The normal seed contains 256 original annotations and three alternative products. All 79 seeded PMID abstracts and all 31 cached Reactome summaries were read, along with the UniProt function, interaction, domain and localization sections. Abstract access is distinguished below from actual Methods/Results reading. Individual pairs in large interaction datasets have not all been inspected. The configured Falcon attempt and its fallback failed; no provider research report was fabricated. These are manual research notes.
APC is a multidomain tumor suppressor with separable beta-catenin-regulatory and cytoskeletal functions. Its armadillo repeats recruit numerous partners, central repeats bind beta-catenin, SAMP motifs bind AXIN, and the basic/C-terminal regions interact with microtubules, actin and plus-end proteins. The kinases and ubiquitin ligase perform the respective chemistry of beta-catenin destruction; APC contributes binding, scaffolding and regulated substrate handling. Neither its phosphorylation nor the presence of an enzyme in its complex makes APC itself a kinase, protease or ubiquitin ligase.
PMID:19116273(https://pubmed.ncbi.nlm.nih.gov/19116273/) cites activated protein C binding ApoER2/LRP8. The original author-posted paper was read at its actual binding Results and Methods, not just its title. It measures human activated protein C, including APC-V5, against soluble ApoER2 and EPCR with receptor-associated protein and soluble VLDLR controls. The coagulation protease, its Gla region and signaling through protease-activated receptors establish the reagent identity. These are PROC-derived reagents, not the adenomatous polyposis coli tumor suppressor P25054. The seeded GO:0005515 IPI record with LRP8 therefore has a demonstrable wrong-protein source. This is a source-specific citation correction, not a universal claim that tumor-suppressor APC can never interact with LRP8. An independent reviewer checked this same distinction. The original source identifier and partner must remain in the audit record.
The seeded abstracts of PMID:10811618(https://pubmed.ncbi.nlm.nih.gov/10811618/), PMID:15327768(https://pubmed.ncbi.nlm.nih.gov/15327768/), PMID:15327769(https://pubmed.ncbi.nlm.nih.gov/15327769/), PMID:16293619(https://pubmed.ncbi.nlm.nih.gov/16293619/) and PMID:16753179(https://pubmed.ncbi.nlm.nih.gov/16753179/) explicitly report structural or quantitative binding experiments. Phosphorylation greatly strengthens APC-repeat binding to beta-catenin; different repeats have different affinities. AXIN's RGS-domain interaction with APC uses a surface distinct from classical RGS G-protein recognition. These results support specific binding assertions without making APC an RGS enzyme.
PMID:16188939(https://pubmed.ncbi.nlm.nih.gov/16188939/) distinguishes soluble beta-catenin-targeting and microtubule-binding pools; its abstract supports dynamic distribution between those pools. PMID:9601641(https://pubmed.ncbi.nlm.nih.gov/9601641/) reports human AXIN binding APC, beta-catenin and GSK3 and promoting phosphorylation in a complex. The 2011 and 2012 studies PMID:21471006(https://pubmed.ncbi.nlm.nih.gov/21471006/) and PMID:22682247(https://pubmed.ncbi.nlm.nih.gov/22682247/) offer differing mechanistic accounts of destruction-complex remodeling. Their presence should not be converted into a single settled sequence of handoffs. The 2021 all-human reconstitution PMID:34352208(https://pubmed.ncbi.nlm.nih.gov/34352208/) is an additional selected primary source; its full recovered source still awaits inspection before detailed mechanistic use.
PMID:7890674(https://pubmed.ncbi.nlm.nih.gov/7890674/) explicitly reports purified-protein and two-hybrid tests: APC binds beta-catenin and plakoglobin, while alpha-catenin can associate through beta-catenin. APC and E-cadherin did not associate under the tested conditions. Preserve the original NOT cadherin-binding annotation. Adherens-junction localization and catenin binding do not reverse this experimentally scoped negative assertion.
The Reactome summaries were read as pathway-model evidence. The cytosolic destruction-complex reactions agree with the independent localization literature. R-NUL entries explicitly describe human APC phosphorylation by murine kinases; they do not establish a human-kinase reagent or APC catalytic activity. The promoter-exchange summaries explicitly note uncertainty about recruitment of APC to a complex whose beta-catenin and TCF interactions are competitive.
PMID:20566685(https://pubmed.ncbi.nlm.nih.gov/20566685/), author PDF, was read at purified-fragment assembly Results, nucleotide/latrunculin-binding experiments, deletion analysis and protein-purification Methods. APC-B/C fragments nucleate unbranched actin filaments, with two required regions (ANS1 and ANS2). The assays distinguish nucleation from severing, barbed-end capping and accelerated elongation. NIH3T3 microinjection provides a cellular observation but does not identify the species of an expressed fragment. The Methods inherit APC-B/C plasmids from an earlier study and use bacterial GST/His expression with dual affinity purification and size exclusion. The PMID:17293347(https://pubmed.ncbi.nlm.nih.gov/17293347/) abstract explicitly identifies a human APC basic domain that binds/bundles F-actin and can crosslink actin and microtubules. That predecessor's complete plasmid provenance remains to be checked before asserting the species of every later construct. The 2012 collaborative actin-assembly study PMID:22654058(https://pubmed.ncbi.nlm.nih.gov/22654058/) is selected for additional source reading. No actin annotation has yet been added on this provisional synthesis.
PMID:11166179(https://pubmed.ncbi.nlm.nih.gov/11166179/) directly reports APC-dependent microtubule stabilization in its abstract; PMID:10773885(https://pubmed.ncbi.nlm.nih.gov/10773885/) and PMID:19632184(https://pubmed.ncbi.nlm.nih.gov/19632184/) distinguish direct basic-region microtubule binding from EB1-dependent plus-end targeting. PMID:17192415(https://pubmed.ncbi.nlm.nih.gov/17192415/) links APC depletion to unstable peripheral microtubules and impaired migration. PMID:17227893(https://pubmed.ncbi.nlm.nih.gov/17227893/) reports altered kinetochore tension, checkpoint recruitment and mitotic progression in human cells after APC depletion. These are distinct from a claim that APC catalyses motor movement.
Actual construct Methods, interaction Results and Figure 6 captions of PMID:28057765(https://pubmed.ncbi.nlm.nih.gov/28057765/) were read from the complete cache. The nAPC and cAPC constructs derive from human APC; the fluorescent intermediate-chain construct is rat, cell transfections use COS-7, and purified dynein is bovine. Endogenous coimmunoprecipitation also uses mouse brain. The human C-terminal APC fragment associates with dynein in COS-7 cells and in the biochemical experiment, where APC is supplied in bacterial lysate rather than as an independently purified full-length human protein. Methods state that bovine dynein preparations lack detectable LIS1, p150, NDEL1, EB1 and tubulin. Dynein phosphorylation increases APC-fragment association. Insulin-dependent dynein redistribution is not evidence that APC is itself a microtubule motor, nor does localization of dynein at centrosomes automatically place APC there. Figure-caption and Results residue numbering differs for the intermediate-chain mutant; no exact mutant identity is inferred beyond the paper's stated phosphorylation perturbation.
PMID:11972058(https://pubmed.ncbi.nlm.nih.gov/11972058/) reports APC association with CK2 and inhibition by a recombinant C-terminal region. This is distinct from APC's being a substrate of GSK3/CK1. Its actual full Methods/Results remain to be inspected before a specific inhibitor refinement. PMID:17145773(https://pubmed.ncbi.nlm.nih.gov/17145773/) identifies activation of Asef2 through release of autoinhibition and argues for Cdc42 rather than Rac specificity; PMID:17599059(https://pubmed.ncbi.nlm.nih.gov/17599059/) reports both Rac1/Cdc42 responses, with truncated APC in some cellular experiments. The discrepancy and truncation boundaries must remain explicit. APC is a regulator of these GEFs, not the nucleotide-exchange catalyst.
The actual signalosome discussion in the review PMID:24115276(https://pubmed.ncbi.nlm.nih.gov/24115276/) focuses on LRRK2 recruitment with DVL/LRP6 and destruction-complex interactions. It does not, in the inspected passages, resolve specific APC residence in the receptor signalosome. Such a location should remain uncertain pending direct evidence. PMID:26948948(https://pubmed.ncbi.nlm.nih.gov/26948948/) reports cold intolerance and brown-fat defects in tumor-bearing Apc-min mice; this is a pleiotropic physiological phenotype, not direct thermogenic chemistry by APC. PMID:14728717(https://pubmed.ncbi.nlm.nih.gov/14728717/) reports reduced APC expression during DNA-damage-induced senescence; association with the response alone does not establish a repair enzyme.
Large-scale interaction publications retain their original source fields. A paper's generic abstract, a compatible partner name, or another paper's mechanism cannot establish which assay an individual screen pair passed. When source-local evidence is uninspected, uncertainty should be recorded rather than manufacturing an experimental replacement or rejecting the pair.
The recovered normal caches for PMID:20566685, PMID:22654058 and PMID:34352208 contain XML-derived full text. In the 2012 single-molecule study, APC-C remains at the nucleation site while mDia1 tracks growing barbed ends; APC does not accelerate elongation. Its supplementary Methods are not included in the recovered body. The 2010 plasmid Methods explicitly inherit APC-B/C constructs from Moseley 2007. The author institution's 2007 primary record identifies the purified human APC basic domain. The 2017 study, PMID:28663347 further separates actin nucleation from microtubule binding using the m4 mutant. Its indexed primary Methods identify RNAi-resistant full-length human APC from Addgene 16507 and connect the basic-domain constructs to the same 2007 plasmid. Normal-cache requests for those two additional sources have been initiated once. These facts distinguish the APC construct from its bacterial expression host, mouse NIH3T3 assay host, and rabbit actin reagent.
The 2021 destruction-complex study's actual Results and cloning/purification Methods were read. It reconstitutes full-length human AXIN1, APC, CK1alpha, GSK3beta and beta-catenin using insect-cell expression. Removing APC reduces beta-catenin recruitment and degron phosphorylation; truncated cancer variants retain some complex association and substrate recruitment. Purified APC shifts the SCF-beta-TrCP complex in analytical gel filtration, whereas the complex lacking beta-TrCP does not shift. The experiment establishes dependence on beta-TrCP but does not alone distinguish direct APC-beta-TrCP contact from another beta-TrCP-dependent interface. APC increases the fraction of beta-catenin ubiquitinated while limiting chain length/processivity. Neither this scaffolding role nor APC's own phosphorylation makes APC a kinase or ubiquitin-transfer catalyst. The existing specific beta-catenin-binding and destruction-complex annotations capture this mechanism without importing catalytic annotations.
The actual PMID:21311754(https://pubmed.ncbi.nlm.nih.gov/21311754/) body identifies a human brain/testis-library APC fragment, residues 222–653. Active versus GDP-state Cdc42 controls, an in-vitro-translated APC fragment with GST-Cdc42, and cellular interaction assays support small-GTPase binding. This supports refining that paper's own generic binding row; it does not resolve the separately cited PMID:17145773 APC–Cdc42 screen/interaction assertion or confer GEF activity on APC.
Targeted indexed Results of PMID:11972058(https://pmc.ncbi.nlm.nih.gov/articles/PMC122884/) show dose-dependent CK2 inhibition by C-terminal APC fragments. The full-length preparation was pig brain; the recombinant-fragment species remains unresolved in the Methods read so far. The source supports regulation of kinase activity but should not be presented as a purified full-length human APC experiment. Targeted indexed Figure 5/Results of PMID:17145773(https://pmc.ncbi.nlm.nih.gov/articles/PMC1800726/) show APC-armadillo-region stimulation of Asef2-mediated Cdc42 GDP release; APC alone lacks GEF activity. The APC-armadillo construct's source species remains unconfirmed. These observations inform conservative review of the existing rows; neither is a reason to label APC as the catalytic nucleotide-exchange enzyme.
The official actin-nucleation term was checked directly, including its parent actin filament organization and the separate de-novo/Arp2/3 children. The same-role comparator check found actin-nucleation GO annotations in the primary UniProt tables for human SPIRE2, Q8WWL2 and mouse Fmn2, Q9JL04; both displayed annotations are InterPro-derived, not new experimental evidence for APC. Human SPIRE1's J3KNG6 entry also displays this term. A direct QuickGO API request failed and is not represented as a successful all-annotations census. These positive comparators provide no indication that nucleators are systematically excluded from this process term. They do not establish an APC annotation gap by themselves: the relevant APC evidence is its own purified-fragment nucleation and monomer-recruitment experiments. No APC-target GO-CAM activity was found in the local index, and no existing actin-process ancestor or descendant occurs in the seeded 256 assertions. The originally considered Q53SF7 comparator was resolved as COBLL1, not COBL, and was excluded from the comparison.
All 256 original assertions are reviewed: 114 ACCEPT, 43 KEEP_AS_NON_CORE, 80 UNDECIDED, 18 MODIFY and one REMOVE. The two conservative additions are actin binding (GO:0003779) and actin nucleation (GO:0045010); no additional actin-process ancestor or descendant is proposed. Every original source object, qualifier, partner, tested isoform and the experimental NOT cadherin-binding assertion is preserved, as are the three alternative products.
Four core functions distinguish beta-catenin recruitment/processing, microtubule binding and organization, direct actin nucleation, and Asef-family GEF activation. APC does not perform the kinase, ubiquitin-transfer, GTPase-exchange or motor chemistry of its partners. Existing secondary physiological and nuclear contexts remain represented without manufacturing new activities. The actin process passes the participation test because APC directly recruits monomers and assembles the seed; purified single-molecule and human separation-of-function studies support that step. The GO parent, same-role SPIRE2/FMN2 comparator, redundancy and GO-CAM checks described above apply to this single process proposal.
The initial species uncertainties in this journal are superseded where actual Methods resolved them: PMID11972058 APC C-terminal fragments were cloned from a human fetal-fibroblast cDNA library, separately from pig-brain full-length APC. PMID17145773 APC ARM205–885 was amplified from pancreatic tumor-cell cDNA, with human-cell endogenous association and purified fusion activation treated separately. This does not establish every modern Asef isoform or every other interaction in those papers. The official GO:0160124 label is guanyl nucleotide exchange factor activator activity.
The specific PDZ refinements use their own papers: PMID16611247 SCRIB PDZ1/4, PMID17240990 weak PTP1E PDZ2 binding, PMID21858148 and PMID23185543 DLG1 structures/affinities, and PMID8638125 the DLG PDZ/DHR region. Unread PDZ screen pairs remain uncertain. PMID18281465 explicitly supports Trabid binding/deubiquitylating APC; APC is the substrate, not the deubiquitinase. PMID22000517 directly describes the APC-armadillo/Sam68 complex, without assigning APC RNA-processing chemistry. The single removal is supported by actual PMID19116273 assay Methods/SPR identifying activated protein C (PROC), rather than an inference from a differently focused abstract.
The recovered PMID17293347 abstract explicitly identifies purified human APC basic domain binding and bundling actin. PMID28663347 Methods establish RNAi-refractory full-length human APC/Addgene16507 rescue, separately from rabbit skeletal-muscle actin and E.coli-expressed fragments. That paper labels m4 L2539A/I2541A in Results but L2539A/P2541A twice in Methods. This within-paper inconsistency supersedes the earlier single-residue shorthand in these notes; the review uses only m4 and makes no residue-specific claim. Targeted assembly/rescue evidence, not a blanket complete-paper read, supports the two actin additions.
All 84 normal publication records and 31 Reactome summaries are now present and preserved. Literal findings quote their own caches; actual full-text flags are distinguished from targeted external Methods reads. Reactome reaction summaries are not presented as new microscopy assays. Validation, rendering and an independent final all-row consultation follow this authored draft.
The authored review passed schema, term, reference and best-practice validation, and HTML rendering. The 24 advisories consist of 23 inspected, valid generic interactions deliberately retained as non-core under the explicit user ActionEnum, plus one source-specific plasma-membrane uncertainty alongside independently supported membrane records. Genericity alone does not demonstrate a false interaction; an unread target-specific localization assay is not settled by another source. Integrity checks preserve all 256 source objects, three products, the NOT assertion, all 115 normal source cache hashes, 84 exact PMID titles and every literal supporting quotation/availability flag. COMPLETE means the review decisions are filled, including explicitly unresolved evidence; it does not mean those 80 uncertainties were resolved or that publication/merge has occurred. Independent final scientific consultation and root acceptance remain required before publication.
The final independent review identified a narrower meaning of GO:0016342 catenin complex: the cadherin-associated actin-linkage assembly, not any complex containing beta-catenin. The official parent is plasma membrane protein complex, with an extrinsic plasma-membrane relation. The available PMID7806582 abstract distinguishes alternative APC and E-cadherin assemblies; its complete source-local localization experiment has not been inspected. Accordingly, original rows 200–202 (IDA, IBA and ARBA) now remain UNDECIDED, with the source, node and rule retained. Neither the negative direct cadherin-binding result nor an unresolved PAINT topology establishes that indirect adhesion-complex membership is impossible. This supersedes the earlier broad catenin-complex acceptance, and the final counts are 111 ACCEPT, 43 KEEP_AS_NON_CORE, 83 UNDECIDED, 18 MODIFY, one REMOVE and two NEW. The four cores and two actin additions are unchanged.
The same review strengthened three literal finding attachments: the WTX-complex quotation now includes APC; the SxIP quotation states APC's EB1-dependent tip-targeting context; and the dynein quotation identifies reciprocal APC association in mouse-brain extracts. These are claim-matching repairs, not additional functions or changes to the source caches.
The AXIN/APC structural study provides a justified refinement of its own generic-binding assertion. Targeted original binding Results and Methods resolve an APC SAMP repeat interacting with the isolated AXIN RGS-like domain; the 1EMU record identifies human AXIN and maps the APC peptide to human P25054; the peptide organism cell itself is unassigned. The reviewed domain-specific binding term applies without assigning APC an RGS catalytic function. The immutable normal cache remains abstract-only.
The cold-induced thermogenesis row is now UNDECIDED. The complete PMID26948948 abstract reports mutant-mouse cold intolerance and brown-fat abnormalities. Full Methods and control experiments were not recovered; the publisher route failed and the author page offered request-only full text. These observations do not distinguish an APC regulatory contribution from secondary disease effects, and they do not establish tumor burden as the sole cause. The GO:0120162 definition covers positive regulation, so lack of heat-production chemistry is not an exclusion criterion. The earlier non-core action and chemistry-based rationale are superseded, without declaring a failed mouse-to-human transfer.
Unread interaction-pair reasons now state the unresolved source-local evidence directly, without a generic-binding policy slogan. Other positive, inspected interactions remain non-core when no better source-supported molecular function has been established. This follows the explicit user ActionEnum: genericity alone does not establish an incorrect interaction, and inspected positive evidence is not returned to uncertainty solely to satisfy an advisory. Summary acronyms retain the capitalization of the original terms. An actin-filament question separates filament bundling from monomer recruitment while retaining exactly the existing two actin additions.
All 256 original source objects, three products, the NOT assertion, both existing NEW rows, all four cores and every source-cache/published-history byte are preserved. Only row 5 changes to MODIFY and row 251 to UNDECIDED; final counts are 111 ACCEPT, 41 KEEP_AS_NON_CORE, 84 UNDECIDED, 19 MODIFY, one REMOVE and two NEW. Prior counts and validation-advisory totals above describe the earlier review state.