PICALM / CALM - domain architecture, residue coordinates, and which parts the fusion keeps
also called CALM domain map · PICALM domain map
This is the parts list for the PICALM protein, with the numbers of the amino acids each part occupies. It matters because the leukaemia fusion keeps almost all of PICALM and throws away only the last four amino acids - so the interesting question is which of the kept parts actually does the damage.
The same thing, for a scientist
PICALM (UniProt Q13492, 652 aa) is a two-part protein: a folded N-terminal **ANTH** membrane/cargo-binding module and a long, largely unstructured C-terminal tail carrying short linear motifs. Coordinates here are in **Q13492-1 (652 aa) numbering**, which is also the frame Conway et al. state they use (NCBI NP_009097.2) [E4]. The commonest CALM::MLLT10 chimera retains **CALM 1-648** and discards only the last four residues [E4] - so essentially the whole protein is retained and a domain map cannot, by itself, tell you what matters. Deletion mapping does: **CALM 400-648** is sufficient for transformation [E5], and within it the decisive element is the **CRM1-recognised nuclear export signal at 544-553** [E4]. **Two annotation conflicts are recorded below and neither should be silently resolved**: UniProt calls the N-terminal module 'ENTH 14-145' while Pfam/InterPro and the entire literature call it **ANTH 23-283** [E1][E2]; and there are **no human PICALM structures at all** - every PDB deposit is rat [E3].
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.
Reading the coordinates safely
Three numbering frames are in circulation for this fusion and mixing them silently produces confident wrong residues.
| entity | frame | length |
|---|---|---|
| CALM/PICALM | UniProt Q13492-1 = NCBI NP_009097.2 | 652 aa |
| CALM/PICALM (short isoform) | lacks 49 aa, ~66 kDa | 603 aa |
| AF10/MLLT10 | UniProt P55197-1 | 1,068 aa |
| AF10/MLLT10 | the frame used by the 2026 review (AF10 'contains 1027 amino acid residues') | 1,027 aa |
| AF10/MLLT10 | NCBI NP_001182555, as used by Conway et al. ('exon 11 = aa 234') | third frame |
This entry asserts CALM coordinates only. AF10 coordinates live in protein-mllt10-domains and are in P55197-1 numbering; any AF10 residue number read out of a CALM-AF10 paper must have its frame checked first. See breakpoint-picalm-mllt10-transcript-variants.
What deletion mapping says, as against what the domain map says
A domain map is nearly uninformative here because the fusion keeps almost everything. The informative experiments are truncations:
- CALM 400-648 + AF10 OM-LZ is sufficient to induce AML in mice - the 'minimal fusion' 5.
- Within that, CALM 540-557, or a heterologous NES from ABL1, Rev, PKIA or APC, confers transforming potential on AF10 on its own 4. So the sufficient element is a ten-residue export signal.
- The last 55 residues (593-648) were reported to mediate dimerisation, an alternative account of the C-terminal requirement that Conway et al. explicitly weigh against their NES model 4.
That last point is a real, unresolved alternative and is recorded as counter-evidence rather than smoothed away.
| uniprot | Q13492 |
|---|---|
| length AA | 652 |
| gene symbol | PICALM |
| alphafold ID | AF-Q13492-F1 |
| PDB species note | All eight deposits are RATTUS NORVEGICUS CALM (UniProt O55012). `pdb_ids` is deliberately empty because the graph's structure viewer should not be handed a rat structure as though it were the human protein. |
| isoform note | Coordinates are in UniProt Q13492-1 (652 aa) numbering, which equals NCBI NP_009097.2 - the frame Conway et al. state they use. A shorter ubiquitously expressed isoform lacking 49 aa also exists. |
| assumed fusion for fate | fusion-picalm-mllt10 (PICALM::MLLT10, t(10;11)(p13;q14)), PICALM as the 5' partner |
| domain coordinate caveat | PICALM is the OPPOSITE case from MLLT10: MLLT10's informative modules are crystallographically defined and sit on either side of the break, whereas PICALM is retained almost whole, has no human structure at all, and its only load-bearing element is a ten-residue linear motif. A domain map of PICALM therefore explains very little on its own and must be read alongside the deletion series. |
PDB IDS
[]
PDB IDS non human
[ "3ZYK", "3ZYL", "3ZYM", "7JXV", "1HF8", "1HFA", "1HG2", "1HG5" ]
domains — the numbers behind the chart
[
{
"name": "ANTH (AP180 N-terminal homology domain)",
"start": 23,
"end": 283,
"pfam": "PF07651",
"note": "Binds PtdIns(4,5)P2 and the SNARE motif of VAMP-family R-SNAREs; senses and drives membrane curvature. UniProt instead calls an ENTH domain at 14-145 over the same N-terminus - see conflicting_annotations. Retained in the fusion, and no experiment in this beat's reading assigns it a role in transformation.",
"fate": "retained",
"evidence_basis": "curated",
"boundary_precision": "Pfam/InterPro model boundaries; the rat crystallised construct is residues 1-289 in rat numbering"
},
{
"name": "PIMREG-interaction region",
"start": 221,
"end": 294,
"note": "UniProt functional region; PIMREG binding is proposed to target PICALM to the nucleus in some cells. Soft-edged functional call, not a folded domain.",
"fate": "retained",
"evidence_basis": "curated",
"boundary_precision": "functional region, soft edges"
},
{
"name": "Clathrin-binding / transformation-sufficient C-terminal region",
"start": 400,
"end": 648,
"note": "The C-terminal 248 residues. Sufficient, when fused to the AF10 OM-LZ, to induce AML in mice. Largely unstructured; contains the clathrin- and AP2-binding short linear motifs and the NES.",
"fate": "retained",
"evidence_basis": "functional",
"boundary_precision": "deletion-mapping boundary, +/- tens of residues"
},
{
"name": "NES (CRM1-dependent nuclear export signal)",
"start": 544,
"end": 553,
"note": "THE decisive element from the 5' partner. Necessary for CALM-AF10 transformation; CALM 540-557, or a heterologous NES, is sufficient to confer transforming potential on AF10.",
"fate": "retained",
"evidence_basis": "functional",
"boundary_precision": "+/- 4 aa; the sufficient construct is 540-557"
},
{
"name": "C-terminal dimerisation region",
"start": 593,
"end": 648,
"note": "Reported to facilitate dimerisation of CALM-AF10. A rival explanation for why the C-terminus is required; drawn here so the alternative is visible rather than buried in prose.",
"fate": "retained",
"evidence_basis": "functional",
"boundary_precision": "reported span, not independently re-derived here"
}
]conflicting annotations
[
{
"feature": "N-terminal module identity and extent",
"claim_a": "UniProt Q13492: Domain 'ENTH', residues 14-145.",
"claim_b": "Pfam PF07651 / InterPro IPR011417: 'ANTH domain', residues 23-283; the literature and every PDB construct call it ANTH.",
"our_reading": "Claim B. We draw ANTH 23-283. The UniProt ENTH call covers roughly the first half of the module; ENTH and ANTH are related folds and UniProt appears to have kept the narrower ENTH model. Drawing 14-145 would misrepresent the module by about 140 residues."
},
{
"feature": "Why the CALM C-terminus is required for transformation",
"claim_a": "It carries the clathrin-binding domain; disrupted endocytosis is the mechanism (Deshpande et al.).",
"claim_b": "It carries a CRM1-dependent NES; nuclear export is the mechanism, and a heterologous NES substitutes (Conway et al.).",
"claim_c": "It mediates dimerisation (cited and argued against within Conway et al.).",
"our_reading": "Claim B has the strongest direct test, because heterologous NES substitution swaps the proposed function while changing nothing else. It does not exclude A or C, and the three proposals are all about the same stretch of protein."
}
]- The only load-bearing element from PICALM is a generic leucine-rich NES that many proteins carry, so a CALM-AF10-selective agent cannot be built against it - blocking it means blocking CRM1, which is a whole-cell export pathway.
- No human structure exists, so structure-guided work on PICALM starts on a rat template and inherits whatever the species difference is.
- Does the retained ANTH domain contribute anything to transformation? Nothing in this beat's reading tests an ANTH-deleted CALM-AF10, even though the minimal fusion (400-648) shows the ANTH domain is dispensable for transformation - dispensable is not the same as inert.
- Do the NES, dimerisation and clathrin-binding proposals for the CALM C-terminus describe one event or three? A construct carrying a heterologous NES but no CALM sequence at all would separate them.
- What are the coordinates of the CALM clathrin boxes, DPF and NPF motifs in Q13492 numbering? No coordinate-level annotation was found in UniProt on 2026-09-20 and they are therefore omitted here rather than guessed.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt Q13492 annotates a single domain, 'ENTH', at residues 14-145, plus a region 221-294 for interaction with PIMREG and a disordered region 559-580; it annotates no clathrin box, DPF or NPF motif with coordinates.
UniProtKB Q13492 (PICAL_HUMAN) - feature table (Domain ENTH 14-145; Region 221-294 interaction with PIMREG) · UniProt Consortium 2026What this evidence cannot show
Fetched live 2026-09-20. The absence of coordinate-level clathrin-box/DPF/NPF annotation is why this entry does not draw those motifs: per the brief, a domain with a null boundary is invisible to every figure, and a guessed boundary is worse than an omitted one.
Evidence that argues against this (1)
- SOURCE LIMITATION WITH A CONSEQUENCE, not a bare curation gap. The absence of coordinate-level clathrin-box, DPF and NPF annotation is a curation state, and it hides the region the mouse genetics turns on: the C-terminal portion of CALM required for optimal clathrin binding determines which disease the fusion causes - excluded, a myeloproliferative disease; included, acute myeloid leukaemia with changes in Pim1 and Crebbp. So the one part of PICALM with a demonstrated, boundary-dependent phenotype in this leukaemia has no coordinate in the feature table this entry is built from. Declining to draw a motif without a boundary is the right call; concluding from the drawing that the annotated ENTH region is where the biology is would be the wrong one. Stoddart A et al. 2012
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Pfam PF07651 / InterPro IPR011417 place the ANTH domain of Q13492 at residues 23-283; InterPro additionally calls an ENTH domain (IPR013809) at 14-145 and an ANTH-domain superfamily match at 147-300.
InterPro/Pfam matches on UniProt Q13492: ANTH domain IPR011417 / PF07651 at residues 23-283 · InterPro / EMBL-EBI 2026What this evidence cannot show
Fetched live 2026-09-20. ENTH and ANTH are related folds and InterPro carries both models over this protein; the literature and the PDB constructs use ANTH, and the functional module is the longer one. Drawing only the 14-145 ENTH call would cut the module roughly in half.
Evidence that argues against this (1)
- BOTH CALLS ARE PREDICTIONS, and no experimentally determined human boundary exists to arbitrate. IPR011417/PF07651 at 23-283 and IPR013809 at 14-145 are model matches on the same sequence, and the experimental support behind the longer module is rat: the curvature-sensing work that establishes what the ANTH domain does was largely on isolated ANTH constructs and reconstituted membranes, with rat structures underneath. Choosing 23-283 because 'the literature and the PDB constructs use ANTH' is choosing the model that agrees with a non-human structure, which is a defensible call and is not a measurement. The entry should say that no human-numbered, experimentally determined ANTH boundary exists. Miller SE et al. 2015
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
There are no human PICALM structures in the PDB; 3ZYK, 3ZYL, 3ZYM, 7JXV, 1HF8, 1HFA, 1HG2 and 1HG5 are all rat CALM (UniProt O55012), and 3ZYM is a single-chain chimera of mouse VAMP8 11-41 fused to rat CALM 1-264.
RCSB PDB entry and polymer-entity records for 3ZYK, 3ZYL, 3ZYM, 7JXV, 1HF8, 1HFA, 1HG2, 1HG5 · RCSB PDB 2026What this evidence cannot show
Queried live 2026-09-20. The accession-scoped RCSB query on Q13492 returned HTTP 204 with an empty body - the real answer, not a failure. The PDB construct annotations state 'ANTH DOMAIN, RESIDUES 1-289' (3ZYK) and 'RESIDUES 1-264' (3ZYM) in RAT numbering, which is not guaranteed to align to human numbering residue-for-residue; do not read a human span off these deposits.
Evidence that argues against this (1)
- THE MISSING STRUCTURE IS NOT THE ONE THIS ITEM NAMES. 'No human PICALM structures in the PDB' understates what is known, because the rat deposits carry the field's mechanistic model and the curvature work rests on them. What is genuinely absent is a structure of anything the FUSION contains: the rat constructs are annotated as ANTH, residues 1-289 and 1-264, i.e. the N-terminal module, while the transforming part of CALM in the fusion is the C-terminal 248 residues, roughly aa 400-648. That region is disordered and motif-bearing and has no structure in any organism. So the structural gap relevant to this graph is not species, it is which half of the protein. Deshpande AJ et al. 2011
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In a murine bone-marrow retroviral system, the CALM-AF10 chimera used fuses all but the last four residues of CALM (numbering referenced to the longest CALM isoform, NP_009097.2) to AF10, and CALM residues 544-553 constitute a CRM1-dependent NES that is necessary, with CALM 540-557 sufficient, for transformation.
A CALM-derived nuclear export signal is essential for CALM-AF10-mediated leukemogenesis. · Conway AE, et al. · Blood 2013What this evidence cannot show
Murine bone marrow and HEK293 cells. The paper states its numbering frame explicitly, which is why it is usable; most CALM-AF10 papers do not.
Evidence that argues against this (1)
- The same paper records a competing explanation for the C-terminal requirement: the last 55 residues of CALM within CALM-AF10 (aa 593-648) were reported to facilitate dimerisation, which would make the C-terminal requirement an oligomerisation requirement rather than an export requirement. Conway et al. argue against it but do not exclude it. Conway AE et al. 2013
- E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The C-terminal 248 residues of CALM (which include the clathrin-binding region, i.e. approximately aa 400-648) fused to the AF10 OM-LZ domain generate a minimal fusion that induces AML in mice and recapitulates Hoxa cluster upregulation.
The clathrin-binding domain of CALM and the OM-LZ domain of AF10 are sufficient to induce acute myeloid leukemia in mice. · Deshpande AJ, et al. · Leukemia 2011What this evidence cannot show
Murine bone-marrow transplantation. The authors attribute the effect to the clathrin-binding domain and to disrupted endocytosis; Conway et al. [E4] reattribute it to the NES that lies inside the same span. Both claims are about the same piece of protein and the experiments do not separate them cleanly.
Evidence that argues against this (1)
- PUBLISHED CONFLICTING RESULT that rules against both attributions this item records. The dispute in the caveat is clathrin-binding versus NES; a third experiment on the same span finds against the first and adds a third candidate. Excluding the C-terminal portion of CALM required for optimal clathrin binding changed the mouse disease from AML to a myeloproliferative disease, yet the leukaemia COULD NOT BE ATTRIBUTED DIRECTLY to interference with endocytosis, full-length CALM-AF10 in leukaemia cells showed no consistent effect on growth-factor endocytosis, and the fusion's measured propensity to HOMO-OLIGOMERISE is proposed as what the region supplies. Three candidate functions, one span, no experiment that separates them. Stoddart A et al. 2012
domain map of UniProt Q13492 (652 aa)
- ← related toPICALM::MLLT10 (CALM-AF10), t(10;11)(p13;q14) - the DOT1L-dependent HOXA fusion that contains no KMT2A
CALM domain coordinates in Q13492-1 numbering
- ← related toWhere the break falls, and the three incompatible numbering frames that make it easy to get wrong
CALM coordinates asserted in Q13492-1
the structural gap is which HALF, not which species: the rat deposits cover ANTH (aa 1-289/1-264) while the transforming part of the fusion is the C-terminal 248 residues