Oncofusion

start / graph / interface-menin-ledgf-ibd-menin-side-91-105

Interfacestrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q34-menin-surface-map

The menin side of the LEDGF clamp: residues 91-105 and 209-212, on the opposite face of menin from the drug pocket and 26-38 A from every other surface in this beat

also called menin-LEDGF interface · menin 91-105 · 3U88 menin-IBD contact

LEDGF, the protein that anchors the KMT2A fusion to chromatin, grips menin on a patch made of two short stretches near menin's start. That patch is on the far side of menin from where the drugs bind, which is why a menin drug is not expected to break the LEDGF half of the clamp directly.

The same thing, for a scientist

In PDB 3U88 (menin + a 75-residue engineered KMT2A chain + the LEDGF/PSIP1 integrase-binding domain, X-ray 3.0 A) the LEDGF IBD is modelled at PSIP1 residues 348-430 and contacts menin through exactly two short stretches: menin 91-105 (fifteen residues, the dominant patch) and menin 209-212. In the same coordinate frame the LEDGF IBD is 37.9 A from menin Val371, 26.6 A from Asp418, 27.8 A from Trp436 and 22.8 A from Ile360 - the farthest of any partner measured in this beat. Fourteen LEDGF residues are within 5 A of menin (378, 384, 410, 411, 414, 415, 417, 418, 421, 422, 424, 425, 428, 429) and there are 24 LEDGF-KMT2A residue pairs within 5 A in the same structure, so the IBD bridges both partners simultaneously. This entry supplies the menin side of a clamp whose KMT2A side this graph already holds.

What it looks like

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.

Go deeper
Why the menin side was missing

This graph holds interface-psip1-ibd-kmt2a, interface-psip1-ibd-kmt2a-ibm1-ibm2, interface-psip1-ibd-kmt2a-phosphoswitch and interface-psip1-ibd-hiv1-integrase-competition - four entries about the IBD, all of them describing what the IBD grips on the other protein. None records which menin residues the IBD touches, and 3U88 has contained that information since 2012.

Measured from the deposited coordinates, per chain pair 1:

sideresidues within 5 A
menin91, 92, 94, 95, 96, 98, 99, 100, 101, 102, 103, 104, 105, 209, 210, 212
LEDGF/PSIP1 IBD378, 384, 410, 411, 414, 415, 417, 418, 421, 422, 424, 425, 428, 429

Identities read back against the fetched O00255 and O75475 sequences; the IBD chain begins M348-D349-S350-R351-L352-Q353, matching PSIP1 exactly.

The distance that matters for this beat

The LEDGF contact is the control surface for every distance argument in this beat. From the LEDGF IBD to the other menin surfaces, in the same structure 1:

menin I360 (SvID boundary)      22.8 A
menin D418 (SUV39H1-critical)   26.6 A
menin W436 (SUV39H1-critical)   27.8 A
menin V371 (gammaH2A.X motif)   37.9 A

So when this beat reports that D418 is 16.6 A from revumenib and calls that far, the calibration is that a genuinely unrelated surface on the same protein is 27 A away. 16 A is far, but it is not as far as separate. That is the honest reading and it is why the adjudication in ana-q34-menin-surface-geometry does not claim the SvID and the drug site are independent.

Facts and figures
partner agene-men1
partner bgene-psip1
motif amenin 91-105 (dominant patch) and 209-212
motif bLEDGF/PSIP1 integrase-binding domain; contacting residues in 3U88 are 378, 384, 410, 411, 414, 415, 417, 418, 421, 422, 424, 425, 428, 429
kd NM1400
kd status1,400 nM is the ITC affinity of the IBD for the PREFORMED menin-KMT2A 1-160 complex (PMID 25305204), carried from `interface-psip1-ibd-kmt2a`; no affinity has been measured for the menin-IBD contact in isolation.
method that mapped itX-ray crystallography of the ternary complex (3U88), residue contacts computed here from the deposited coordinates
ligand noteNo menin inhibitor contacts any of menin 91-105 or 209-212 in any of the five inhibitor co-crystals measured.
druggabilityThe IBD side is already a chemical-matter target in this graph (`target-ledgf-pwwp-fragment-chemistry-and-its-ceiling`, `interface-psip1-ibd-hiv1-integrase-competition`). The MENIN side of the same contact has never been proposed as one, and a 15-residue contiguous patch is at least a conventional shape.
shown for fusionThe ternary structure uses a spliced KMT2A N-terminal construct, not a fusion protein. No fusion co-structure exists.
relationship to existing entriesComplements the four existing PSIP1/IBD interface entries, all of which describe the IBD's grip on KMT2A or on HIV-1 integrase and none of which records the menin-side residues. It also corrects the modelled spans both `interface-psip1-ibd-kmt2a` and `interface-menin-kmt2a-mbm2` give for 3U88; reported to the coordinator rather than edited.
interface residues a
[
  "menin 91",
  "menin 92",
  "menin 94",
  "menin 95",
  "menin 96",
  "menin 98",
  "menin 99",
  "menin 100",
  "menin 101",
  "menin 102",
  "menin 103",
  "menin 104",
  "menin 105",
  "menin 209",
  "menin 210",
  "menin 212"
]
interface residues b
[
  "PSIP1 378",
  "PSIP1 384",
  "PSIP1 410",
  "PSIP1 411",
  "PSIP1 414",
  "PSIP1 415",
  "PSIP1 417",
  "PSIP1 418",
  "PSIP1 421",
  "PSIP1 422",
  "PSIP1 424",
  "PSIP1 425",
  "PSIP1 428",
  "PSIP1 429"
]
domains
[
  {
    "name": "menin LEDGF-binding patch I",
    "start": 91,
    "end": 105
  },
  {
    "name": "menin LEDGF-binding patch II",
    "start": 209,
    "end": 212
  }
]
structures
[
  "3U88"
]
distance to other menin surfaces angstrom
{
  "to_menin_I360": 22.8,
  "to_menin_D418": 26.61,
  "to_menin_W436": 27.81,
  "to_menin_V371": 37.91,
  "why_this_is_here": "this is the calibration for every 'far from the drug' claim in this beat - a genuinely separate surface on menin is 23-38 A away, so 10-17 A is near, not separate"
}
ligands
[]
How this could bite us back
  • Cooperativity means the clamp can fail from either side, so a LEDGF-directed agent and a menin-directed agent are not independent interventions and combining them may be less additive than the distance suggests.
  • The menin-LEDGF patch survives every pocket inhibitor, so a fusion complex that keeps its LEDGF anchor after the MBM contact is blocked is a structurally specified residual-tethering mechanism, alongside the MBM2 one.
  • A menin degrader removes this patch too, and LEDGF has its own essential roles in HR end resection - so removing the menin-LEDGF contact in a child on genotoxic therapy compounds the same repair concern this beat raises for the gamma-H2A.X motif.
Still open
  1. Is the menin 91-105 patch druggable? It is a contiguous fifteen-residue surface engaging a folded IBD, and nobody has asked - all LEDGF-directed chemistry in this graph aims at the IBD or the PWWP.
  2. Does a menin pocket inhibitor reduce LEDGF recruitment indirectly, through the cooperativity Murai et al. measured, even though the two surfaces are 23-38 A apart? That is a testable prediction of the 10-fold cooperative enhancement and it has never been tested with a clinical compound.
  3. Do the two existing entries' differing accounts of 3U88's KMT2A span matter for any claim built on them? Both describe the construct rather than the model, and the model is what a figure draws.
Every source behind this entry (2)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    In PDB 3U88 the LEDGF/PSIP1 IBD is modelled at residues 348-430 and contacts menin through residues 91-105 and 209-212 only; the same IBD is 22.8 A from menin Ile360, 26.6 A from Asp418, 27.8 A from Trp436 and 37.9 A from Val371, and makes 24 residue pairs within 5 A with the KMT2A chain in the same structure.

    What this evidence cannot show

    Computed 2026-09-22 from live RCSB mmCIF, minimum heavy-atom distances at a 5.0 A cutoff, per chain pair. 3U88 is 3.0 A, so the residue list is reliable and individual distances are not. The KMT2A entity in 3U88 is an engineered chain splicing three separated segments of Q03164, so this is a model of the assembled state rather than a measurement of association. A contact list from one crystal form is one state.

    Evidence that argues against this (1)
    • Murai et al. report that the menin-MLL-LEDGF assembly is cooperative - the IBD binds the preformed menin-MLL 1-160 complex ~10-fold tighter (1.4 uM by ITC) than it binds MLL 1-160 alone (14.7 uM by NMR) - so the menin-IBD contact is not independent of the KMT2A contact even though the two surfaces are 23-38 A apart. Distance does not equal thermodynamic independence. [Citation identifiers - PMID, title, DOI, venue, year - re-fetched from PubMed on 2026-09-22 rather than copied from the sibling entry that carries this result; the two affinity NUMBERS (1.4 uM ITC, 14.7 uM NMR) are inherited from `interface-psip1-ibd-kmt2a` and were not read from the paper by this beat.] Murai MJ et al. 2014
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    PDB 3U88 is a 3.0 A crystal structure titled 'Crystal structure of human menin in complex with MLL1 and LEDGF', containing three polymer entities: menin (O00255, 550-residue construct), KMT2A (Q03164, 75-residue construct) and LEDGF/PSIP1 (O75475, 89-residue construct), in two copies each.

    PDB 3U88: Crystal structure of human menin in complex with MLL1 and LEDGF · RCSB Protein Data Bank depositors · RCSB Protein Data Bank 2026
    What this evidence cannot show

    Entity descriptions, accessions, construct lengths, resolution and method fetched from data.rcsb.org on 2026-09-22. The RCSB primary-citation title is not the PubMed title and is not used as one. PSIP1 accession O75475 confirmed live, primary gene name reads back PSIP1, 530 aa.

    Evidence that argues against this (1)
    • This graph's `interface-psip1-ibd-kmt2a` records 3U88's contents as 'MEN1 2-550 + KMT2A 103-153 + PSIP1 347-435' while `interface-menin-kmt2a-mbm2` records the KMT2A chain as mapping to '5-15 + 23-39 + 109-153'. Measured from the coordinates today, the KMT2A chain is MODELLED at 6-15, 23-39 and 107-135 and the PSIP1 chain at 348-430 - so both existing entries describe the construct rather than the model, and they disagree with each other about the KMT2A span. Q34-menin-surface-map 2026
Connected to