DCLRE1C

Summary: Artemis cleaves internal phosphodiester bonds in exposed single-stranded DNA and at single-strand/double-strand junctions, including DNA hairpins and overhangs.

Reason: PMID:11955432 reports hairpin and overhang endonucleolysis by Artemis with DNA-PKcs. Its original full body remained inaccessible after bounded retrieval, so its detailed experiments are not claimed as read. The activity is independently established by the full original PMID:20117966: purified recombinant human Artemis cleaved 5′- and 3′-labelled ssDNA, with time-course products consistent with internal cleavage rather than 3′ exonucleolysis, and the activity tracked purification and antibody effects. This matches the current GO definition. PMID:16874298 additionally demonstrates human Artemis-mediated hairpin and overhang cleavage with DNA-PK. DNA-PKcs autophosphorylation can enable the reaction without requiring phosphorylation of Artemis itself; salt, Ku, substrate and construct affect the observed requirements. These are catalytic experiments, not an inference from repair necessity or family membership. Retain the source assertion and its unqualified product scope, without assigning the activity to all splice products.

Summary: Artemis contributes to telomere stability in particular cellular contexts.

Reason: The PAINT assertion is placed at PTN002757677 on the actual human lineage. Its literal descendant evidence includes mouse Dclre1c as well as Apollo and MRT-1 homologs; donor count or target/paralog naming does not invalidate the placement. Independently, human Artemis-defective fibroblasts show mild telomeric fusion and population-doubling-dependent telomere-dysfunction phenotypes. The younger cells did not show the same TIF defect, and the study did not identify a direct Artemis telomere-incision reaction. Retain the contextual contribution without elevating telomere maintenance to the principal nuclease role.

Summary: Damaged-DNA recognition is consistent with the inherited repair-nuclease function.

Reason: The literal IBD lies at family root PTN000581728, ancestral to the exact human sequence, and is grounded in the experimental yeast PSO2 annotation. This is a curator-inferred ancestral function, not a pairwise similarity transfer. The term denotes binding to damaged DNA and does not require recognition of a chemically modified base. Artemis processes DNA break-end structures in NHEJ; the inference is retained with the original evidence code, without claiming that the yeast source directly assayed human Artemis or established a damage-specific affinity constant. Separately read structural work observes a DNA overhang contacting the Artemis active site, with partial-occupancy and disorder limitations; it does not measure affinity for chemically modified bases. This broad inherited DNA-recognition descriptor is retained as non-core; the established catalytic end-processing activity is represented by the endonuclease core.

Summary: Refine the documented nuclease reaction to DNA endonuclease activity.

Reason: The actual Reactome reaction is DNA double-strand-break end processing. Catalyst 175587 reaches the 1–692-residue Q96SD1 component through complex 5687161, and the human biochemical evidence concerns internal DNA cleavage, including hairpin opening and overhang processing. DNA endonuclease activity is more informative than the nucleic-acid-wide parent and complements the separately retained ssDNA-specific annotation. The phosphorylated Reactome forms are modelled proteoforms, not evidence that all splice products are equivalent; DNA-PKcs activation should not be conflated with a universal requirement to phosphorylate Artemis itself.

Proposed replacements: DNA endonuclease activity

Summary: Artemis directly binds the DNA ligase LIG4 through its C-terminal region.

Reason: The literal source partner P49917 is human LIG4, an experimentally characterized ATP-dependent DNA ligase. The original combines human cellular co-immunoprecipitation with bacterial-protein pull-downs and binding of the Artemis 485–495 peptide to the LIG4 DNA-binding domain. It separates LIG4 binding from the PRKDC-binding interface. Enzyme binding accurately refines the measured partner class; it does not assign DNA ligase activity to Artemis. The peptide and engineered-mutant experiments do not establish equal binding by all official splice products.

Proposed replacements: enzyme binding

Summary: Artemis associates with the serine/threonine kinase DNA-PKcs.

Reason: The literal partner P78527 is PRKDC, whose curated function and reaction objects establish serine/threonine protein kinase activity. The source-specific human cellular co-immunoprecipitation and Artemis-interface mutants support this binding class independently of the LIG4 interaction. Refinement names the bound enzyme class; it does not attribute kinase catalysis to Artemis or prove that Artemis phosphorylation is the activating event.

Proposed replacements: protein serine/threonine kinase binding

Summary: The Artemis C-terminal peptide forms a defined complex with LIG4.

Reason: The original structure and calorimetry directly measure the human LIG4 DNA-binding domain with synthetic Artemis residues 485–495, with subsequent cellular interface-mutant tests. The actual source-specific interaction records also retain these constructs. LIG4 is a catalytically established DNA ligase, so enzyme binding is the supported informative class. The isolated peptide interface is not evidence of catalytic activity by Artemis or of identical interactions for every splice product.

Proposed replacements: enzyme binding

Summary: A direct Artemis-fragment/LIG4 complex supports enzyme binding.

Reason: The current interaction query returned no literal pair for this PMID, but the recovered original directly reports human LIG4 residues 1–609 crystallized with synthetic Artemis residues 485–495. The original source, rather than a query nonmatch, establishes the interaction. LIG4 is an ATP-dependent DNA ligase, supporting enzyme binding; the experiment is a fragment complex and does not establish an intact Artemis–LIG4 structure or equivalent binding by short splice products.

Proposed replacements: enzyme binding

Summary: Artemis is associated with DNA-PKcs in the nuclear repair machinery.

Reason: The current literal interaction query did not recover this source pair. The original nevertheless explicitly detects Artemis in nuclear FLAG–DNA-PKcs immunoprecipitates from human HEK293S cells, with an empty-vector control and ethidium-bromide treatment; the association persisted independently of DNA and was not increased by the tested irradiation condition. P78527 is a serine/threonine protein kinase, supporting the narrower binding class. This does not transfer the kinase activity or the paper’s XLS perturbation results to Artemis.

Proposed replacements: protein serine/threonine kinase binding

Summary: The BioPlex interaction associates Artemis with the protein serine/threonine kinase DNA-PKcs/PRKDC.

Reason: The actual positive IntAct record for this source pair is Q96SD1–P78527, PMID:33961781, EBI-54776522/IM-29278-14071. It records C-terminally tandem-tagged Artemis bait and PRKDC prey in human HEK293T cells, detected by anti-tag co-immunoprecipitation, with spoke-expanded association semantics. Both sequence/taxon ROGIDs match the official human proteins. The original Methods describe FLAG-HA-tagged human ORFs, anti-HA affinity purification and mass spectrometry; the network analysis aggregates at gene level. PRKDC is an experimentally established protein serine/threonine kinase, as recorded in its official function and catalytic reaction objects. GO:0120283 specifically denotes binding to that kinase class, making it an informative partner-class refinement of this positive interaction. This does not establish a purified binary contact, kinase activation by Artemis, a kinase activity for Artemis, or interaction by every Artemis splice product. The imported row's absent isoform is preserved. The interaction accession is shared with other spoke-expanded pairs and was not used alone as the identity key.

Proposed replacements: protein serine/threonine kinase binding

Summary: The primary Artemis transcript encodes a nuclear protein; nucleoplasm is the more informative retained location.

Reason: The complete cached abstract explicitly reports nuclear localization of the wild-type primary transcript and correction of the coding-joint defect. The full original remained inaccessible, so its unexamined localization controls and alternative-transcript details are not inferred. Retain that positive source observation while using the independently supported nucleoplasm annotation for the more informative cellular location; no source isoform field is retrofitted.

Summary: Human Artemis was directly localized within the nucleus outside nucleoli.

Reason: The separately recovered original describes C-terminal myc–His-tagged human wild-type Artemis and engineered mutants in HEK293T cells. Results and immunofluorescence Methods report nuclear distribution excluding nucleoli. The broader nucleus assertion is correct but less informative than the independently retained nucleoplasm location. The original cache remains abstract-only; the external full-text access and exact source locators are documented separately.

Summary: The phylogenetic nucleus assertion is valid but broader than nucleoplasm.

Reason: The literal IBD at PTN000581728 is on the human lineage. Its donors include direct human EXP nucleus records for Q96SD1 as well as homolog evidence; the target in its own WITH/FROM is legitimate descendant evidence, not circularity. The broad inherited location is retained as positive evidence but marked too general relative to supported nucleoplasm.

Summary: The UniProt subcellular mapping is correct but broad.

Reason: The source is the UniProt SubCell SL-0191 nucleus mapping, not an InterPro, EC or ARBA inference. Nuclear localization is supported by the target experiments, while the retained nucleoplasm annotation is more informative. The original deterministic mapping and source metadata are preserved.

Summary: Nucleoplasm is an experimentally supported site of Artemis localization.

Reason: The actual HPA record for ENSG00000152457/Q96SD1 identifies nucleoplasm as the main subcellular location, with supported immunofluorescence reliability and antibody HPA069295. This review inspected the data record, not the microscopy images. The same record lists Golgi as an additional location; that is not erased or elevated to a new core. Nuclear repair-complex and tagged-protein evidence independently corroborate nucleoplasmic residence.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during ATM-mediated phosphorylation of Artemis. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during recruitment of Artemis into an ATM-associated complex. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during assembly with PRKDC and Ku at DNA breaks. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during the modelled PRKDC-associated phosphorylation step. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during the repair assembly during POLL/POLM gap filling. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during Artemis-mediated DNA-end processing. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during assembly with LIG4/XRCC4 and accessory NHEJ factors. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during the repair assembly during TDP1/TDP2 end processing. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis is represented in a nucleoplasmic NHEJ assembly.

Reason: The exact Reactome source places Q96SD1 in nucleoplasm during the repair assembly during LIG4-dependent ligation. Independently traversed physical-entity paths reach the human 1–692-residue component, including unmodified or explicitly phosphorylated proteoforms. This supports location in the repair assembly without transferring the catalytic activity of ATM, PRKDC, polymerases, phosphodiesterases or LIG4 to Artemis. The location assertion does not require accepting every historical phosphorylation-necessity statement in the pathway summary, and the model does not establish equivalence of all splice products.

Summary: Artemis contributes a DNA-end-processing step to nonhomologous end joining.

Reason: The NHEJ IBD at PTN000581728 is on the reconstructed human lineage and has a literal Dictyostelium Dclre1 experimental donor. That distant donor is a valid basis for the curated ancestral inference, not a reason to reject it. Independent human Artemis end-processing, interaction and repair studies establish the same performed role. Artemis cleaves appropriate DNA ends before ligation; it is not merely a substrate consumed by this process.

Summary: Artemis is an enzymatic participant in the NHEJ pathway.

Reason: The source pathway includes Artemis-containing DNA-end processing, with complete positive physical-entity paths to human Q96SD1. Artemis cleaves obstructive DNA-end structures before downstream ligation. The positive target paths support participation without implying that the entire pathway graph was inspected, that Artemis performs polymerase, kinase or ligase chemistry, or that every NHEJ event requires it.

Summary: Purified full-length Artemis supports 5′-directed exonucleolytic removal from single-stranded DNA, with conflicting purification results and uncertain physiological importance.

Reason: Narrow the nucleic-acid-wide term to the experimentally tested single-stranded DNA substrate. The original PMID:11955432 abstract explicitly reports this activity; its full original was not recovered. In the separately read full original PMID:20347402, hydroxyapatite fractionation separated most exonuclease activity from Artemis while retaining endonuclease activity, raising a genuine contamination concern. This is weighed against the later original PMID:24500713: recombinant human WT Artemis and H115A were purified from insect cells; the mutation reduced both activities, WT and mutant had comparable circular-dichroism profiles and retained DNA-PKcs association, and cofractionation, poly(dA) substrates, 3′-labelled substrate time courses and inhibitor controls supported intrinsic, nonprocessive 5′ exonuclease activity. The independent full-length preparation in PMID:34387696 also showed weak exonuclease activity, whereas its engineered catalytic-domain construct did not. Taken together these data support retaining and specifying the biochemical activity rather than removing it as contamination or treating it as equivalent to the dominant endonuclease function. Its contribution to repair in vivo remains unresolved and it is not made a separate core function. No activity is transferred to the four natural products merely from their sequences.

Proposed replacements: single-stranded DNA 5'-3' DNA exonuclease activity

Summary: The inherited DNA-exonuclease term is entailed by the supported single-stranded-DNA-specific replacement already asserted elsewhere.

Reason: The exact IBD at PTN000581728 lies on the human lineage and is grounded in the literal human DCLRE1A/Q6PJP8 IDA annotation. This is a phylogenetic assertion, not direct testing of Artemis in that donor paper. Later target-specific assays support weak full-length Artemis exonuclease activity, despite an earlier purification study separating most of that activity. GO:0045145 single-stranded DNA 5'-3' DNA exonuclease activity is_a GO:0035312 and is already asserted as the MODIFY replacement on the GO:0008409 row (zero-based row 26). That specific supported assertion entails this broader term, so TOO_GENERAL avoids a separate broad assertion while retaining the valid inherited inference. The historical disagreement and uncertain physiological contribution remain explicit; neither the donor's paralog identity nor absence of activity in an engineered catalytic-domain construct establishes universal target inactivity.

Summary: The inherited interstrand-cross-link repair assignment remains uncertain for human Artemis.

Reason: The IBD at family root PTN000581728 is genuinely ancestral to human Artemis, with literal PSO2, Apollo and trypanosome donor evidence; its placement is not rejected because of donor count or naming. Target-specific evidence is mixed: chicken DT40 Artemis loss did not increase cisplatin/MMC sensitivity, whereas an Artemis-null human HCT116 clone was more sensitive to these agents. The human study used an hMLH1-defective background and discusses alternative damage-response mechanisms. Neither sensitivity nor the chicken negative identifies or excludes a human Artemis cross-link-removal step under all conditions. The valid phylogenetic inference therefore remains unresolved at the target-process level; this is not a wrong-gene or nonbinding claim.

Summary: Artemis is a component of the NHEJ repair machinery.

Reason: The literal IBD at PTN000581730 lies on the human lineage and is supported by Q96SD1’s own experimental complex annotation. Target self-evidence is legitimate grounding for a phylogenetic assertion. The source paper independently detects Artemis in nuclear DNA-PKcs-associated repair complexes. Membership is compatible with regulated recruitment and does not mean that every reaction uses a permanently fixed stoichiometric assembly.

Summary: Human Artemis is experimentally associated with the NHEJ complex.

Reason: The actual original identifies Artemis among the nuclear FLAG–DNA-PKcs-associated NHEJ proteins in human HEK293S cells, and the Western blot analysis distinguishes its DNA-independent association from other DNA-dependent copurifications. Although the paper foregrounds XLS, it directly examines Artemis; title emphasis is not misattribution. Complex membership does not transfer XLS-specific functional perturbation results or kinase/ligation activity to Artemis.

Summary: Human Artemis directly contributes coding-end hairpin incision to V(D)J recombination.

Reason: The full original PMID:14744996 tests human Artemis in patient-derived, Artemis-deficient fibroblasts carrying a chromosomal RAG1/RAG2-dependent recombination reporter. Full-length Artemis restores recombination frequency and coding-joint quality, and active-site substitutions compromise this function while retaining DNA-PKcs association. Independently read direct hairpin-cleavage experiments in PMID:16874298 link this cellular phenotype to DNA incision performed by Artemis itself; it is an active enzyme in the process rather than a consumed substrate. The latter study also tests human Artemis in mouse embryonic fibroblasts, a separate explicitly cross-species complementation experiment. Current is_a/part_of ancestry shows that V(D)J recombination is neither an ancestor nor a descendant of any term in the 31-source-row set. The frozen local GO-CAM index has no DCLRE1C/Q96SD1 match; this does not establish absence from all public models. A current human comparator query recovered V(D)J annotations for RAG1 (DNA cleavage) and LIG4 (DNA joining), but none for DNTT (DNA-end nucleotide addition). These proteins perform different chemical steps within the same process; this mixed pattern is not used to infer an annotation gap. The proposal rests on directly measured target catalysis and chromosomal recombination, with no broader immune-development or lymphocyte-differentiation NEW.