUX4 (double homeobox 4) domain architecture
also called DUX4 · double homeobox 4
DUX4 is the 5' partner of DUX4::IGH B-ALL. This entry lists where each of its functional parts sits along the protein, measured as residue numbers, so the fusion diagrams can draw which parts survive the break and which are thrown away.
The same thing, for a scientist
Reference domain architecture for DUX4 (UniProt Q9UBX2, 424 aa). A double-homeobox transcription factor: two DNA-binding homeoboxes at 19-78 and 94-153, and a C-terminal 327-424 region UniProt annotates as required for EP300/CREBBP interaction and for transactivation. Created so the 3D fusion viewer can resolve fusion-dux4-igh, which previously had no coordinates for this partner and silently dropped the lane. Domain fate in the fusion is asserted by the fusion entry, not here.
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.
Why this entry exists
The join checker reported fusion-dux4-igh: no protein entry for DUX4. Without coordinates the comparative 3D viewer cannot draw the lane at all, and it was omitting it without warning rather than flagging it. This is a reference-coordinate entry: it asserts the architecture and nothing about the fusion.
Folded domains versus functional regions. UniProt mixes two very different annotation kinds. Domain, Zinc finger and DNA binding are fold calls with real boundaries. Region is a functional assertion that often spans disorder and can overlap other regions or run for hundreds of residues. Both are recorded with boundary_precision marking which is which, because drawing a soft functional region as a crisp box is exactly the kind of false precision this graph exists to avoid.
Architecture
| Part | Residues | Kind | Precision |
|---|---|---|---|
| Homeobox 1 | 19-78 | DNA binding | firm |
| Homeobox 2 | 94-153 | DNA binding | firm |
| Required for interaction with EP300 and CREBBP, and for transcriptional activation of target genes | 327-424 | Region | soft |
| Important for transcriptional activation of target genes | 405-424 | Region | soft |
11 PDB entries reference this protein.
| uniprot | Q9UBX2 |
|---|---|
| gene symbol | DUX4 |
| length AA | 424 |
| protein name | Double homeobox protein 4 |
| alphafold ID | AF-Q9UBX2-F1 |
| n PDB total | 11 |
| uniprot verified | true |
| accession resolution | exact gene-name search, reviewed human only, gene symbol confirmed in the response |
| created for | fusion-dux4-igh |
domains — the numbers behind the chart
[
{
"name": "Homeobox 1",
"start": 19,
"end": 78,
"feature_type": "DNA binding",
"note": "UniProt folded-domain call.",
"boundary_precision": "firm",
"aliases": [
"double homeodomain",
"DUX4 double homeodomain"
],
"alias_basis": "The two tandem homeoboxes are together called the double homeodomain, which is what the gene name \"double homeobox 4\" refers to. Literature naming the double homeodomain means both boxes, not one."
},
{
"name": "Homeobox 2",
"start": 94,
"end": 153,
"feature_type": "DNA binding",
"note": "UniProt folded-domain call.",
"boundary_precision": "firm",
"aliases": [
"double homeodomain",
"DUX4 double homeodomain"
],
"alias_basis": "Second of the two tandem homeoboxes that together form the double homeodomain."
},
{
"name": "Required for interaction with EP300 and CREBBP, and for transcriptional activation of target genes",
"start": 327,
"end": 424,
"feature_type": "Region",
"note": "UniProt functional region, not a fold boundary. Soft edges; render as low confidence.",
"boundary_precision": "soft"
},
{
"name": "Important for transcriptional activation of target genes",
"start": 405,
"end": 424,
"feature_type": "Region",
"note": "UniProt functional region, not a fold boundary. Soft edges; render as low confidence.",
"boundary_precision": "soft"
}
]PDB IDS
[ "5Z2S", "5Z2T", "5Z6Z", "5ZFW", "5ZFY", "5ZFZ", "6A8R", "6DFY", "6E8C", "6U81", "6U82" ]
- Nothing here licenses targeting any part of DUX4. These are coordinates, not a druggability claim; a domain with crisp boundaries is not thereby a ligandable pocket.
- Drawing UniProt `Region` features as boxes in the viewer risks implying structural knowledge that does not exist. They are marked soft and should be rendered differently from fold calls.
- Where does the fusion-dux4-igh breakpoint fall relative to these boundaries, and is the junction codon actually mapped or only inferred from cDNA?
- Which of these parts are retained in the fusion, and is retention ever being confused with requirement for transformation?
- Do InterPro and the deposited crystal constructs agree with these UniProt boundaries? Where they do not, the disagreement should be recorded rather than silently resolved in UniProt's favour.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
Human DUX4 (Double homeobox protein 4) is 424 residues (UniProt Q9UBX2) with the domain and region boundaries tabulated in this entry, and is referenced by 11 PDB entries.
UniProtKB Q9UBX2 (DUX4_HUMAN): sequence length, domain, zinc-finger, DNA-binding, repeat, motif and region features, and PDB cross-references · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Accession resolved by exact gene-name search restricted to reviewed human entries and confirmed against the gene symbol returned, because a previous lookup in this project silently resolved to the wrong gene. Coordinates are for the canonical isoform only; several of these proteins have isoforms that shift numbering. UniProt `Region` features are curator functional calls, not folds.
Evidence that argues against this (1)
- Domain boundaries are not a settled fact and differ between sources. InterPro, Pfam, SMART and individual crystal-structure constructs routinely disagree with UniProt by tens of residues, and published figures often use construct boundaries rather than database ones. Any downstream claim that a breakpoint falls just inside or just outside a domain is therefore only as firm as the boundary definition it assumes, which for every `Region` feature here is soft. InterPro Consortium 2026
- ← related toMechanism class 5 — enhancer hijacking (inv(3) GATA2/EVI1 — and DUX4-r, which only half belongs here)
Cell-type assumption in the truncation argument (Shadle 2017, PMID 28273136): DUX4 apoptosis runs through MYC-mediated death and the dsRNA innate immune response, measured in rhabdomyosarcoma cells and myoblasts, never in a B progenitor.
Counter-evidence: Mitsuhashi 2018 (PMID 29618456) maps DUX4 toxicity and transactivation to the same C-terminal 80 residues, so C-terminal truncation disables rather than selectively detoxifies DUX4
- ← analogous toGLIS2 (GLI-similar 2) domain architecture
both are DNA-binding transcription factors contributed as the oncogenic moiety of a paediatric fusion
reference domain coordinates for the DUX4 side of this fusion