EWSR1::FLI1 has no residue-resolved, fusion-demonstrated PTM in its retained portion - every mapped site falls past the breakpoint, and the one site shown on the fusion has no residue
also called EWS-FLI1 PTM · EWSR1 arginine methylation not retained · EWS S266
For the Ewing sarcoma fusion protein, every chemical modification anyone has pinned to a specific amino acid sits in the part the fusion throws away. The only modification actually demonstrated on the fusion itself has never been traced to a particular spot.
The same thing, for a scientist
In the commonest (type 1) EWSR1::FLI1 the retained EWSR1 moiety is aa 1-264 and the retained FLI1 moiety is roughly aa 220-452. UniProt Q01844 annotates 38 modified residues; the earliest is S266 and the arginine-methylation block runs from R300. Q01543 (FLI1) annotates exactly one modified residue, at position 39, in the LOST N-terminal portion. So: every residue-resolved EWSR1 or FLI1 modification is outside the retained sequence of the dominant fusion. The one PTM demonstrated on the fusion protein itself - O-GlcNAcylation of the EWS transactivation domain - has no mapped residue and no identified enzyme.
Every mapped site, against the breakpoint
| protein | site | what it is | position vs breakpoint |
|---|---|---|---|
| EWSR1 Q01844 | S266 | Phosphoserine, by PKC, in the IQ domain (256-285) | breakpoint is 264 → LOST in type 1, retained only in variants breaking at EWSR1 exon 8 or later |
| EWSR1 Q01844 | R300, R302, R304, R309, R314, R317, R321 | asymmetric dimethylarginine (PRMT1, and PRMT8 at lower level) | all ≥ 300 → LOST |
| EWSR1 Q01844 | R455-R506, R563-R636 | further methylarginine blocks | LOST |
| EWSR1 Q01844 | K439 | N6-acetyllysine - the protein's only acetylation site | LOST |
| FLI1 Q01543 | 39 | the protein's only annotated modified residue | FLI1 aa 1-219 is lost → LOST |
1 Positions verified in this beat: Q01844 is 656 aa with Q at 264, S at 266, R at 300, K at 439; Q01543 is 452 aa with S at 39.
🔴 So the arginine-methylation hypothesis is excluded by arithmetic, not by weak evidence. EWSR1's RGG methylation is real, well characterised and enzymatically assigned - and it is entirely C-terminal to the fusion junction. Wild-type EWSR1 is a heavily methylated protein; EWS::FLI1 is not.
The one thing shown on the fusion, and why it is weak
O-GlcNAcylation of the retained EWS transactivation domain 2. Immunopurified EWS-FLI1 is recognised by an O-GlcNAc-specific antibody; mutational analysis localises the signal to the N-terminal EWS TAD, which is retained; inhibiting the hexosamine pathway abolishes the signal and specifically blocks transactivation of the EWS-FLI1 target Id2.
Weaknesses, all of them structural: no residue, no enzyme demonstrated (OGT is presumed), a single paper from 2009, and no replication surfaced. It is exactly the profile of a finding that is probably true and cannot be built on.
S266 is the near-miss worth recording. It is a mapped PKC site, and a 2001 paper tested S266A on EWS/FLI1 and EWS/ATF1 constructs, finding reduced DNA binding and reporter activation with partial rescue by a phosphomimetic 3. But the authors say plainly that "the IQ domain is incorporated into only a subset of fusion transcripts", and 266 > 264, so the dominant type 1 fusion does not have it. A retained-site result that applies to a minority fusion variant is a different claim from a retained-site result, and the literature has often blurred the two.
What this family contributes
EWSR1::FLI1 is the case where the fusion subtracts a PTM layer wholesale. Wild-type EWSR1 carries at least 20 methylarginines; the fusion carries none of them. That is a larger, cleaner subtraction than anything in KMT2A - and nobody appears to treat it as a mechanism.
For this beat's purposes the lesson is directional: asking 'what regulated modification does the fusion KEEP' is only half the question. The other half - what regulated modification does it LOSE - is where EWSR1::FLI1 has the striking answer, and where KMT2A's own answer (all four SUMO sites, and K2528 in particular) is the thing this beat nearly missed.
| script | worker-agent literature search plus direct UniProt flat-file parse in this beat, 2026-09-24 |
|---|---|
| status | run |
| fusion family | EWSR1::FLI1 |
| uniprot 5prime | Q01844 |
| length 5prime | 656 |
| uniprot 3prime | Q01543 |
| length 3prime | 452 |
| residue resolved retained ptms | 0 |
| ptm shown on fusion | O-GlcNAcylation of the EWS transactivation domain - domain-level only, no residue, no enzyme |
| corrects brief | the brief proposed EWSR1 RGG arginine methylation as the likely retained site; every annotated methylarginine is C-terminal to the breakpoint |
| result summary | No residue-resolved, fusion-demonstrated retained PTM. The striking fact about this family is subtraction: wild-type EWSR1 carries 20+ methylarginines and the fusion carries none. |
retained 5prime AA
[ 1, 264 ]
retained 3prime AA
[ 220, 452 ]
excluded candidates
{
"arginine methylation (PRMT1/PRMT8)": "all sites R300 and above - LOST",
"K439 acetylation": "LOST",
"FLI1 position 39": "in the lost FLI1 N-terminus",
"S266 PKC phosphorylation": "266 > 264 - LOST in type 1, retained only in EWSR1 exon-8-or-later variants",
"DNA-PK phosphorylation of EWS-FLI1": "no primary paper found; the hypothesis was not supported"
}measurements
[
{
"quantity": "residue-resolved modification sites in the retained EWSR1 portion",
"subject": "Q01844 aa 1-264",
"value": 0,
"units": "sites",
"n": 38,
"universe": "all 38 MOD_RES features on Q01844",
"null": null,
"note": "the retained portion is 40.2% of EWSR1 by length; under a length null 15.3 sites would be expected"
},
{
"quantity": "annotated methylarginines lost at the EWSR1::FLI1 breakpoint",
"subject": "Q01844 R300 and above",
"value": 20,
"units": "sites",
"n": 38,
"universe": "all 38 MOD_RES features on Q01844",
"null": null,
"note": "counted as the asymmetric-dimethylarginine and omega-N-methylarginine entries at positions 300-636; wild-type EWSR1 is heavily methylated and the fusion is not"
}
]- Two of the five PTM hypotheses in the dispatch brief for this family were excluded by nothing more than comparing a residue number to a breakpoint number. That is a two-minute check and it should precede any literature search, not follow it.
- Does losing 20+ methylarginines change the phase behaviour of the EWS low-complexity domain in the fusion? Arginine methylation is a known regulator of RGG-domain condensation, EWS::FLI1 condensate biology is an active field, and the two literatures do not appear to have been joined.
- Which residue carries the O-GlcNAc on the EWS transactivation domain? Seventeen years after the observation there is still no site, and without one the finding cannot be mutated, cannot be quantified and cannot be targeted.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProt Q01844 (EWSR1, 656 aa) annotates 38 modified residues with the earliest at S266 (phosphoserine, by PKC) and asymmetric dimethylarginines beginning at R300; UniProt Q01543 (FLI1, 452 aa) annotates exactly one modified residue, at position 39.
What this evidence cannot show
Parsed directly from the UniProt flat files in this beat, independently of the worker agent. Canonical isoforms only. The EWSR1 aa 1-264 / FLI1 aa 220-452 boundaries for the type 1 fusion were reported by the worker and NOT independently verified against a primary breakpoint paper in this beat - they are the standard textbook boundaries but they are the one unverified link in the arithmetic.
Evidence that argues against this (1)
- EWSR1's retained region is a low-complexity prion-like SYGQ-rich domain, the same sequence class that makes NUP98's FG repeats hard to measure. An absence of annotated sites there is partly an absence of tryptic peptides. UniProt also does not systematically curate O-GlcNAc, which is precisely the modification reported on that domain.
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Immunopurified EWS-FLI1 is O-GlcNAcylated on the retained N-terminal EWS transactivation domain; hexosamine-pathway inhibition abolishes the modification and blocks transactivation of the target gene Id2.
O-GlcNAcylation is involved in the transcriptional activity of EWS-FLI1 in Ewing's sarcoma · Bachmaier R, Aryee DN, Jug G, Kauer M, Kreppel M, Lee KA, Kovar H · Oncogene 28:1280-4 2009What this evidence cannot show
No residue mapped, no enzyme demonstrated, a single paper from 2009, and no replication was surfaced. The localisation to the EWS TAD is from deletion mutants, which places the modification in a domain rather than at a site. Read at abstract level by this beat; the full text was read by a worker agent.
Evidence that argues against this (1)
- Hexosamine-pathway inhibition is a global metabolic intervention that changes O-GlcNAc on thousands of proteins, so the Id2 transactivation effect need not be mediated by O-GlcNAc on EWS-FLI1 itself.
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Phosphorylation of the EWS IQ domain at S266 regulates DNA binding and transcriptional activity of EWS/ATF1 and EWS/FLI1 constructs, but the IQ domain is incorporated into only a subset of fusion transcripts and lies past the type 1 breakpoint at EWSR1 aa 264.
Phosphorylation of the EWS IQ domain regulates transcriptional activity of the EWS/ATF1 and EWS/FLI1 fusion proteins · Olsen RJ, Hinrichs SH · Oncogene 20:1756-64 2001What this evidence cannot show
Exogenous constructs in NIH3T3, not endogenous fusion in Ewing cells. The critical qualifier - that the IQ domain is in only a subset of fusion transcripts - was read by a worker agent from the full text and is quoted at second hand. S266 is verified as a serine in Q01844 by this beat; the 264 breakpoint is not independently verified here.
Evidence that argues against this (1)
- Because S266 sits two residues past the type 1 breakpoint, this result describes a minority fusion variant. Anyone citing it as 'a regulated PTM in the retained portion of EWS-FLI1' is generalising from the variant to the common form.
one of five families in the scorecard
both families subtract a coherent PTM layer; EWSR1 loses 20+ methylarginines, KMT2A loses all four SUMO sites