ERCC5

Summary: The inherited single-stranded DNA endonuclease activity is directly corroborated on purified human XPG.

Reason: The current PAINT tree places human leaf PTN002497156 below the Eukaryota assertion node PTN000792422 in PTHR16171; the target MSA sequence is an exact 1,186-residue match to P28715. The three relevant IBD assertions are present and there is no target-path NOT. The SGD:S000003490/Rad2 donor has experimental GO:0000014 support from PMID:8247134; its publisher abstract was read but the original Methods and figures remain unavailable. Independently, the actual human original PMID:8078765, Methods pp.3312–3313 and Results/Figures 1–3 pp.3314–3315, expresses the complete XPG coding frame in Sf9 cells, purifies XPG, and demonstrates cleavage of covalently closed single-stranded M13 DNA with activity coeluting with the protein. Thus this decision does not infer isolated-ssDNA cleavage from a bubble-junction assay. The one-donor IBD is valid phylogenetic evidence; these assays do not establish activity of both shorter natural products.

Summary: XPG participates in a functional human nucleotide-excision repair assembly.

Reason: PMID:11259578, Results Figures 2–6 and dual-incision Methods, recovers endogenous HeLa XPG with TFIIH and other NER factors and tests activity on a cisplatin-adduct substrate. The associations are salt-sensitive and do not establish a permanently preassembled repairosome; that limitation does not negate XPG participation in the repair complex.

Summary: XPG belongs to the conserved TFIIH-associated nucleotide-excision repair factor 3 assembly.

Reason: The current PAINT tree places human leaf PTN002497156 below the Eukaryota assertion node PTN000792422 in PTHR16171; the target MSA sequence is an exact 1,186-residue match to P28715. The three relevant IBD assertions are present and there is no target-path NOT. GO:0000112 describes an endonuclease/helicase NER assembly and gives yeast Rad2–core TFIIH-Ssl2 as its example; it is not defined as yeast-only. The actual donor original PMID:8855246, Methods/Results pp.10718–10721 and Figures 2–3, purifies a stable yeast Rad2–TFIIH complex and reconstitutes UV-DNA incision/excision with other NER factors. Human corroboration is provided by PMID:31253769, Results “NER factors stimulate the TFIIH helicase XPD” and “XPG in lesion scanning,” and the protein-purification/crosslinking Methods: recombinant full-length human XPG assembles with core TFIIH and XPA on DNA and crosslinks to TFIIH subunits. XPG was absent from the final cryo-EM density, so it is the biochemical assembly/crosslinking evidence, not direct visualization in that density, that supports human conservation. This does not make XPG a constitutive transcriptional TFIIH subunit or confer helicase activity on XPG itself. PMID:11259578 Figure 6 further recovers functionally active human XPG with TFIIH after washing at 150 mM KCl, but not 300 mM KCl. This supports a recoverable repair subassembly with condition-dependent stability; the absence of a permanent all-factor repairosome does not contradict that smaller NEF3-class association.

Summary: Full-length human XPG directly binds bubble DNA.

Reason: The original PMID:16246722 main experiments remain inaccessible. Independently, PMID:32522879 Results Figures 3–4 and Supplementary Online Methods use separately purified full-length human XPG for bubble-DNA binding, incision and electron microscopy. These assays establish the same molecular function without treating the engineered XPGcat hybrid as the natural protein or claiming the inaccessible transcription-complex assays were verified.

Summary: Human XPG recognizes bubble substrates, with construct-specific structural limits.

Reason: PMID:32821917 directly studies an engineered human catalytic core on a splayed-arm substrate; its crystal is not a full-length XPG–repair-bubble structure. The matching bubble-binding activity is independently demonstrated with full-length human XPG in PMID:32522879 Figure 4 and Supplementary Methods. Preserve the annotation with that explicit corroboration.

Summary: XPG contributes non-catalytically to homologous-recombination repair.

Reason: PMID:26833090 Results Figures 5–6 combine human-cell DR-GFP repair measurements with XPG association with BRCA2/PALB2/RAD51 and reduced recruitment of those proteins after XPG depletion. This supports participation through repair-protein assembly or recruitment, beyond a nonspecific survival phenotype. Retain this additional repair role as non-core; it does not assign strand-exchange catalysis to XPG.

Summary: The source-specific RNA polymerase II interaction remains unresolved.

Reason: PMID:16246722 has a positive abstract describing transcription-bubble recognition, but its operative interaction experiments and construct controls were not recovered. Neither bubble-DNA binding nor membership in an NER assembly independently verifies binding to the RNA polymerase II complex. Preserve the curator assertion as undecided rather than disputing it from incomplete access.

Summary: Refine broad DNA binding to the supported bubble, duplex and single-stranded DNA interactions.

Reason: IPR008918 is genuinely present in human P28715 and supports DNA interaction, but its GO:0003677 mapping does not express XPG substrate recognition. In PMID:32522879, Results Figures 1F, 3D and 4 and Online Methods “Electrophoretic Mobility Shift Analysis,” full-length human XPG and separately identified engineered catalytic constructs are tested on bubble/Y substrates; full-length XPG localizes to the central unpaired bubble in the EM experiments. Refine to all three already represented substrate terms: bubble, double-stranded and single-stranded DNA binding. PMID:12644470 Figure 4 and the phosphate-ethylation-interference Results resolve contacts with duplex DNA and short single-stranded overhangs at the junction; Figure 2 competition shows preference for junction-containing substrates rather than unrestricted isolated-strand binding. The XPGcat construct contains an archaeal FEN substitution and does not cleave bubble DNA, so it is not treated as equivalent to full-length XPG or either natural short product.

Proposed replacements: bubble DNA binding double-stranded DNA binding single-stranded DNA binding

Summary: The structural substrate does not by itself resolve the damaged-DNA binding assertion.

Reason: The actual PMID:32821917 Methods and Results use synthetic splayed-arm DNA with an engineered XPG catalytic core; the resolved structure chiefly shows duplex contacts. GO:0003684 means binding to damaged DNA and does not require preferential affinity over undamaged DNA. Nevertheless, this experiment does not directly identify the annotated damaged substrate. Retain uncertainty about this source-specific claim without denying XPG incision in NER. The supplement and figure pixels were not inspected, so this partial access does not establish that the entire study lacks an appropriate binding experiment. UNDECIDED records that unresolved evidence boundary; it does not require lesion-selective affinity or reject the claim from an artificial substrate alone.

Summary: XPG contacts duplex DNA at a single-/double-strand junction.

Reason: PMID:12644470, Results and Figures 3–4/8, maps phosphate-backbone contacts on both strands of the duplex using human XPG footprinting. This is junction-associated duplex binding, not a claim of unrestricted affinity for isolated duplex DNA; the competition assays favor junction-containing substrates.

Summary: Direct duplex contacts are independently established for human XPG.

Reason: The PMID:8090225 original remains unavailable beyond its positive abstract. Human XPG footprinting in PMID:12644470 Figures 3–4 and the engineered human-core DNA complex in PMID:32821917 establish duplex contacts at a junction. Accept that molecular claim while keeping the source-access and construct limits explicit.

Summary: XPG contacts a single-stranded segment at a DNA junction.

Reason: PMID:12644470 Figure 4 identifies phosphate contacts 1–4 nucleotides into the 5′ single-stranded overhang. Figure 6 tests overhang-dependent binding. Retain ssDNA binding with this junction-associated scope; Figure 2B shows poor competition by isolated ssDNA, so this is not a general ssDNA-coating activity.

Summary: Human XPG directly contacts single-stranded DNA adjacent to a single-strand/double-strand junction.

Reason: The target genuinely matches IPR001044. Independently read original PMID:12644470 uses recombinant human C-terminally His-tagged WT and E791A XPG expressed in Sf9 cells (Experimental Procedures, XPG expression and EMSA; Results Figure 1). Its phosphate-ethylation interference experiment maps WT XPG contacts to positions 4–1 nucleotides into the 5-prime single-stranded overhang, as well as duplex contacts (Results “Phosphate Ethylation Interference DNA Footprinting,” Figure 4A–B). Figure 2B shows strong preference for junction-containing substrates over isolated single-stranded or duplex competitors, and Figure 6 shows loss of junction binding as the complementary 3-prime overhang is shortened. These are positive single-stranded-DNA contacts within a junction-binding mechanism, not a claim of unrestricted high-affinity binding to isolated ssDNA. Retain the family mapping with this experimentally resolved scope; neither cleavage alone nor modeled contacts are substituted for the binding experiments, and neither shorter natural product was tested here.

Summary: The mapped catalytic activity is correct but can be specified as DNA endonuclease activity.

Reason: The cited InterPro signature is present on P28715; the mapping describes a broader activity class than the human experiments establish. PMID:8078765, Methods pp.3312–3313 and Results/Figures 2–3 pp.3314–3315, directly demonstrates internal cleavage of covalently closed ssDNA and the resulting termini by purified human XPG. PMID:32522879 separately demonstrates incision of structured DNA substrates by full-length XPG (Results Figure 1 and Online Methods “Incision assays”); the engineered XPGcat chimera has a different substrate range. Replace catalytic activity with the already present GO:0004520 DNA endonuclease activity. This covers the directly tested DNA cleavage without assigning RNA cleavage, exonucleolysis, or every activity of a nuclease superfamily to the target.

Proposed replacements: DNA endonuclease activity

Summary: The mapped nuclease activity is correct but can be specified as DNA endonuclease activity.

Reason: The cited InterPro signature is present on P28715; the mapping describes a broader activity class than the human experiments establish. PMID:8078765, Methods pp.3312–3313 and Results/Figures 2–3 pp.3314–3315, directly demonstrates internal cleavage of covalently closed ssDNA and the resulting termini by purified human XPG. PMID:32522879 separately demonstrates incision of structured DNA substrates by full-length XPG (Results Figure 1 and Online Methods “Incision assays”); the engineered XPGcat chimera has a different substrate range. Replace nuclease activity with the already present GO:0004520 DNA endonuclease activity. This covers the directly tested DNA cleavage without assigning RNA cleavage, exonucleolysis, or every activity of a nuclease superfamily to the target.

Proposed replacements: DNA endonuclease activity

Summary: Refine endonuclease activity to DNA endonuclease activity.

Reason: PMID:32522879 assays DNA incision by full-length human XPG and, separately, an engineered XPGcat–archaeal FEN hybrid. The full-length protein cuts bubble and splayed-arm substrates, whereas the hybrid has different substrate requirements. DNA endonuclease activity is the supported common catalytic class; a generic nucleic-acid term is unnecessarily broad, and a flap-only term would misstate the substrate range.

Proposed replacements: DNA endonuclease activity

Summary: Refine the generic nuclease term to DNA endonuclease activity.

Reason: PMID:32821917 Methods and Figure 5 directly test DNA incision and catalytic-site mutants of an engineered human XPG core. Independent full-length assays in PMID:32522879 establish the same reaction class. DNA endonuclease activity captures the chemistry without equating the engineered fragment with any natural shorter product.

Proposed replacements: DNA endonuclease activity

Summary: The mapped endonuclease activity is correct but can be specified as DNA endonuclease activity.

Reason: The cited InterPro signature is present on P28715; the mapping describes a broader activity class than the human experiments establish. PMID:8078765, Methods pp.3312–3313 and Results/Figures 2–3 pp.3314–3315, directly demonstrates internal cleavage of covalently closed ssDNA and the resulting termini by purified human XPG. PMID:32522879 separately demonstrates incision of structured DNA substrates by full-length XPG (Results Figure 1 and Online Methods “Incision assays”); the engineered XPGcat chimera has a different substrate range. Replace endonuclease activity with the already present GO:0004520 DNA endonuclease activity. This covers the directly tested DNA cleavage without assigning RNA cleavage, exonucleolysis, or every activity of a nuclease superfamily to the target.

Proposed replacements: DNA endonuclease activity

Summary: Human XPG catalyzes internal DNA cleavage at structural junctions.

Reason: PMID:12644470 Figure 1 and nuclease Methods demonstrate incision of splayed-arm, bubble and stem-loop DNA by purified human XPG. Cleavage geometry differs among substrates, so the DNA endonuclease term is retained rather than restricting every reaction to duplex-only or flap-only cleavage.

Summary: Human XPG DNA endonuclease activity is independently established.

Reason: PMID:8090225 is abstract-only in the recovered material. The exact catalytic class is independently demonstrated by human XPG in PMID:12644470 Figure 1 and full-length assays in PMID:32522879. This accepts the well-established activity without claiming the 1994 original experiments were inspected.

Summary: The matched ARBA rule correctly identifies the experimentally demonstrated DNA endonuclease activity.

Reason: ARBA00029205 condition set 2 (zero-based) requires IPR006085, IPR029060 and Primates; all three are present on the official P28715 record. This is a verified rule match, not independent experimental proof. The original human ssDNA-cleavage experiments in PMID:8078765 (Methods pp.3312–3313; Results/Figures 2–3 pp.3314–3315) and full-length XPG bubble/Y incision in PMID:32522879 (Results and Online Methods) independently substantiate GO:0004520. The broader DNA-endonuclease term appropriately covers these tested substrate configurations without requiring all three natural products to share their activity.

Summary: Reactome identifies ERCC5 itself as the active DNA endonuclease for the 3-prime incision in global-genome NER.

Reason: The current human event R-HSA-5690988 points to catalyst 5691069; its activeUnit is entity 67447, whose referenceEntity is UniProt P28715 and whose modeled coordinates are 1–1186 in nucleoplasm. The catalyst activity is GO:0004520. This explicit active-unit chain attributes the 3-prime incision to XPG itself, independently corroborated by the human nuclease experiments in PMID:8078765 and PMID:32522879. Neighboring ERCC1–ERCC4 5-prime incision and polymerase gap filling are distinct reactions and are not assigned to ERCC5. The model does not establish activity of the 419- or 232-residue products.

Summary: Reactome identifies ERCC5 itself as the active DNA endonuclease for the 3-prime incision in transcription-coupled NER.

Reason: The current human event R-HSA-6782224 points to catalyst 6782216; its activeUnit is entity 67447, whose referenceEntity is UniProt P28715 and whose modeled coordinates are 1–1186 in nucleoplasm. The catalyst activity is GO:0004520. This explicit active-unit chain attributes the 3-prime incision to XPG itself, independently corroborated by the human nuclease experiments in PMID:8078765 and PMID:32522879. Neighboring ERCC1–ERCC4 5-prime incision and polymerase gap filling are distinct reactions and are not assigned to ERCC5. The model does not establish activity of the 419- or 232-residue products.

Summary: Refine NTHL1 association to its experimentally supported activating role.

Reason: The PMID:15358233 abstract explicitly reports recombinant FLAG-XPG binding to NTH1 and stimulation of its thymine-glycol DNA glycosylase/AP-lyase readout; the main assays were not recovered. Independent PMID:9927729 Figures 1–3 demonstrate XPG–NTHL1 association and stimulation. Enzyme activator activity is a more informative replacement; this does not assign NTHL1 glycosylase catalysis to XPG.

Proposed replacements: enzyme activator activity

Summary: The precise human XPG–CSB interaction remains source-limited.

Reason: The positive PMID:16246722 abstract is consistent with XPG–CSB cooperation, but its operative interaction assays and construct controls were not accessed. DNA binding or shared repair-pathway membership alone does not verify this pair. Retain uncertainty without asserting an erroneous curator assignment.

Summary: The original supplement supports human XPG–GADD45A association.

Reason: PMID:17268471 Supplementary Figure S5 and its Methods report co-immunoprecipitation from HEK293T nuclear extracts after expression of human XPG and FLAG-GADD45A, with IgG and FLAG-Smad1 controls. The main article remains unavailable. Retain the measured association without turning it into an XPG demethylase or transcription-factor activity.

Summary: XPG associates with BRCA1 in the examined repair systems.

Reason: PMID:26833090 Figure 7A and Results demonstrate BRCA1–XPG co-immunoprecipitation in human U2OS extracts and in co-infected insect cells expressing recombinant proteins. XPG depletion changes BRCA1 chromatin retention. These data support the interaction, but do not isolate a sufficiently specific molecular adaptor or enzymatic activity to replace this binding assertion.

Summary: XPG associates with the BRCA2/PALB2/RAD51/DSS1 repair machinery.

Reason: PMID:26833090 Figure 5B–F reports reciprocal nuclear co-immunoprecipitation and recombinant coexpression/purification of the named human proteins. Association strengths differ, and the assembled complex is distinct from proof of every binary direct contact. Retain the supported interaction context without assigning RAD51 strand-exchange chemistry to XPG.

Summary: The exact ERCC5–BRAF screening record was not recovered.

Reason: PMID:35512704 main Methods and Results were read, including the BRET-based screening design and mutant/WT comparisons. The ERCC5–BRAF pair is not identified in the inspected main text, and the operative supplementary spreadsheet could not be accessed. Do not generalize a selected BRAF-mutant mechanism or assume a wild-type interaction; leave this imported pair undecided.

Summary: Refine the independently supported XPG–XPD association to TFIIH-class factor binding.

Reason: The original PMID:8652557 remains inaccessible. Independently, PMID:31253769 Results “XPG in lesion scanning” and crosslinking Methods identify human XPG contacts with XPD in a purified TFIIH–XPA–DNA assembly. GO:0001097 captures the verified TFIIH partner class. XPG was absent from the final cryo-EM density; crosslinks are not relabeled as resolved atomic contacts, and the detailed supplementary crosslink table was not independently parsed.

Proposed replacements: TFIIH-class transcription factor complex binding

Summary: Refine the independently supported XPG–XPB association to TFIIH-class factor binding.

Reason: The original PMID:8652557 remains inaccessible. Independently, PMID:31253769 Results “XPG in lesion scanning” and crosslinking Methods identify human XPG contacts with XPB in a purified TFIIH–XPA–DNA assembly. GO:0001097 captures the verified TFIIH partner class. XPG was absent from the final cryo-EM density; crosslinks are not relabeled as resolved atomic contacts, and the detailed supplementary crosslink table was not independently parsed.

Proposed replacements: TFIIH-class transcription factor complex binding

Summary: The source-specific XPG–p62/GTF2H1 interaction remains unresolved.

Reason: PMID:8652557 has positive interaction context but its full original assays were not recovered. The independently inspected TFIIH work does not establish this exact p62/GTF2H1 pair in its main Results. Complex co-membership or contacts with other TFIIH subunits are insufficient to resolve it. This is an access-limited judgment, not a claim that the experimental annotation is incorrect.

Summary: The source-specific XPG–CSB/ERCC6 interaction remains unresolved.

Reason: PMID:8652557 has positive interaction context but its full original assays were not recovered. The independently inspected TFIIH work does not establish this exact CSB/ERCC6 pair in its main Results. Complex co-membership or contacts with other TFIIH subunits are insufficient to resolve it. This is an access-limited judgment, not a claim that the experimental annotation is incorrect.

Summary: The source-specific XPG–p44/GTF2H2 interaction remains unresolved.

Reason: PMID:8652557 has positive interaction context but its full original assays were not recovered. The independently inspected TFIIH work does not establish this exact p44/GTF2H2 pair in its main Results. Complex co-membership or contacts with other TFIIH subunits are insufficient to resolve it. This is an access-limited judgment, not a claim that the experimental annotation is incorrect.

Summary: Human XPG binds PCNA in the reported biochemical assays.

Reason: PMID:9305916 Methods and Figures 4–6 test the human XPG 981–1009 region and full-length human XPG expressed in insect cells. PCNA-bead binding, competition, GST/BSA controls and DNase treatment support association. The proposed coupling of excision to repair synthesis is a mechanistic interpretation, not a demonstrated new catalytic activity of XPG.

Summary: Refine NTHL1 binding to the demonstrated enzyme-activating role.

Reason: PMID:9927729 Figures 1–3 show reciprocal association and increased NTHL1 substrate binding and activity. Enzyme activator activity expresses this measured functional interaction; it does not make XPG the glycosylase.

Proposed replacements: enzyme activator activity

Summary: Refine nuclear localization to the supported nucleoplasmic compartment.

Reason: The original PMID:16246722 assays remain unavailable beyond abstract material. PMID:26833090 Figure 4C–E and Results identify endogenous human XPG nuclear foci in S phase and after replication stress, using XPG-deficient cells, a second antibody and depletion controls. These data support nucleoplasmic localization without assigning the same behavior to the untested shorter natural products.

Proposed replacements: nucleoplasm

Summary: Refine nuclear localization to the supported nucleoplasmic compartment.

Reason: The original PMID:7651464 assays remain unavailable beyond abstract material. PMID:26833090 Figure 4C–E and Results identify endogenous human XPG nuclear foci in S phase and after replication stress, using XPG-deficient cells, a second antibody and depletion controls. These data support nucleoplasmic localization without assigning the same behavior to the untested shorter natural products.

Proposed replacements: nucleoplasm

Summary: Refine nuclear localization to the supported nucleoplasmic compartment.

Reason: The original source itself was read. PMID:26833090 Figure 4C–E and Results identify endogenous human XPG nuclear foci in S phase and after replication stress, using XPG-deficient cells, a second antibody and depletion controls. These data support nucleoplasmic localization without assigning the same behavior to the untested shorter natural products.

Proposed replacements: nucleoplasm

Summary: Refine nuclear localization to the supported nucleoplasmic compartment.

Reason: The original PMID:8710877 assays remain unavailable beyond abstract material. PMID:26833090 Figure 4C–E and Results identify endogenous human XPG nuclear foci in S phase and after replication stress, using XPG-deficient cells, a second antibody and depletion controls. These data support nucleoplasmic localization without assigning the same behavior to the untested shorter natural products.

Proposed replacements: nucleoplasm

Summary: Refine the supported nuclear-localization inference to nucleoplasm.

Reason: The official P28715 record genuinely contains IPR001044 and the nucleus mapping SL-0191. Controlled human XPG localization in PMID:26833090, Results Figure 4C–E, resolves nuclear repair foci using XPG-deficient cells, a second antibody and depletion controls. Together with the explicit human P28715 nucleoplasm entity in Reactome and the HPA nucleoplasm record, this supports the more informative GO:0005654 already represented in rows 41–51. HPA cautions that its antibodies can recognize proteins from multiple genes, so that source is corroborative rather than the sole basis. Refine the broad location as in the independently reviewed human nuclear-source rows 36–39; do not infer that each shorter natural product was individually localized.

Proposed replacements: nucleoplasm

Summary: The HPA nucleoplasm localization is supported, with its antibody-specificity caution retained.

Reason: The official ERCC5 HPA entry (ENSG00000134899, ICC/IF section) records nucleoplasm with supported reliability: HPA045845 in U-251MG and HPA050374 in A-431, U-251MG and U2OS. The entry explicitly cautions that the antibodies target proteins from multiple genes, so these data alone are not treated as uniquely identifying every ERCC5 product. Independent human XPG nuclear-foci controls in PMID:26833090 (Results Figure 4C–E) and the human P28715 nucleoplasm entity 67447 traced through Reactome NER assemblies corroborate this localization. This is a context-dependent nuclear localization, not an assertion that chromosome-associated XPG is excluded or that each splice product was resolved.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during formation of the pre-incision complex in gg-ner.

Reason: The source event is human and its current forward participant route is 5689317 input 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during dna polymerases delta, epsilon or kappa bind the gg-ner site.

Reason: The source event is human and its current forward participant route is 5690213 input 5691059; 5691059 hasComponent 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during 3'-incision of dna by ercc5 (xpg) in gg-ner.

Reason: The source event is human and its current forward participant route is 5690988 output 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during 5'- incision of dna by ercc1:ercc4 in gg-ner.

Reason: The source event is human and its current forward participant route is 5690990 output 5691059; 5691059 hasComponent 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during binding of ercc1:ercc4 (ercc1:xpf) to pre-incision complex in gg-ner.

Reason: The source event is human and its current forward participant route is 5690991 input 5691043; 5691043 hasComponent 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during ercc5 and rpa bind tc-ner site.

Reason: The source event is human and its current forward participant route is 6782138 input 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during binding of ercc1:ercc4 (ercc1:xpf) to pre-incision complex in tc-ner.

Reason: The source event is human and its current forward participant route is 6782141 input 6782134; 6782134 hasComponent 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during 5' incision of damaged dna strand by ercc1:ercc4 in tc-ner.

Reason: The source event is human and its current forward participant route is 6782204 output 6782205; 6782205 hasComponent 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during dna polymerases delta, epsilon or kappa bind the tc-ner site.

Reason: The source event is human and its current forward participant route is 6782211 input 6782205; 6782205 hasComponent 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The human Reactome participant chain places ERCC5 in nucleoplasm during 3' incision by ercc5 (xpg) in tc-ner.

Reason: The source event is human and its current forward participant route is 6782224 output 67447. Entity 67447 resolves to UniProt P28715, modeled residues 1–1186, with GO:0005654 nucleoplasm. Thus the location belongs to the target protein rather than merely to the pathway title. Controlled human nuclear XPG localization in PMID:26833090, Results Figure 4C–E, is corroborative. Retaining this location does not assign neighboring polymerase chemistry or ERCC1–ERCC4 5-prime incision to XPG, and does not extend the full-length participant model to both shorter natural products.

Summary: The exact XPG association with the RPA complex remains source-limited.

Reason: The operative PMID:7700386 original experiments were not recovered. The source association must not be restated as a claim that XPG is a constitutive RPA subunit. Independent human NER reconstitution establishes XPG repair participation but does not settle the precise RPA-associated localization or interaction in this paper. Preserve this annotation as undecided rather than substituting a different complex on incomplete evidence.

Summary: Refine the broad chromosome location to repair-associated chromatin.

Reason: PMID:26833090 Figure 6D/S6F and Results analyze XPG in the chromatin-associated fraction and its relationship to BRCA2/PALB2/RAD51 loading. The Methods P100 fraction also contains nuclear matrix and insoluble material, so fractionation alone is not a pure-chromatin binding assay; the accompanying nuclear repair-foci evidence supports the localized chromatin context. No constitutive chromosome-wide coating is implied.

Proposed replacements: chromatin

Summary: Refine the supported chromosome association to repair-associated chromatin.

Reason: The current target genuinely carries the UniProt chromosome location SL-0468, traced to PMID:26833090. Its controlled human nuclear repair foci in Figure 4C–E and the repair-factor loading analysis in Figure 6D/S6F support a chromatin-associated repair context, matching the independently reviewed source row 53. Refine the broad chromosome term to GO:0000785 chromatin. The Methods define P100 as chromatin, nuclear matrix and other insoluble material; that fraction alone is not a purified-chromatin binding assay. The localization judgment combines the fractionation and repair-site evidence and does not claim constitutive chromosome-wide coating, an untested chromosome subdomain, or equivalent localization of all three natural products.

Proposed replacements: chromatin

Summary: XPG executes an incision step in transcription-coupled nucleotide-excision repair.

Reason: The current curated human Reactome pathway R-HSA-6781827 contains dual incision R-HSA-6782135 and the XPG 3′-incision event R-HSA-6782224. Its catalyst explicitly identifies full-length human ERCC5/P28715 as the active DNA endonuclease, rather than merely a neighboring pathway member. Independently read human incision and repair-reconstitution experiments in PMID:12644470 and PMID:11259578 establish the relevant catalytic work. Accept this broad repair participation while retaining the access limit on PMID:16246722: its upstream transcription-bubble recognition and CSB-interaction assays were not independently verified.

Summary: ERCC5 executes the 3-prime incision step within human transcription-coupled NER.

Reason: The actual human Reactome pathway R-HSA-6781827 contains the dual-incision branch R-HSA-6782135 and event R-HSA-6782224. That event uses catalyst 6782216 with activeUnit 67447 (P28715, residues 1–1186) for GO:0004520 DNA endonuclease activity. This is participation through a defined repair step rather than a phenotype-only requirement. XPG acts downstream of stalled-transcription recognition; the transcription machinery, XPB/XPD helicase activities, DNA synthesis and ligation remain assigned to their own components. The event does not establish identical function for the two shorter source-described ERCC5 products.

Summary: XPG supports AP-site formation through NTHL1 stimulation.

Reason: PMID:9927729 Figures 1–3 show XPG-enhanced NTHL1 substrate binding and increased product in a combined glycosylase/AP-lyase incision assay. NTHL1 performs base removal; XPG supplies a measured activating cofactor role. The assay does not separately resolve the rate of AP-site formation. Retain this additional BER contribution as non-core, without assigning glycosylase chemistry or the discussion’s transcription-coupled model to XPG.

Summary: The ancestral NER assertion is correctly inherited by human ERCC5.

Reason: The current PAINT tree places human leaf PTN002497156 below the Eukaryota assertion node PTN000792422 in PTHR16171; the target MSA sequence is an exact 1,186-residue match to P28715. The three relevant IBD assertions are present and there is no target-path NOT. The IBD is experimentally grounded in yeast Rad2 and multiple other descendants, including human P28715. The target appearing in WITH/FROM is expected and not circular. Human biochemical DNA cleavage (PMID:8078765) and the explicit P28715 active-unit chain for the 3-prime incision in Reactome R-HSA-5690988 show that ERCC5 performs a step in NER. No target-specific loss or altered clade placement was found; this acceptance does not assign every repair-complex activity or all natural product variants to XPG.

Summary: Human XPG is an active component of nucleotide-excision repair.

Reason: PMID:11259578 dual-incision Methods and Figures 3–6 test XPG-containing human repair assemblies with a defined cisplatin-adduct substrate. Reconstitution and functional immunoprecipitates establish participation, not merely UV sensitivity or disease association.

Summary: Full-length XPG catalyzes a DNA-incision step relevant to NER.

Reason: PMID:32522879 separates full-length human XPG assays from the engineered XPGcat hybrid and shows bubble/junction incision. The same repair participation is directly grounded by human reconstituted dual incision in PMID:11259578 and catalytic-mutant experiments in PMID:10026181. Engineered structural fragments are not equated with the natural shorter products.

Summary: The structural and biochemical work supports XPG’s established NER role.

Reason: PMID:32821917 identifies the human catalytic core and tests DNA incision, while its NER-complex positioning is a model. Functional human repair reconstitution in PMID:11259578 independently establishes XPG participation in the pathway. Accept the process without treating the fragment crystal as an intact preincision-complex experiment.

Summary: Human XPG participation in nucleotide-excision repair is independently established.

Reason: PMID:8090225 was available only as a positive abstract. Actual human repair reconstitution in PMID:11259578 and the wild-type/catalytic-mutant incision experiments in PMID:10026181 establish the same pathway role. This preserves a clearly established function without claiming verification of the inaccessible source-specific assay.

Summary: Both the family inference and matched ARBA rule agree with ERCC5 execution of NER.

Reason: P28715 contains IPR001044. ARBA00086510 condition set 2 requires FunFam 1.10.150.20:FF:000037, 3.40.50.1010:FF:000022 and 3.40.50.1010:FF:000023; all three are present on the target record. These condition matches are corroborated by the explicit human 3-prime-incision active-unit model in Reactome R-HSA-5690988 and the actual human DNA-endonuclease experiment PMID:8078765. The product catalyzes a repair step; it is not merely a substrate or a downstream damage readout. No separate claim about the two short natural products follows from these canonical-sequence inferences.

Summary: Human XPG participation in nucleotide-excision repair is independently established.

Reason: PMID:7657672 was available only as a positive abstract. Actual human repair reconstitution in PMID:11259578 and the wild-type/catalytic-mutant incision experiments in PMID:10026181 establish the same pathway role. This preserves a clearly established function without claiming verification of the inaccessible source-specific assay.

Summary: The source reaction directly places ERCC5 catalysis within nucleotide-excision repair.

Reason: R-HSA-5690988 is the human ERCC5 3-prime-incision reaction in global-genome NER. Its catalyst 5691069 names P28715 entity 67447 as activeUnit for DNA endonuclease activity; the event process is GO:0006289. The target therefore executes an incision rather than simply being present in a repair-containing complex. The original human DNA-cleavage experiments in PMID:8078765 are independently consistent with this biochemical role. The subsequent polymerase and ligase steps are not attributed to ERCC5.

Summary: The matched broad recombination inference can be refined to the demonstrated homologous double-strand-break repair role.

Reason: ARBA00028026 condition set 11 requires IPR006085, IPR029060 and Primates; P28715 matches all three. The mechanistic evidence is independent: PMID:26833090, Results Figures 5B–G and 6A–F, combines human nuclear-extract interactions, recombinant HRR-protein assemblies, a DR-GFP gene-conversion assay and XPG-dependent loading of RAD51/BRCA2/PALB2. The P100 fraction includes nuclear matrix and insoluble material as stated in Methods, and the evidence does not make XPG a strand-exchange enzyme. It supports participation in repair through the HRR assembly/loading step, making the already represented GO:0000724 double-strand break repair via homologous recombination more informative than unrestricted DNA recombination. Depletion phenotypes alone are not used to infer this participation.

Proposed replacements: double-strand break repair via homologous recombination

Summary: XPG activates NTHL1 in a defined biochemical system.

Reason: PMID:9927729 Figures 1–3 directly establish stimulation rather than inference from knockout necessity. Retain the activator activity as an additional, non-core repair function; it is separable from XPG nuclease catalysis.

Summary: XPG contributes to the cellular response to ultraviolet damage.

Reason: PMID:8710877 has positive abstract evidence for UV-dependent nuclear redistribution, but its main figures were not recovered. Independently, PMID:16167068 tests human fibroblast UV responses and complementation with human XPG. Retain the broad response as non-core alongside the more mechanistic NER annotations, without claiming the 1996 redistribution experiment was directly inspected.

Summary: The UV-response inference is supported as a broad consequence of XPG-mediated damage repair.

Reason: ARBA00028103 condition set 5 matches the three target FunFams 1.10.150.20:FF:000037, 3.40.50.1010:FF:000022 and 3.40.50.1010:FF:000023. XPG has a direct incision role in the human NER machinery that repairs UV-induced lesions, as modeled by Reactome R-HSA-5690988 and mechanistically examined in PMID:31253769. The broad GO response-to-UV term is biologically consistent but contributes less functional information than the established NER step, so retain it as non-core. PMID:8078765 Figure 4 did not find preferential nicking of UV-damaged supercoiled DNA by isolated XPG; this decision therefore does not mistake a cellular repair response for direct UV-lesion sensing by the purified nuclease.

Summary: Human XPG alters the measured response to UV exposure.

Reason: PMID:16167068 Results and Figures 1–5 examine primary human fibroblasts, UVC dose/time responses and lentiviral XPG complementation. The observed response is genuine but broader than XPG’s core repair chemistry; retain it as non-core.

Summary: Human XPG contributes to UV-C responses in fibroblasts.

Reason: The recovered PMID:17208056 Methods identify primary human fibroblasts and 254-nm irradiation, but only partial Results were accessible. Independent PMID:16167068 tests human XPG complementation in the same exposure class and resolves the broad UV-C response claim. This does not independently verify all CD95/caspase/p53 conclusions of the 2007 paper.

Summary: The mapped ester-bond hydrolase activity is correct but can be specified as DNA endonuclease activity.

Reason: The cited InterPro signature is present on P28715; the mapping describes a broader activity class than the human experiments establish. PMID:8078765, Methods pp.3312–3313 and Results/Figures 2–3 pp.3314–3315, directly demonstrates internal cleavage of covalently closed ssDNA and the resulting termini by purified human XPG. PMID:32522879 separately demonstrates incision of structured DNA substrates by full-length XPG (Results Figure 1 and Online Methods “Incision assays”); the engineered XPGcat chimera has a different substrate range. Replace ester-bond hydrolase activity with the already present GO:0004520 DNA endonuclease activity. This covers the directly tested DNA cleavage without assigning RNA cleavage, exonucleolysis, or every activity of a nuclease superfamily to the target.

Proposed replacements: DNA endonuclease activity

Summary: The generic complex category adds no specificity beyond established repair-complex membership.

Reason: PMID:26833090 Figure 5 demonstrates XPG association with BRCA2/PALB2/RAD51/DSS1 using nuclear extracts and recombinant coexpression/purification. The imported generic protein-containing-complex term is TOO_GENERAL because the independently accepted nucleotide-excision repair complex is already its descendant. This is a judgment of term specificity, not a rejection of the additional HR assembly. No named constitutive BRCA complex is assigned, and the HR experiment is not relabeled as evidence for the NER assembly.

Summary: XPG self-association is experimentally supported under defined conditions.

Reason: PMID:10026181 Figure 7 tests coexpressed wild-type and engineered N-terminally truncated XPG; separately expressed material mixed afterwards is a negative control. High-salt hydrodynamic measurements favor monomeric protein, so the oligomeric state is condition-dependent. Independent full-length coexpression and SAXS in PMID:32522879 Figure 3 support a dimeric assembly. Retain homodimerization as non-core without asserting a universal dimer in every NER intermediate.

Summary: XPG suppresses UV-induced apoptosis in the tested human-cell context.

Reason: PMID:16167068 compares primary human fibroblasts and complements XPG deficiency with wild-type or E791A XPG, measuring Annexin V/propidium iodide responses. Wild-type rescue supports the regulatory effect; the mutant does not fully restore protection. Retain this contextual regulation as non-core. The study does not establish a dedicated anti-apoptotic enzymatic mechanism, and uncertainty about that mechanism does not negate the measured regulation.

Summary: Specify binding to the TFIIH holo complex.

Reason: PMID:11259578 Results Figures 1–6 recover XPG with endogenous HeLa TFIIH captured through CDK7; the recovered TFIIH retains core and CAK subunits and repair activity. GO:0062058 is more informative than unspecified complex binding. This association does not establish XPG as a constitutive transcriptional TFIIH subunit or convert every co-precipitated factor into a direct binary contact.

Proposed replacements: transcription factor TFIIH holo complex binding