AFF1 AF4_int motif to AF9/ENL AHD: the index fusion's only documented elongation interface, and it competes with DOT1L
also called AF4-AF9 interface · PDB 2LM0 · PF18875 to PF17793
The infant fusion keeps a short piece of AF4 that sticks to a floppy tail on the AF9 and ENL proteins. That same floppy tail is where the H3K79-writing enzyme binds, so the fusion and the enzyme want the same spot.
The same thing, for a scientist
KMT2A::AFF1 retains AFF1's **AF4 interaction motif** (Pfam PF18875 / InterPro IPR043639); this graph's `protein-aff1-domains` places the conserved core at aa 761-775 with the NMR construct of **PDB 2LM0** spanning 738-779, bound to the AF9/MLLT3 AHD at 490-568. This is the structurally defined interface by which the index fusion reaches the ENL family [E1], and through them P-TEFb and the super elongation complex [E2]. **The same AHD surface binds DOT1L.** The AHD recruits AF4, AFF4 or DOT1L, and a chemical probe against those interactions reduces SEC assembly and H3K79 methylation together [E3]. Reconstitution of minimal complexes is explicit: *DOT1 forms a stable complex with AF9 and does not interact with AF9-AF4-P-TEFb complexes* [E4]. DOT1L engages AF9 at three sites with graded functional consequence, so the competition is quantitative rather than binary [E5]. **The consequence for this graph is an inverted prediction.** In MLL-AF9 the fusion *supplies* the AHD and recruits DOT1L onto its targets. In KMT2A::AFF1 the fusion *occupies* somebody else's AHD with an AF4 motif, which on the reconstitution evidence should **exclude** DOT1L rather than deliver it. **The interface is functionally required in the index disease.** A peptide (PFWT) derived from the mapped AF9-binding domain of AF4 disrupts the AF4-AF9 interaction and kills t(4;11) MLL-AF4 leukaemia cells by apoptosis while sparing normal haematopoietic progenitors [E6]. That is a human, index-fusion, separation-of-function-flavoured result for this interface - and it reads out viability, not H3K79 methylation, so it does not say whether the interface delivers DOT1L or excludes it.
The interface, as far as it is defined
| side | protein | element | residues | structure |
|---|---|---|---|---|
| ligand | AFF1 (P51825) | AF4_int, PF18875 / IPR043639 | conserved core ~761-775; NMR construct 738-779 | 2LM0 |
| receptor | MLLT3 (P42568) | AHD, PF17793 / IPR040930 | 490-568 in the 2LM0 entity | 2LM0 |
| receptor | MLLT1 (Q03111) | AHD, PF17793 | not separately mapped here | - |
Residue ranges are taken from this graph's existing protein-aff1-domains entry, which resolved them from Pfam/InterPro and the RCSB entity annotations; this beat did not re-derive them and flags them as inherited rather than independently checked 1.
What would have to be true for the index fusion to reach DOT1L anyway
Three escape routes, each with a distinguishing prediction:
- Two AHD molecules. The fusion occupies one AF9/ENL and DOT1L occupies another nearby. Predicts DOT1L occupancy at index-fusion targets should be insensitive to AF4_int point mutation, and should scale with total AF9/ENL abundance rather than with fusion binding.
- Multivalency within DOT1L. DOT1L's three AF9-binding sites might permit one to remain engaged while AF4_int occupies the dominant surface 5. Predicts a ternary complex should be detectable at high concentration in vitro - the experiment Biswas' reconstitution ran and did not see 4.
- No AHD route at all. H3K79me2 at index-fusion targets is deposited by transcription-coupled endogenous DOT1L. Predicts AF4_int mutation should not change H3K79me2, and MLLT10 knockout should.
All three are separable by a single degron-plus-calibrated-ChIP series (exp-ahd-occupancy-competition-in-t411-cells).
The one index-fusion functional test this interface has
PFWT is the only experiment this beat found in which the AF4-AF9 interface is perturbed in t(4;11) cells 6. Its value and its limits are both worth stating.
Value: human, index-fusion, with a normal-progenitor control, published as a therapeutic lead - so the interface is not a structural curiosity, it is load-bearing for survival in the index disease.
Limits: 2004, cell-permeable peptide, no concentration in the abstract, and the readout is apoptosis. Since the AHD is a shared surface, disrupting it removes the fusion's P-TEFb route and any putative DOT1L route simultaneously, so the experiment cannot attribute the death to either. Repeating it with a calibrated H3K79 readout would convert a 20-year-old viability result into the discriminating measurement this entry needs.
Correction: 2LM0 is a single engineered chain, and the affinity is 0.17 nM
This entry described PDB 2LM0 as AFF1 738-779 "bound to" AF9 AHD 490-568, which reads as a two-chain complex. It is one chain. The RCSB polymer-entity record (fetched 2026-09-20) shows a single 125-residue polymer annotated to both P42568 and P51825 - an engineered single-chain chimera. Nothing in the entry title says so, and this deposit is one of at least six central to this disease built that way.
The affinity was also missing and is remarkable. PMID 23260655 reports the AF9 AHD binding AF4 at KD 0.17 nM, against DOT1L 1.6 nM, BCOR 32 nM and CBX8 >0.9 uM. The AHD is an intrinsically disordered region that folds on binding, and the AF4-family ligand outcompetes every other partner by one to four orders of magnitude.
That matters for the index fusion. This is KMT2A::AFF1's only documented elongation interface, and the number says the AF4-AHD contact is close to irreversible on the timescale of the others. A chemical probe that blocks it (PMID 34373735) works at micromolar - four orders of magnitude weaker than the ligand it displaces.
| partner a | gene-aff1 |
|---|---|
| partner b | gene-mllt3 |
| motif b | AF4_int / PF18875 / IPR043639, conserved core ~761-775 of AFF1 (P51825) |
| druggability | targeted by a micromolar chemical probe (PMID 34373735) that blocks AF9/ENL binding to DOT1L, AF4 and AFF4 together; selectivity between the index fusion's use of the surface and DOT1L's use of it has not been demonstrated |
| note on residues | ranges inherited from this graph's protein-aff1-domains entry, not independently re-derived in this cycle |
| functional validation in index disease | PMID 15269783 (PFWT peptide, t(4;11) cells, apoptosis readout, normal progenitors spared) |
| structures note | 2LM0 is a SINGLE 125-residue polymer chain annotated to both P42568 (MLLT3/AF9) and P51825 (AFF1) - an engineered fusion construct, not a two-chain complex. Reading it as a co-structure overstates what has been solved. |
| kd NM | 0.17 |
interface residues a
[]
structures
[ "2LM0" ]
competing ligands
[ "DOT1L", "AFF4", "BCOR", "CBX8" ]
- An AHD-interaction inhibitor developed against MLL-AF9 would, in t(4;11) disease, strip the fusion's own P-TEFb route rather than its DOT1L route - efficacy by an unintended mechanism, optimised against the wrong biomarker.
- P-TEFb is required by essentially every transcribed gene, so any compound that works in t(4;11) by detaching AEP has a window set by general elongation toxicity, not by fusion selectivity.
- If the index fusion competes with DOT1L for the AHD, then killing the fusion could increase H3K79 methylation somewhere in the genome, and a pharmacodynamic assay expecting a decrease would read a successful treatment as a failure.
- Is the AF4_int-AHD interaction retained at physiological stoichiometry in a t(4;11) cell, or is endogenous AFF4 the dominant AHD ligand there? If AFF4 wins, the index fusion may not occupy the AHD at all and the inverted prediction dissolves.
- Has anyone co-immunoprecipitated DOT1L with KMT2A::AFF1 from a human t(4;11) line under the same conditions used for MLL-AF9? This beat found no such side-by-side.
- Could the AF4_int-AHD interface be drugged selectively against DOT1L's use of the same surface, or is the shared hydrophobic core too similar to discriminate?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The structure of the AF9 AHD in complex with the elongation factor AF4 shows the AHD is intrinsically disordered and folds on binding, with conserved aliphatic residues of each of its partners forming an integral part of the complex's hydrophobic core.
Leukemia fusion target AF9 is an intrinsically disordered transcriptional regulator that recruits multiple partners via coupled folding and binding. · Leach BI, Kuntimaddi A, Schmidt CR, Cierpicki T, Johnson SA, Bushweller JH · Structure 2013What this evidence cannot show
Abstract fetched from PubMed and read on 2026-09-19; full text not read. NMR of isolated domains and peptides; the residue ranges quoted in this entry's table are inherited from this graph's protein-aff1-domains entry rather than re-derived here, and are flagged as such. The construct is AFF1 peptide plus AF9 AHD, not a fusion protein in chromatin.
Evidence that argues against this (1)
- A disordered hub that refolds around each partner need not use an identical surface for all of them, so 'the same site' is an inference from shared hydrophobic-core chemistry rather than a superposition of two structures. Kuntimaddi A et al. 2015
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
AF4 and ENL family proteins form a higher-order complex with P-TEFb (AEP) that is normally recruited to MLL-target chromatin, and MLL oncoproteins fused with AEP components constitutively form MLL/AEP hybrid complexes driving sustained target expression and progenitor transformation.
A higher-order complex containing AF4 and ENL family proteins with P-TEFb facilitates oncogenic and physiologic MLL-dependent transcription. · Yokoyama A, Lin M, Naresh A, Kitabayashi I, Cleary ML · Cancer Cell 2010What this evidence cannot show
Abstract fetched from PubMed and read on 2026-09-19; full text not read. Biochemical purification plus murine myeloid transformation assays; MLL-AF4 is an AEP-component fusion by this classification but the transformation readout is myeloid.
Evidence that argues against this (1)
- AEP recruitment establishes a route to P-TEFb, which is what this interface delivers; it says nothing about DOT1L, and the same body of work shows the DOT1L-containing and AF4-containing assemblies are distinct. Biswas D et al. 2011
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The AHD of AF9/ENL binds AF4, AFF4 or DOT1L, and a small-molecule inhibitor of those protein-protein interactions (IC50 0.9-3.5 uM) reduces both super-elongation-complex assembly and DOT1L-mediated H3K79 methylation in leukaemia cells.
Small-molecule inhibitor of AF9/ENL-DOT1L/AF4/AFF4 interactions suppresses malignant gene expression and tumor growth. · Wu F, Nie S, Yao Y, Huo T, Li X, Wu X, Zhao J, Lin YL, Zhang Y, Mo Q, Song Y · Theranostics 2021What this evidence cannot show
Abstract fetched from PubMed and read on 2026-09-19; full text not read, identifiers copied from that record. A micromolar probe with a broad transcriptional footprint and no separation-of-function control described in the abstract; one compound reducing two outputs cannot distinguish blocking one shared site from blocking two independent ones.
Evidence that argues against this (1)
- If the same compound removes SEC and DOT1L, then in KMT2A::AFF1 cells - where the fusion itself uses this surface - it would detach the fusion's own elongation machinery, so its therapeutic reading in the index disease is the opposite of its reading in MLL-AF9. Biswas D et al. 2011
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In reconstitution from defined components, DOT1 forms a stable complex with AF9 and does not interact with AF9-AF4-P-TEFb complexes, while AF4 and AFF4 independently bridge P-TEFb to AF9 through direct contacts with AF9 and cyclin T1.
Function of leukemogenic mixed lineage leukemia 1 (MLL) fusion proteins through distinct partner protein complexes. · Biswas D, Milne TA, Basrur V, Kim J, Elenitoba-Johnson KS, Allis CD, Roeder RG · Proc Natl Acad Sci U S A 2011What this evidence cannot show
Abstract fetched from PubMed and read on 2026-09-19; full text not read. Purified minimal complexes 'in the absence of other factors': a stringent sufficiency test that can miss interactions stabilised in cells by chromatin, by the KMT2A moiety, or by a bridging factor. The paper's functional validation is MLL-AF9 immortalisation, not MLL-AF4.
Evidence that argues against this (1)
- The same study describes these assemblies as dynamic and interconverting, so a steady-state negative does not exclude transient DOT1L access to an AF4-occupied AHD in cells - the timescale on which a distributive methyltransferase would work anyway. Leach BI et al. 2013
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
DOT1L has three AF9-binding sites, and structure-guided point mutations give graded reduction of DOT1L recruitment to MLL-AF9, graded loss of H3K79me2/me3 at target genes and graded loss of transformation.
Degree of recruitment of DOT1L to MLL-AF9 defines level of H3K79 Di- and tri-methylation on target genes and transformation potential. · Kuntimaddi A, Achille NJ, Thorpe J, Lokken AA, Singh R, Hemenway CS, Adli M, Zeleznik-Le NJ, Bushweller JH · Cell Rep 2015What this evidence cannot show
Abstract fetched from PubMed and read on 2026-09-19; full text not read. Entirely MLL-AF9; the quantitative backbone of the AHD-DOT1L axis exists only for the partner the index fusion does not carry.
Evidence that argues against this (1)
- Three binding sites make a strict either/or occupancy model harder to sustain, and would be the mechanistic basis for escape route 2 - one DOT1L site engaged while AF4_int occupies the dominant surface. Biswas D et al. 2011
- E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
A synthetic peptide designed from the mapped AF9-binding domain of AF4 disrupts the AF4-AF9 interaction in vitro and in vivo, inhibits proliferation of t(4;11) MLL-AF4 leukaemia cells through apoptosis, and does not affect the proliferative capacity of haematopoietic progenitor cells.
The synthetic peptide PFWT disrupts AF4-AF9 protein complexes and induces apoptosis in t(4;11) leukemia cells. · Srinivasan RS, Nesbit JB, Marrero L, Erfurth F, LaRussa VF, Hemenway CS · Leukemia 2004What this evidence cannot show
Abstract fetched from PubMed and read on 2026-09-19; full text not read, every identifier copied from that record. 2004 cell-permeable peptide work: no concentration is given in the abstract, peptides of that class frequently act through membrane effects, and the normal-progenitor sparing is asserted without a quantitative window. The readout is apoptosis, not chromatin, so nothing about DOT1L follows from it.
Evidence that argues against this (1)
- Because the AHD is a shared surface, a peptide occupying it removes the fusion's route to P-TEFb and any route to DOT1L at the same time; the experiment therefore cannot say which loss the cells died from, and a later micromolar small molecule against the same interactions shows the same ambiguity. Wu F et al. 2021
- E-2lm0strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
PDB 2LM0 is a single 125-residue polymer chain annotated to both MLLT3 (P42568) and AFF1 (P51825) - an engineered single-chain construct rather than a two-chain complex. The AF9 AHD binds AF4-family ligand with KD 0.17 nM, against DOT1L 1.6 nM, BCOR 32 nM and CBX8 above 0.9 uM.
RCSB PDB entry 2LM0 polymer-entity record: a single 125-residue chain annotated to both P42568 and P51825 · RCSB Protein Data Bank 2026What this evidence cannot show
Polymer-entity composition read from the RCSB REST record on 2026-09-20. A single-chain construct is a legitimate and common way to solve a weak or transient interface and does not make the structure wrong - it makes it a model of the bound state rather than a measurement of association. The affinities are from a separate paper on AF9, not AFF1's own binding, and are peptide affinities.
Evidence that argues against this (1)
- A 0.17 nM affinity measured on isolated domains may not survive in the SEC, where AFF1 and AF9 are both embedded in a larger assembly with competing contacts - and the same paper's point is that the AHD is disordered and folds on binding, which makes fragment affinities particularly sensitive to construct boundaries. Structure 2013
the other retained AFF1 interface in the fusion
adds the AF9 affinity ladder and flags that 2LM0 is a single-chain chimera
named in the audit as asserting an interface with no residue on either side
predicted AFF1 peptide 2.1 A from the best NMR model; shuffles 12-14 A
AFF2/3/4 carry the AFF1 contact residues of the AF9 AHD surface
same AF9 surface, competing partner
AF4_int motif, the ligand side
the AF9 AHD, the receptor side (PDB 2LM0)
ENL carries the same AHD
AF4 occupancy of the AHD excludes DOT1L in reconstitution
the index fusion's only documented elongation interface
route to P-TEFb via AEP/SEC
the interface the route argument turns on
- ← testsDoes KMT2A::AFF1 deliver DOT1L to the AHD or displace it? A competition experiment in t(4;11) cells
tests the interface's sign in cells
- treated by →KMT2A-rearranged acute lymphoblastic leukaemia
PFWT peptide disrupting AF4-AF9 kills t(4;11) cells and spares normal progenitors
scaffold identity bears on which AHD interface is present