Menin residue numbers: the resistance literature (G331R) and the crystal structures (residue 326) disagree by exactly five
also called MEN1 615 vs 610 numbering · G331R is residue 326 · menin numbering offset of 5
The menin resistance mutations everyone quotes, such as G331R, are numbered on a protein five amino acids longer than the one in the crystal structures, so the same change is residue 326 in a structure file. Looking up the quoted number in a structure returns the wrong amino acid.
The same thing, for a scientist
UniProt O00255 has three isoforms. O00255-1 is 615 aa and carries M327, G331 and T349 - the residues the resistance literature names. The canonical entry O00255-2 is 610 aa, and PDB coordinate files number on it: there the same residues are 322, 326 and 344, and positions 327, 331 and 349 hold Y, N and Q. This was tested both ways for all six hotspots the corpus names (M327 (residue 322 in coordinates), G331 (residue 326 in coordinates), T349 (residue 344 in coordinates), C334 (residue 329 in coordinates), E368 (residue 363 in coordinates), V372 (residue 367 in coordinates)); every one is a residue of the stated identity at literature number in the 615-aa sequence AND at number minus five in the 610-aa sequence. 9C4S differs from wild-type 4GQ6 by one substitution, Gly->Arg, at residue 326, though RCSB titles it G331R.
What goes wrong if you do not know this
A pocket-contact table, a variant-effect lookup or a structure-based distance computed by reading the quoted residue number off a PDB file returns Tyr for M327, Asn for G331 and Gln for T349, and the analysis proceeds with no error. Nothing in the file marks the mismatch. The correction is to subtract five, and to say which numbering a number is in whenever it is written down.
What was actually measured
Residue identity was read back from the fetched UniProt sequences at all six literature positions in both isoforms (table in data), and the Gly->Arg change in 9C4S was located by superposing it on 4GQ6 and comparing every residue: the only same-number mismatch in chain A is residue 326. Why the literature uses the longer numbering is inferred from that concordance, not read from the resistance papers' methods sections.
| offset | 5 |
|---|---|
| isoform 615 | O00255-1 |
| isoform 610 canonical and PDB numbering | O00255-2 |
hotspots tested both ways — the numbers behind the chart
[
{
"literature": "M327",
"isoform_615_has": "M",
"coordinate_number": 322,
"isoform_610_at_that_number": "M",
"isoform_610_at_literature_number": "Y",
"consistent_with_offset_5": true
},
{
"literature": "G331",
"isoform_615_has": "G",
"coordinate_number": 326,
"isoform_610_at_that_number": "G",
"isoform_610_at_literature_number": "N",
"consistent_with_offset_5": true
},
{
"literature": "T349",
"isoform_615_has": "T",
"coordinate_number": 344,
"isoform_610_at_that_number": "T",
"isoform_610_at_literature_number": "Q",
"consistent_with_offset_5": true
},
{
"literature": "C334",
"isoform_615_has": "C",
"coordinate_number": 329,
"isoform_610_at_that_number": "C",
"isoform_610_at_literature_number": "V",
"consistent_with_offset_5": true
},
{
"literature": "E368",
"isoform_615_has": "E",
"coordinate_number": 363,
"isoform_610_at_that_number": "E",
"isoform_610_at_literature_number": "A",
"consistent_with_offset_5": true
},
{
"literature": "V372",
"isoform_615_has": "V",
"coordinate_number": 367,
"isoform_610_at_that_number": "V",
"isoform_610_at_literature_number": "I",
"consistent_with_offset_5": true
}
]PDB 9C4S vs 4GQ6
{
"mutated_residue_in_coordinates": 326,
"wild_type": "GLY",
"mutant": "ARG",
"residue_331_in_both": "ASN",
"ca_rmsd_A": 0.153,
"ca_atoms_aligned": 433
}- A resistance residue looked up by its literature number in a PDB file or a UniProt-canonical-indexed table returns the wrong amino acid, so a structure-based explanation of resistance built that way would be silently wrong. Subtract five, and state the numbering whenever a menin residue number is written down.
- Which numbering does each source paper use? 44 entries name these mutations by their literature numbers; none was found mapping them onto coordinates, but a systematic read of each source's methods was not done.
- Does the +5 hold for menin residues outside 322-367 that the corpus cites, such as the MBM-binding groove or the H2AX motif? Only the tested stretch is claimed.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt O00255-1 (615 aa) carries Met327, Gly331, Thr349, Cys334, Glu368 and Val372, the residues named in the menin-inhibitor resistance literature, while the canonical O00255-2 (610 aa) carries those same residues five positions earlier, at 322, 326, 344, 329, 363 and 367, and holds Y, N and Q at positions 327, 331 and 349.
UniProtKB O00255 (MEN1_HUMAN) isoforms O00255-1, -2 and -3 · UniProtKB 2026What this evidence cannot show
COMPUTED from the fetched sequences, not typed: all six hotspots pass the two-way test (identity at the literature number in the 615-aa sequence and at number minus five in the 610-aa sequence), so the offset is constant across the 322-367 stretch tested, not extrapolated beyond it. WHY the literature uses the 615-aa numbering is inferred from that concordance and not read from the papers' methods; individual papers may differ.
Evidence that argues against this (1)
- Isoform O00255-3 (574 aa) is NOT the literature's numbering - its residue 331 is not Gly and 327 is not Met - which rules out one alternative but not a third convention used by some individual paper. Three concordant residues make +5 overwhelmingly likely; they are not a reading of any paper. UniProtKB 2026
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
In the deposited coordinates, 9C4S differs from wild-type 4GQ6 by a single substitution in chain A, Gly->Arg at residue 326, over 433 aligned C-alpha atoms at 0.15 A RMSD; residue 331 is Asn in both files. RCSB titles 9C4S 'G331R', the depositors' literature numbering.
RCSB PDB entries 9C4S (menin G331R with MLL peptide) and 4GQ6 (wild-type menin with MLL peptide) · RCSB PDB 2026What this evidence cannot show
Located by superposing 9C4S on 4GQ6 in PyMOL 3.2.0 and comparing residue identity at every shared residue number; residues 325-337 read back as AGYHCRNRNVREA in 4GQ6 and ARYHCRNRNVREA in 9C4S, matching UniProt 325-337. One mutant against one wild-type structure.
Evidence that argues against this (1)
- The RCSB title is not an error. It uses the numbering the clinical literature uses, and the coordinates use UniProt's; the two conventions simply coexist without a flag. Anyone who treats the title as a coordinate will highlight Asn331 in the structure and see nothing unusual. RCSB PDB 2026
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
This corpus's own interface-menin-h2ax-val371-gln400 range is in coordinate (610) numbering: Val371 and Gln400 are V and Q in O00255-2, whereas residue 371 of the 615-aa isoform is Glu.
UniProtKB O00255 (MEN1_HUMAN) isoforms O00255-1, -2 and -3 · UniProtKB 2026What this evidence cannot show
Agrees independently with that entry's own statement that author numbering was proved equal to UniProt numbering at eight positions per deposit. Gln400 is Gln in BOTH isoforms, so only the 371 end discriminates the two numberings.
Evidence that argues against this (1)
- A reader holding the 615-aa numbering would find Glu, not Val, at 371 and might conclude the entry's boundary is wrong. It is not: it is stated in the other convention. The entry is correct and the hazard is that nothing on it says which convention it uses. UniProtKB 2026
- related to →MEN1 drug-interface mutations — the first proof that a chromatin drug can drive Darwinian escape in patients
names the resistance hotspots in literature numbering
that entry's range is in coordinate numbering (E3)
numbering convention for the menin gene product
the drug whose resistance mutations are numbered this way
only the +5 offset places the expected residue at all six sites