ERCC6L2

Summary: Human long-form ERCC6L2 localizes to centromeric regions during interphase.

Reason: PMID:37014751 Results Figure 1A–C and Methods for constructs and immunofluorescence directly show gene-synthesized human 1,561-residue YFP-ERCC6L2 colocalizing with CENP-A/B in U2OS cells. The two C-terminal fragments tested separately also localize there. This is positive evidence for the chromosomal region named by the term. It does not establish the same localization for the natural 712-residue product, and the centromeric signal is not inferred merely from a damage-response phenotype.

Summary: Centromeric localization is directly demonstrated for long human ERCC6L2 constructs.

Reason: The actual SL-0047 mapping cites PMID:37014751. Its Results ERCC6L2 is a centromere-associated protein, Figure 1A–C and Methods Plasmids/Immunofluorescence show YFP-tagged human ERCC6L2 colocalizing with CENP-A/CENP-B in U2OS cells throughout interphase. C-terminal fragments 1053–1247 and 1248–1561 also localize to these foci. These observations support the chromosome-centromeric compartment, while retaining the tested tagged long-product context; they do not establish centromeric localization of the 712-residue splice product.

Summary: HuRI supports ERCC6L2 association with CYREN/MRI using a historical C-terminal reagent.

Reason: PMID:32296183 HuRI PSI records at lines 81748, 127420 and 159970 connect target ORF 8144 (historical H7BXQ9; ENSG00000182150) with CYREN Q9BWK5-1, ORF 20, in screens 1 and 3. These are multiple method records for one pair, not three independent biological replications. The original describes pairwise retesting and sequence confirmation; pair-specific mammalian validation is not inferred from the global benchmark. The archived H7BXQ9 sequence has 522 residues, with one unresolved N-terminal residue and the remaining 521 matching displayed ERCC6L2 residues 1041–1561; exact clone sequence was not independently recovered. PMID:32355287 Figure 4B/C and Methods independently establish DNA-independent association of human ERCC6L2 and MRI/CYREN in HEK293T cells. Retain the supported generic interaction as non-core under campaign policy without assigning a new enzymatic/adaptor activity or equating the historical reagent with a natural product.

Summary: HuRI supports a historical ERCC6L2-reagent interaction with TCP11L1.

Reason: PMID:32296183 HuRI PSI lines 108109, 111672 and 123364 identify target ORF 8144 (H7BXQ9; ENSG00000182150) paired with TCP11L1 Q9NUJ3, ORF 71146, in screen 3. Official UniProt verifies the partner identity. The three records describe screening/validation methods for the same pair. A single screen is not grounds to reject a verified HuRI interaction, and the study’s general orthogonal benchmark is not claimed as validation of this individual pair. The target accession maps to a historical 522-residue sequence largely corresponding to the long protein’s C terminus, rather than to an independently sequence-verified present-day natural product. Retain this supported interaction as non-core; no more specific molecular activity or physiological TCP11L1 mechanism is established by the inspected evidence.

Summary: Human ERCC6L2 and CYREN/MRI associate in cell extracts.

Reason: PMID:32355287 Figure 4B/C, Results describing the interaction and Methods for co-immunoprecipitation demonstrate association of tagged ERCC6L2 and MRI/CYREN in HEK293T cells with Benzonase and ethidium bromide present. Mapping implicates ERCC6L2’s C-terminal region and a conserved MRI motif. Endogenous-tagged mouse B-cell experiments provide separate corroboration. The human full-length construct and mouse 1,537-residue constructs are distinguished in Supplementary Table S2. Association supports the imported binding claim without proving an isolated binary complex or transferring ligase activity to ERCC6L2. Retain as supported non-core binding under campaign policy.

Summary: ERCC6L2 contains an experimentally verified PCNA-binding aPIP motif.

Reason: PMID:37014751 Figure 4D/F/J–N, Results on the aPIP motif, and PCNA-pull-down/ITC/crystallography Methods establish direct interaction. Human ERCC6L2 residues 753–819 as a GST fragment and 790–811 as a biotinylated peptide bind recombinant PCNA with Q798A controls; the 795–816 peptide was used for ITC and crystallography. These engineered reagents differ from the natural 712-residue product. The study explicitly did not establish functional consequences of the aPIP mutation beyond localization. Current bounded ontology searches did not identify an appropriate PCNA-specific binding MF; PCNA-complex membership, activation and polymerase chemistry are not substitutes for the measured association. Retain supported generic binding as non-core under campaign policy.

Summary: ATP binding is supported by the target domain and nucleotide-dependent purified-enzyme behavior.

Reason: The current Q5T890 record and InterPro protein match contain IPR000330. Independently, PMID:41741470 Results Figure 5f–i and Methods ERCC6L2 purification/Helicase assay test purified human ERCC6L2 from Sf9 cells with ATP, AMP-PNP or no ATP. ATP enhances melting of short-overhang DNA, and K154A in the nucleotide-binding pocket loses that enhancement while retaining DNA binding. The actual Supplementary Table3 wild-type protein sequence exactly matches the 1550-residue Q5T890-1 product; the engineered mutant differs only at K154A (equivalent to displayed-product K165). This supports ATP engagement by the tested motor without treating the domain name as proof of broad helicase activity. Mg-dependent residual melting with AMP-PNP and the RPA/mtSSB requirement are separate assay properties. No ATP-binding equivalence among all three natural products is inferred.

Summary: Human long-form ERCC6L2 is nuclear and recruited to nuclear DNA damage.

Reason: PMID:32355287 Figure 3B/G and laser-microirradiation Methods show full-length GFP-tagged ERCC6L2 and its C-terminal fragment in human U2OS nuclei and at induced damage sites. The N-terminal fragment behaves differently, and mouse experiments remain separately identified. Independent PMID:27185855 Figure 4C/9 and cloning Methods distinguish nuclear 1,561-residue Hebo from mostly cytoplasmic 712-residue ERCC6L2-SF. Accept the broad nuclear location for the gene product while retaining these construct/product boundaries; the more specific HPA nucleoplasm context is already represented independently.

Summary: Nuclear localization is supported by original short-construct and independent long-product experiments.

Reason: The SL-0191 mapping cites PMID:24507776 and PMID:32355287. Newly recovered PMID:24507776 Results and Figure 2A show the historical 712-residue GFP construct in both cytoplasm and nucleus; Figure 4 and fractionation Methods also support stress-associated nuclear redistribution of endogenous ERCC6L2. Independent PMID:27185855 Figure 4C and cloning Methods show predominantly nuclear 1,561-residue Hebo-GFP, whereas its short construct is mainly cytoplasmic. PMID:32355287 Figure 3 documents tagged long ERCC6L2 at nuclear damage sites. These product- and condition-specific observations support the gene-level nuclear annotation without making nuclear localization universal for all products.

Summary: The recovered 2014 original supports nuclear localization in its tested human-cell contexts.

Reason: PMID:24507776 Results and Figure 2A show diffuse cytoplasmic and nuclear GFP-tagged historical 712-residue ERCC6L2 in A549 cells. Its genotoxic-stress Results, Figure 4 and Subcellular Fractionation Methods support nuclear redistribution of endogenous ERCC6L2 after MMC or Irofulven; the endogenous antibody signal is not independently resolved to each current natural product. The normal cache remains incomplete, but the relevant original Methods and Results were recovered separately through the indexed PMC article. Independent PMID:27185855 and PMID:32355287 establish nuclear long-product localization and damage recruitment. Retain the general nuclear location with these distinct reagent and condition limits.

Summary: HPA independently supports nucleoplasm for the long ERCC6L2 products.

Reason: Current HPA ERCC6L2 ICC/IF uses HPA021260 in A-431/U-251MG and HPA022422 in Rh30/U2OS, reporting nucleoplasm in all four cell contexts. The complete HPA021260 antigen maps to residues 1244–1332 of Q5T890-4 and 1233–1321 of Q5T890-1; HPA022422 maps 730–809 and 719–798, respectively. Neither complete antigen occurs in Q5T890-2 (712 residues). The two selected official images were inspected and are consistent with the recorded nuclear signal, but they do not distinguish the 1561- versus 1550-residue products. Accept the long-product-supported nucleoplasmic assertion without assigning it to the short splice product.

Summary: The recovered original supports a cytosolic ERCC6L2 compartment.

Reason: PMID:24507776 actual genotoxic-stress trafficking Results explicitly identify the diffuse basal cytosolic and nuclear distribution and movement out of the cytosolic fraction after MMC or Irofulven. Its Subcellular Fractionation Methods describe differential extraction and GAPDH, TATA-binding protein and BiP compartment controls. This provides original-source support for refining cytoplasm to GO:0005829 cytosol. Figure 2A separately shows the historical 712-residue GFP construct in cytoplasm and nucleus; the endogenous Abcam-antibody fractionation does not uniquely assign that signal to every current product. Independent HPA021260 cytosol/nucleoplasm staining in A-431 and U-251MG corroborates long-product contexts, but its antigen is absent from the 712-residue product and its western assessment includes additional bands. The original refinement therefore no longer depends on transferring the HPA long-product signal to the short construct.

Proposed replacements: cytosol

Summary: Stress-associated mitochondrial localization is supported and retained as non-core.

Reason: The official SL-0173 location traces to PMID:24507776. PMID:24507776 actual Results on genotoxic-stress trafficking and Figure 4, with Subcellular Fractionation Methods, show endogenous ERCC6L2 redistribution from the cytosolic fraction toward membrane and nuclear fractions after three hours of MMC or Irofulven in human A549 cells. Mitochondria-enriched extracts and MitoTracker colocalization support the mitochondrial location. The study describes a 712-residue protein and separately tests its GFP-tagged short construct, but the endogenous Abcam-antibody localization experiments do not independently distinguish all current products. This is not evidence that every ERCC6L2 product constitutively resides in mitochondria or performs mitochondrial DNA repair. Retain this conditional compartment as non-core; later long-product nuclear studies neither negate it nor make it constitutive.

Summary: The original supports conditional mitochondrial localization in human A549 cells.

Reason: PMID:24507776 actual Results on genotoxic-stress trafficking and Figure 4, with Subcellular Fractionation Methods, show endogenous ERCC6L2 redistribution from the cytosolic fraction toward membrane and nuclear fractions after three hours of MMC or Irofulven in human A549 cells. Mitochondria-enriched extracts and MitoTracker colocalization support the mitochondrial location. The study describes a 712-residue protein and separately tests its GFP-tagged short construct, but the endogenous Abcam-antibody localization experiments do not independently distinguish all current products. This is not evidence that every ERCC6L2 product constitutively resides in mitochondria or performs mitochondrial DNA repair. The normal publication cache was preserved; actual operative Methods and Results were read separately. The positive stress-specific localization is retained as non-core rather than being treated as disproven by independent long-product nuclear/cytosolic experiments.

Summary: The final-version centrosomal localization claim remains unresolved despite recovered binding experiments.

Reason: UniProt explicitly records centrosomal localization supported by PMID:20873783, and the final published abstract reports partner colocalization. These are positive curated source statements. The recovered HAL author manuscript and separate figure set establish RAD26L binding and phosphorylation by NEK6, but differ from the final published version and contain only database-based localization analysis rather than the operative target-specific cellular colocalization experiment. The inspected final supplement Table S1 identifies historical RAD26L A4D997.3, a 1–531/531 reagent; A4D997 is an ERCC6L2 secondary accession. Neither its reagent length nor omission from the earlier manuscript establishes a wrong target or absence of centrosomal localization. Keep UNDECIDED until the relevant final-version assay can be assessed, with no universal negative localization claim.

Summary: The curated centrosome mapping is positive but its final experimental context remains unresolved.

Reason: The official SL-0048 mapping and UniProt location comment cite PMID:20873783. The final published abstract reports NEK6-partner centrosomal colocalization, whereas the recovered HAL manuscript/figures are an earlier version with biochemical RAD26L binding/phosphorylation and database-derived localization analysis. Those experiments do not independently identify the final target-specific centrosomal assay. The final supplement establishes the historical A4D997.3 reagent without demonstrating its equivalence to a current full-length product. Retain UNDECIDED for this source-specific compartment, acknowledging the positive curated claim and the version boundary rather than inferring that the location is false.

Summary: ERCC6L2 regulates DNA double-strand break end processing in a context-dependent manner.

Reason: PMID:37014751 Figure 5A–G and phleomycin/native-BrdU Methods show increased pRPA and exposed ssDNA in ERCC6L2-deficient human U2OS cells without a corresponding increase in gamma-H2AX; CtIP, BLM or BRCA1 depletion reduces the excess processing. PMID:41741470 Figures 3–4, SMART/AsiSI/ChIP Methods, NBN loss and MRE11/CtIP perturbations independently connect ERCC6L2 loss to increased resection and MRE11 spreading at the tested breaks. This supports the more informative regulatory process GO:1903775, not ERCC6L2 exonuclease chemistry. Direction is conditional: PMID:40913148 Figures 3 and 5–7 and SMART/AsiSI, decay and co-immunoprecipitation Methods show that under prolonged ATM inhibition, ERCC6L2-containing condensates preserve CtIP and support resection, while ERCC6L2 loss decreases it. That study uses declared 1,561-residue full-length and C-terminal constructs; exact clone sequence equivalence was not independently established. The replacement is a lateral correction from broad damage response to regulation of DSB processing, not an is_a-child assertion. No universal negative-regulator, replication-fork protection or coupling of CtIP stabilization to the purified helicase activity is claimed.

Proposed replacements: regulation of DNA double-strand break processing

Summary: A historical RAD26L reagent directly binds NEK6 and is phosphorylated in the recovered author manuscript.

Reason: The final original Supplementary Table S1 identifies RAD26L A4D997.3, a historical 1–531/531 reagent, in the human NEK6 two-hybrid screen. The recovered HAL author manuscript Results pages 6 and 9, Methods pages 27–28 and Figures 2/4B additionally show bead-immobilized GST-RAD26L from bacterial extracts binding purified full-length His-NEK6 and phosphorylation by wild-type NEK6 but not its S206A mutant, with GST and bacterial-lysate controls. This is positive support for protein kinase binding and agrees with the curated UniProt NEK6 phosphorylation statement. Table S2 contains predicted phosphorylation sites; that statement applies only to the table and does not deny the observed phosphorylation experiment. The deposited manuscript/figure set differs from the final published version, so it is not claimed to recover every final-version assay. Retain the specific kinase interaction as non-core, without equating the historical reagent with each current natural product or inferring a stable physiological complex.

Summary: Stable functional complex membership remains unresolved after recovery of direct biochemical interaction evidence.

Reason: GO:0032991 describes a stable assembly whose components function together. The final PMID:20873783 supplement supports a RAD26L–NEK6 two-hybrid interaction, and the recovered earlier HAL manuscript Figures 2/4B support that bead-immobilized GST-RAD26L from bacterial extracts binds purified full-length His-NEK6 and undergoes NEK6-dependent phosphorylation of GST-RAD26L. These observations establish the interaction and substrate relationship but do not independently settle the stable functional assembly underlying this component annotation. The final published version contains additional localization claims not recovered in the author-manuscript version. UniProt phosphorylation is positive evidence, not evidence against the claim, but being a kinase substrate alone does not identify a stable assembly. Retain UNDECIDED without treating an uninspected final assay as absent or reconstructing the original complex from unrelated PCNA/CYREN interactions.

Summary: ERCC6L2 has a supported contextual response to genotoxic-stress-associated ROS.

Reason: PMID:24507776 actual Cell Viability and ROS Evaluation Methods and Results on ROS-dependent genotoxic-stress response distinguish increased DHE signal after knockdown from the NAC intervention. In human A549 cells, NAC reduces Irofulven-associated cytotoxicity and inhibits ERCC6L2 trafficking to mitochondria and nucleus (Results referencing Figure 5C/D and Figure S7). These observations support a ROS-dependent cellular response beyond ROS being only a downstream knockdown readout. The separate gamma-H2AX/NAC observation is reported as data not shown and is not described as an inspected figure. The historical study describes a 712-residue protein, but endogenous siRNA/antibody responses do not uniquely discriminate all current products. Retain the response as non-core without assigning ROS scavenging chemistry, a direct ROS receptor or a universal mitochondrial repair function to ERCC6L2.

Summary: The ICL assertion has a valid PAINT chain but unresolved source-level mechanistic and product scope.

Reason: The actual PTHR45629 tree places Q5T890 leaf PTN002473830 below PTN000085120; the current IBD at that node is GO:0036297 with Q5T890 as its experimental descendant. The target MSA exactly matches the displayed 1561-residue Q5T890-4 product, and no target-path NOT is present. Self-WITH and a single donor are legitimate grounding, not circularity. The experimental source PMID:24507776 now has recovered operative Methods and Results: A549 siRNA depletion increases sensitivity after 48 hours of MMC exposure, with associated damage/stress readouts. These endpoints support a genotoxic-stress contribution but do not resolve the particular ERCC6L2 work in crosslink removal/restoration rather than tolerance or repair of secondary breaks. This remaining uncertainty is mechanistic, not an access-only objection or a blanket demand for isolated biochemical reconstitution. Actual PMID:27185855 shows mild MMC sensitivity/chromosome aberrations with preserved FANCD2 and a long-versus-short complementation difference for phleomycin; PMID:32355287 reports no cisplatin sensitivity in its mouse screening context. These later endpoints neither universally refute ICL participation nor resolve which ERCC6L2 product performs an ICL-repair step. Retain uncertainty about the specific process while preserving the sound ancestry proof.

Summary: The recovered original leaves the particular ERCC6L2 role in interstrand-crosslink repair unresolved.

Reason: The current term concerns removal of an interstrand crosslink and restoration of DNA. PMID:24507776 actual siRNA/viability Methods and Results show reduced A549 survival after 48 hours of MMC exposure, together with damage-response and ROS-associated endpoints. This supports a genotoxic-stress role but does not settle whether ERCC6L2 performs work in crosslink repair itself, tolerance or repair of secondary DNA breaks. PMID:27185855 documents mild MMC sensitivity/chromosomal aberrations with non-classic Fanconi responses and a long-form phleomycin rescue; PMID:32355287 reports no cisplatin sensitivity in its tested mouse screen while establishing end-joining functions. Neither result proves universal absence of an ICL role. Retain mechanistic uncertainty despite actual original access, consistently with the valid PAINT ancestry; an isolated biochemical assay is not treated as a universal prerequisite.

Summary: ERCC6L2 contributes to classical nonhomologous end joining without performing ligation itself.

Reason: PMID:32355287 Results Figures 2–5 combine mouse B-cell Cas9-directed joining, XLF-redundant V(D)J repair, mutant complementation and repair-factor recruitment with human U2OS damage sensitivity/recruitment and HEK293T interaction evidence. In mouse cells, ERCC6L2 promotes MRI/CYREN and XRCC4–ligase-IV loading and proper joining; human U2OS cells also show a smaller XRCC4 recruitment effect. These assays support participation in the end-joining machinery, not merely a generic survival consequence, while species and construct boundaries remain explicit. Human 1,561-residue and mouse 1,537-residue constructs are separately named in Table S2. The normal signal-joint fidelity reported under different conditions in PMID:27185855 does not contradict a context-dependent, partly redundant NHEJ role. Accept this existing process assertion without assigning DNA ligase activity or blanket function to the 712-residue product.

Summary: Classical end joining is supported beyond the matching electronic rule.

Reason: The actual ARBA00093742 rule requires target FunFams 3.40.50.10810:FF:000019 and 3.40.50.300:FF:000821 plus Eukaryota; all match Q5T890. Independent PMID:32355287 Results Figures3–7 and Methods demonstrate NHEJ-factor association, ligase-complex recruitment, XLF-dependent genetic redundancy and altered end-joining orientation, chiefly in mouse B cells/MEFs; human U2OS damage recruitment/sensitivity and HEK293T association are distinct experiments. These support an active end-joining role rather than disease association alone. PMID:41741470 additionally identifies a purified human nucleotide-dependent short-overhang DNA-melting activity compatible with processing staggered ends. Neither source makes ERCC6L2 the DNA ligase, an obligatory factor for every NHEJ substrate, or a functionally equivalent enzyme in all three natural products.

Summary: Purified human Q5T890-1 has an ATP-dependent component of short-overhang duplex-unwinding activity.

Reason: PMID:41741470 Figure 5 and protein-purification/helicase Methods directly measure strand separation by the purified human 1,550-residue sequence independently verified as Q5T890-1. ATP enhances unwinding relative to AMP-PNP/no ATP, and K154A loses that enhancement while retaining DNA binding. These controls support an ATP-hydrolysis-coupled component by inference, not a separately quantified ATP-hydrolysis assay. Residual ATP-independent melting and the requirement for RPA or mtSSB under the tested conditions remain explicit. This proposed MF records experimentally established activity absent from the imported annotation set; it is not added to remove a validation warning. No directional motor, general translocase, exonuclease activity or extension to the 1,561-/712-residue products is inferred. The dsDNA-dependent ATPase specificity of GO:0036121 is not independently resolved by these short-overhang unwinding experiments.