Oncofusion

start / graph / interface-psip1-ighmbp2-declined-compartment-compatible-single-crosslink

Interfaceweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q129-the-never-discussed-partners-of-kmt2a-and-psip1

PSIP1-IGHMBP2 is a single DSSO crosslink in one detergent fraction of U2OS nuclei with no bait; unlike the other Fasci rows the compartments match, so this is declined on evidence depth alone - and IGHMBP2 is a fourth ribosome-biogenesis-adjacent protein on this list

also called PSIP1-IGHMBP2 · LEDGF-IGHMBP2 · SMUBP-2 · SMBP2 · EBI-20906152

A database lists IGHMBP2, a nuclear and cytoplasmic helicase that unwinds DNA and RNA, as a partner of the nuclear protein LEDGF/PSIP1. Both proteins really are in the nucleus, so the usual objection does not apply; the problem is that the whole claim rests on one chemical crosslink seen in one preparation.

The same thing, for a scientist

**Verdict: DECLINED on evidence depth, not on compartment. 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-20906152, from PMID 30021884, host U2OS osteosarcoma, intact-miscore 0.4, annotated to Table S2-pink and *'interaction detected in: TX100 soluble fraction (fractionated nuclei)'* - **one preparation only, against 58% replicate overlap in the source dataset** [E1, E2]. IGHMBP2 is UniProt P38935, 993 aa, **Nucleus [ECO:0000269] and Cytoplasm [ECO:0000269]**, a 5'-to-3' ATP-dependent helicase that unwinds both RNA and DNA duplexes, specific for 5'-phosphorylated single-stranded guanine-rich sequences; loss of function causes spinal muscular atrophy with respiratory distress type 1 [E3]. **There is no compartment mismatch and no lineage mismatch**, which removes the argument used against GDF5, TMCO1 and TMEM214 and leaves only the thinness of the record. **Two things are worth flagging beyond the verdict.** First, UniProt's own function comment says IGHMBP2 *'may play a role in RNA metabolism, **ribosome biogenesis** or initiation of translation'* [E3] - making it a **fourth** ribosome-biogenesis-adjacent protein among the thirteen never-discussed partners, alongside ABT1, KRR1 and UTP3, and it arrives by a completely different method. **It is NOT an SSU processome component and UniProt gives it no nucleolar location at any evidence level, so calling it a fourth member is a claim about a hedged annotation clause and not about shared machinery** - KRR1 and UTP3 are bona fide SSU processome proteins and IGHMBP2 is not, which is why this entry says ribosome-biogenesis-ADJACENT throughout and does not count it inside the triple. Second, a helicase specific for G-rich single-stranded nucleic acid crosslinked to a chromatin-binding coactivator is a pair that could be produced by shared occupancy of one locus rather than by contact.

Go deeper
The hypothesis this row bears on, and why it does not rescue it

This beat was dispatched to test whether ABT1, KRR1 (on KMT2A) and UTP3 (on PSIP1) - three nucleolar ribosome-biogenesis factors on two different baits - indicate real nucleolar biology or an affinity-purification artefact.

IGHMBP2 is a fourth candidate for that class and it changes the shape of the question rather than the answer. Its UniProt function comment names ribosome biogenesis explicitly, and unlike the other three it arrives by crosslinking rather than by affinity purification 31. On the face of it, a fourth member by an orthogonal method is exactly what would strengthen reading (a).

It does not, for three reasons that were pre-registered before the data were pulled.

  1. The ribosome-biogenesis attribution is hedged in the source annotation itself - 'may play a role in RNA metabolism, ribosome biogenesis or initiation of translation', three alternatives in one clause 3. IGHMBP2's characterised biology is as a disease-causing helicase in motor neurons, not as a nucleolar factor, and UniProt gives it no nucleolar location comment at any evidence level.
  2. The pre-registered second uninformative condition applies. If the 'ribosome-biogenesis factor' label can only be supported from the same annotation database that supplied the grouping, the grouping is an annotation artefact. Here the label is a hedged clause in a UniProt function comment and nothing independent was read.
  3. A shared nucleic acid is a live alternative. IGHMBP2 binds G-rich single-stranded nucleic acid; PSIP1 binds chromatin through its PWWP domain. Two nucleic-acid-binding proteins crosslinked in a nucleus may be close because they occupy the same template, which is a real spatial statement but not a protein-protein interaction.
What would admit this row, in order of cost
  1. The crosslinked lysine positions and PSM score, in supplemental Tables S1/S2 of PMID 30021884 and in that paper's PRIDE deposition. Neither was opened. If the crosslink lands in PSIP1's PWWP or integrase-binding domain it becomes a constraint on a characterised surface; if it lands in the long disordered central region it is uninformative. This is an afternoon's work and it is the single highest-value missing datum for all five Fasci rows in this beat.
  2. A second preparation or a second dataset. This record is one fraction of one experiment in one cell line.
  3. Any experiment with PSIP1 as the bait. IntAct records PSIP1 as an experimental bait in only 11 rows across 4 publications, the most recent being PMID 22872640 (J Biol Chem, October 2012) 4. No PSIP1-bait interaction experiment has been curated into IntAct for fourteen years, so every partner added since is PSIP1 appearing in somebody else's experiment.
Facts and figures
partner agene-psip1
partner a uniprotO75475
partner a length AA530
partner bIGHMBP2
partner b uniprotP38935
partner b length AA993
verdictDECLINED
adjudicated byQ129, cycle 92, 2026-09-23
decline strengthdeclined on evidence depth only - no compartment mismatch, no lineage mismatch; one of the two most plausible pairs in the beat
source publication pmid30021884
source datasetFasci 2018 nuclear XL-MS (DSSO, U2OS intact nuclei)
detection methodchemical crosslinking mass spectrometry (DSSO XL-MS)
experimental role partnerneutral component
experimental role psip1neutral component
bait exists in datasetfalse
intact miscore0.4
n publications1
n methods1
orthogonal method existsfalse
compartment mismatchfalse
nucleic acid bridge riskmoderate - IGHMBP2 binds G-rich single-stranded nucleic acid and PSIP1 binds chromatin; no nuclease control is recorded
ribosome biogenesis candidatehedged - UniProt says 'may play a role in RNA metabolism, ribosome biogenesis or initiation of translation'; no nucleolar location comment at any evidence level
psip1 bait publications total4
psip1 most recent bait publicationPMID 22872640, J Biol Chem, October 2012
druggabilityNot assessable from the record. If the crosslinked lysines fall in PSIP1's PWWP or integrase-binding domain - both druggable surfaces, the IBD being the LEDGIN target - the row would become a constraint worth having. Those positions are in supplemental Table S1/S2 and were not retrieved.
absence classmeasured-not-searched
absence class noteThe decisive missing datum is the crosslinked lysine positions and PSM score for EBI-20906152, in supplemental Tables S1/S2 of PMID 30021884 and in that paper's PRIDE deposition, neither opened here. That is an afternoon. The separate never-measured absence is that no experiment has used PSIP1 as a bait against IGHMBP2, in any cell - and IntAct records no PSIP1-bait experiment at all since 2012.
shown for fusionNO - nobody has looked in any KMT2A fusion. U2OS osteosarcoma nuclei; no fusion protein present.
epigenetic readoutnone - no histone mark, occupancy or chromatin measurement exists for this pair.
host cell lines
[
  "U2OS osteosarcoma"
]
preparations detected in
[
  "TX100 soluble fraction (fractionated nuclei)"
]
records
[
  "EBI-20906152"
]
partner compartments
[
  "Nucleus [ECO:0000269]",
  "Cytoplasm [ECO:0000269]",
  "Cell projection, axon [ECO:0000250]"
]
interface residues a
[]
structures
[]
How this could bite us back
  • PSIP1/LEDGF tethers the menin-KMT2A complex to chromatin and is the target of the LEDGIN allosteric inhibitor class, so a genuine PSIP1 partner that also touches translation would mean a LEDGIN has effects outside transcription that nobody is measuring.
  • IGHMBP2 loss causes a severe infantile motor-neuron disease, so if it were a real node in this pathway it would be an intervention target with a known catastrophic loss-of-function phenotype in children - which is the population this project's disease is in. That is an argument against targeting it, and it should be recorded before anyone is tempted by the row rather than after.
Still open
  1. Where do the crosslinked lysines fall on PSIP1? This is the same question for all five Fasci rows and it is answerable from a supplementary table in one sitting; for IGHMBP2, whose compartments already match, it is the difference between a discardable row and a structural constraint.
  2. Is IGHMBP2 genuinely a ribosome-biogenesis factor in a human cell, or is the UniProt clause carrying yeast and in-vitro inference? Nothing independent was read, and the answer decides whether the 'nucleolar block' on this list has three members or four.
  3. Has any PSIP1-bait interaction experiment been published since 2012? IntAct has curated none, and whether that reflects the field or the curation queue was not checked - the same question arises for KMT2A, whose most recent curated bait publication is 2015.
Every source behind this entry (4)
Connected to