MLLT10 / AF10 — domain architecture and residue coordinates
also called AF10 · MLLT10 · P55197
AF10 is the KMT2A partner that dominates infant myeloid leukemia, and it works by handing the fusion a grip on an enzyme called DOT1L. This entry gives the residue numbers for that grip, which is one of the few partner interfaces in this whole project with a real crystal structure behind it.
The same thing, for a scientist
MLLT10 (UniProt P55197, 1,068 aa) contributes to KMT2A fusions a C-terminal octapeptide-motif leucine-zipper (OM-LZ) that binds the DOT1L coiled-coil domain. Unusually for a fusion partner, both ends of the protein are structurally characterised: the N-terminal PHD-Zn-knuckle-PHD (PZP) reader module has crystal structures spanning residues 1-208 (PDB 5DAG, 5DAH, 7MJU), and the OM-LZ has crystal structures in complex with DOT1L spanning 704-779 (PDB 6CKN, 6CKO) and 699-782 (6JN2). The fusion retains the OM-LZ and discards the PZP, so the two structurally solved modules land on opposite sides of the break.
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 break separates two solved modules2 deeper
AF10 is the best-resolved partner in this beat, and the interesting thing about its coordinates is where the break falls relative to them.
Lost (N-terminal): the PZP module. UniProt calls three separate zinc-finger features — PHD-type 1 at 22-74, a C2HC pre-PHD Zn knuckle at 79-112, and PHD-type 2 at 135-198 1 — but the crystallography shows these fold as a single module, not three beads on a string. Chen and colleagues crystallised the PZP and showed it recognises histone H3 residues 22-27 and that the interaction is abrogated by H3K27 modification 2. The crystallised constructs run 1-208 3.
Retained (C-terminal): the OM-LZ. UniProt annotates 703-784 as "transactivation domain; required for DOT1L-binding" with a leucine zipper at 750-778 1; InterPro's AF10-specific coiled-coil model (IPR049773) gives 715-778; and the DOT1L co-crystals use AF10 704-779 4 or 699-782 5.
That is a rare piece of luck for the viewer. Both sides of the AF10 lane can be drawn from experimental coordinates, and the break sits in the ~500 residues of predicted disorder between them.
Boundary spread on the OM-LZ, and why the small spread matters
Four sources give four slightly different OM-LZ extents:
- UniProt functional region: 703-784 (82 aa)
- InterPro IPR049773 coiled coil: 715-778 (64 aa)
- PDB 6CKN / 6CKO construct: 704-779 (76 aa)
- PDB 6JN2 construct: 699-782 (84 aa)
The three crystallography-derived ranges cluster tightly around 700-780. The InterPro coiled-coil model is narrower because it calls only the heptad-repeat core and excludes the octapeptide motif that gives the module half its name.
We draw 704-779 because it comes from the structure of the actual complex with DOT1L, and record the others. The practical consequence of picking the wrong one is small in a figure but large in a construct: a peptide truncated to the InterPro coiled coil (715-778) would omit the octapeptide motif, and Zhang and colleagues showed that disrupting the DOT1L-AF10 interface abrogates MLL-AF10 transformation 4, so getting the competing peptide's boundaries right is the difference between a tool compound and a negative result.
The lost module is not a bystander: losing the PZP is what permits transformation
The clean version of the story is: the fusion keeps the DOT1L grip and throws away the chromatin reader, so the reader does not matter. That reading is wrong, and the correction is the most important thing in this entry.
Klein and colleagues showed that the AF10 PZP domain — consistently impaired or deleted in leukemogenic AF10 translocations — actively blocks malignant transformation 6. Putting a functional PZP back into the CALM-AF10 fusion prevented transformation of bone-marrow HSPCs in vitro and in vivo, abrogated CALM-AF10-mediated leukemogenesis in vivo, downregulated Hoxa genes and reversed cellular transformation. Their crystallographic and biochemical work shows PZP engaging the nucleosome core particle through multivalent contacts with the H3 tail and DNA, discriminating against H3K27me3, and it also promotes nuclear localisation of CALM-AF10.
This inverts what the fate: "lost" label implies. A lost domain in this graph normally means removed, therefore its function is absent. Here it means removed, and the removal is itself the oncogenic event. On a domain-architecture figure the PZP block should be read as a deleted brake, not as a deleted function.
It also makes the PZP a target in a way a lost module usually is not: restoring or mimicking PZP-nucleosome engagement is a gain-of-restraint strategy rather than a blockade. That is a genuinely different pharmacological ask from blocking the OM-LZ-DOT1L interface, and the two are not interchangeable.
The main caveat is fusion identity. Klein and colleagues worked on CALM-AF10, not KMT2A::MLLT10. Both fusions retain the OM-LZ and both lose the PZP, so the argument transfers on architecture, but it has not been demonstrated for the KMT2A fusion. We mark that as an extrapolation.
| uniprot | P55197 |
|---|---|
| length AA | 1068 |
| gene symbol | MLLT10 |
| alphafold ID | AF-P55197-F1 |
| isoform note | Coordinates are in UniProt P55197-1 (1,068 aa) numbering. |
| assumed fusion for fate | fusion-kmt2a-mllt10 (KMT2A::MLLT10, t(10;11)), MLLT10 as the 3' partner |
| domain coordinate caveat | MLLT10 is the best-resolved partner in this set: both the lost PZP module and the retained OM-LZ have experimental crystal structures, and the break falls in the predicted-disordered ~500 residues between them. The two soft calls are UniProt's self-association region (80-287) and its DNA-binding region (311-674); the latter spans 364 residues across predicted disorder and should be drawn as a low-confidence band, not a domain. The OM-LZ boundary varies by about five residues across three crystallographic constructs and by ~11 at the N-terminus if the narrower InterPro coiled-coil model is used instead — the narrower model omits the octapeptide motif, so do not use it to design a competing peptide. |
PDB IDS
[ "5DAG", "5DAH", "7MJU", "6CKN", "6CKO", "6JN2" ]
domains — the numbers behind the chart
[
{
"name": "PZP (PHD-Zn knuckle-PHD reader module)",
"start": 22,
"end": 198,
"note": "Folds as ONE module, not three. UniProt splits it into PHD1 22-74, C2HC pre-PHD Zn knuckle 79-112 and PHD2 135-198; crystal constructs span 1-208 (5DAG, 5DAH, 7MJU). Reads unmodified H3K27 via H3 residues 22-27 and engages the nucleosome through multivalent H3-tail and DNA contacts. Lost in KMT2A::MLLT10 — and its loss is not incidental: restoring PZP to CALM-AF10 abolishes transformation, so this is a deleted brake, not a deleted function.",
"pfam": "PF13831",
"fate": "lost",
"evidence_basis": "structure",
"boundary_precision": "module core 22-198 firm; crystallised construct extends to 208"
},
{
"name": "Self-association region",
"start": 80,
"end": 287,
"note": "UniProt functional region; overlaps the PZP and extends past it. Not a folded domain call and not separately structurally characterised.",
"fate": "lost",
"evidence_basis": "curated",
"boundary_precision": "functional region, soft edges"
},
{
"name": "DNA-binding region",
"start": 311,
"end": 674,
"note": "UniProt functional region spanning 364 residues that overlaps two predicted disordered regions (291-505, 583-612). Drawn here because UniProt annotates it, but it is emphatically NOT a folded DNA-binding domain and should be rendered as a low-confidence band.",
"fate": "variable",
"evidence_basis": "curated",
"boundary_precision": "very poor; a large functional call over predicted disorder"
},
{
"name": "OM-LZ (octapeptide motif-leucine zipper)",
"start": 704,
"end": 779,
"note": "THE oncogenic module: binds the DOT1L coiled-coil domain. Coordinates taken from the DOT1L co-crystal constructs 6CKN/6CKO (704-779). Compare UniProt 703-784, InterPro IPR049773 715-778 (coiled-coil core only, omits the octapeptide motif) and PDB 6JN2 699-782. Retained in KMT2A::MLLT10 and in CALM::AF10.",
"fate": "retained",
"evidence_basis": "structure",
"boundary_precision": "+/- ~5 aa across three crystallographic constructs; the InterPro coiled-coil call at 715-778 is narrower and excludes the octapeptide motif"
},
{
"name": "LZ (leucine zipper core)",
"start": 750,
"end": 778,
"note": "UniProt leucine-zipper call, the heptad-repeat core inside the OM-LZ. Drawn as a nested sub-feature.",
"fate": "retained",
"evidence_basis": "curated",
"boundary_precision": "+/- 2 aa"
}
]conflicting annotations
[
{
"feature": "OM-LZ extent",
"claim_a": "UniProt 703-784 (transactivation domain required for DOT1L binding); PDB 6CKN/6CKO 704-779; PDB 6JN2 699-782.",
"claim_b": "InterPro IPR049773 AF10-like coiled-coil domain: 715-778.",
"our_reading": "We draw 704-779 from the DOT1L co-crystal. The InterPro model calls only the heptad core and excludes the octapeptide motif, so it is a coiled-coil annotation rather than a functional-module annotation."
},
{
"feature": "PZP as one module or three",
"claim_a": "UniProt: three separate zinc fingers (22-74, 79-112, 135-198).",
"claim_b": "Crystallography: a single integrated PZP module that rearranges on H3 binding.",
"our_reading": "Claim B. We draw one PZP block 22-198 and keep UniProt's sub-boundaries in the note, because drawing three separate fingers would misrepresent a module that acts as a unit."
}
]- DOT1L catalytic inhibition (pinometostat) has already underperformed clinically in KMT2A-r leukemia; blocking the OM-LZ-DOT1L interface instead of the active site may inherit the same problem if the bottleneck is target-gene reversibility rather than enzyme occupancy.
- A coiled-coil heterodimer interface is extended and hydrophobic, with no pocket. Small-molecule tractability is genuinely doubtful, and the honest alternatives are a stapled peptide with the usual delivery problems or degradation of the fusion.
- AF10's PZP reader is needed for normal H3K79 methylation, so any strategy that hits both the fusion's OM-LZ and wild-type AF10 function risks broad epigenetic toxicity.
- Does restoring PZP function suppress KMT2A::MLLT10 as it does CALM-AF10? The experiment has been done for one AF10 fusion and not the other, and the answer decides whether 'restore the brake' is a general AF10-translocation strategy or a CALM-specific one.
- If deleting the PZP is what permits transformation, is there any other KMT2A partner where the LOST domain — rather than the retained one — is the oncogenic event? KMT2A's own PHD3/CYP33 switch is the obvious candidate, and the parallel has not been drawn explicitly in this graph.
- Do the 6CKN/6CKO heterodimer and the 6JN2 higher-order configuration describe the same interface in solution, or is one a crystal-packing artefact? The answer changes whether a competing peptide needs to block a dimer or an oligomer.
- The OM-LZ-DOT1L interface is the only partner interface in the KMT2A family with a co-crystal and a demonstrated transformation requirement. Why has no small molecule against it been reported, and is that a druggability verdict or just neglect?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt P55197 (AF10_HUMAN) is 1,068 aa with PHD-type zinc finger 1 at 22-74, a C2HC pre-PHD-type zinc finger at 79-112, PHD-type zinc finger 2 at 135-198, a self-association region at 80-287, a histone-H3-interaction region at 106-190, a DNA-binding region at 311-674, a transactivation domain required for DOT1L binding at 703-784 and a leucine zipper at 750-778.
UniProtKB entry P55197 (AF10_HUMAN), protein AF-10 / MLLT10 · UniProt Consortium · UniProtKB 2026What this evidence cannot show
Retrieved 2026-09-18. The 'DNA-binding region 311-674' is a very large call that overlaps two annotated disordered regions (291-505, 583-612) and should not be drawn as a folded DNA-binding domain. UniProt's three separate zinc-finger calls at 22-74, 79-112 and 135-198 understate the fact that crystallography shows them folding as one module.
Evidence that argues against this (1)
- Three of UniProt's largest AF10 calls have no corroboration from any profile method. The live InterPro call returns 22 entries for P55197 and every one lies in the PHD/ePHD region (10-206) or in the coiled coil; not one member database annotates anything between 207 and 714, so the 311-674 DNA-binding region - 364 residues, a third of the protein - has no independent support at all. Where the two sources do overlap they disagree: InterPro IPR049773 and CDD cd20901 place the AF10 coiled coil at 715-778, 35 residues further N-terminal than UniProt's leucine zipper at 750-778, and the ePHD is 79-198 in PROSITE PS51805 but 78-206 in CDD cd15708. InterPro Consortium; EMBL-EBI 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
An AF10 PHD finger-Zn knuckle-PHD finger (PZP) region folds into a single module that recognises histone H3 residues 22-27, the interaction is abrogated by H3K27 modification, and PZP recognition of H3 is required in cells for H3K79 dimethylation, DOT1L-target gene expression and proliferation of DOT1L-addicted leukemic cells.
The PZP domain of AF10 senses unmodified H3K27 to regulate DOT1L-mediated methylation of H3K79 · Chen S, Yang Z, Wilkinson AW, Deshpande AJ, Sidoli S, Krajewski K, Strahl BD, Garcia BA, Armstrong SA, Patel DJ, Gozani O · Mol Cell 60:319-327 2015What this evidence cannot show
Crystallography plus cell-line experiments; the cellular requirement is shown for wild-type AF10 in DOT1L-addicted lines, not for the KMT2A::MLLT10 fusion, which lacks the PZP. The paper does not state PZP boundaries as a single residue range, so the 1-208 figure below comes from the deposited construct boundaries rather than from the text.
Evidence that argues against this (1)
- In the leukaemias this node is about, H3K79 hypermethylation happens with no PZP present at all. DOT1L binds AF10 through the OM-LZ region - the region retained in MLL-AF10, which discards the PZP entirely - and, decisively, fusing hDOT1L directly to MLL, a construct containing no AF10 sequence whatsoever, transforms cells in a methyltransferase-activity-dependent manner with Hoxa9 upregulation and H3-K79 hypermethylation. The PZP requirement is therefore a property of wild-type AF10 in DOT1L-addicted lines, not a general requirement for H3K79me2 or for DOT1L-target expression, and it cannot be carried into a KMT2A::MLLT10 argument. Okada Y et al. 2005
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Crystal structures of the human AF10 PZP domain, alone (PDB 5DAG, 5DAH) and bound to the histone H3 tail (PDB 7MJU), use constructs spanning AF10 residues 1-208 or 19-208.
PDB entries 5DAG (Crystal structure of PZP domain of human AF10 protein) and 7MJU (Crystal structure of human AF10 PZP bound to histone H3 tail), with UniProt P55197 chain mappings · RCSB Protein Data Bank; UniProt Consortium · RCSB Protein Data Bank 2021What this evidence cannot show
Entry titles and chain ranges were confirmed against the RCSB data API on 2026-09-18. Construct boundaries are chosen for expression and crystallisation; residue 1 and residue 208 are not claims about where the fold begins and ends. Disordered termini are frequently unmodelled in the deposited coordinates even when present in the construct.
Evidence that argues against this (1)
- The constructs are 1-208 but the coordinates are much less. Converting the RCSB polymer-entity-instance records to author numbering: 5DAG chain A leaves residues 1-23, 179-188 and 205-208 unobserved, so it models 24-178 plus 189-204 - a 10-residue hole in the middle of the second PHD region and 37 of 208 residues missing in total; 7MJU chain A leaves 13-22, 180-188 and 204-208 unobserved. 5DAH is also not what the framing implies: its RCSB title is 'Crystal structure of PZP domain of human AF10 protein fused with Histone H3 peptide', a covalent fusion construct at 2.611 A rather than a co-complex, and it leaves 1-20 and 200-208 unobserved. Any PZP = 1-208 bar is a cloning boundary with about 18% of it never seen. RCSB Protein Data Bank 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Crystal structures of apo AF10 OM-LZ and of its complex with the DOT1L coiled-coil domain (PDB 6CKN, 6CKO; AF10 chains 704-779) show the recognition interface, and disrupting that interface abrogates MLL-AF10-associated leukemic transformation.
Structural and functional analysis of the DOT1L-AF10 complex reveals mechanistic insights into MLL-AF10-associated leukemogenesis · Zhang H, Zhou B, Qin S, Xu J, Harding R, Tempel W, Nayak V, Li Y, Loppnau P, Dou Y, Min J · Genes Dev 32:341-346 2018What this evidence cannot show
Transformation readout is a retroviral MLL-AF10 colony/immortalisation assay in mouse cells, not a patient-derived model. Interface-disrupting mutations show the interaction is necessary; they do not show a small molecule could achieve the same block, and a coiled-coil heterodimer interface is a hard small-molecule target.
Evidence that argues against this (1)
- The functional unit is not the heterodimer the 2018 structures describe. A later structure (PDB 6JN2, 3.6 A, coiled-coil domains of human DOT1L in complex with AF10, mapped by SIFTS to P55197 699-782) shows that the C-terminal end of the AF10 leucine zipper plus the OM domain drive tetramerisation of the binary complex into a DOT1L-AF10 OCTAMER, and that this oligomerisation ability - not merely the 1:1 interface - is essential for MLL1-AF10's leukaemogenic function. Interface mutations that abolish transformation are therefore ambiguous between blocking the heterodimer and blocking the octamer, and the druggability argument changes accordingly. Song X et al. 2019
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database
A separate crystal structure of the DOT1L-AF10 coiled-coil heterodimer (PDB 6JN2, AF10 chain 699-782) reports a higher-order configuration of the two coiled-coil domains that potentiates leukemogenic activity.
PDB entry 6JN2 — Crystal structure of the coiled-coil domains of human DOT1L in complex with AF10; primary citation: A higher-order configuration of the heterodimeric DOT1L-AF10 coiled-coil domains potentiates their leukemogenenic activity · RCSB Protein Data Bank · RCSB Protein Data Bank / Proc Natl Acad Sci USA 2019What this evidence cannot show
We verified this entry's title, method and primary-citation title against the RCSB data API on 2026-09-18 but did not independently verify a PMID for the primary citation, so we cite the structure rather than the paper. A 'higher-order configuration' seen in a crystal may be a lattice artefact; this claim needs solution-state corroboration before it is treated as mechanism.
Evidence that argues against this (1)
- The 6CKN/6CKO structures from the Min group describe a straightforward heterodimeric coiled-coil interface without invoking a higher-order assembly. Two crystallographic descriptions of the same interface differ on oligomeric state. Zhang H et al. 2018
- E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The AF10 PZP domain is consistently impaired or deleted in leukemogenic AF10 translocations and actively blocks malignant transformation: restoring a functional PZP to CALM-AF10 prevents transformation of bone-marrow HSPCs in vitro and in vivo, abrogates leukemogenesis in vivo, downregulates Hoxa genes and reverses cellular transformation. PZP binds the nucleosome core particle through multivalent H3-tail and DNA contacts and discriminates against H3K27me3.
The role of the PZP domain of AF10 in acute leukemia driven by AF10 translocations · Klein BJ, Deshpande A, Cox KL, Xuan F, Zandian M, Barbosa K, Khanal S, Tong Q, Zhang Y, Zhang P, Sinha A, Bohlander SK, Shi X, Wen H, Poirier MG, Deshpande AJ, Kutateladze TG · Nat Commun 12:4130 2021What this evidence cannot show
The fusion studied is CALM-AF10, not KMT2A::MLLT10. Both retain the OM-LZ and lose the PZP, so the argument transfers on architecture, but it has not been demonstrated for the KMT2A fusion and we mark the transfer as an extrapolation. The restoration experiments use engineered fusions with PZP added back, which is not a physiological configuration, and mouse HSPC transformation assays reward self-renewal specifically.
Evidence that argues against this (1)
- A competing account of the same fusion puts the decisive element on the CALM side, not the AF10 side. CALM contains a nuclear export signal that drives cytoplasmic localisation of CALM-AF10 and is necessary for transformation and for Hoxa upregulation and aberrant H3K79 methylation; decisively, fusing nuclear export signals from four unrelated proteins (ABL1, HIV Rev, PKIA, APC) in frame with AF10 is SUFFICIENT to immortalise murine haematopoietic progenitors. Adding a nucleosome-binding PZP back onto such a fusion plausibly blocks transformation by re-tethering it to chromatin and defeating the export-dependent mechanism, which the restoration experiments cannot exclude. Conway AE et al. 2013
- E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
MLLT10 is a recurrent KMT2A translocation partner within the recombinome of 3,401 acute leukemia patients analysed between 2003 and 2022.
The KMT2A recombinome of acute leukemias in 2023. · Meyer C, Larghero P, Almeida Lopes B, Burmeister T, Gröger D, Sutton R, et al. · Leukemia 37:988-1005 2023What this evidence cannot show
A referral-based diagnostic cohort, so partner frequencies reflect referral patterns. The lineage skew (MLLT10 commoner in myeloid than lymphoid infant disease) comes from other sources in this graph, not from this citation.
Evidence that argues against this (1)
- NO PUBLISHED OPPOSITION, and this is the bounded statement of that. The Frankfurt diagnostic series is the only dataset of its kind and no independent cohort has published a competing MLLT10 share of KMT2A rearrangements; the referral-ascertainment objection is this node's own caveat, not a published result. WHAT WOULD FALSIFY IT: a population-based or consecutively enrolled (non-referral) cohort screened by RNA sequencing reporting an MLLT10 share of KMT2A rearrangements materially different from the recombinome's, or showing that a substantial fraction of MLLT10 rearrangements in an unselected population have a partner other than KMT2A. SEARCH RECORD, run 2026-09-19: PubMed `("KMT2A recombinome"[tiab] OR "MLL recombinome"[tiab])`, translated by NCBI to `"MLL recombinome"[Title/Abstract]`, returns 9 records (PMIDs 37295428, 32294539, 30815012, 28701730, 23628958, 22927255, 19262598, 19144982, 16511515) and every one is the same group's serial report. A second query `MLLT10[tiab] AND (rearrangement[tiab] OR rearranged[tiab] OR fusion[tiab]) AND (PICALM[tiab] OR "non-KMT2A"[tiab] OR partners[tiab])` returns 46 records, none giving an alternative KMT2A partner-frequency denominator. NCBI PubMed 2026
Counter-evidence for the PZP subtraction: H3K79 hypermethylation and Hoxa9 upregulation occur with no AF10 sequence at all - fusing hDOT1L directly to MLL transforms in a methyltransferase-dependent manner (Okada 2005, PMID 15851025). And a competing account puts the decisive element on the CALM side: NES motifs from four unrelated proteins fused to AF10 are SUFFICIENT to immortalise progenitors (Conway 2013, PMID 23487024).
Counter-evidence bearing on DOT1L-AF10 druggability: the functional unit is not the 1:1 heterodimer. PDB 6JN2 shows the AF10 leucine zipper plus OM domain driving tetramerisation of the binary complex into a DOT1L-AF10 OCTAMER, and that oligomerisation - not the heterodimer interface alone - is essential for MLL1-AF10 leukaemogenesis (Song 2019, PMID 31527241). Interface mutants are ambiguous between the two.
domain architecture of the 3' partner; OM-LZ (704-779) retained, PZP (22-198) lost
AF10 OM-LZ 704-779 binds the DOT1L coiled-coil domain (PDB 6CKN, 6CKO)
gives residue coordinates for the AF10 surface that recruits DOT1L and for the PZP reader that licenses H3K79me2
- supports →Do different breaks give different biology? A verdict on partner identity vs breakpoint position
losing the PZP is itself the oncogenic event: restoring it to CALM-AF10 abolishes transformation
the OM-LZ-DOT1L interface is an alternative to catalytic DOT1L inhibition, with coordinates from a co-crystal
- ← related toPICALM::MLLT10 (CALM-AF10), t(10;11)(p13;q14) - the DOT1L-dependent HOXA fusion that contains no KMT2A
AF10 domain coordinates in P55197-1 numbering
- ← related toWhere the break falls, and the three incompatible numbering frames that make it easy to get wrong
AF10 coordinates asserted in P55197-1, which is not the frame the CALM-AF10 literature uses
OM-LZ retained and PZP lost in both MLLT10 fusions
uses the retained OM-LZ versus discarded PZP composition
positions placed against the PZP and OM-LZ coordinates
a crystallised OM-LZ span is chosen to diffract; the functional boundary was set by deletion in a transformation assay and need not coincide
- ← related toMLLT10 (AF10): the discarded PZP reader — the one partner where loss is proven to be the oncogenic event
domain coordinates and fate assignments for AF10
full domain map for the AF10 protein