Oncofusion

start / graph / localization-xpo1-chromatin-bound-versus-soluble-pool

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q61-export-axis

Two pools of one protein: the chromatin-bound XPO1 that matters here has never been separated biochemically from the soluble export machine

also called chromatin-bound CRM1 pool · XPO1 subnuclear distribution · prebound Crm1

The same export protein appears in two places at once - shuttling cargo through the nuclear pore, and sitting still on specific genes. Everything interesting for leukaemia is about the sitting-still pool, and no experiment has ever purified it away from the other one.

The same thing, for a scientist

**What has been localised, and by what assay.** | pool | where | assay | system | |---|---|---|---| | soluble/shuttling | nucleoplasm, cytoplasm, nucleolus, Cajal bodies | immunofluorescence, curated | human cells generally [E1] | | **chromatin-bound** | **HOXA and HOXB clusters; MEIS1 named in a 2026 perspective** | **ChIP-seq with two independent anti-CRM1 antibodies** | **LOUCY (human T-ALL), MEGAL (human AML), HEL** [E2] | | chromatin-prebound | Hox clusters | ChIP | mouse embryonic stem cells [E3] | | trapped in fusion bodies | discrete intranuclear bodies | immunofluorescence, co-IP | LOUCY, MEGAL, HeLa [E4] | | in a condensate with menin and KMT2A | NPM1c C-bodies | imaging, abstract-level | human lines, mouse, primary patient samples [E5] | **The distinction is a localisation question and this graph's standing priority makes it a first-class one.** The two pools are the same gene product; nothing published assigns a modification, a conformation, a cargo state or a partner that marks one and not the other. Every measurement of the chromatin pool is a **crosslinked ChIP**, which reports proximity to DNA and not residence, and every inhibitor used to test it (leptomycin B, KPT-330) also perturbs the soluble pool and, in the KPT-330 case, degrades the docked fusion as well [E2]. **So the honest statement is:** a locus-restricted XPO1 ChIP signal exists, it is reproducible across two groups and three systems, it is not uniform across the genome or across cell lines, and **it has never been shown to be a biochemically distinct species.**

Go deeper
Why separating the pools is the experiment this whole axis is waiting for

Three published claims all depend on the separation and none of them has it:

  1. 'CRM1 is the recruiter' 2 - true of the ChIP signal; unknown whether the recruiting molecules are the same population as the exporting ones.
  2. 'An XPO1 inhibitor removes the fusion from HOX' 2 - true; but leptomycin B and selinexor both act at Cys528 in the NES groove, which is the export-relevant surface, so the experiment cannot tell a chromatin effect from a transport effect.
  3. 'Export or docking?' - the unresolved CALM-AF10 contradiction in mechanism-calm-af10-crm1-nuclear-export-requirement is, restated, a question about which pool the NES is talking to.

The cheapest separation available is a chromatin-versus-nucleoplasm fractionation (salt-extracted chromatin pellet versus soluble nuclear fraction) followed by mass spectrometry of XPO1 from each fraction, asking whether the chromatin-associated XPO1 differs in post-translational modification or in co-precipitating partners - in particular whether it carries RAN and cargo (transport intermediate) or does not (resident platform). Nothing in the literature read in this beat reports such a fractionation for XPO1 in a leukaemia cell.

The one piece of localisation evidence that is not from this field

A 2026 J Cell Biol paper reports that CRM1 inhibits NPAT condensation and histone locus body formation by competitively occupying NPAT, which is a chromatin-adjacent structural role for CRM1 with the opposite sign to the leukaemia story - CRM1 there prevents an assembly rather than nucleating one 6. It is not about HOX and not about leukaemia, and it is recorded because it is the only independent evidence read in this beat that CRM1 does structural work at a nuclear body at all, and because it shows the sign is not fixed.

Facts and figures
mechanism classsubnuclear localisation
proteingene-xpo1
assay not yet donechromatin-versus-nucleoplasm fractionation of XPO1 with mass spectrometry of each fraction, in any leukaemia cell, to test whether the chromatin pool is a distinct biochemical species
never measured inany KMT2A-rearranged cell line or patient sample - no XPO1 ChIP in a KMT2A-rearranged system was found on 2026-09-20
confoundevery inhibitor used to test the chromatin pool acts at Cys528 in the NES groove and therefore also disables export
pools
[
  "soluble nucleoplasmic / shuttling (the export machine)",
  "chromatin-bound at HOXA, HOXB and reportedly MEIS1",
  "trapped in nucleoporin-fusion nuclear bodies",
  "co-partitioned in NPM1c C-body condensates with KMT2A and menin"
]
assays used
[
  "crosslinked ChIP-seq, two anti-CRM1 antibodies",
  "immunofluorescence",
  "co-immunoprecipitation",
  "RanGTP-dependent binding assay"
]
systems
[
  "LOUCY",
  "MEGAL",
  "HEL",
  "K562",
  "HL60",
  "HeLa",
  "mouse embryonic stem cells",
  "OCI-AML3"
]
How this could bite us back
  • Every drug that has been used to interrogate the chromatin pool also shuts down bulk nuclear export, so a therapeutic argument that specifically invokes the chromatin pool has no agent that tests it.
  • If the chromatin signal is a transport intermediate rather than a resident platform, the whole address model is an artefact of crosslinking, and there is currently no assay in this literature that would tell.
  • A localisation defined only by ChIP cannot be used to argue for a selective therapeutic window, because ChIP cannot count molecules or say what fraction of XPO1 the pool represents.
Still open
  1. Is chromatin-bound XPO1 cargo-loaded or cargo-free? A RAN co-IP from the chromatin fraction answers it and has not been reported.
  2. Does the chromatin pool carry a distinguishing post-translational modification? XPO1 has annotated phosphosites at S391, T448, S450, Y454 and S1031 and acetylation at K446 and K693, and none has been tested for a chromatin-versus-soluble split.
  3. Is there an XPO1 ChIP signal at HOXA in a KMT2A-rearranged cell? The experiment has apparently never been run in this graph's index genotype.
  4. Are the fusion nuclear bodies the same object as the HOX chromatin signal, or two compartments? No published image puts the HOXA locus inside a SET-Nup214 body.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt records XPO1 as localising to the cytoplasm, the nucleoplasm, Cajal bodies and the nucleolus, shuttling between nucleus/nucleolus and cytoplasm.

    What this evidence cannot show

    Curated database annotation fetched live 2026-09-20, aggregating several primary studies in unspecified cell types. It does NOT list chromatin as a location, which is itself informative: the chromatin pool is recent and not yet in the reference annotation.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT, and the direction matters for this entry's whole table. The caveat reads the absence of chromatin from UniProt's subcellular-location list as meaning the chromatin pool is 'recent and not yet in the reference annotation' - i.e. small and specialised. Chen et al. indicate the opposite: XPO1 co-occupies HUNDREDS of loci with p300 and its removal detaches p300 from chromatin genome-wide. If that is right, the missing annotation is missing a LARGE and general pool, not a rare one, and the two-pool framing of this entry's title ('chromatin-bound versus soluble') may be understating how much of the protein sits on chromatin at any time. Chen YF et al. 2024
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    ChIP-seq with two independent anti-CRM1 antibodies shows CRM1 selectively bound at HOXA and especially HOXB cluster regions in the human T-ALL line LOUCY and absent from K562 and HL60; in the AML line MEGAL CRM1 and SET-Nup214 preferentially occupy HOXA9; KPT-330 at 1,000 nM for 24 h removes the CRM1 peaks; and KPT-330-sensitive CRM1 signals are also seen at HOX regions in HEL, which carries none of these lesions.

    PMID 31755865DOI
    What this evidence cannot show

    Crosslinked ChIP reports proximity, not residence, and cannot distinguish a resident platform from a transport intermediate that passes close to DNA. Two antibodies control for epitope but not for the assay's inherent ambiguity. Human leukaemia cell lines only, no patient sample. KPT-330 at this exposure also degrades SET-Nup214.

    Evidence that argues against this (1)
    • Every chromatin measurement here is crosslinked ChIP, which reports proximity rather than residence, and the rival model - a transport intermediate passing close to DNA - is excluded by nothing published. The condensate literature offers a third reading in which XPO1 is co-partitioned into an assembly rather than bound to a locus, and that reading predicts a ChIP signal too. Datar GK et al. 2025
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In mouse embryonic stem cells, Crm1 is originally prebound at Hox cluster regions, and NUP98-HoxA9 is preferentially targeted to those regions; leptomycin B causes loss of NUP98-HoxA9 chromatin binding.

    PMID 26740045DOI
    What this evidence cannot show

    MOUSE EMBRYONIC STEM CELLS, not a haematopoietic cell and not a leukaemia. Abstract-level for this beat. Same group as E2, so the chromatin-pool observation has one laboratory behind it across both papers.

    Evidence that argues against this (1)
    • PUBLISHED CONFLICTING RESULT bearing on what leptomycin B licenses in this row. Loss of NUP98-HoxA9 chromatin binding after leptomycin B is read here as removal of a docking anchor. Leptomycin B is elactocin, taken to a human phase 1 in which dose-limiting profound anorexia and malaise appeared on all five schedules tested, no responses were observed, and the authors advised against further trials - an in-vivo demonstration that inactivating XPO1 globally is globally damaging. Disassembly of a chromatin interaction after such an agent is consistent with docking being disrupted AND with the whole nucleocytoplasmic economy being disrupted, and mouse embryonic stem cells give no way to tell those apart. Newlands ES et al. 1996
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    SET-Nup214 nuclear bodies in LOUCY and MEGAL contain CRM1; in HeLa the bodies recruit CRM1, export cargoes and certain nucleoporins and impair protein and polyA+ RNA export; the CRM1 interaction is RanGTP-dependent; and the bodies disassemble in mitosis and on leptomycin B.

    PMID 27613868DOI
    What this evidence cannot show

    Abstract-level for this beat; recorded in full in `localization-set-nup214-crm1-nuclear-bodies`. Cited here as the THIRD localisation of XPO1 - trapped in a fusion-built body - which is a different compartment from both the soluble pool and the HOX chromatin pool, and the relationship between the body and the locus is unestablished.

    Evidence that argues against this (1)
    • NAMED METHODOLOGICAL DEFEATER for reading the 'dynamic nuclear bodies' row as a distinct compartment. The assays that define these bodies as dynamic are FRAP-type recovery and disassembly under perturbation, and Muzzopappa et al. benchmarked exactly those against well-controlled model systems: 1,6-hexanediol treatment and CLASSICAL FRAP both fail to distinguish liquid-liquid phase separation from the alternative of molecules binding spatially clustered binding sites without phase-separating. Only preferential internal mixing in a half-bleach (their MOCHA-FRAP workflow) discriminates them. 'A fusion-built body that contains CRM1' may therefore be a cluster of CRM1 binding sites rather than a compartment, and the distinction is exactly the one this entry says is unestablished between the body and the locus. Muzzopappa F et al. 2022
  5. E5weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    NPM1c forms nuclear condensates in human cell lines, mouse models and primary patient samples, which recruit NUP98 and KMT2A; a commentary on the work reports XPO1/CRM1 and menin among the co-partitioned components.

    Disparate leukemia mutations converge on nuclear phase-separated condensates. · Datar GK, Khabusheva E, Anand A, et al. · Cell 2025
    PMID 41192422DOI
    What this evidence cannot show

    Abstract-level. The XPO1 co-partitioning statement comes from the commentary (PMID 41481253) and not from the abstract of this paper, so it is a second-hand reading of a figure. This graph's standing caveat on LLPS assays applies. Cited here as the FOURTH reported localisation of XPO1 and the only one with primary patient samples in the system list.

    Evidence that argues against this (1)
    • NAMED METHODOLOGICAL DEFEATER for the condensate row, and it is specific to chromatin. Itoh et al. show by live-cell single-nucleosome imaging that 1,6-hexanediol rapidly and dose-dependently suppresses chromatin motion and HYPER-CONDENSES chromatin in living human cells - chromatin was 'frozen' at 5% or above - and state that this action is distinct from the reagent's droplet-melting activity, instructing that droplet results be carefully reinterpreted when the droplets are chromatin-associated. NPM1c condensates that recruit NUP98 and KMT2A are chromatin-associated by construction. This bites harder here than the generic LLPS caveat because the claim being carried - that XPO1 co-partitions - is read off images of exactly such a structure, and at second hand from a commentary. Itoh Y et al. 2021
  6. E6weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    CRM1 inhibits NPAT condensation and histone locus body formation via a competitive occupation strategy.

    PMID 41481226DOI
    What this evidence cannot show

    TITLE-LEVEL ONLY - the abstract was not read in this beat. Not leukaemia, not HOX. Recorded because it is independent evidence that CRM1 performs structural work at a nuclear body, and because its SIGN is opposite to the leukaemia story: here CRM1 PREVENTS condensation. Carried at weak confidence and nothing is asserted about its mechanism beyond the title.

    Evidence that argues against this (1)
    • PUBLISHED ALTERNATIVE MODEL that weakens the 'opposite sign' reading this row is recorded for. Chong et al. showed by live-cell single-molecule imaging that at PHYSIOLOGICAL expression, low-complexity-domain interactions that control transcription form local high-concentration hubs which stabilise DNA binding and recruit RNA Pol II, are highly dynamic and selective, are DIFFERENTIALLY sensitive to hexanediols - and occur WITHOUT DETECTABLE PHASE SEPARATION. If the endogenous state is a hub rather than a condensate, then 'CRM1 inhibits condensation' and 'CRM1 promotes a body' may be the same competitive-occupancy behaviour reported against two different models, rather than two opposite effects. Recording this row as evidence that CRM1's sign flips between systems assumes the condensate model in both, which is the assumption under dispute. Chong S et al. 2018
Connected to