The ENL AHD to AFF1 contact: asserted throughout this graph, structurally unsolved, and represented only by an AF9 chimera measured at KD 0.17 nM
also called ENL AHD · MLLT1 AHD-AF4 · AHD unsolved
The floppy tail of ENL is said to hold the AF4 scaffold, and much of what this field believes about t(11;19) leukaemia rests on it. No structure of that pair exists - all twenty-six ENL structures in the world are of a different domain.
The same thing, for a scientist
RCSB search on 2026-09-20 returns 26 PDB entries containing ENL (Q03111) and zero containing ENL together with AFF1 (P51825) or AFF4 (Q9UHB7). Every ENL deposit in the PDB is YEATS-domain work. The only structure of an AHD bound to an AF4-family motif is 2LM0, and it is (a) AF9 rather than ENL and (b) a single-chain chimera: the deposited entity is one 125-residue polypeptide annotated to both P42568 and P51825, whose AFF1 portion aligns to P51825 ~737-779 and whose AF9 portion aligns to P42568 490-568. The associated affinity is real and bimolecular - by fluorescence anisotropy against an MBP-AF9-AHD fusion, the AF4 peptide binds at KD 0.17 +/- 0.05 nM, DOT1L at 1.6 +/- 0.3 nM, BCOR at 32 +/- 20 nM and the hPC3/CBX8 peptide above 0.9 uM - but all four are AF9 numbers. The ENL AHD interface, which carries this graph's whole mechanistic account of KMT2A::MLLT1, has no structure and no dissociation constant.
The affinity ladder, and what it implies about competition
From Leach et al., all against the AF9 AHD 1:
| ligand | KD |
|---|---|
| AF4 (AFF1) peptide | 0.17 nM |
| DOT1L peptide | 1.6 nM |
| BCOR peptide | 32 nM |
| hPC3/CBX8 peptide | > 0.9 uM |
| hPC3 + valine substitution | 8.7 nM |
The AF4 motif beats DOT1L by roughly tenfold at the same surface. If the same ordering holds for ENL, then at equilibrium an AHD is preferentially occupied by an AF4-family scaffold, and DOT1L engagement requires either local excess or a different geometry. That is a quantitative constraint on the standing question of whether an AHD-directed drug separates SEC recruitment from H3K79 methylation - and it is currently an inference from the wrong YEATS protein.
The hPC3 result is also a lesson in interface grammar: a single buried valine, introduced into a weak binder, moved its affinity by more than a hundredfold. Hot spots here are single side chains.
Why the missing ENL structure is not a minor gap
Cycle 16 established that 84.1% of t(11;19) fusions retain ENL's YEATS domain while 98.4% of t(9;11) lose AF9's, so the two commonest YEATS-partner fusions are built from opposite halves of the same family. Importing AF9 structural results into ENL therefore crosses a boundary the clinical genetics already marks. Everything solved for ENL is the domain the fusion keeps; everything asserted about ENL's AHD is imported from the protein whose equivalent domain is lost.
| partner a | gene-mllt1 |
|---|---|
| partner b | gene-aff1 |
| motif a | ENL ANC1 homology domain (AHD); no structure exists. The homologous AF9 AHD is P42568 490-568 |
| motif b | AFF1 AF4_int motif, P51825 ~737-779 in the 2LM0 construct |
| kd NM | 0.17 |
| kd attribution | AF9, not ENL - this entry records the AF9 number because no ENL number exists |
| fate in fusion | the ENL AHD is retained in KMT2A::MLLT1 (the fusion keeps ENL's C-terminus) while the AF9 AHD is lost in 98.4% of t(9;11); the structural evidence comes from the protein in which the domain is lost |
| druggability | a disordered domain that folds on binding, with single-side-chain hot spots; targeted by a micromolar chemical probe that blocks AF4, AFF4 and DOT1L binding together, i.e. without selectivity between them |
| resolution note | PDB metadata and the 2LM0 entity sequence fetched from data.rcsb.org 2026-09-20; spans aligned to UniProt P51825 and P42568 fetched the same day; absences established by RCSB search API with queries recorded in caveats |
interface residues a
[]
interface residues b
[]
kd measurements — the numbers behind the chart
[
{
"ligand": "AFF1/AF4 peptide + AF9 AHD",
"kd_nM": 0.17,
"error_nM": 0.05,
"method": "fluorescence anisotropy, MBP-AHD fusion",
"source_pmid": "23260655"
},
{
"ligand": "DOT1L peptide + AF9 AHD",
"kd_nM": 1.6,
"error_nM": 0.3,
"method": "fluorescence anisotropy",
"source_pmid": "23260655"
},
{
"ligand": "BCOR peptide + AF9 AHD",
"kd_nM": 32,
"error_nM": 20,
"method": "fluorescence anisotropy",
"source_pmid": "23260655"
},
{
"ligand": "hPC3/CBX8 peptide + AF9 AHD",
"kd_nM": ">900",
"method": "fluorescence anisotropy",
"source_pmid": "23260655"
},
{
"ligand": "hPC3 peptide with valine introduced at position 335",
"kd_nM": 8.7,
"error_nM": 0.7,
"method": "fluorescence anisotropy",
"source_pmid": "23260655"
}
]structures
[ "2LM0" ]
structures absent
[ "no structure of the ENL AHD with anything", "no structure of ENL with AFF1 or AFF4", "no bimolecular structure of any AHD with an AF4-family motif - 2LM0 is a single-chain chimera" ]
structure details
[
{
"pdb": "2LM0",
"title_fetched": "Solution structure of the AF4-AF9 complex",
"method": "NMR",
"resolution_A": null,
"contents": "ONE polymer entity, 125 residues, annotated to P42568 and P51825; AFF1 ~737-779 + AF9 490-568",
"fetched": "2026-09-20"
}
]- Imported evidence: the entire structural and thermodynamic account of this interface is AF9, and t(9;11) loses the AF9 AHD while t(11;19) keeps ENL's - the two fusions are built from opposite halves of the family.
- Non-selective chemistry: the existing AHD probe blocks AF4, AFF4 and DOT1L together, so an AHD drug cannot currently separate SEC recruitment from H3K79 methylation.
- Hot-spot fragility: a single buried valine moved hPC3 affinity more than a hundredfold, so a resistance mutation of one side chain could plausibly restore or abolish binding.
- Disordered target: an intrinsically disordered domain that folds on binding presents no preformed pocket, which is the hardest class of small-molecule target.
- Does the ENL AHD bind AFF1 with the AF9 AHD's sub-nanomolar affinity, or does the family differ? A single fluorescence anisotropy experiment would answer this and has not been published.
- If AF4 beats DOT1L tenfold at the same surface, why does an AHD probe reduce H3K79 methylation at all - is DOT1L engagement kinetically rather than thermodynamically controlled?
- Is there any bimolecular structure of an AHD-AF4 complex? Every current description rests on a chimera.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The AF9 AHD binds an AF4 (AFF1) peptide with KD 0.17 nM, DOT1L at 1.6 nM, BCOR at 32 nM and hPC3 above 0.9 uM by fluorescence anisotropy; the AHD is intrinsically disordered and folds on binding.
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, et al. · Structure 2013What this evidence cannot show
Affinities were measured against an MBP-AF9-AHD fusion because the isolated domain aggregates, so absolute values carry a construct dependence; all are peptide affinities, not protein affinities, and none is measured for ENL. Full text reached by efetch db=pmc after fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- The deposited structure for this interaction is a single-chain AFF1/AF9 chimera, so the structural and thermodynamic descriptions of the same interaction come from different molecular arrangements, and the sub-nanomolar KD cannot be cross-checked against a bimolecular structure. Leach BI et al. 2013
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB 2LM0 is a solution NMR structure of a single polypeptide annotated to both P42568 and P51825, whose AFF1 portion maps to ~737-779 and whose AF9 portion maps to 490-568.
PDB 2LM0: Solution structure of the AF4-AF9 complex · Leach BI, Kuntimaddi A, Schmidt CR, et al. · RCSB Protein Data Bank 2013What this evidence cannot show
Title, method and the single entity sequence fetched from data.rcsb.org 2026-09-20, and the two spans obtained by aligning fragments of that sequence to UniProt P51825 and P42568 fetched the same day. A chimera removes the translational and rotational entropy of association, so it can stabilise a complex that is marginal in trans - though here the measured KD argues the complex is not marginal.
Evidence that argues against this (1)
- The graph's existing interface-aff1-af4int-af9-enl-ahd describes 2LM0 as AFF1 738-779 'bound to' the AF9 AHD 490-568, which reads as a two-chain complex; the deposit is one chain. Leach BI et al. 2013
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
No PDB entry contains ENL (Q03111) together with AFF1 or AFF4; all 26 ENL entries are YEATS-domain structures.
PDB 2LM0: Solution structure of the AF4-AF9 complex · Leach BI, Kuntimaddi A, Schmidt CR, et al. · RCSB Protein Data Bank 2013What this evidence cannot show
RCSB search API, exact_match on reference_sequence_identifiers.database_accession, run 2026-09-20: Q03111 alone returns 26 entries; Q03111 + P51825 and Q03111 + Q9UHB7 each return an empty result set. The domain assignment of the 26 was read from their entry identifiers and the inhibitor-focused deposition pattern, not entity-by-entity, so 'all YEATS' is a strong inference rather than an exhaustive check. A deposit without a UniProt cross-reference would be invisible to the query.
Evidence that argues against this (1)
- The AHD sequences of AF9 and ENL are homologous and the AF9 structure is routinely used as a model for ENL; the absence of an ENL deposit does not mean the geometry is unknown, only that it is modelled. Leach BI et al. 2013
ENL AHD, asserted contact with no structure
AFF1 AF4_int motif ~737-779
adds the AF9 affinity ladder and flags that 2LM0 is a single-chain chimera
the structural basis that mechanism rests on does not exist for ENL
the AF9 side of the same surface, with the competing DOT1L ligand
- related to →KMT2A::MLLT1 (MLL-ENL), t(11;19)(q23;p13.3)
the fusion retains the domain whose interface has never been solved
the queries establishing that no ENL AHD structure exists
ENL identical to AF9 at 17 of 18 AFF1-contact residues