Oncofusion

start / graph / gene-topbp1

Geneweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q78-interactome-absent

TOPBP1 - a real KMT2A contact that no abstract-level query can see, because it is a control panel in a paper about a different protein

also called TOPBP1 · TopBP1 · DNA topoisomerase 2-binding protein 1 · KIAA0259 · DNA topoisomerase II-beta-binding protein 1

TopBP1 is the protein that switches on the cell's main damage alarm. A 2007 paper about a different protein quietly showed, in one control panel, that the same region of KMT2A that TopBP1's partner uses can bind TopBP1 too - and no literature search on KMT2A can find it.

The same thing, for a scientist

TOPBP1 (Q92547, 1,522 aa) is the eight-BRCT-domain activator of the ATR kinase and a replication-checkpoint mediator. Its single IntAct record against a core bait is a **yeast two-hybrid physical association with KMT2A** [E1] from Sjottem et al. 2007 - a paper about **SPBP/TCF20**, not about KMT2A [E2]. The record is not a curation error and it is not the ATR-CHK1 story the dispatch brief predicted. It is Figure 7D: having found TopBP1 in a two-hybrid screen with SPBP's **extended PHD (ePHD)** as bait, the authors asked whether the homologous ePHD domains of the trithorax proteins bind too, and report that **'the ePHD domains from the trithorax group proteins MLL, MLL2 and -4 can bind to TopBP1'**, while the SPBP homologue RAI1 binds very weakly and the related hRNF4 zinc finger does not bind at all [E3]. IntAct curated all three trithorax rows: KMT2A (Q03164), KMT2B (Q9UMN6) and KMT2D (O14686) [E1]. **`TOPBP1[tiab] AND (MLL OR KMT2A)[tiab]` returns 0** [E4] - the measurement is invisible to every abstract-level query this project runs, and is a clean worked instance of brief rule 80.

Go deeper
What the experiment is, and the two internal controls that make it worth keeping

A bare one-row yeast two-hybrid is normally the weakest thing an interactome can offer: a heterologous organism, an overexpressed fusion, no reciprocal arm. Two features raise this one above that floor 3:

  • A graded negative in the same panel. RAI1, SPBP's own paralogue and an ePHD protein, shows very low affinity for TopBP1 in the same two-hybrid assay, confirmed by GST pull-down. So the assay discriminates between ePHD domains rather than reporting that any zinc finger sticks to any BRCT.
  • A structural negative. hRNF4, a RING-finger protein whose zinc-finger domain is related to the ePHD and which is a known SPBP binder, does not bind TopBP1.

And a third feature raises it further: the contact is domain-resolved on the TopBP1 side. 165 of 280 sequenced positive clones from the original screen encoded BRCT6-7 and part of BRCT8, TopBP1(862-1406), and the SPBP contact was mapped to BRCT6 3.

What is NOT available: the KMT2A construct boundaries. They appear only inside the Figure 7D panel, and the word MLL occurs exactly twice in the paper's 75,142 characters of full text. This entry therefore states the domain by name and asserts no residue numbers (brief: never invent a residue boundary).

Why the brief's prediction failed, and it is worth saying that it did

This beat was dispatched on the hypothesis that TOPBP1 and RBBP8/CtIP were the same DNA-damage story, and that both connected to the ATR-CHK1 thread opened last cycle around menin's chromatin pool after UV 5. They are three different stories.

baitsourceassaywhat it is about
TOPBP1KMT2ASjottem 2007, SPBP/Ets1 transcriptionyeast two-hybrid, ePHD vs BRCT6a domain-family property of ePHDs
RBBP8/CtIPPSIP1Daugaard 2012, LEDGF and HRco-IP + direct in vitroend resection
SKP2KMT2ALiu 2010, ATR and the S-phase checkpointco-IP, both directionsdegradation of wild-type MLL

Only the third is an ATR story, and it was already in the graph. The DDR-shaped cluster in the interactome gap was an artefact of grouping partners by what they are famous for rather than by what the record says. Sorting partners by their canonical pathway is a hypothesis about the data, not a reading of it.

What would make this matter, and the honest statement of what it currently is

TopBP1 is the ATR activator. If the KMT2A ePHD genuinely engages TopBP1 BRCT6 in a human cell, then KMT2A sits one contact away from checkpoint activation, and that would sit beside the S516 result - ATR phosphorylating KMT2A - as a possible reciprocal arm. There is no evidence for any of that. What exists is one two-hybrid panel in a 2007 transcription paper, with good internal controls and no follow-up in nineteen years, in no relevant cell type, on a domain whose boundaries were not recorded.

Graded weak, kept because the record is real and because the absence it sits inside is the point: a partner with zero PubMed reachability from either symbol is precisely the partner an interactome coverage figure will keep flagging and nobody will ever find by reading.

Facts and figures
uniprotQ92547
length AA1522
gene symbolTOPBP1
protein nameDNA topoisomerase 2-binding protein 1 (TopBP1)
uniprot resolutionaccession:Q92547, reviewed human, one record; primary gene name read back as TOPBP1 (entry TOPB1_HUMAN); length 1,522 aa checked against UniProt 2026-09-21.
symbol collision checkdata/symbol-collisions.json greped 2026-09-21 for TOPBP1: no row existed and none is proposed.
baitKMT2A (Q03164)
shown for fusionNO - nobody has looked in any fusion, in any lineage. The ePHD lies C-terminal to the fusion breakpoint region and is LOST in KMT2A fusions, which is the first thing to check before this contact is given any weight in this disease.
fetched2026-09-21
subnuclear localizationDNA-damage foci at DSBs; Treacle-nucleated phase-separated condensates in the NUCLEOLAR FIBRILLAR CENTRE (the nucleolar DNA-damage response); discrete nuclear bodies at steady state on HPA immunofluorescence. All condition-dependent.
localization shown for fusionNO - nobody has localised TOPBP1 in any KMT2A-rearranged cell, in any lineage; the bounded query returned 0. And the caution already in this entry's shown_for_fusion stands and matters more for localisation than for binding: the TOPBP1-contacting ePHD region of KMT2A lies C-terminal to the breakpoint and is LOST in KMT2A fusions, so a fusion-expressing cell may have no reason to place TOPBP1 anywhere KMT2A used to.
interaction with kmt2a
{
  "source_pmid": "17913746",
  "source_class": "focused study of a DIFFERENT protein (SPBP/TCF20); the KMT2A row is a domain-family control panel",
  "n_intact_records": 1,
  "n_spoke_expanded": 0,
  "n_source_publications": 1,
  "intact_acs": [
    "EBI-20560922"
  ],
  "interaction_types": {
    "MI:0915 physical association": 1
  },
  "detection_methods": [
    "2 hybrid"
  ],
  "host_organism": "Saccharomyces cerevisiae",
  "domain_on_kmt2a": "extended PHD (ePHD); residue boundaries NOT recoverable - they appear only inside Figure 7D and the string MLL occurs twice in 75,142 characters of full text",
  "domain_on_topbp1": "BRCT6 (SPBP contact); the original screen's positive clones encoded TopBP1(862-1406), i.e. BRCT6-7 and part of BRCT8",
  "internal_controls": [
    "RAI1 ePHD: very low affinity, confirmed by GST pull-down",
    "hRNF4 RING zinc finger: no binding"
  ],
  "reciprocal_tested": false,
  "human_cell_confirmation": false,
  "independent_sources": 0,
  "pubmed_reachability": "TOPBP1[tiab] AND (MLL OR KMT2A)[tiab] = 0 on 2026-09-21",
  "shared_pdb_entries": []
}
same paper paralogue rows
{
  "KMT2B": "Q9UMN6, 2 hybrid",
  "KMT2D": "O14686, 2 hybrid"
}
localization
{
  "subcellular": "Nucleus and chromosome, with a genuine CYTOPLASMIC pool at the centrosome and spindle pole. UniProt Q92547: Nucleus (ECO:0000269 PubMed:11395493, 12697828, 12773567), Chromosome (fourteen ECO:0000269 citations), and Cytoplasm/cytoskeleton/microtubule-organising-centre/centrosome plus spindle pole (both ECO:0000269 PubMed:15138768). HPA immunofluorescence: main Nucleoplasm and NUCLEAR BODIES, Plasma membrane additional, reliability Supported.",
  "subnuclear": "THREE distinct subnuclear states, all condition-dependent. (1) DNA-damage FOCI at double-strand breaks, the classic pattern, across many ECO:0000269 citations. (2) NUCLEOLAR CONDENSATES in the fibrillar centre: Treacle phosphorylated by CK2 and ATR/ATM nucleates TOPBP1 phase separation inside the FC, building the nucleolar DNA-damage-response platform that guards rDNA (PMID 42152682). (3) Discrete NUCLEAR BODIES at steady state on antibody immunofluorescence (HPA). The nucleolar condensate is the best-characterised subnuclear address of any protein in this tranche.",
  "how_shown": "immunofluorescence focus formation; inducible oligomerisation plus physiological rDNA-damage models for the nucleolar condensate (PMID 42152682); antibody immunofluorescence (HPA)",
  "tissue_of_evidence": "human cancer and immortalised lines; the dendritic-cell work is murine. Nothing leukaemic has been imaged.",
  "why_this_matters_here": "This graph already holds localization-nucleolus-kmt2a-fusion and localization-condensate-evidence-standards. TOPBP1 is a well-evidenced positive control for both: a protein with a real, inducible, mechanistically explained nucleolar condensate. Any condensate claim made elsewhere in this project can be graded against it.",
  "node": "localization-nucleolus-kmt2a-fusion"
}
cell type
{
  "haematopoietic_answer": "CONVENTIONAL DENDRITIC CELLS, and earlier, B AND T LYMPHOCYTE DEVELOPMENT. DC-specific Topbp1 deletion in mice causes cDC deficiency with pre-DC accumulation and blocks terminal differentiation of XCR1- CD24+ emergency progenitors, and Flt3L immunotherapy fails in those animals; the same abstract records TopBP1's established role in V(D)J recombination during B and T cell development (PMID 42104014). The cDC defect is NOT attributable to DNA-damage-driven cell death, which makes it a differentiation requirement rather than a genome-stability one.",
  "expression_atlas": "HPA Blood Atlas: low immune cell specificity, detected in many; RNA tissue specificity low. Expression gives no lineage restriction - the genetics do.",
  "context": "murine conditional knockout (brief rule 73). No human and no KMT2A-fusion system."
}
cell state
{
  "state": "TWO, and they are different in kind. (1) rDNA DAMAGE / nucleolar stress, which is what nucleates the TOPBP1 condensate (PMID 42152682). (2) DIFFERENTIATION TRANSIT - the pre-DC-to-cDC step, where TopBP1 is required independently of DNA damage (PMID 42104014).",
  "relevance": "A protein whose subnuclear address is set by ribosomal-DNA damage is directly relevant to a disease whose lines are reported vulnerable to ribosome-biogenesis inhibition; that link is a hypothesis here, not a result."
}
localization bounded absence
{
  "date": "2026-09-21",
  "tool": "NCBI E-utilities esearch, db=pubmed",
  "shape": "gene clause AND localisation clause AND KMT2A/MLL-fusion clause, one [tiab] tag per term (brief rule 94), every cytogenetic term quoted (brief rule 95)",
  "querytranslation_fusion_clause": "\"KMT2A\"[Title/Abstract] OR \"MLL\"[Title/Abstract] OR \"MLL1\"[Title/Abstract] OR \"MLL-AF4\"[Title/Abstract] OR \"MLL-AF9\"[Title/Abstract] OR \"MLL-ENL\"[Title/Abstract] OR \"KMT2A-rearranged\"[Title/Abstract] OR \"MLL-rearranged\"[Title/Abstract] OR \"t 4 11\"[Title/Abstract] OR \"t 9 11\"[Title/Abstract] OR \"MLL fusion\"[Title/Abstract]",
  "count": 0,
  "controls": {
    "fusion_clause_alone": 8743,
    "gene_clause_alone": 529
  },
  "reading": "Counted in TITLES AND ABSTRACTS only; supporting figures were not searched (brief rule 80)."
}
How this could bite us back
  • TopBP1 is the ATR activator and 'KMT2A binds the ATR activator' is a sentence that would travel much further than its evidence. It is one two-hybrid panel, in yeast, in a 2007 paper about SPBP, with no follow-up.
  • The domain is almost certainly not in the fusion. A contact on a lost domain cannot be a fusion mechanism however well it is eventually confirmed, and the entry should never be cited as though it were.
  • The temptation here is to fold TOPBP1 into the menin ATR-CHK1 thread because both are ATR-adjacent. They share a pathway name and nothing else - no bait, no assay, no paper, no cell type.
Still open
  1. Is the KMT2A ePHD retained or lost in KMT2A::AFF1? KMT2A's PHD cluster sits well C-terminal to the breakpoint cluster region, so the expectation is that the fusion loses it entirely - which would make this contact a property of wild-type KMT2A only, and that is worth stating explicitly rather than assuming.
  2. Does the KMT2A ePHD-TopBP1 BRCT6 contact occur in a human cell? Nineteen years, one yeast two-hybrid panel, no co-IP. A single pull-down from a human nuclear extract would move this from weak to moderate or kill it.
  3. How many other real contacts on the six core proteins are sitting in control panels of papers about other proteins? This one is invisible to every abstract-level query, and the only reason it is in this graph is that IntAct's curator read Figure 7D.
  4. TOPBP1's nucleolar condensate is Treacle- and rDNA-damage-dependent (PMID 42152682), and KMT2A-rearranged lines have been reported vulnerable to ribosome-biogenesis inhibition. Does a KMT2A fusion raise rDNA damage or nucleolar stress enough to occupy the TOPBP1 platform constitutively - and if so, is that a dependency or a liability?
Every source behind this entry (9)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    IntAct's REST record holds exactly one TOPBP1 interaction against any of the six core proteins: EBI-20560922, KMT2A as bait, typed MI:0915 physical association, detection '2 hybrid', host Saccharomyces cerevisiae, no expansion method, from PMID 17913746; the same publication also supplies TOPBP1-KMT2B and TOPBP1-KMT2D two-hybrid rows and five TOPBP1-TCF20 rows across four detection methods.

    What this evidence cannot show

    One database on one day, REST endpoint (brief rule 93). A yeast two-hybrid row is a heterologous-host assay on overexpressed fusion constructs; it is not evidence that the contact occurs in a human cell. The KMT2B/KMT2D assignment carries brief rule 66's ambiguity - the paper writes 'MLL2 and -4', which in 2007 could map either way, and IntAct's choice was not independently verified here.

    Evidence that argues against this (2)
    • The curator says in the record that the TOPBP1 identity is an INFERENCE. Every one of the 17 rows from this publication carries the annotation 'data-processing (Origin of TOPBP1 inferred from sequence length and mutant mapping)'. So the accession Q92547 on the KMT2A row was not read off an identifier in the paper; it was assigned by IntAct from length and from where the mutants mapped. That is the same class of hazard as brief rules 62-66, admitted up front by the database rather than hidden - and it means a claim about WHICH TopBP1 rests on a curator's reconstruction. EBI-20560922 also carries intactMiscore 0.37, the lowest of the pairs in this beat. EMBL-EBI IntAct 2026
    • 'Exactly one against any of the six core proteins' is exactly one at this endpoint. Measured the same day, PSICQUIC id:Q92547 returns 38 records against REST's 85 (45%), so more than half of TOPBP1's curated record is unreachable by the query form this project's earlier interactome analyses used, and a second core-bait row inside that half would not be visible to either instrument's user without the comparison. The completeness of a one-record count is a property of the endpoint that produced it. EMBL-EBI IntAct and PSICQUIC 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    The source publication is about SPBP/TCF20, not about KMT2A: a yeast two-hybrid screen with the C-terminal ePHD3-12 region of SPBP (amino acids 1837-1960) as bait identified TopBP1, the interaction was confirmed by co-immunoprecipitation and GST pull-down, mapped to SPBP's ePHD and TopBP1's BRCT6, and SPBP and TopBP1 together co-operate to stimulate Ets1-mediated transcription on the c-myc and MMP3 promoters.

    PMID 17913746DOI
    What this evidence cannot show

    Full text read this session via Europe PMC fullTextXML (PMC2095823, 75,142 characters). HeLa cells and human fibroblasts; no haematopoietic system, no KMT2A fusion (brief rule 73). SPBP is the protein under study; KMT2A appears only in the domain-family comparison.

    Evidence that argues against this (1)
    • IntAct records an explicit limitation in the SPBP-TopBP1 arm that this claim presents as settled. The curator's comment on the co-immunoprecipitation reads: 'Both GFP-tagged and endogenous TCF20 were present in the cell, it is not possible to tell in this experiment which was present in the immunoprecipitate.' So 'confirmed by co-immunoprecipitation' is confirmed with a tagging ambiguity the curator thought worth writing down, and the 17 rows from this paper include ChIP rows against ENSG identifiers rather than proteins - which are protein-DNA records sitting in the same publication's interaction set. EMBL-EBI IntAct 2026
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In the same paper's Figure 7D, a yeast two-hybrid assay reports that the ePHD domains of the trithorax group proteins MLL, MLL2 and MLL4 can bind TopBP1, while the SPBP homologue RAI1 shows very low affinity (confirmed by GST pull-down) and the related RING-finger protein hRNF4 does not bind at all; the KMT2A construct boundaries appear only inside the figure panel and are not stated in the text.

    PMID 17913746DOI
    What this evidence cannot show

    Read from the full text, but the panel itself was not viewed - the claim rests on the authors' sentence describing their own figure. No quantitation is given for the trithorax arm, no residue boundaries for the MLL construct, no reciprocal arm, no human-cell confirmation, and no follow-up publication in nineteen years. The string 'MLL' occurs twice in 75,142 characters of full text, so nothing further about the construct is recoverable from this paper. Brief rule 66 applies to 'MLL2 and -4'.

    Evidence that argues against this (1)
    • The authors themselves read the result as a general property of some ePHD domains rather than as a specific MLL-TopBP1 interaction, and the paper's own conclusion section makes no claim about MLL. Reading it as a KMT2A partner is a stronger statement than the source makes. Sjøttem E et al. 2007
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    As of 2026-09-21, PubMed returns 0 records for TOPBP1[tiab] AND (MLL OR KMT2A)[tiab], while TOPBP1[tiab] AND leukemia[tiab] returns 4 - so the instrument can return a one and the zero is a real bound on titles and abstracts, not a broken query.

    What this evidence cannot show

    Zero in titles and abstracts; supporting figures not searched (brief rule 80) - and the measurement that generated this entry IS in a supporting figure of a paper whose abstract names neither protein, so this is a worked demonstration of that rule rather than a counted absence of biology. Name list tried: TOPBP1, MLL, KMT2A. The strings TopBP1, KIAA0259 and the trithorax phrase were not run as separate operands; the paper found here would not have been reached by any of them either, because the abstract contains none of them.

    Evidence that argues against this (1)
    • The unrun strings in this caveat were run, and the zero holds; the field tag, however, is not doing what the caveat says. Re-run 2026-09-21: '(TOPBP1 OR "TopBP1" OR KIAA0259 OR "topoisomerase II binding protein 1" OR "topoisomerase IIbeta binding protein 1")[tiab] AND (MLL OR KMT2A OR trithorax OR "mixed lineage leukemia" OR "mixed-lineage leukemia")[tiab]' still returns 0, and TOPBP1[tiab] AND leukemia[tiab] still returns 4 - so the widened name list changes nothing and the positive control holds. But PubMed's querytranslation shows KIAA0259 silently collapsing into 'topbp1' and, more importantly, the parenthesised (MLL OR KMT2A) group losing its [tiab] and expanding to [All Fields]. The bare TOPBP1[tiab] does translate to Title/Abstract. So this zero is TOPBP1-in-title-or-abstract against MLL/KMT2A ANYWHERE in the record, which is a broader and therefore stronger absence than the claim states. NCBI PubMed 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Menin's chromatin-bound pool rises 2.8-fold 4 hours after UV in HEK293 cells, is blocked by caffeine and is ATR-CHK1 dependent - the thread this beat was dispatched to test TOPBP1 against, and which the TOPBP1 record turns out not to belong to.

    PMID 16690369DOI
    What this evidence cannot show

    Cited here only to record that the predicted connection was tested and did not hold. HEK293 cells; no haematopoietic system, no KMT2A fusion (brief rule 73). TOPBP1 is the canonical ATR activator, so the prediction was reasonable; the record simply comes from a different experiment about a different question.

    Evidence that argues against this (1)
    • This 2006 account is no longer the best available account of menin in the DNA-damage response, and the current one has no ATR-CHK1 step at all. Lian et al. 2026 report menin directly recognising gamma-H2A.X through a Val371-Gln400 motif, recruiting SETD2 to write H3K36me3, which LEDGF reads to deliver CtIP for end resection and RPA2/RAD51 activation, with Men1 ablation producing spontaneous lung carcinogenesis and severe genomic instability. Two consequences: the prediction this beat was dispatched on failed for a better reason than 'the record belongs to a different experiment' - the ATR-CHK1 framing itself is superseded; and a 2.8-fold chromatin-loading measurement in HEK293 from J Surg Res 2006 has had twenty years without replication, which is a weaker foundation than the narrative use this item is put to. Lian A et al. 2026
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt Q92547 places TOPBP1 in the nucleus and on chromosome with fourteen coded experimental citations, and additionally in the cytoplasm at the centrosome and spindle pole (both ECO:0000269 PubMed:15138768), noting recruitment to double-strand breaks.

    What this evidence cannot show

    Database comment on one day. The chromosome/foci address is damage-conditional, not constitutive. The centrosome and spindle-pole calls rest on a single publication. Nothing haematopoietic or leukaemic in the coded set.

    Evidence that argues against this (1)
    • 'Nucleus and chromosome' is a mitotic-interphase summary rather than a general address, and the record's own weakest lines say so. The single publication behind the centrosome and spindle-pole calls reports TopBP1 at centrosomes in MITOSIS and on chromosome cores in MEIOSIS - two cell-cycle- and germline-specific distributions that no steady-state compartment list captures. A protein whose address changes with cell-cycle phase and with meiosis cannot be summarised by a static field, and this entry's own subnuclear content shows three further condition-dependent states. Chromosoma 2004
  7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Treacle, phosphorylated by CK2 and ATR/ATM, nucleates TOPBP1 phase separation within the nucleolar fibrillar centre, creating a spatially confined signalling platform that amplifies nucleolar DNA-damage-response signalling and directs rDNA repair pathway choice.

    PMID 42152682DOI
    What this evidence cannot show

    Non-haematopoietic human cells, including an inducible oligomerisation system that forces condensation and is therefore partly synthetic. Abstract-level reading. The condensate is inducible and rDNA-damage-dependent, so 'TOPBP1 is nucleolar' without that condition is wrong.

    Evidence that argues against this (1)
    • Two objections, and the first is methodological: the study uses an INDUCIBLE TOPBP1 OLIGOMERISATION system, which forces the behaviour being measured, so the physiological threshold for condensation is not established by it. The second is that the nucleolus is not special - TOPBP1 biomolecular condensates are reported and pursued as a therapeutic target in advanced colorectal cancer, outside any nucleolar framing. If TOPBP1 condenses wherever it is concentrated, then 'nucleolar fibrillar centre' names where the damage was, not where the protein lives. Elife 2025
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    TopBP1 is required for conventional dendritic cell differentiation in mice - DC-specific deletion causes cDC deficiency with pre-DC accumulation and abolishes Flt3L-driven tumour immunity - and the defect is not attributable to DNA-damage-driven cell death; TopBP1 also has an established role in V(D)J recombination during B and T cell development.

    PMID 42104014DOI
    What this evidence cannot show

    Murine conditional knockout, dendritic lineage, no KMT2A fusion and no human data. Abstract-level reading. Brief rule 73 applies: this is mouse cDC development, not human infant B-lineage KMT2A::AFF1.

    Evidence that argues against this (1)
    • A differentiation step with high proliferative demand can manufacture a lineage-looking phenotype from a general requirement. TopBP1's canonical job is DNA replication and damage-checkpoint signalling, which every dividing progenitor needs; the authors exclude DNA-damage-driven cell DEATH but not replication-stress-driven failure to complete a proliferative transition. The pre-DC-to-cDC step is exactly such a transition, so 'required for cDC differentiation' and 'required wherever replication is stressed' are not separated by this experiment. J Biol Chem 2001
  9. E9moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    Human Protein Atlas v25.1 assigns TOPBP1 main subcellular locations Nucleoplasm and Nuclear bodies (Plasma membrane additional) at reliability Supported, and low immune cell specificity in the Blood Atlas.

    What this evidence cannot show

    Antibody immunofluorescence in non-haematopoietic lines and in undamaged cells; HPA's Nuclear bodies class does not identify which bodies, so its agreement with the condensate literature is suggestive only. The Plasma membrane additional call is not supported by anything else and is probably noise.

    Evidence that argues against this (1)
    • This row adds little and one of its calls is not credible: the same HPA run scores Plasma membrane for a BRCT-domain nuclear scaffold. More importantly its 'Nuclear bodies' class does not identify which bodies, so it cannot corroborate the nucleolar condensate - that had to be found by dedicated imaging with a named nucleator and a phospho-dependent trigger. Treating an unresolved 'Nuclear bodies' call as agreement with a specific compartment is the kind of loose join this project has been bitten by. Nucleic Acids Res 2026
Connected to