Menin's nucleosome grip (8GPN): a three-patch surface at menin 353-362, 428-442 and 556-587 that holds H3, H4 and H2B - and touches H2A at exactly one residue
also called menin H3K79me2 reader surface · menin fingers and palm nucleosome interface · 8GPN interface
A cryo-electron microscopy structure shows how menin grips a whole nucleosome: it holds histones H3, H4 and H2B firmly but barely touches H2A. That matters because a 2026 paper says menin binds the damage-marked version of H2A, so whatever that contact is, it is not the grip seen here.
The same thing, for a scientist
Lin et al. identified menin as a reader of H3K79me2 using a nucleosome-based photoaffinity probe, and solved a 3.2 A cryo-EM structure (PDB 8GPN) of menin on an H3K79me2 nucleosome, reporting engagement through the fingers and palm domains and a pi-cation recognition of the mark. Measured here from the deposited coordinates, menin's nucleosome-facing surface is three patches: 353-362 and 428-434 against histone H3; 199-200, 355-358 and 433-434 against H4; 434 and 556-586 against H2B; and 135, 441-442, 571-576 against nucleosomal DNA. The contact with H2A is **a single residue, menin 577**. Menin Trp436 - the SUV39H1-critical patient-mutation position - sits 7.5 A from H3 and 8.1 A from H4, i.e. on this face rather than near the drug pocket. Menin residues 386-400, the core of the gamma-H2A.X motif, are unmodelled here as they are in all 69 menin deposits, and residue 371 is 14 A from H3.
No three-dimensional model is available for this protein, so the map above is the whole picture. That is itself informative: AlphaFold does not model chains this long, and the parts that have been crystallised are held under the entries for the individual domains.
The three patches, measured
Menin residues within 5.0 A of each nucleosome component in 8GPN, computed from the deposited coordinates with the menin chain identified by reading the entity's UniProt cross-reference (menin is entity 7, not entity 1 - entity 1 is histone H3.2, and an early run of this analysis reported H3 residues as menin residues before that was checked) 2:
| component | n menin residues within 5 A | which |
|---|---|---|
| histone H3 | 12 | 353, 355, 358, 361, 362, 428, 429, 430, 433, 434, 582, 587 |
| histone H4 | 9 | 199, 200, 355, 356, 358, 433, 434, 580, 582 |
| histone H2B | 16 | 434, 556, 560, 563, 564, 571, 574, 577, 578, 580-586 |
| histone H2A | 1 | 577 |
| nucleosomal DNA | 6 | 135, 441, 442, 571, 572, 576 |
Menin is modelled at 451 residues over the span 2-587 in this deposit.
The single-residue H2A contact is the result that matters for this beat. If menin's constitutive nucleosome grip barely touches H2A, then the damage-induced menin-gamma-H2A.X contact described in PMID 42578372 cannot be this grip seen again - it is either a different mode, or it requires the Ser139 phosphorylation and the H2A.X-specific C-terminal tail that a canonical H2A type 1 nucleosome does not have. 8GPN contains H2A type 1 (P06897), not H2A.X.
Where the SvID sits on this face
Menin 428-434 and 441-442 are inside the SvID window (360-445), and Trp436 is 7.5 A from H3 and 8.1 A from H4 while being 9.9-11.1 A from every menin inhibitor 2. So the C-terminal helix of the SvID is nucleosome-facing.
That raises a question nobody has asked: W436R is catalogued as a SUV39H1-binding mutant, but it sits on the nucleosome grip. A patient substitution of a buried-to-charged tryptophan on a histone-contacting helix could be a nucleosome-binding defect that presents as a co-IP defect for whatever chromatin-associated partner is being blotted. Neither the 2013 SUV39H1 paper nor the 2023 structure paper could have noticed this, because they are ten years apart and neither cites the other.
| partner a | gene-men1 |
|---|---|
| partner b | mark-h3k79me2 |
| motif a | three menin patches: 353-362, 428-442 and 556-587, plus 199-200 |
| motif b | the H3K79me2 nucleosome; the mark itself is read by a pi-cation interaction per the source abstract |
| method that mapped it | nucleosome photoaffinity probe plus quantitative proteomics for identification; single-particle cryo-EM at 3.2 A for the interface |
| ligand note | 8GPN contains no inhibitor. The nucleosome face and the MBM drug pocket are on opposite sides of menin: the nucleosome-contacting residues are 7.5-20.8 A from the drug-pocket residues measured in the same coordinate frame in other deposits, and no clinical compound contacts any of the 16 nucleosome-contacting residues listed here. |
| druggability | Not assessed and structurally unpromising - three dispersed patches over a 230-residue span engaging a 200 kDa particle is a protein-protein interface of the kind small molecules do not block. |
| epigenetic readout | READER of H3K79me2 in a nucleosomal context, via pi-cation recognition; this is in addition to menin's role as the recruiter of the H3K4 writer KMT2A and, separately, of SUV39H1 and SETD2 |
| shown for fusion | NO. The probe work and the structure are biochemical and in unspecified cultured cells per the abstract; nothing in a KMT2A-rearranged cell, nothing in a fusion system. |
| relationship to existing entries | New. The graph holds `mark-h3k79me2` and a large DOT1L layer but has no entry placing menin on that mark, and no menin-nucleosome interface of any kind. It also supplies the structural context that `interface-menin-suv39h1-svid-360-445` needs: the SvID's C-terminal helix is nucleosome-facing. |
interface residues a
[ "menin 353", "menin 355", "menin 358", "menin 361", "menin 362", "menin 428", "menin 429", "menin 430", "menin 433", "menin 434", "menin 441", "menin 442", "menin 577", "menin 580", "menin 582", "menin 587" ]
interface residues b
[]
domains — the numbers behind the chart
[
{
"name": "nucleosome patch I (H3/H4)",
"start": 353,
"end": 362
},
{
"name": "nucleosome patch II (H3/H4/H2B)",
"start": 428,
"end": 442
},
{
"name": "nucleosome patch III (H2B/DNA)",
"start": 556,
"end": 587
}
]structures
[ "8GPN" ]
structure details
[
{
"pdb": "8GPN",
"title_fetched": "Human menin in complex with H3K79Me2 nucleosome",
"method": "ELECTRON MICROSCOPY",
"resolution_A": 3.2,
"polymer_entities": 7,
"menin_entity_id": 7,
"menin_modelled": "2-587, 451 residues",
"histone_h2a_species": "H2A type 1 (P06897) - canonical, NOT H2A.X",
"fetched": "2026-09-22"
}
]contacts per component within 5A — the numbers behind the chart
{
"H3": 12,
"H4": 9,
"H2B": 16,
"H2A": 1,
"DNA_strand_I": 5,
"DNA_strand_J": 1
}ligands
[]
- If menin is a chromatin reader as well as a scaffold, then evicting the fusion from chromatin with a pocket inhibitor does not evict menin from chromatin, and residual menin at H3K79me2-marked gene bodies is a mechanism for incomplete response that the pocket model does not contain.
- A degrader removes the reader function too. That is a larger perturbation of gene-body chromatin than an inhibitor, in every cell in the body, and its consequences are unmeasured.
- Essential-function exposure: H3K79me2 is broadly distributed over transcribed gene bodies, so a menin function tied to it is not restricted to the ~150 fusion-target loci the leukaemia argument is built on.
- The H2A single-residue contact cuts both ways as a warning: it is evidence that the 2026 gamma-H2A.X mode is separate, and it is also a reason to doubt any model that treats 'menin binds nucleosomes' as a single event.
- Does menin bind an H2A.X nucleosome differently from a canonical one? 8GPN gives one menin-H2A contact; the 2026 HR paper needs many. A cryo-EM structure of menin on a gamma-H2A.X nucleosome is the single experiment that would resolve this beat's central question structurally.
- Is W436R a nucleosome-binding mutant rather than a SUV39H1-binding mutant? It sits 7.5 A from H3 on this face, and the two possibilities have never been tested side by side.
- If menin reads H3K79me2 and KMT2A fusions drive H3K79me2 through DOT1L, is there a feed-forward loop in which the fusion's own mark recruits more menin? Nothing in this graph's H3K79 layer models menin as a reader.
- Does a menin inhibitor change menin's nucleosome residence? The drug site and the nucleosome face are on opposite sides of the protein, so the naive answer is no - and `gene-suv39h1` already records a mechanism (Ser487 phosphorylation) by which menin leaves chromatin without any change at the SvID.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Menin is a reader of H3K79me2 in a nucleosomal context: a nucleosome-based photoaffinity probe plus quantitative proteomics identified it, a cryo-EM structure of menin bound to an H3K79me2 nucleosome shows engagement through the fingers and palm domains with pi-cation recognition of the mark, and in cells menin is selectively associated with H3K79me2 on chromatin, particularly in gene bodies.
Menin "reads" H3K79me2 mark in a nucleosomal context. · Lin J, Wu Y, Tian G, Yu D, Yang E, Lam WH, Liu Z, Jing Y, Dang S, Bao X, Wong JWH, Zhai Y, Li XD · Science 2023What this evidence cannot show
Cited at ABSTRACT level: the efetch abstract was fetched and read on 2026-09-22; the full text was not opened, so the pi-cation residue assignment and the fingers/palm boundaries are the authors' words as given in the abstract and are not independently checked here. The structural numbers in this entry come from the deposited coordinates (E2), not from the paper. The reconstituted nucleosome is a designed substrate; H3K79me2 is installed chemically, and the physiological stoichiometry of menin on H3K79me2 chromatin is not addressed by a structure.
Evidence that argues against this (1)
- This graph already holds an extensive H3K79 layer built on DOT1L as the writer and on H3K79me2 as a fusion-associated mark. A menin READER function on the same mark would make menin both the recruiter of KMT2A and a reader of the mark DOT1L deposits downstream of it, which is a feedback architecture none of this graph's H3K79 mechanism entries model. That is a gap, not a contradiction, and it is recorded as one. RCSB Protein Data Bank depositors 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
In PDB 8GPN (cryo-EM, 3.2 A) menin is modelled over residues 2-587 (451 residues) and contacts histone H3 through menin 353, 355, 358, 361, 362, 428, 429, 430, 433, 434, 582 and 587; H4 through 199, 200, 355, 356, 358, 433, 434, 580 and 582; H2B through 434 and 556-586; nucleosomal DNA through 135, 441, 442, 571, 572 and 576; and histone H2A through the single residue 577. Menin Trp436 lies 7.5 A from H3 and 8.1 A from H4. Menin residues 386-400 are unmodelled.
analyses/q34-menin-surface-geometry/run.py - menin surface geometry over all 69 O00255 PDB deposits · Q34-menin-surface-map · Oncofusion Knowledge Graph 2026What this evidence cannot show
Computed 2026-09-22 from the live RCSB mmCIF; minimum heavy-atom distances at a 5.0 A cutoff; auth numbering verified equal to UniProt O00255 numbering by residue-identity readback at I360, E366, D370, V371, A385, D418, W436 and G445. At 3.2 A cryo-EM, side-chain placement is a model: the residue LISTS are reliable at this cutoff, individual sub-angstrom distances are not. The nucleosome is a reconstituted particle containing Xenopus/human canonical histones including H2A type 1 (P06897) - NOT H2A.X - so the single-residue H2A contact is a statement about a canonical nucleosome. IMPORTANT METHOD NOTE: the first run of this measurement assumed the menin polymer was entity 1 and therefore reported histone H3 residues 37-134 as 'menin residues'. Entity identity is read back from the UniProt cross-reference in the shipped script.
Evidence that argues against this (1)
- A single cryo-EM model of one menin-nucleosome complex is one state. Menin's contact with H2A could be larger in a nucleosome carrying H2A.X, whose C-terminal tail differs from canonical H2A and carries the Ser139 that becomes gamma-H2A.X - and no structure of that particle with menin exists. Lian A et al. 2026
menin 353-362, 428-442 and 556-587, the nucleosome-contacting face
menin reads H3K79me2 in a nucleosomal context by pi-cation interaction (8GPN, 3.2 A)
the SvID's C-terminal helix 428-445, carrying W436, is part of this nucleosome face
- related to →H2AX - filed as a gene and not as a mark, and the reason that decision changes what the graph can say
menin contacts canonical H2A at ONE residue here, so the damage-induced gammaH2A.X mode cannot be this grip
8GPN is a different conformer and models to residue 587; the parse here is of the free protein