FANCD2 binds DNA and functions with FANCI in the ID2 repair complex. DNA binding, regulated clamp formation and recruitment of other repair proteins support interstrand-crosslink repair and additional genome-maintenance roles. The biochemical performer remains explicit: FANCL/UBE2T transfer ubiquitin, nucleases cut DNA and polymerases synthesize it; FANCD2 is not assigned those activities from complex membership.
The core uses GO:0003677 DNA binding to encompass two directly distinguished states: purified human FANCD2 preferentially binding ssDNA in PMID:19609304 and human ubiquitinated ID2 encircling duplex DNA in PMID:32269332. The latter structure places much of the DNA-contact groove on FANCI but also identifies FANCD2 DNA contacts. It is not evidence that isolated FANCD2 has every property of the heterodimer. The more specific ssDNA annotation remains intact; the core's broader activity preserves the experimentally supported substrate/state breadth rather than creating a redundant second unit.
GO:0003689 DNA clamp loader activity requires ATP-hydrolysis-driven opening and closing of a clamp. It is not inferred from ID2 ring closure or from ATP present in the ubiquitination reaction. No independently established FANCD2 ATPase, nuclease, polymerase or polymerase-processivity reaction is assigned. The finite official term searches did not establish a more suitable dedicated ID2-clamp activity or complex term; the supported DNA repair complex term is retained.
The three core process facets are interstrand cross-link repair, double-strand break repair via homologous recombination and replication fork processing. Current official is_a/part_of closures show that none is an ancestor of another. They summarize the DNA-associated repair role, not three independently measured catalytic reactions. The chicken DT40 HR inference, Xenopus ICL experiments and human reconstitution/restart experiments remain distinct below. Aphidicolin restart fits the restoration/restart definition; hydroxyurea fork restraint is retained separately without asserting a demonstrated physical reversal mechanism.
The previous core's checkpoint process link is omitted only from this activity-to-process grouping. Its accepted annotation remains unchanged. Human transformed-cell complementation supports a checkpoint contribution, whereas primary mouse fibroblasts can retain an IR/PUVA checkpoint; these data do not establish that the same DNA-contact or clamp state directly performs every checkpoint response. Chromatin and nucleoplasm remain distinct locations: the latter excludes chromosomes and nucleolus. Neither location is used to erase the other.
PMID:19965384(https://pubmed.ncbi.nlm.nih.gov/19965384/) directly tests Xenopus ID2 depletion and add-back in replication-coupled interstrand-crosslink repair. Wild-type versus ubiquitination-site mutant experiments support chromatin recruitment and repair participation. They do not make FANCD2 the incision enzyme or polymerase. Human PMID:19995904(https://pmc.ncbi.nlm.nih.gov/articles/PMC2815579/) independently supplies chromatin-associated FANCD2/POLN evidence; selected Methods and Results distinguish benzonase-treated association from a resolved binary interface.
PMID:32269332(https://pubmed.ncbi.nlm.nih.gov/32269332/) reconstitutes full-length human ID2 and resolves the FANCI–FANCD2 heterodimer and its ubiquitinated DNA-clamp state. FANCL and UBE2T catalyze ubiquitin transfer. The resolved interface independently supports the heterodimerization refinement of the older FANCI interaction records; it does not retrospectively turn the older coimmunoprecipitations into structural assays. The selected original Methods, Results and captions were read; no new cryo-EM images or coordinates were analyzed.
Human PMID:30335751(https://pubmed.ncbi.nlm.nih.gov/30335751/) follows recruitment to nuclear psoralen/UVA lesions, while PMID:26323318(https://pubmed.ncbi.nlm.nih.gov/26323318/) separates nuclear/chromatin fractions. The encompassing nucleus term is retained across these pools. An undetectable untreated chromatin signal is an assay limit, not proof of absence. No inference of identical localization of all natural products follows.
The interaction studies distinguish individual partners and assay resolutions. Positive association is retained without assigning the partner’s catalytic activity to FANCD2.
Current official definitions of enzyme binding, protein kinase binding, protein heterodimerization activity and DNA-binding transcription factor binding were inspected. Binding to these partner classes does not require regulation of the partner's activity. The specific refinements therefore do not assign the partner's chemistry to FANCD2. Where no defensible specific molecular activity was established, supported generic binding remains non-core under the standing project instruction.
PMID:20603015 Figure 1A explicitly identifies human KIAA1018/MTMR15/FAN1 as Q9Y2M0. Its Figure 4 experiments and Methods examine that 114-kDa nuclease. The saved current IntAct query for PMID:20603015 instead assigns Q16658/FSCN1, with FAN1 as an alias, in FANCD2 interaction records IM-15159-2, -3, -6, -7, -9 and -10. This is an affirmative identifier discrepancy grounded in the original target and current source records, not a guess from a paper title.
The Q16658-specific current context is UNDECIDED and has a CONFIRMED source-defect CHECK request. The independently imported Q9Y2M0 context remains positive. No source field is silently changed, no report has been sent upstream, and no claim is made that FANCD2 cannot associate with FSCN1 in another setting.
These additional functional observations have distinct experimental and inference scopes.
The actual PANTHER version 19 PTHR32086 tree and alignment place the human node PTN002528859 below PTN000788426. Its ungapped 1,451-residue sequence matches the displayed Q9BXW9 sequence. Seven non-negated IBD assertions occur on that ancestral node: nucleus, condensed chromosome, DNA polymerase binding, interstrand cross-link repair, intra-S DNA-damage checkpoint, meiotic chromosome pairing and meiotic double-strand-break repair. These are ancestral assertions, not independent similarity transfers. Human FANCD2 appearing among the descendant evidence for nucleus and polymerase binding is expected and is not circular support.
The donor identities were checked against current official records: mouse Fancd2/Q80V62/MGI:2448480, fly Fancd2/Q8MYL1/FBgn0038827, Arabidopsis FANCD2/F4JIJ1/AT4G14970 and worm fcd-2/Q9U2H4/WBGene00012767. Other protein products were not silently equated with these records.
Human checkpoint corroboration includes the actual cached full text of PMID:14988723(https://pubmed.ncbi.nlm.nih.gov/14988723/): transformed FA-D2 cells show partial loss of cross-link-induced DNA-synthesis inhibition, restored by complementation, with CHK1 and NBS1/FANCD2 branches distinguished. PMID:12086603(https://pubmed.ncbi.nlm.nih.gov/12086603/) supplies abstract-level ATM/S222 checkpoint support. These contexts are explicitly distinguished from the primary-mouse negative result.
For polymerase binding, the actual original main of PMID:19995904(https://pubmed.ncbi.nlm.nih.gov/19995904/) reports Flag-POLN/endogenous FANCD2 association in human 293T cells and endogenous, cell-cycle-dependent associations in HeLa cells. The IP Methods include benzonase; Discussion still allows indirect association. This supports an accessory binding function, not a purified binary interface or FANCD2 polymerase activity. Its immutable local publication cache is abstract-only; external original availability and the selected passages actually read are separate facts.
The Human Protein Atlas FANCD2 record reports nucleoplasm with CAB016117 in U-251 MG and HPA063742 in A-549, SiHa and U2OS. HPA063742 also reports A-549 nucleoli and SiHa/U2OS nuclear bodies. Actual saved A-549 and U2OS multichannel images were viewed without quantification. Nuclear puncta were not assigned a DNA-damage identity from morphology alone. The HPA063742 antigen begins at residue 253 in products Q9BXW9-1, -2 and -3 and is not present as a complete sequence in the short -4 product. Shared antigen sequence prevents isoform-specific localization claims.
The encompassing nucleus annotation retains chromosome-associated and other nuclear pools. It is not replaced by nucleoplasm merely because a narrower location is also supported. A-549 nucleolar localization is retained as non-core without inferring ribosome-biogenesis activity.
The actual InterPro IPR029448 entry maps to DNA repair and covers residues 1–1444 of the displayed 1,451-residue sequence. The broad repair term preserves ICL and other supported repair contexts; a signature name containing “nuclease” does not establish FANCD2 nuclease chemistry. Official UniProt nucleus mapping SL-0191 and current mouse localization support the nuclear electronic assertion. The inherited GO_REF:0000044 and current GO_REF:0000120 versions remain explicitly distinguished rather than silently substituted.
Eleven source events were traced through 25 explicit input/output/component or member paths to FANCD2 Q9BXW9 entities. The target entities themselves carry nucleoplasm and 1–1451 coordinates: free FANCD2 (419522), phosphorylated FANCD2 (6788481), and monoubiquitinated/phosphorylated FANCD2 (6785318). The location does not depend only on the reaction's compartment.
The events are ID2/core binding, ID2 monoubiquitination, FANCD2–FANCI binding, nuclease recruitment, cross-link unhooking, POLN recruitment, POLN synthesis, ID2 deubiquitination, ATR recruitment, FANCD2 phosphorylation, and CDK12-dependent expression.
FANCD2 is a substrate or structural component in these assemblies. Membership in a composite catalyst does not confer every component's chemistry. In the CDK12 event, FANCD2 is a member of the protein output set (6797713), not a protein performing transcription. Saved prior objects were reused where refreshes returned rate-limit bodies; those failed requests are not counted as successful new graph retrievals or as biological absence.
The four literal products remain Q9BXW9-2 (1451 residues), Q9BXW9-1 (1471), Q9BXW9-3 (1249) and Q9BXW9-4 (241). The exon-44-containing purified construct, human full-length ID2 preparations, engineered fragments and mutants, and endogenous mouse tag are separate reagents. The S/L bands in older papers describe posttranslational ubiquitination states, not an automatic assignment to these splice products. Neither the broad DNA-binding unit nor the HPA antigen establishes equivalence of all four products.
The primary-study assessment includes the actual source and supplement/pixel scope specified above. The database assessment includes selected original text and HPA images, but no new article-pixel or supplement read. The additional synthesis reading revisited selected human ID2 structural/EMSA text, purified FANCD2 preparation/construct Methods, and the contrasting human/mouse checkpoint Results; it did not add new article pixels. Exact private read locators distinguish these scopes rather than claiming every paper was read in full.
Reference full_text_unavailable flags describe immutable normal-cache metadata. External complete originals, selected passages and actual reading are separately recorded. All 38 inherited reference identities, titles and findings survive. The 22 current whole-reference assessments and ten added donor/corroborating references are retained without overwriting either source scope; the database assessment left the 38 old reference objects unchanged, so no conflicting shared reference-review fields required adjudication. The separate database reading of PMID:19995904 remains described above. No new publication quotation is introduced.
The current GOA partner contexts are represented separately. NBS1 association remains non-core; MRE11 and FAN1 are enzyme partners; FANCI is the heterodimer partner. The imported Q16658/FSCN1 context remains unresolved with the confirmed IntAct identifier CHECK, while the positive Q9Y2M0/FAN1 context remains separate. The historical FAN1 quotation retained with the older record is provenance for the studied FAN1 protein, not supporting evidence for FSCN1. The source identifier is not silently repaired.
The older GO_REF:0000044 nuclear annotation is archived as a superseded source version, with its positive review preserved. The current GO_REF:0000120 nuclear assertion remains active. The historical authored R-loop-disassembly proposal is withdrawn from active assertions while retaining its UNDECIDED review, source and quotations. This curator withdrawal is distinct from source disappearance and does not deny the supported RNA-export/R-loop-homeostasis biology. There are 51 stored annotations, 49 active (45 current imported contexts and four inherited proposals); these counts supersede the earlier partial-stage counts in the notes. All original quotation associations and natural products remain preserved.
The two newly adopted source contexts require GO_REF:0000107 and GO_REF:0000120 in the reference closure. The genuine seeder emitted placeholder titles; these were not accepted. Their complete official GO reference pages and existing HTTP 200 receipts were independently read, and two separate whole metadata objects with exact official titles were appended. All 48 accepted reference objects remain exact; the final total is 50. These are source-method references, not new biological experiments or normal-cache files.
Six citation-location fields from PR #5481 are retained after verification against the unchanged UniProt record; they do not alter the biological review or quotation text.