A residue axis for menin, derived from coordinates: three spatial lobes, not four named ones, and the sharpest seam in the protein runs through the drug pocket

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.

• If the clinical menin inhibitors occupy 43-46% of their footprint in the same lobe that carries the SvID and the gamma-H2A.X motif, then some of the class's on-target effects are effects on surfaces the KMT2A displacement rationale never named. Escalating dose or residence time to deepen KMT2A displacement escalates those too.
• A parse is a design tool, and this one is from a single crystal form of a loop-deleted construct. Designing a partner-selective menin ligand against boundaries whose largest uncertainty is +/- 11 residues, at a boundary that cuts a helix, will produce a compound whose selectivity argument rests on a convention.
• Menin is a single compact domain with no separable module. A degrader or a truncation-based tool cannot remove one function and leave the others, which is the usual escape route when a scaffold protein turns out to have more jobs than the programme accounted for.

1. Does the palm|fingers seam after residue 317 move when menin is on a nucleosome? 8GPN models to 587 and the same parse could be run on it; if the seam closes or shifts, the drug pocket's three-lobe geometry is a property of the free protein only.
2. Is there a conformational reason the two clinical compounds reach further into the fingers lobe than the KMT2A peptide does, or is it just that a small rigid molecule can occupy a subpocket a peptide backbone cannot? A shape-complementarity calculation on the same three deposits would separate these.
3. Residues 462-519 are modelled in no deposit and contain NLS-1. Does the predicted structure (AF-O00255-F1) place them anywhere reproducible, and if so with what pLDDT? This analysis deliberately used only experimental coordinates and so cannot say.
4. Does the transglutaminase-like assignment of the thumb (Huang et al.'s fold-recognition claim, untested here) survive a modern structure search, and if the thumb is a degenerate transglutaminase, is the catalytic triad position inside or outside the drug pocket?

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The founding structural description of menin as four associated domains - an NTD with a long beta-hairpin, a transglutaminase-like thumb, a helical palm with three TPR motifs and a C-terminal fingers domain - gives no residue boundaries in its main text or figure legends, deferring them to a supplementary figure and supplementary text; the same paper records that an unstructured loop, residues 460-519, was genetically deleted to obtain the crystals.

   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, et al. · Nature 2012
   PMID 22327296DOI
   ▸What this evidence cannot show

   Full text read here through the PMC HTML route (45,775 characters after tag-stripping) because `fullTextXML` returned HTTP 500 for PMC3983792. The supplementary material was NOT retrieved and is not cited. The absence of numbers is therefore bounded to the main text, figure legends and Methods of the article body - the boundaries may well be stated in Supplementary Fig. 4, which is exactly the problem, because nobody in this graph can read it.

   ▸Evidence that argues against this (1)

   • A published table of numeric MEN1 domain coordinates does exist, described as crystal-structure-identified, so the claim is about retrievability rather than about whether anyone ever wrote numbers down. Its rendering is corrupt and its isoform frame is internally inconsistent - see E4 - but a reader with the print edition may have usable boundaries this analysis did not. Nelakurti DD et al. 2020
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset

   Across all 69 O00255 PDB deposits with per-residue numbering verified against the UniProt sequence, 83 of menin's 610 residues are observed in no deposit; a 5 A contact graph on the best-covered deposit 3U84 (505 verified residues) clusters into three substantial spatial lobes plus a detached C-terminal helix with 92.8% of contacts internal, the NTD and palm falling in one lobe; the deepest sequence-contiguous boundary is after residue 317, cutting 64 contacts against a random-partition median of 109; and the revumenib and ziftomenib footprints share only 40.5% and 41.0% of their union with the KMT2A MBM1 peptide footprint while placing 43.5% and 46.2% of their contacts in the fingers lobe against the peptide's 17.2%.

   analyses/q36-menin-residue-axis/run.py - contact-graph domain parse, solvent accessibility and ligand footprints over all 69 RCSB deposits referencing UniProt O00255, run 2026-09-22 · RCSB PDB and UniProt, computed 2026
   ▸What this evidence cannot show

   `strong`, not `established`, and the reasons are structural rather than statistical. (1) The parse is of ONE crystal form of a construct missing residues 460-519; a different conformer could move a boundary. (2) A contact-graph cut is a statement about packing, not about folding independence - no thermodynamic or evolutionary criterion was applied. (3) The 40-residue minimum block size in the contiguous parse is a choice, and the k=4 contiguous solution ([2-317], [318-439], [440-552], [553-608]) is NOT the four named domains; its last two blocks are the sparsely-modelled C-terminal half and the program is exploiting that sparsity, which is why the spectral parse is the one reported. (4) The NTD|thumb boundary cuts helix 82-101 in half and should be read as a convention. (5) Solvent accessibility is computed on the isolated chain with no partners bound. (6) Deposits are not independent: 27 of the 69 share a construct and many are the same crystal form with different ligands, so 'observed in n of 69' is a statement about the deposit set, not about n independent experiments.

   ▸Evidence that argues against this (1)

   • A cryo-EM structure of menin on a nucleosome exists and is a different conformational state from the crystal form parsed here; if menin's lobes reorient on chromatin, a parse built from 3U84 describes the free protein and may not describe the chromatin-bound one. That deposit, 8GPN, models to residue 587 but is not the best-covered deposit and was not used as the reference frame. Lin J et al. 2023
3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   The 2026 paper that maps the menin-gamma-H2A.X contact describes the mapped region qualitatively as lying in the Palm and Fingers domains and locates its truncation constructs in a supplementary figure, giving as its only numeric coordinates the 371-400 motif itself, the Pep-MH peptide covering 371-400 and the Pep242 peptide covering 231-248.

   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, et al. · Nucleic Acids Research 2026
   PMID 42578372DOI
   ▸What this evidence cannot show

   Full text read here (96,530 characters from Europe PMC `fullTextXML` for PMC13458390). The supplementary figures S2M and S2N that contain the truncation construct boundaries were not retrieved, so what the authors' Palm and Fingers constructs actually spanned is unknown. Their system is A549 lung adenocarcinoma and Men1-knockout mouse embryonic fibroblasts, not a KMT2A-rearranged haematopoietic cell, and nothing in the paper is a leukaemia experiment.

   ▸Evidence that argues against this (1)

   • That paper's own domain language is inherited, not measured: it cites the nucleosome cryo-EM paper for 'menin binds to nucleosomes through its Palm and Fingers domains', so both papers use a vocabulary whose boundaries neither of them states. A reader who takes 'Palm and Fingers' as a coordinate has been handed a name twice and a number never. Lin J et al. 2023
4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

   A 2020 review tabulates numeric coordinates for the NTD, thumb, palm and fingers as crystal-structure-identified domains, but the served text of that table is corrupt - the cells read 1-100, 101-1230, 231-1300, 301-1305, 479-1497, 588-1608 and 219-1395 - and the numbering frame is internally inconsistent between the 610-residue and 615-residue isoforms.

   Comprehensive Analysis of MEN1 Mutations and Their Role in Cancer. · Nelakurti DD, Pappula AL, Rajasekaran S, Miles WO, Petreaca RC · Cancers 2020
   PMID 32937789DOI
   ▸What this evidence cannot show

   `moderate` because the source cannot be read as intended. The corruption was confirmed at the byte level - the Europe PMC `fullTextXML` for PMC7565326 serves the literal bytes `101\xe2\x80\x931230`, a valid UTF-8 en dash followed by `1230` - so this is not a decoding fault at this end. The publisher's own HTML at mdpi.com returned HTTP 403 and the PMC HTML page returned a 20,406-character bot-detection stub, so the printed table was never seen. The decoding rule (a spurious `1` follows the dash) is confirmed on three rows whose true values are independently known and FAILS on the Fingers row, which is the row this beat most wanted. Nothing in this analysis's parse depends on this citation.

   ▸Evidence that argues against this (1)

   • The same review's own body text places the palm|fingers boundary elsewhere than its table does - it says the P325L mutation 'sits right at the boundary of the palm and fingers domains' and notes that P325 in its numbering is P320 in the 610-residue isoform, i.e. a boundary near 320 rather than near 300. The internal disagreement is the reason this is evidence about the state of the literature and not evidence about menin. Biancaniello C et al. 2022
5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   Menin's four-domain description is used with a fifth name in at least one source, which calls the N-terminal domain the heel, and maps isoform-specific insertions and deletions onto the thumb and palm without giving those domains numeric boundaries either.

   Alternative splicing generates isoform diversity in MEN1. · Ramachandran A, Le Quesne Stabej P, Boyle V, Elston MS, Pattison S, Lawrence B, et al. · Endocrine Oncology 2024
   PMID 39649118DOI
   ▸What this evidence cannot show

   Bibliographic fields taken from a PubMed esummary fetched in this session and the DOI cross-checked against `efetch_doi()`, which reads only the article's own ArticleIdList. The substantive point is a vocabulary observation, not a measurement: menin's N-terminal domain is called the NTD by Huang et al. and the heel here, so a beat searching on one name will miss the other. The isoform mapping in that paper is a prediction onto a crystal structure, not a new structure.

   ▸Evidence that argues against this (1)

   • A naming variant in one paper is not evidence that the field is divided; 'heel' may simply be that group's usage, and no count of how often each term is used was made here. Huang J et al. 2012

• related to →MEN1 / menin — the adaptor the fusion cannot work without

  supplies the structural residue axis gene-men1's data.domains explicitly says it does not have

• supports →ADJUDICATION: the SvID and the gamma-H2A.X motif are OVERLAPPING BUT NOT THE SAME SURFACE - and the premise that made the question urgent, that the SvID is distinct from the MLL cavity, is FALSE at the SvID's own boundary

  re-runs the coverage sweep chain-wide with per-residue numbering verification; the 371-400 counts agree

• supports →The SvID, menin 360-445: a truncation boundary the field reads as a domain, whose functional residues D418 and W436 are 10-20 A from every menin drug - but whose first ten residues ARE the drug pocket

  all 71 observed SvID residues fall in one spatial lobe (the fingers)

• supports →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

  all 16 observed residues of the gamma-H2A.X motif fall in the fingers lobe

• related to →The menin acidic shelf, Glu366-Asp370-Val371: the counter-charge that MBM2 binds, resolved in 3U88 after all - and the one patch where the drug pocket, the SvID and the gamma-H2A.X motif all touch

  places E366/D370/V371 on the fingers side of the deepest seam in menin, not in the palm

• refutes →Menin's two nuclear localisation signals, 479-497 and 588-608: mapped in 1998, verified residue-for-residue today, annotated by UniProt nowhere - NLS-1 in no structure at all, NLS-2 resolved in one

  3U84 models menin 596-608 as a helix, so NLS-2 is NOT outside every coordinate set; the NLS-1 claim stands

• related to →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

  8GPN is a different conformer and models to residue 587; the parse here is of the free protein

• related to →The menin–KMT2A grip (MBM1), the interface the drugs actually block

  the KMT2A MBM1 footprint is 29 menin residues, 17.2% of them in the fingers lobe

• related to →Menin-SETD2, mapped to the NTD and Palm and tooled with a peptide at menin 231-248: the third partner assigned to the same half of menin, and the writer arm of the HR scaffold

  Pep242 covers menin 231-248, which this parse places at the thumb-palm boundary rather than inside either

• ← depends onThe menin small-molecule pocket as a coordinated surface: four clinical-stage compounds, 21-26 contact residues each, and bleximenib sitting on a different part of it from the other three

  lobe assignment of every pocket residue comes from that parse

• ← depends onUniProt's menin 214-390 FANCD2 region is not a surface: it spans three spatial lobes and 29.9% of menin's solvent-accessible area, so its twenty-residue overlap with the gamma-H2A.X motif is what chance gives you

  the lobe assignment and the accessibility numbers come from that parse

• ← depends onMenin stimulates homology-directed repair through Chk1 and the region required is 1-228: a 2010 paper that is cited in no entry of this graph and in no reference list of the 2026 paper that reports the same conclusion

  lobe composition of the 1-228 region comes from that parse

• supports →The menin small-molecule pocket as a coordinated surface: four clinical-stage compounds, 21-26 contact residues each, and bleximenib sitting on a different part of it from the other three

  supplies the lobe decomposition of every pocket residue