PSIP1-TMEM214 is a DSSO crosslink in two preparations of isolated U2OS nuclei with no bait on either side; TMEM214 is an ER-membrane mediator of CASP4-dependent ER-stress apoptosis - declined
also called PSIP1-TMEM214 · LEDGF-TMEM214 · EBI-20922014
A database lists TMEM214, a protein embedded in the membrane of the endoplasmic reticulum, as a partner of the nuclear protein LEDGF/PSIP1. The supporting measurement is a chemical crosslink inside isolated nuclei, in a paper whose authors say part of their nucleus-to-ER signal comes from contamination during nuclei purification.
The same thing, for a scientist
**Verdict: DECLINED. Detection method is chemical crosslinking mass spectrometry (DSSO XL-MS) with no bait: both participants are recorded as 'neutral component'** [E1]. One IntAct record, EBI-20922014, from PMID 30021884, host U2OS osteosarcoma, intact-miscore 0.4, annotated to Table S2-white and Table S1 and *'interaction detected in: TX100 unsoluble fraction (fractionated nuclei), unfractionated nuclei'* - **two preparations, and one of them is the Triton-insoluble fraction in which the paper reports the nuclear share of crosslinks rises to 85%** [E1, E2]. That makes this the **best-supported of the five Fasci rows in this beat** on internal grounds. TMEM214 is UniProt Q6NUQ4, 689 aa, with a single subcellular-location comment: **Endoplasmic reticulum membrane [ECO:0000269]**, and a single function - critical mediator, with CASP4, of ER-stress-induced apoptosis [E3]. **The decline therefore rests entirely on compartment, and the paper's own figure that 4% of its interactions are nucleus-endoplasmic-reticulum within a 30% cross-compartment fraction it attributes partly to contamination during nuclei isolation** [E2]. Neither 'PSIP1', 'LEDGF' nor 'TMEM214' appears anywhere in the paper's text [E2].
The one row in this beat where the internal evidence is above the dataset's own floor, and what that is worth
Of the five Fasci crosslinks in this beat, the annotations record:
| pair | preparations detected in |
|---|---|
| PSIP1-TMEM214 | TX100 INSOLUBLE fraction + unfractionated nuclei |
| PSIP1-GDF5 | TX100 soluble fraction + unfractionated nuclei |
| KMT2A-TMCO1 | TX100 soluble fraction only |
| PSIP1-IGHMBP2 | TX100 soluble fraction only |
| PSIP1-PPP1R13B | unfractionated nuclei only |
The paper reports that histones were largely detected in the Triton X-100 insoluble fraction and that the identifiable-nuclear share of crosslinks rose to 85% in the insoluble fraction 2. So an insoluble-fraction crosslink is, by the paper's own accounting, in the cleanest part of the experiment - and two preparations is better than one against a 58% replicate overlap 2.
And it is still an ER membrane protein. The two facts do not cancel; they locate the disagreement precisely. Either TMEM214 has an undocumented nuclear or inner-nuclear-membrane pool, or a crosslink in the cleanest fraction of a good dataset is nonetheless wrong. Nothing available here decides between those, which is why the verdict is DECLINED rather than REFUTED, and why this entry names the supplementary table that would settle it.
Why this one matters more than a floor-score AP-MS row
PSIP1/LEDGF is not a peripheral protein in this project. It is the chromatin tether that holds the menin-KMT2A complex at its target genes through the PWWP domain and the integrase-binding domain, and it is the target of the LEDGIN class of allosteric inhibitors.
So a claimed PSIP1 partner has consequences: a genuine ER-stress protein in PSIP1's neighbourhood would connect the tether to the unfolded-protein response, which is active in plasma-cell and some leukaemic contexts and is a live drug target.
That is exactly why the row gets an argued decline rather than silence. A reader who finds TMEM214 in IntAct's PSIP1 list and no comment in this graph has no way to tell whether it was considered. It was, on 2026-09-23, and the reason it was not admitted is written above along with the specific document - supplemental Table S1/S2 of PMID 30021884, plus that paper's PRIDE deposition - that would overturn it.
| partner a | gene-psip1 |
|---|---|
| partner a uniprot | O75475 |
| partner a length AA | 530 |
| partner b | TMEM214 |
| partner b uniprot | Q6NUQ4 |
| partner b length AA | 689 |
| verdict | DECLINED |
| adjudicated by | Q129, cycle 92, 2026-09-23 |
| decline strength | the best-supported of the five Fasci rows on internal grounds - two preparations including the Triton-insoluble one - and declined on compartment alone |
| source publication pmid | 30021884 |
| source dataset | Fasci 2018 nuclear XL-MS (DSSO, U2OS intact nuclei) |
| detection method | chemical crosslinking mass spectrometry (DSSO XL-MS) |
| experimental role partner | neutral component |
| experimental role psip1 | neutral component |
| bait exists in dataset | false |
| intact miscore | 0.4 |
| n publications | 1 |
| n methods | 1 |
| orthogonal method exists | false |
| dataset cross compartment PCT | 30 |
| dataset nucleus er PCT | 4 |
| druggability | Not assessable. The crosslinked lysine positions are in supplemental Table S1/S2 and were not retrieved; PSIP1's druggable surfaces - the PWWP and integrase-binding domains - are not implicated by anything in the record. |
| absence class | measured-not-searched |
| absence class note | The decisive missing datum is the crosslinked lysine positions and PSM score for EBI-20922014, in supplemental Tables S1/S2 of PMID 30021884 and in that paper's PRIDE deposition, neither opened here. If the crosslink falls in PSIP1's PWWP or IBD it becomes a testable constraint on a druggable surface. The separate never-measured absence is that nobody has tested PSIP1 against TMEM214 with PSIP1 as bait or in any haematopoietic cell. |
| shown for fusion | NO - nobody has looked in any KMT2A fusion. U2OS osteosarcoma nuclei; no fusion protein present. |
| epigenetic readout | none - no histone mark, occupancy or chromatin measurement exists for this pair. |
host cell lines
[ "U2OS osteosarcoma" ]
preparations detected in
[ "TX100 unsoluble fraction (fractionated nuclei)", "unfractionated nuclei" ]
records
[ "EBI-20922014" ]
compartment mismatch
{
"partner_compartments": [
"Endoplasmic reticulum membrane [ECO:0000269]"
],
"partner_nuclear_annotation": "none at any evidence level",
"partner_annotation_depth": "sparse - one location comment and one function comment in total",
"caveat": "the ER membrane is continuous with the outer nuclear membrane, so the mismatch is not absolute"
}interface residues a
[]
structures
[]
- PSIP1/LEDGF is the chromatin tether for the menin-KMT2A complex and the target of the LEDGIN allosteric inhibitor class, so any PSIP1 partner has intervention consequences and a floor-level row still deserves a recorded decline rather than silence.
- If PSIP1 did have an ER-stress arm, agents that induce the unfolded-protein response - proteasome inhibitors among them, which this graph already discusses in pro-B MLL leukaemia - could modulate the tether by a route no chromatin assay would detect. The row is far too weak to support that, and it is written down so the idea stays traceable to its evidence rather than circulating without it.
- What do the 'pink' and 'white' colour codes in Fasci supplemental Table S2 denote? IntAct carries them into its annotations and this beat used them without knowing what they mean, which is a live interpretive gap across all five Fasci rows adjudicated here.
- Where do the crosslinked lysines fall on PSIP1? If they fall in the PWWP or integrase-binding domain the row becomes a constraint on a druggable surface rather than a compartment argument.
- Does PSIP1 have any documented role in the unfolded-protein response? Nothing in this graph addresses it, and a genuine ER-stress connection for the menin-KMT2A tether would be a new axis rather than an extension of an existing one.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
IntAct curates one record of a PSIP1-TMEM214 physical association, EBI-20922014, from PMID 30021884. Detection method 'crosslink' (XL-MS), interaction type 'physical association', NO expansion method, host Homo sapiens U2OS osteosarcoma cell. Both TMEM214 (Q6NUQ4) and PSIP1 (O75475) are recorded as 'neutral component' - no bait, no prey. intact-miscore 0.4. Annotations: 'figure legend (Table S2-white, Table S1)' and 'comment (interaction detected in: TX100 unsoluble fraction (fractionated nuclei), unfractionated nuclei)'. Within this publication PSIP1 has six crosslink partners: GDF5, HSPA8, IGHMBP2, PPP1R13B, SCAF11 and TMEM214.
IntAct molecular interaction database, REST findInteractions on O75475 (PSIP1), 114 records retrieved 2026-09-23 by Q129 · IntAct (EMBL-EBI) 2026What this evidence cannot show
Read live from IntAct REST on 2026-09-23, not PSICQUIC (rule 93: PSICQUIC returns about 44% of IntAct's PSIP1 rows). Supplemental Tables S1 and S2 were not opened, so the crosslinked residue positions and PSM score are unknown, and the meaning of the 'white' versus 'pink' colour coding in Table S2 was not determined - it is quoted from the IntAct annotation without knowing what it denotes.
Evidence that argues against this (1)
- The colour codes matter and were not resolved. If 'pink' marks previously unreported interactions and 'white' marks database-supported ones, or the reverse, that changes how this row should be read relative to the GDF5 and TMCO1 rows, which are both annotated 'pink'. Recording the annotation without its legend is a real limitation of this entry. Fasci D et al. 2018
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In isolated intact human U2OS nuclei, Fasci et al. crosslinked with DSSO and identified ~8,710 unique crosslinks at 1% FDR, ~58% overlapping between two independent biological replicates. Histones were largely detected in the Triton X-100 insoluble fraction and the identifiable-nuclear share of crosslinks rose to 85% in that fraction; from the soluble and insoluble fractions together 7,095 unique crosslinks were identified, 4,606 of them connecting distinct proteins. The paper states: '20% of the interactions were nucleus-cytosol/plasma membrane, 6% nucleus-mitochondria and 4% nucleus-endoplasmic reticulum. This result is in line with the multiple subcellular localization reported for many proteins but could also be partly influenced by contamination with other subcellular compartments during our nuclei isolation.'
Histone Interaction Landscapes Visualized by Crosslinking Mass Spectrometry in Intact Cell Nuclei. · Fasci D, van Ingen H, Scheltema RA, Heck AJR · Molecular & cellular proteomics : MCP 2018What this evidence cannot show
Full text read 2026-09-23 through the PMC HTML route (PMC6166682, 92,127 stripped characters) after fullTextXML returned HTTP 500 and efetch db=pmc returned a body-less document. The strings 'PSIP1', 'LEDGF' and 'TMEM214' each appear ZERO times in the paper's text; this row exists only in a supplementary table that was not opened. The paper's 85%-in-the-insoluble-fraction figure is a property of that fraction as a whole and says nothing about this particular crosslink.
Evidence that argues against this (1)
- Reading 'insoluble fraction' as 'cleanest' is an inference. Triton-insoluble material in a nuclei preparation includes the nuclear lamina, chromatin AND any residual membrane that the detergent did not dissolve, so an integral membrane protein appearing in the insoluble fraction is not automatically evidence that it was chromatin-associated rather than membrane-associated. This cuts directly against the paragraph above. Fasci D et al. 2018
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProtKB Q6NUQ4 is 689 aa, primary gene name TMEM214, recommended name 'Transmembrane protein 214'. Its ONLY subcellular-location comment is Endoplasmic reticulum membrane [ECO:0000269]; there is no nuclear comment at any evidence level and no other compartment is listed. Its only function comment is: critical mediator, in cooperation with CASP4, of endoplasmic-reticulum-stress-induced apoptosis, required for the activation of CASP4 following ER stress. PSIP1 (O75475) is 530 aa with a single subcellular-location comment, Nucleus [ECO:0000269], and is a transcriptional coactivator.
UniProtKB entries Q6NUQ4 (TMEM214) and O75475 (PSIP1): primary gene name, sequence length, subcellular location with evidence codes and function, retrieved 2026-09-23 · UniProt Consortium 2026What this evidence cannot show
UniProt location comments are curated pointers, not measurements (rule 91); the cited paper was not opened. TMEM214 is a sparsely annotated protein with one function comment and one location comment, so the absence of a nuclear line carries less weight than it would for a heavily studied protein - absence of annotation and absence of the thing are not the same, and no PubMed query on nuclear TMEM214 was run in this beat.
Evidence that argues against this (1)
- The ER membrane is continuous with the outer nuclear membrane, and several ER-annotated proteins have documented inner-nuclear-membrane pools. For a protein with only one curated location comment, a second undocumented compartment is entirely possible, so the compartment argument here is weaker than it is for the secreted proteins elsewhere in this beat. UniProt Consortium 2026
adjudicates the IntAct PSIP1-TMEM214 row; DECLINED on compartment alone - the best-supported of the five crosslinks on internal grounds
same dataset, both non-nuclear partners of PSIP1; TMEM214 appears in two preparations including the Triton-insoluble one, GDF5 in two including the soluble one
the same compartment-mismatch verdict, reached for the first time on a BioID2 proximity row rather than a crosslink