GAN

Summary: MAP1B association is real, but the bare binding term does not describe GAN function.

Reason: The original study reports human-brain yeast two-hybrid recovery of the MAP1B light chain, association in transfected COS-7 cells and colocalization in mouse sensory neurons. Its Kelch region recognizes the light-chain C terminus, and the reported microtubule association is indirect through MAP1B. Removing the uninformative GO:0005515 assertion preserves this positive interaction; the adaptor activity and cytoskeletal role are represented separately. The later degradation study, not the 2002 experiment alone, establishes the ubiquitin-scaffolding interpretation.

Summary: Original tandem-affinity proteomics identifies GAN in a 14-3-3 sigma-associated protein pool.

Reason: The recovered original Table I explicitly lists GAN/gigaxonin in the 14-3-3 sigma tandem-affinity/MudPIT preparation from DLD1 cells. The selected Methods, opening Results, table caption and footnote establish positive co-association, while allowing indirect association; sequence-coverage percentages are not binding affinities or interaction magnitudes. The duplicate EGF-condition columns are not independent biological conditions, and footnote c refers to an earlier yeast-isoform affinity study. This source-specific evidence supersedes the earlier inaccessible-target-table rationale. Remove the uninformative generic protein-binding annotation without denying the positive GAN–SFN association or asserting a purified direct interface. The mechanistic ligase-substrate adaptor function is reviewed separately.

Summary: Human UBA1 association supports scaffolding, not an informative bare binding annotation.

Reason: The authentic source abstract explicitly assigns ubiquitin-activating E1 binding to the GAN BTB region and MAP1B light-chain recognition to the Kelch region. Exact human UBA1/Q9H2C0 IntAct records preserve the reported two-hybrid and pull-down assays. Retain the association and degradation mechanism while removing generic protein binding; GAN is not thereby an E1 enzyme, and its substrate-adaptor activity is already represented. Full original construct Methods were not inspected.

Summary: The mouse Uba1 interaction preserves an experimental partner context without adding a specific function.

Reason: The retained IntAct records explicitly pair human GAN with mouse Uba1 Q5U300 in coimmunoprecipitation assays from PMID:16227972; they are distinct from the human UBA1 context. The source abstract supports the BTB/E1 association and MAP1B degradation mechanism. Remove only the uninformative protein-binding assertion, without denying the heterologous interaction or inferring identical natural-product constructs.

Summary: TBCB association is mechanistically relevant but the bare binding term is redundant in functional content.

Reason: The authentic abstract explicitly reports GAN-dependent TBCB degradation through the ubiquitin-proteasome pathway and its disruption by GAN-associated variants. Exact target-pair IntAct records retain two-hybrid and coimmunoprecipitation evidence. This supports substrate-recognition biology, which is captured by the existing ligase-substrate adaptor and protein-catabolism annotations; remove generic protein binding without treating GAN as a tubulin-folding cofactor itself.

Summary: The NUDCD3 LUMIER association does not define a distinct GAN molecular function.

Reason: The source describes a human chaperone/cochaperone interaction survey, and the exact retained IntAct target pair assigns GAN–NUDCD3 to LUMIER. The available native extraction contains the assay framework but not a complete target workbook. Independent later GAN–NUDCD3 AP-MS rows and targeted 2025 results corroborate the association. Remove bare protein binding as uninformative; NUDCD3 being a chaperone does not confer chaperone activity on GAN.

Summary: BioPlex supports GAN–NUDCD3 association but not a new informative binding function.

Reason: The authentic retained BioPlex2 table contains the exact Q9H2C0–Q8IVD9 pair at line 42499, with reported interaction probability 0.980008863787163. This affinity-purification result is positive association evidence; it does not on its own establish a direct interface or chaperone function for GAN. Remove the bare binding term while preserving the original pair and assay provenance.

Summary: The PRKN-5 screen interaction is retained as evidence, not as an informative functional annotation.

Reason: The exact IntAct records from the neurodegeneration interactome preserve the GAN–PARKIN pair with the literal upstream partner O60260-5, with pooled, array and validated two-hybrid assay records. The study’s authentic abstract describes a systematic human interaction map, not proof of a particular shared enzymatic function. Remove generic protein binding without denying the observed interaction or assigning a GAN isoform from the partner suffix.

Summary: The VHL-2 screen association does not establish a specific GAN molecular activity.

Reason: The exact retained PMID:32814053 records pair GAN with P40337-2 (VHL) in pooled, array and validated two-hybrid experiments. Preserve that partner-product context and the positive assay provenance while removing uninformative protein binding. The interaction does not by itself assign VHL substrate recognition, oxygen sensing or a VHL-complex role to GAN.

Summary: BioPlex3 independently records GAN–NUDCD3 association.

Reason: The original retained BioPlex3 table includes Q9H2C0–Q8IVD9 at line 71889, with reported pInt 0.988639668235094. The target-specific association is real screen evidence, with affinity-purification rather than purified direct-binding scope. Remove generic GO:0005515 while retaining this independent source and its original context.

Summary: The GAN–SFN association is positive but generic protein binding adds no functional specificity.

Reason: Exact current IntAct records from the 14-3-3 study preserve GAN with SFN/P31947 in coimmunoprecipitation and cross-linking assays. The source’s authentic abstract concerns 14-3-3 effects on client localization; it does not make each client a chaperone. Remove the bare binding term while retaining the assay-derived association. Full target-panel images and Methods were not inspected.

Summary: The YWHAE interaction is retained without assigning a 14-3-3-like activity to GAN.

Reason: The retained PMID:36931259 interaction records explicitly pair GAN and YWHAE/P62258 in cross-linking assays. This establishes the source-specific association at the curated assay-record level, not a demonstrated chaperone function for GAN. Remove uninformative protein binding while preserving the exact distinct partner; complete original target-panel interpretation remains outside the inspected scope.

Summary: The U2OS AP-MS map records GAN–NUDCD3 association without defining a new molecular function.

Reason: The authentic retained 2025 U2OS network has GAN node 547898, NUDCD3 node 555054 and their explicit edge 624764; GAN is marked as bait. This is original experimental-network evidence, not an inference from the GO annotation itself. It does not establish a direct interface, natural isoform specificity or GAN chaperone activity. Remove generic protein binding while retaining this positive source context.

Summary: The inherited cytoplasmic location can be specified as cytosol.

Reason: The actual cytoplasm IBD lies on GAN’s verified ancestral path, and the target appearing among descendant evidence is legitimate. Diffuse cytoplasmic GAN and cytoskeletal associations, together with the exact Reactome GAN entity in cytosolic CRL3 complexes, support a cytosolic setting. GO:0005829 is directly part_of GO:0005737 in the inspected ontology. Refine the location without asserting a failed phylogenetic transfer or excluding transient association with other structures.

Proposed replacements: cytosol

Summary: GAN’s reported cytoplasmic localization is consistent with a more specific cytosolic placement.

Reason: The original full text explicitly reports cytoplasmic GAN-1 localization, with the GAN observation described as data not shown; the ENC1 title does not make this a wrong-gene annotation. That IDA supports the broader cytoplasm claim. The additional cytosol precision comes from the target-specific Reactome CRL3 complex model, corroborated by diffuse cellular GAN staining, rather than from resolving cytosol in the unpublished IDA image. Refine along the verified GO:0005829 part_of GO:0005737 edge while retaining the separate microtubule-associated pool and the original experimental source and its presentation limit.

Proposed replacements: cytosol

Summary: The broad UniProt cytoplasm mapping is refined to cytosol.

Reason: The SL-0086 mapping states a true cytoplasmic location. Direct localization and the exact human GAN-bearing CRL3 model support cytosol, a verified part_of descendant of cytoplasm. This is a supported compartment refinement, not an inference solely from absence of a signal peptide or membrane helix.

Proposed replacements: cytosol

Summary: The Reactome event contains GAN as a cytosolic CRL3 adaptor component.

Reason: The actual event-to-component chain 8955227 → 976101 → 976145 → 976014 ends in the human GAN entity referencing UniProt Q9H2C0, with cytosol GO:0005829 explicitly recorded. This supports the cellular-component assertion as a curated model, consistent with direct cytoplasmic GAN evidence. The event describes cullin-complex regulation; GAN membership does not make GAN the NEDD8-conjugating enzyme, the deneddylase or the CAND1-exchange catalyst.

Summary: The Reactome event contains GAN as a cytosolic CRL3 adaptor component.

Reason: The actual event-to-component chain 8952578 → 8955227 → 976101 → 976145 → 976014 ends in the human GAN entity referencing UniProt Q9H2C0, with cytosol GO:0005829 explicitly recorded. This supports the cellular-component assertion as a curated model, consistent with direct cytoplasmic GAN evidence. The event describes cullin-complex regulation; GAN membership does not make GAN the NEDD8-conjugating enzyme, the deneddylase or the CAND1-exchange catalyst.

Summary: The Reactome event contains GAN as a cytosolic CRL3 adaptor component.

Reason: The actual event-to-component chain 8955222 → 976101 → 976145 → 976014 ends in the human GAN entity referencing UniProt Q9H2C0, with cytosol GO:0005829 explicitly recorded. This supports the cellular-component assertion as a curated model, consistent with direct cytoplasmic GAN evidence. The event describes cullin-complex regulation; GAN membership does not make GAN the NEDD8-conjugating enzyme, the deneddylase or the CAND1-exchange catalyst.

Summary: The Reactome event contains GAN as a cytosolic CRL3 adaptor component.

Reason: The actual event-to-component chain 8955224 → 8955222 → 976101 → 976145 → 976014 ends in the human GAN entity referencing UniProt Q9H2C0, with cytosol GO:0005829 explicitly recorded. This supports the cellular-component assertion as a curated model, consistent with direct cytoplasmic GAN evidence. The event describes cullin-complex regulation; GAN membership does not make GAN the NEDD8-conjugating enzyme, the deneddylase or the CAND1-exchange catalyst.

Summary: The Reactome event contains GAN as a cytosolic CRL3 adaptor component.

Reason: The actual event-to-component chain 8956021 → 8952606 → 8952604 → 976145 → 976014 ends in the human GAN entity referencing UniProt Q9H2C0, with cytosol GO:0005829 explicitly recorded. This supports the cellular-component assertion as a curated model, consistent with direct cytoplasmic GAN evidence. The event describes cullin-complex regulation; GAN membership does not make GAN the NEDD8-conjugating enzyme, the deneddylase or the CAND1-exchange catalyst.

Summary: The cytoskeletal location can be specified as microtubule cytoskeleton.

Reason: The UniProt cytoskeleton mapping is a true but broad location. PMID:12147674 directly reports MAP1B-dependent localization of tagged GAN along microtubules in cotransfected COS-7 cells and endogenous gigaxonin/MAP1B colocalization in mouse DRG neurons and sciatic nerves. Together with the retained HPA microtubule-cytoskeleton annotation, this supports GO:0015630, a verified is_a descendant of GO:0005856. The term includes associated proteins: GAN need not bind tubulin directly. Intermediate-filament binding and turnover are not used as evidence that intact GAN localizes along those filaments; the source also tests an isolated Kelch-domain construct, which is kept distinct from full-length GAN.

Proposed replacements: microtubule cytoskeleton

Summary: GAN supports intermediate-filament and microtubule cytoskeleton organization.

Reason: The broad GAN-specific InterPro mapping is refined to two supported is_a descendants. Intermediate-filament organization is supported by substrate recruitment and proteasome-sensitive filament clearance, altered soluble vimentin pools and loss/rescue effects on filament arrangement (PMID:23585478, PMID:31944090, PMID:37043392); purified CRL3–GAN ubiquitination of selected intermediate-filament proteins establishes an active adaptor contribution beyond disease necessity alone (PMID:37903270). Microtubule cytoskeleton organization has separate experimental support: coexpressed gigaxonin and MAP1B light chain protect the microtubule network against colchicine-induced disassembly, whereas the tested truncated GAN construct does not (PMID:12147674). This is a context-specific stabilization result, not direct tubulin binding. Both GO:0045104 and GO:0000226 descend directly from GO:0007010. Kelch-mutant clearance and morphology effects can separate, and none of these findings establishes GAN as a motor or a universal direct filament-disassembly enzyme.

Proposed replacements: intermediate filament cytoskeleton organization microtubule cytoskeleton organization

Summary: GAN can localize with the microtubule cytoskeleton.

Reason: The HPA source reports a microtubule-cytoskeleton location. Independent original work detects GAN along MAP1B-containing cytoskeletal structures and concludes that association with microtubules is indirect through MAP1B. The GO definition includes associated proteins, so direct tubulin binding is not required. Preserve the IDA context without claiming that the HPA antibody images were independently inspected or that GAN is an obligate structural microtubule subunit.

Summary: The original GAN1 ubiquitination context is retained alongside independently established adaptor participation.

Reason: The authentic PMID:15983046 abstract explicitly includes GAN1 among BTB-Kelch proteins ubiquitinated by a CUL3-dependent complex; the Keap1 title does not imply a wrong-gene citation. Being modified alone would not establish that GAN performs the ubiquitination step. Independent human CUL3/RBX1/GAN reconstitution demonstrates its substrate-adaptor contribution, supporting the existing whole biological claim while preserving the original IDA source and its self-ubiquitination scope.

Summary: GAN participates in protein ubiquitination as the substrate-recognition adaptor of CRL3.

Reason: The UniPathway mapping is biologically supported by recruitment of substrates to a CUL3–RBX1 ubiquitin ligase and by purified human reconstitution. GAN supplies the substrate-recognition step, so participation is more than being necessary for downstream protein loss. The RING/E2 machinery performs ubiquitin transfer; no autonomous GAN ligase activity is asserted.

Summary: GAN is the substrate-recognition component of a CUL3–RING ligase complex.

Reason: The exact incoming AN25 IBD and its donor set map to GAN’s verified ancestor; the human target’s own source evidence is valid descendant grounding. Independent human reconstitution and cell-biochemical studies support GAN together with CUL3 and RBX1. The current GO definition explicitly distinguishes the cullin/RING catalytic core from the BTB-containing specificity component, matching GAN’s role.

Summary: The CUL3-dependent GAN complex assignment is supported.

Reason: The original abstract explicitly includes GAN1 in CUL3-dependent ubiquitination experiments, although the paper foregrounds Keap1. Independent human CUL3/RBX1/GAN reconstitution provides direct support for the complex membership. Retain the original IDA provenance; do not transfer every Keap1-specific degradation or stress-response result to GAN.

Summary: The curated CUL3-complex role is independently established for GAN.

Reason: The inspected review places gigaxonin among cytoskeleton-associated BTB proteins in the CRL3 literature. Its general paragraph is secondary context; the actual human CUL3/RBX1/GAN experiments independently demonstrate the complex. Retain this NAS annotation as a correct core complex-membership claim without treating a family-wide differentiation thesis as a GAN-specific developmental function.

Summary: The inherited proteasomal catabolism role is independently supported in human GAN.

Reason: The actual AN25 IBD and literal donor set are ancestral to the exact GAN sequence, with no retained target-path loss. GAN recruits selected substrates to ubiquitination machinery; substrate accumulation after loss and proteasome-sensitive turnover support its participation in proteasome-mediated ubiquitin-dependent protein catabolism. This is a recognition step within the pathway, not intrinsic protease activity.

Summary: GAN-mediated substrate turnover is a core proteasomal catabolic role.

Reason: The inspected review discusses gigaxonin among cytoskeleton-related BTB proteins and cites the intermediate-filament literature. Its broad review paragraph is not a new primary assay. Independent MAP1B/TBCB studies and the human CRL3-GAN reconstitution and degradation experiments support the existing process directly. Preserve the original NAS provenance while grounding the biological decision in those primary results.

Summary: GAN is a CUL3 ligase-substrate adaptor.

Reason: The actual AN25/PTN001561260 IBD lies on the verified human GAN target path in the incoming PTHR24412 book. Independent human experiments establish the same function: the Kelch region recognizes substrates and GAN supports their ubiquitination with CUL3–RBX1. The 2023 purified reconstitution uses human GST-GAN, neddylated CUL3/RBX1 and E1/E2 components. This supports adaptor activity, not autonomous catalytic ubiquitin-ligase activity. The 2009 isolated BTB/3-box structure informs the cullin-recruiting architecture without defining a natural truncated product.