Menin-JunD: the same pocket, the same drug, the opposite job
also called menin-JunD interface · JunD MBM · shared menin peptide pocket
The groove on menin that leukaemia drugs plug is the same groove that holds JunD, a growth-driving protein that menin keeps switched off in hormone glands. One groove, two jobs, and a drug cannot tell them apart.
The same thing, for a scientist
Human menin has one deep peptide-binding pocket. Crystal structures of the same protein bind an MLL1 peptide (PDB 3U85) and a JunD peptide (PDB 3U86) in the same manner and at the same site, with opposite transcriptional consequences: menin-MLL1 promotes transcription, menin-JunD blocks JNK-mediated JunD phosphorylation and suppresses JunD-induced transcription [E1][E2][E3]. Every clinical menin inhibitor is a ligand for this pocket, occupying the MBM1 site. It follows structurally - though it has never been shown pharmacologically for any approved compound - that a menin inhibitor should displace JunD as readily as it displaces MLL1, and that the leukaemic cofactor function and the endocrine tumour-suppressor function cannot be separated by dose or schedule. Several MEN1 missense mutations found in patients disrupt the menin-JunD interaction specifically, which is what links this interface to tumour suppression [E1][E4].
Tell me more about this picture
PDB 3U86, X-ray at 2.84 A; the deposited JunD peptide is SPGRLFPGAPPTAKK, approximately JunD residues 27-38.
Tell me more about this picture
The companion structure 3U85 shows the same menin pocket holding the MLL1 MBM1 peptide SRWRFPARPGTTGGGGGGGRR, with RWRFP at KMT2A residues 6-10.
Tell me more about this picture
Both peptides carry an aromatic-proline anchor, which is consistent with the published finding that they bind in the same manner.
The structural evidence, as deposited
| PDB | Complex | Method | Resolution | Peptide as deposited |
|---|---|---|---|---|
| 3U84 | human menin, apo | X-ray | - | - |
| 3U85 | human menin + MLL1 | X-ray | 3.0 A | SRWRFPARPGTTGGGGGGGRR (MBM1, RWRFP = KMT2A 6-10) |
| 3U86 | human menin + JunD | X-ray | 2.84 A | SPGRLFPGAPPTAKK (approx. JunD 27-38) |
| 3U88 | human menin + MLL1 + LEDGF | X-ray | - | - |
Sources: RCSB entity records, read on 2026-09-18 23. Both peptides carry an aromatic-proline anchor - RWRFP and RLFPG - which is consistent with the published conclusion that they bind in the same manner 1.
The fourth structure, 3U88, is worth noting alongside: the same pocket region also accommodates LEDGF/PSIP1 in a ternary complex with MLL1, which is the tethering mechanism this project's therapeutic hypothesis runs through (mechanism-menin-ledgf-chromatin-tethering).
What follows, and what does not
Follows from the structures. One pocket serves two partners with opposite transcriptional outputs. A competitive pocket ligand has no structural basis on which to discriminate. Any claim that a menin inhibitor is 'selective for the leukaemic function' is therefore a claim about kinetics or cellular context, not about structure, and needs its own evidence.
Does not follow. That a given inhibitor at a given clinical concentration actually displaces JunD. Relative affinities matter: if menin-JunD is substantially tighter than menin-MBM1, a drug titrated to block MLL1 could leave JunD largely bound. Neither the Kd for menin-JunD nor the displacement of JunD by any approved inhibitor has been published, which is the largest single hole in this argument and a one-experiment hole.
Also does not follow. That displacing JunD is harmful on a therapeutic timescale. Relief of menin repression is what MEN1 mutations do over decades of germline haploinsufficiency; a reversible drug does it for months. See risk-menin-inhibitor-exposure-duration-vs-men1-latency.
| partner a | gene-men1 |
|---|---|
| partner b | gene-jund |
| motif b | JunD approx. aa 27-38, core LFPG; crystal peptide SPGRLFPGAPPTAKK |
| comparator motif | KMT2A MBM1 RWRFP (aa 6-10); crystal peptide SRWRFPARPGTTGGGGGGGRR (PDB 3U85) |
| druggability | the pocket is the validated drug site; the problem is not whether it can be occupied but that occupying it cannot distinguish the two partners on any structural basis |
| displacement of jund by approved inhibitors | never published for revumenib, ziftomenib or bleximenib |
structures
[ "3U86", "3U84", "3U85", "3U88" ]
ligands
[ "drug-revumenib", "drug-ziftomenib", "drug-bleximenib" ]
transcriptional consequence
{
"menin_MLL1": "promotes transcription",
"menin_JunD": "blocks JNK-mediated JunD phosphorylation and suppresses JunD-induced transcription"
}- This interface is the reason the endocrine liability cannot be dosed away, and it is also the reason the concern might be empty: nobody has shown that any approved menin inhibitor displaces JunD in a cell at clinical exposure.
- The structures establish inseparability of binding, not equivalence of consequence. MLL1 displacement kills leukaemia cells in days; JunD displacement would have to be sustained for a long time in a slowly dividing endocrine tissue to matter.
- A partner-selective ligand would dissolve the problem, and the same structures that raise the concern are the ones that would be used to attempt it.
- What is the Kd of menin-JunD relative to menin-MBM1? This single number decides whether a therapeutic concentration of a menin inhibitor displaces JunD at all.
- Do the MEN1 patient missense mutations that disrupt menin-JunD sit on the inhibitor-contact surface? Both structures and the mutation list are public; the superposition has not been reported.
- Can a ligand occupy the pocket in an MBM1-blocking orientation while sparing JunD? The two peptides bind in the same manner, so this may be impossible - and a rigorous demonstration that it is impossible would be as valuable as a demonstration that it is not.
- Does menin degradation reported downstream of menin-inhibitor treatment remove JunD repression more completely than occupancy alone? Degradation removes every menin function, not only pocket-mediated ones.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Human menin contains a single deep peptide-binding pocket that binds short peptides of MLL1 and of JUND in the same manner but with opposite effects on transcription: menin-JUND blocks JNK-mediated JUND phosphorylation and suppresses JUND-induced transcription, while menin-MLL1 promotes transcription; several MEN1 missense mutations disrupt the menin-JUND interaction, and menin loss or inactivation causes multiple endocrine neoplasia type 1.
The same pocket in menin binds both MLL and JUND but has opposite effects on transcription · Huang J, Gurung B, Wan B, Matkar S, Veniaminova NA, Wan K, Merchant JL, Hua X, Lei M · Nature 482:542-6 2012What this evidence cannot show
Read as abstract. Crystal structures and cell-based transcription assays. The paper does not test any menin inhibitor and measures no endocrine outcome after pharmacological occupancy; the inference from a shared pocket to clinical endocrine risk is this graph's, not the paper's.
Evidence that argues against this (1)
- The same shared pocket is the basis of a proposal in this graph for a partner-selective menin ligand that separates the two mandates. If such a ligand is achievable, the inseparability argued here is an artefact of current chemistry rather than of structure; if it is provably unachievable, that is equally worth knowing. Either way the structures alone do not settle the pharmacology. Huang J et al. 2012
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB entry 3U86 is the 2.84 A X-ray structure of human menin in complex with a human JunD peptide, sequence SPGRLFPGAPPTAKK.
PDB 3U86: Crystal structure of human menin in complex with JunD · Huang J, Gurung B, Wan B, Matkar S, Veniaminova NA, Wan K, Merchant JL, Hua X, Lei M · RCSB Protein Data Bank 2012What this evidence cannot show
Entity sequence, method and resolution retrieved from the RCSB data API on 2026-09-18. Interface residue contacts were not extracted from the coordinates; the claim that this is the same pocket as the MLL1 site rests on the Huang 2012 paper, not on an independent structural superposition performed here.
Evidence that argues against this (1)
- A peptide-bound structure at 2.84 A defines the binding mode. It carries no affinity, no stoichiometry and no information about occupancy in a cell, so the shared-pocket inference is structural rather than quantitative. PNAS 2003
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB entry 3U85 is the 3.0 A X-ray structure of human menin in complex with an MLL1/KMT2A MBM1 peptide, sequence SRWRFPARPGTTGGGGGGGRR; the companion depositions are 3U84 (apo human menin) and 3U88 (human menin with MLL1 and LEDGF).
PDB 3U85: Crystal structure of human menin in complex with MLL1 · Huang J, Gurung B, Wan B, Matkar S, Veniaminova NA, Wan K, Merchant JL, Hua X, Lei M · RCSB Protein Data Bank 2012What this evidence cannot show
Entity sequences, methods and resolutions retrieved from the RCSB data API on 2026-09-18. 3U84 and 3U88 titles were read from the same API but their entity details were not extracted.
Evidence that argues against this (1)
- At 3.0 A the MLL1 peptide complex resolves backbone and pocket contacts, not water structure or side-chain rotamers at the precision needed to reason about relative affinity. Paired with the JunD structure it establishes a shared site; neither gives the occupancy ratio that a displacement argument needs. Nature 2012
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
A yeast two-hybrid screen identified JunD as a direct menin partner, menin did not bind other Jun or Fos family members directly, menin repressed JunD-activated transcription, and several naturally occurring clustered MEN1 missense mutations disrupted the menin-JunD interaction.
Menin interacts with the AP1 transcription factor JunD and represses JunD-activated transcription · Agarwal SK, Guru SC, Heppner C, Erdos MR, Collins RM, Park SY, Saggar S, Chandrasekharappa SC, Collins FS, Spiegel AM, Marx SJ, Burns AL · Cell 96:143-52 1999What this evidence cannot show
Read as abstract. Predates the structures; the mutations are described as clustered but the abstract does not map them onto pocket residues, so the correspondence between MEN1 patient mutations and the inhibitor-contact surface remains unverified here.
Evidence that argues against this (1)
- The specificity claim - menin binding JunD but not other Jun or Fos members - is a NEGATIVE result in a yeast two-hybrid screen, the assay class most prone to false negatives from folding and expression. Later structural work confirmed the JunD interaction but did not test the others, so the exclusivity rests on the 1999 screen alone. Cell 1999
the JunD side of the shared menin peptide pocket, crystal peptide SPGRLFPGAPPTAKK (PDB 3U86)
one deep peptide-binding pocket serving two partners with opposite transcriptional outputs
same pocket, same binding manner; PDB 3U85 (MLL1) and 3U86 (JunD) are companion depositions of the same protein
a pocket ligand has no structural basis on which to discriminate MLL1 from JunD; displacement of JunD by any approved inhibitor has never been published
the structures that raise the inseparability problem are the ones a partner-selective ligand would have to defeat
only the direct-binding arm of the repression is clearly pocket-mediated, which is the arm a drug removes
- ← depends onHypothesis: the PTH rise on menin inhibitors is on-target and exposure-dependent, not renal
the hypothesis requires that a clinical inhibitor actually displaces JunD, which has never been shown
tests the untested pharmacological assumption the entire beat rests on
JunD's equivalent C-terminal basic pair is ordered where KMT2A's Arg24/Arg25 are not
- ← analogous toMenin-FANCD2: one 2003 co-immunoprecipitation, a curated sufficient region of menin 214-390, and nothing since
a second non-KMT2A menin partner interface; unlike JunD, no evidence it uses the drug pocket
- ← depends onHow menin holds the beta cell still: JunD recruits it, with HDAC3, onto the Pbk promoter
MI-503 and MI-463 disrupt menin-JunD co-IP at 30-300 nM, i.e. the same pocket occupancy, with a proliferative rather than neoplastic readout in islet tissue
- ← refutesA second brake, unconnected to JunD: menin is a TEAD1 corepressor with VGLL4 in the beta cell
Menin restrains the beta cell through at least two independent partners (JunD and TEAD1); the single-pocket account of endocrine liability is incomplete, and whether the TEAD1 surface is drug-occupiable is untested
the same shelf is 2.6-4.0 A from the JUND peptide in 3U86
the same cavity accepts JunD; this entry coordinates the small-molecule footprint specifically
- ← depends onJUND liberation — the drug's mechanism of action and one of its mechanisms of resistance are the same event
the single shared peptide pocket is what makes this mechanism obligatory