The menin-gamma-H2A.X motif, menin Val371-Gln400: necessary, sufficient as a peptide, and the nine residues at its centre have never been seen in any of the 69 menin structures
also called Pep-MH · menin 371-400 · Val371-Gln400 motif · menin-gammaH2AX interface
Menin grabs a damage-marked histone, gamma-H2A.X, using a thirty-residue stretch of itself, and that grab is what lets cells repair broken DNA accurately. The middle of that stretch is floppy and has never been seen in any menin crystal or cryo-EM structure, so nobody knows what it looks like when it is doing its job.
The same thing, for a scientist
Lian et al. (Nucleic Acids Res 2026) map a direct menin-H2A.X contact to menin residues Val371-Gln400 by truncation of the Palm/Fingers region, show that an internal deletion of exactly those residues (menin delta-371-400) abolishes gamma-H2A.X binding and fails to rescue HR in Men1-knockdown A549 cells, and show that a synthetic biotinylated 30-mer covering 371-400 (Pep-MH) precipitates both wild-type H2A.X and the phosphomimetic H2A.X-S139E, blocks the menin-H2A.X interaction, suppresses IR-triggered HR while potentiating NHEJ, and blocks aphidicolin-induced menin binding at a common fragile site. So 371-400 is both necessary and, as an isolated peptide, sufficient for the contact. Structurally the motif is bipartite: residues 371-385 form an ordered helix at the rim of the central MBM cavity and are observed in 68-69 of 69 menin PDB deposits, while residues 389-397 are observed in ZERO of 69 and are absent from the construct altogether in 42 of them - measured in `ana-q34-menin-surface-geometry` this cycle. Val371 itself lies 2.3-4.3 A from ziftomenib, revumenib and DS-1594b in their co-crystals, so the clinical menin-inhibitor class is in van der Waals contact with the first residue of this motif, while the rest of the motif is 10-27 A away.
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.
What was actually done, in what system
System: human A549 lung adenocarcinoma cells, Men1-knockout mouse embryonic fibroblasts, and Men1-floxed mice irradiated at 10 Gy. Not a leukaemia, not a KMT2A fusion, not a B-lineage cell. Rule 73 applies and the first clause of every claim here says so.
| step | result | figure |
|---|---|---|
| co-IP with menin domain truncations | menin binds gamma-H2A.X through the Palm and Fingers domains | Fig 2F |
| finer truncations of Palm/Fingers | the contact narrows to 371-400 | Fig 2G-H |
| internal deletion delta-371-400 | abolishes gamma-H2A.X binding | Supp Fig S3A |
| delta-371-400 rescue | fails to reduce IR-induced gamma-H2A.X foci; fails to rescue HR after Men1 siRNA | Supp Fig S3B, S4D |
| biotinylated Pep-MH (371-400) pulldown | precipitates WT H2A.X and H2A.X-S139E | Supp Fig S3D |
| Pep-MH in cells | blocks menin-H2A.X, suppresses HR, potentiates NHEJ, blocks menin at the CFS by ChIP | Fig 2K-L, Supp Fig S3E-F, S4C |
The reciprocal side is coarse. Menin was mapped to a 30-residue window; H2A.X was mapped only to "both the N- and C-terminal domains" (Fig 2I). There is no H2A.X residue-level map and no structure of the complex 1.
The structural half, measured here rather than quoted
All 69 RCSB entries whose polymer entity references O00255 were fetched and their _pdbx_poly_seq_scheme read, distinguishing present in the construct from observed in the coordinates. Auth numbering was proved equal to UniProt numbering by reading the residue identity back at eight positions per deposit (I360, E366, D370, V371, A385, D418, W436, G445) - never by assuming 3.
residue 371 ... 385 386 387 388 389-397 398 399 400
observed 69 68 54 9 2 0 1 39 42 (of 69)
in construct 69 68 68 27 27 27 28 31 69 69
Residues 389-397 are observed in none of the 69 deposits, and 42 of the 69 constructs do not contain 387-398 at all. The most-cited menin-MLL co-crystal, 4GQ6 at 1.55 A, carries an engineered deletion of 387-398, so residues 386 and 399 are covalently joined in that model - which is why a naive all-atom distance between them reads 1.3 A. The three lowest-energy statements that follow:
- The motif is two objects: an ordered helical N-terminal third (371-385/386) and a Gln/Ser/Gly-rich disordered C-terminal two-thirds (
EERPGEQSQGTQSQ, 387-400) that no experiment has ever resolved. - Any docking or co-folding of menin with H2A.X built on a deposited menin model is building on a chain that is missing the part the 2026 paper says does the binding.
- The alanine-substituted control peptide Pep-Mut is described only as carrying substitutions "at key interfacial residues", and those positions are in Supplementary Fig. S3C, which was not retrievable. So the residue-level interface within 371-400 is unpublished in any form I could read.
Distance to the drug site, per compound
Minimum heavy-atom distance from menin residues to the bound ligand, computed per chain pair so that a crystal-packing contact between two copies is not reported as an interface 3:
| deposit | bound species | resolution | Val371 | residues of 371-400 within 4.5 A |
|---|---|---|---|---|
| 8VA6 | ziftomenib | 1.57 A | 2.3 A | 371 |
| 8IG0 | DS-1594b | 2.6 A | 3.1 A | 371 |
| 6E1A | M-89 | 3.1 A | 3.7 A | 371 |
| 7UJ4 | revumenib | 1.96 A | 4.3 A | 371 |
| 9WKU | bleximenib | 1.9 A | 10.7 A | none |
| 3U88 | KMT2A 1-160 | 3.0 A | 3.7 A | 371, 374, 375, 378 |
| 3U86 | JUND peptide | 2.84 A | 4.0 A | 371 |
| 4GQ6 | KMT2A MBM1 12-mer | 1.55 A | 9.5 A | none |
Three of the four clinical menin inhibitors touch Val371. Bleximenib, at the same resolution, does not - a compound-level difference of 8.4 A at the same residue, which is the kind of separation hyp-partner-selective-menin-ligand-separates-mandates predicts should exist for JUND and which nobody has looked for here. The longer KMT2A 1-160 construct engages four residues of the motif where the 12-residue MBM1 peptide engages none, so how much of this motif the fusion itself occupies depends on which KMT2A construct you ask.
What this does and does not mean for a menin inhibitor
It does not mean revumenib blocks HR. A 4.3 A contact with one residue is not evidence that the drug competes a 30-residue motif off a nucleosome, and nobody has measured menin-gamma-H2A.X binding, gamma-H2A.X foci resolution, H3K36me3 at breaks or HR/NHEJ balance under any menin inhibitor in any system - see the falsifier in risk-a-menin-degrader-is-not-a-pocket-inhibitor-for-the-hr-scaffold.
It does mean the innocence argument has to be made residue by residue. The graph's existing protective reading - stated in gene-suv39h1 as "on this mapping the inhibitor's pocket and the SvID are different surfaces" - is a statement about domains. At the level of residues, the pocket and this motif share their first residue. See ana-q34-menin-surface-geometry for the adjudication.
| partner a | gene-men1 |
|---|---|
| partner b | gene-h2ax |
| motif a | menin Val371-Gln400 (O00255 numbering, 610-aa canonical), within the Palm/Fingers region as the source paper describes it |
| motif b | not mapped below domain level; the source reports only that menin binds both the N-terminal and C-terminal domains of H2A.X |
| interface residues note | NO residue inside 371-400 has been individually implicated in any readable source. The alanine-substituted control peptide Pep-Mut names the interfacial positions only in Supplementary Fig. S3C, which was not retrievable. |
| kd status | NEVER MEASURED. No ITC, SPR, FP or NMR affinity is reported for menin-H2A.X or menin-gammaH2A.X by any group. |
| method that mapped it | co-immunoprecipitation of menin domain truncations, then an internal deletion (delta-371-400), then a synthetic biotinylated 30-mer pulldown - three orthogonal perturbations, no structure |
| phospho dependence | Pep-MH precipitates BOTH wild-type H2A.X and the S139E phosphomimetic, so the isolated motif is not phospho-selective in vitro even though DNA damage promotes the interaction in cells |
| ligand note | revumenib 4.3 A, ziftomenib 2.3 A and DS-1594b 3.1 A from Val371; bleximenib 10.7 A. The contact is with ONE residue and does not by itself predict displacement of the motif. |
| druggability | Not a drug site. The functional core is a 14-residue Gln/Ser/Gly-rich segment that no structure has ever resolved, i.e. an intrinsically disordered peptide-in-groove contact of the class that resists small molecules. A stapled or cell-penetrating peptide is the demonstrated modality - the authors' own Pep-MH is one. |
| fate in fusion | not applicable - this is menin, which is not part of the fusion protein; the relevance is that every KMT2A-fusion leukaemia therapy in this class targets the same menin molecule |
| shown for fusion | NO for any fusion. The entire paper is A549 lung adenocarcinoma, Men1-knockout MEFs and Men1-floxed mice; no KMT2A-rearranged cell, no B-lineage cell and no leukaemia of any kind appears in it. |
| epigenetic readout | reads the gamma-H2A.X (H2A.X Ser139-phosphorylated) damage mark; downstream, menin recruits SETD2 to write H3K36me3 and LEDGF to read it |
| cell type | A549 human lung adenocarcinoma; mouse embryonic fibroblasts; mouse lung |
| cell state | DNA damage. The interaction is reported to be markedly promoted by DNA damage, so this is a damage-state contact rather than a constitutive one, and any attempt to detect it in an unstressed cell may find nothing. |
| relationship to existing entries | Distinct from `interface-menin-kmt2a-mbm1` and `interface-menin-jund-pocket`, which describe the central cavity from the partner side; this entry describes a menin region that ABUTS that cavity at its first residue and extends away from it. Distinct from `interface-menin-fancd2`, whose curated menin region 214-390 overlaps residues 371-390 of this motif - that overlap is stated as an open question, not resolved. The adjudication against the SUV39H1 SvID is `ana-q34-menin-surface-geometry`. |
interface residues a
[]
interface residues b
[]
domains — the numbers behind the chart
[
{
"name": "gammaH2A.X motif (menin 371-400)",
"start": 371,
"end": 400
},
{
"name": "gammaH2A.X motif, ordered helical part",
"start": 371,
"end": 386
},
{
"name": "gammaH2A.X motif, never-observed segment",
"start": 387,
"end": 400
}
]structures
[]
structures absent
[ "no deposit contains menin with H2A.X or with any nucleosome carrying H2A.X", "menin residues 389-397 are observed in ZERO of 69 O00255 deposits", "42 of 69 menin constructs delete residues 387-398 outright, including 4GQ6 at 1.55 A" ]
structural measurement 2026 09 22
{
"n_pdb_entries_queried": 69,
"observed_count_by_residue": {
"385": 68,
"386": 54,
"387": 9,
"388": 2,
"398": 1,
"399": 39,
"400": 42,
"371_to_384": 69,
"389_to_397": 0
},
"min_distance_to_ligand_angstrom": {
"ziftomenib_8VA6_V371": 2.33,
"DS1594b_8IG0_V371": 3.11,
"M89_6E1A_V371": 3.67,
"revumenib_7UJ4_V371": 4.33,
"bleximenib_9WKU_V371": 10.72,
"KMT2A_1_160_3U88_V371": 3.67,
"JUND_3U86_V371": 3.95,
"KMT2A_MBM1_12mer_4GQ6_V371": 9.52,
"LEDGF_IBD_3U88_V371": 37.91
},
"script": "analyses/q34-menin-surface-geometry/run.py"
}ligands
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib" ]
- Repair-competence toxicity: if menin is the scaffold that licenses accurate HR at breaks, then removing menin function in a child receiving etoposide, doxorubicin or mitoxantrone is removing a repair licence at exactly the moment TOP2 breaks are being made. The same argument this graph already makes against prophylactic H3K9 lowering applies here with menin as the node.
- Modality split: an MBM-pocket inhibitor leaves menin protein present and this motif mostly untouched; a menin DEGRADER removes the motif with the protein. Those are different interventions for HR and the graph should stop treating them as interchangeable - see `risk-a-menin-degrader-is-not-a-pocket-inhibitor-for-the-hr-scaffold`.
- Compound heterogeneity as a hidden variable: three of four clinical compounds contact Val371 and one does not, so a class-level statement about HR safety cannot be inferred from whichever compound was tested.
- Undruggable by design: the functional core is disordered, so anyone proposing to block this contact selectively is proposing a peptide or a degrader, not a pocket binder - and the authors' own Pep-MH suppresses HR, i.e. the tool for testing the biology is itself a potential genotoxic sensitiser.
- Which residues inside 371-400 actually touch H2A.X? Pep-Mut's alanine substitutions are the only published answer and they are in a supplementary figure that could not be retrieved; an alanine scan of the 30-mer would settle it in one experiment.
- Is the contact phospho-selective in a cell? The isolated peptide binds H2A.X and H2A.X-S139E equally, yet the cellular interaction is damage-induced - so the selectivity must come from something other than the motif itself, most simply from gamma-H2A.X abundance or from a second reader.
- Does UniProt's curated menin 214-390 FANCD2 region describe the same contact as 371-400? Both are coarse truncation maps in a DNA-damage context on overlapping menin residues. CORRECTED 2026-09-22: this question previously ended 'from papers 23 years apart that do not cite each other'. That is false. Lian 2026 (PMID 42578372) cites Jin 2003 (PMID 12874027) as its reference 18, verified in the PubMed reference list. So the two literatures are not independent, which WEAKENS the coincidence reading the question was built on. Partly answered the same day: the 214-390 region is 173 observed residues spread across all three of menin's spatial lobes and 29.9% of its accessible surface, and a random 30-mer overlaps a 177-residue region by 20 or more residues with p = 0.289 - so the overlap alone is no evidence of a shared contact. What remains genuinely open is residue-level: no experiment has mapped FANCD2 binding more finely than a 177-residue truncation.
- Does bleximenib, which does not reach Val371, behave differently from revumenib and ziftomenib in any gamma-H2A.X or HR readout? This is a free comparison in cells that already exist.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In human A549 cells and Men1-knockout MEFs, menin directly recognises histone H2A.X through its Val371-Gln400 motif: truncation of the Palm/Fingers region narrows the contact to 371-400, internal deletion of 371-400 abolishes gamma-H2A.X binding and abolishes rescue of HR after Men1 knockdown, and a synthetic 30-mer spanning 371-400 (Pep-MH) precipitates both H2A.X and the phosphomimetic H2A.X-S139E, disrupts the interaction in cells, suppresses IR-triggered HR while potentiating NHEJ, and blocks aphidicolin-induced menin binding at a common fragile site.
The tumor suppressor menin is a key scaffold mediator for homologous recombination repair. · Lian A, Zhu X, Jin B, Gu G, Xu B, Zhang L, Lin X, Zhu X, Jin G · Nucleic Acids Res 2026What this evidence cannot show
Human lung adenocarcinoma and mouse fibroblast/lung systems; no haematopoietic cell, no KMT2A fusion, no B-lineage material anywhere in the paper. The 371-400 boundary comes from a truncation series plus one internal deletion, not from a structure or a scanning mutagenesis, so it is the smallest fragment TESTED rather than the minimal fragment. Pep-MH carries an octa-arginine cell-penetrating tag, which is itself chromatin-binding and polyanion-binding, and the mutant control Pep-Mut's substituted positions are in Supplementary Fig. S3C, which was not retrievable - so no residue inside the motif is individually implicated. No affinity constant is reported for menin-H2A.X by any method. Full text read from Europe PMC fullTextXML (96,530 characters) on 2026-09-22.
Evidence that argues against this (1)
- UniProt O00255 curates menin residues 214-390 as the region sufficient for binding FANCD2, from PMID 12874027 - so a second DNA-damage-response partner is already assigned to a region that contains residues 371-390 of this motif, and neither paper cites the other. Two partners mapped to overlapping menin regions by coarse truncation is the expected result of coarse truncation, not proof of a shared site, and it is a reason to treat 371-400 as a region rather than an interface until a structure exists. UniProt Consortium 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Menin's canonical UniProt sequence O00255 is 610 aa and its residues 371-400 read VIPNLLKEAASLLEAGEERPGEQSQGTQSQ, beginning with Val371 and ending with Gln400 exactly as PMID 42578372 names them - so that paper's coordinates are in the 610-residue numbering and are directly comparable with PMID 23579270's SvID coordinates, whose residues 411 (Ala), 413 (Leu), 418 (Asp) and 436 (Trp) also read back correctly.
UniProtKB O00255 (MEN1_HUMAN), Menin, 610 aa, primary gene name MEN1 · UniProt Consortium · UniProt Knowledgebase 2026What this evidence cannot show
Fetched 2026-09-22 with gene_exact:MEN1 AND organism_id:9606 AND reviewed:true, which returned EXACTLY ONE reviewed record whose gene_primary reads back MEN1 (rules 62/65 satisfied by readback, not assumed). Note UniProt DISPLAYS the 610-residue form as canonical and describes a longer 615-residue isoform; a paper using the long numbering would be offset by five residues and this readback is the only thing that rules that out. UniProt annotates NO nuclear localisation signal on this entry and no domain features at all beyond two Regions - see `interface-menin-nls1-479-497-and-nls2-588-608`.
Evidence that argues against this (1)
- The same UniProt entry lists three isoforms, with O00255-2 displayed and O00255-1 described as 'Long'. Menin numbering in the older literature is therefore not guaranteed to be this numbering, and the only defence is residue-identity readback of the specific positions a paper names - which is what was done here and which no accession-plus-length check would have caught. UniProt Consortium 2026
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
Across all 69 RCSB deposits whose polymer entity references O00255, menin residues 389-397 are observed in zero, 387 in nine and 388 in two, while 371-384 are observed in all 69; 42 of the 69 constructs do not contain residues 387-398 at all, including 4GQ6 at 1.55 A. In the co-crystals, Val371 lies 2.3 A from ziftomenib (8VA6), 3.1 A from DS-1594b (8IG0), 3.7 A from M-89 (6E1A), 4.3 A from revumenib (7UJ4), 4.0 A from the JUND peptide (3U86) and 3.7 A from the KMT2A 1-160 chain (3U88, which also puts 374, 375 and 378 within 4.5 A), but 10.7 A from bleximenib (9WKU) and 9.5 A from the 12-residue MBM1 peptide (4GQ6).
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 against live RCSB mmCIF and live UniProt. Distances are minimum heavy-atom distances computed PER CHAIN PAIR and then minimised over pairs, so a contact between two crystallographic copies is not reported as an interface. A distance in a crystal is not an affinity and a 4.3 A contact with one residue does not establish competition. Hydrogens are absent from these models, so 'contact' means heavy atom to heavy atom. Two errors were made and corrected while computing this and are recorded in the script's docstring: entity 1 is histone H3.2 rather than menin in 8GPN, and an early all-chain minimisation produced a physically impossible 1.3 A between menin 386 and 399 in 4GQ6, which turned out to be the signature of that construct's engineered 387-398 deletion.
Evidence that argues against this (1)
- A PDB entry existing is not a structure suitable for design, and an unobserved residue is not necessarily disordered in the cell - it may be ordered only in a complex nobody has crystallised. The correct reading of 389-397 being unobserved in 69 of 69 is that the region is flexible in every state so far crystallised, all of which are free menin, menin plus a short peptide, or menin plus a small molecule. None of them contains H2A.X. Lin J et al. 2023
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB 7UJ4 is a 1.96 A crystal structure of human menin bound to SNDX-5613 (revumenib, chemical component OQ4), and PDB 8VA6 is a 1.57 A structure of wild-type menin bound to ziftomenib (component K5O) - so the Val371 contact is measured in the two deposits that correspond to the two clinically approved compounds rather than to tool molecules.
PDB 7UJ4: Inhibition of Human Menin by SNDX-5613 · RCSB Protein Data Bank depositors · RCSB Protein Data Bank 2026What this evidence cannot show
Titles, methods, resolutions and chemical-component identities fetched from data.rcsb.org on 2026-09-22. 7UJ4's menin entity is annotated to 'Isoform 2 of Menin' and 8VA5 (the other ziftomenib deposit) is a T349M resistance mutant, which is why 8VA6 was used for the wild-type number. The RCSB primary-citation title is not the PubMed title and is not used as one here.
Evidence that argues against this (1)
- Bleximenib, measured the same way in 9WKU at 1.9 A, does not reach Val371 at all (10.7 A). Any claim that 'menin inhibitors contact the gamma-H2A.X motif' is therefore a claim about three specific chemotypes and is falsified for the fourth, which is precisely why the class should not be treated as one object here. RCSB Protein Data Bank depositors 2026
that entry's range is in coordinate numbering (E3)
menin residues 371-400, the gamma-H2A.X recognition motif
direct recognition, damage-promoted; deletion of 371-400 abolishes it
shares its first residue, Val371, with the wall of the MBM cavity - 2.3-4.3 A from three of four clinical inhibitors
371-400 is nested inside 360-445, but the two sets of functional residues are 16-26 A apart
the writer arm of the same HR scaffold whose histone-reading arm is 371-400
adjudicates this motif against the SvID
do menin D418N and W436R still bind gammaH2A.X after 10 Gy?
the HR scaffold arm a pocket inhibitor leaves and a degrader removes
all 16 observed residues of the gamma-H2A.X motif fall in the fingers lobe
answers that entry's stated open question: the 20-residue overlap is what chance gives you
V371 is the first residue of the gamma-H2A.X motif
a third DNA-repair region claim, 1-228, with zero overlap with the gamma-H2A.X motif