Human ANTXR1/TEM8 is UniProt Q9H6X2 (HGNC:21014). The original seed contains 31 annotations and six alternative products. Original GOA, UniProt, annotation source fields and all existing publication/Reactome caches are preserved. The initial source inventory is tmp/ANTXR1-initial/source-census.json. A configured deep-research attempt failed through its actual providers; no provider-labelled file was fabricated. The separate ordinary four-PMID cache attempt timed out after 45 seconds with no output; it is one interrupted batch, not four completed individual failures.
PMID:16762926(https://doi.org/10.1074/jbc.M603676200) supports extracellular attachment coupled to intracellular actin-dependent spreading. In addition to the unchanged abstract cache, the author-uploaded original primary paper was read at Experimental Procedures, targeted Results/Discussion and Figures 1, 3, 5 and 6. The tail/domain and IL2R chimera experiments separate attachment from spreading. Direct collagen binding was left unresolved in that paper; the HEK293 collagen response depends on beta1-integrin, whereas primary rabbit synovial fibroblasts differ. Figure 6 places receptor in protrusive membrane pools, without establishing focal-adhesion membership. Co-IP supports cytoskeletal association; a later purified experiment must not be attributed to this paper.
PMID:19817382(https://pmc.ncbi.nlm.nih.gov/articles/PMC2847348/) was read at complete Experimental Procedures and Results, Figures 1–5 legends and targeted Discussion. A synthetic human ANTXR1-sv1 tail segment and purified human beta-actin provide direct binding evidence, with Y383A controls. The peptide organizes preformed filaments into bundles; monomer-associated aggregates do not establish actin polymerization. Historical construct names are not assigned a current UniProt isoform ID without sequence reconciliation.
PMID:36400786(https://pmc.ncbi.nlm.nih.gov/articles/PMC9674701/) was read at binding/uptake Results, Figures 3, 4 and 7 legends, and the vector, cell, purification, ELISA and internalization Methods. Human TEM8 constructs derive from AF279145.2. AP fusion/MIDAS controls and receptor-dependent cell uptake support collagen I/VI recognition and lysosomal delivery in the tested systems. Mouse stromal experiments and receptor mutations are distinguished from human constructs. Quenching and washing discriminate internalized fluorescence from surface material. ANTXR1 carries cargo; it does not itself catalyse collagen proteolysis.
PMID:36185380(https://pmc.ncbi.nlm.nih.gov/articles/PMC9515600/) presents contrary evidence. Official indexed Results, Figures 1 and 3 legends, and plasmid, SPR and binding Methods were inspected; direct page access showed a CAPTCHA. The receptor domains and full-length constructs are murine, although HEK293T is the host. PA is a positive control. Neither purified C5/domain assays nor conditioned-medium collagen VI binding to receptor-expressing cells reproduced the proposed interaction; NG2 provided a positive control for assembled collagen VI. Species, constructs and ligand preparations differ from PMID:36400786. This remains a substantive unresolved specificity conflict, not grounds to conceal either result or generalize a negative to every human collagen-I assay.
The broad existing collagen-binding assertion can be assessed separately from a universal collagen-VI/C5 receptor claim. Cargo-receptor GO:0038024 includes a cargo/coat-adaptor bridge in its definition; no direct coat-adaptor interface experiment has yet been inspected, so no NEW cargo-receptor or collagen-catabolism annotation is proposed at this stage.
PMID:16564009(https://www.sciencedirect.com/science/article/pii/S0092867406001991) remains abstract-only in the normal cache. Original publisher-indexed Results and full Figures 4/5 legends were read separately. Reciprocal cellular co-IP supports TEM8 association with full-length LRP6 and its extracellular domain; the authors explicitly allow accessory proteins. This supports an IPI-level receptor-binding refinement without claiming a purified interface or universally essential LRP6-dependent toxin entry. The original Reactome PA-binding entry itself retains uncertainty about LRP6 necessity.
The actual full cached PMID:30241478(https://pmc.ncbi.nlm.nih.gov/articles/PMC6151050/) Methods and relevant binding/signaling Results distinguish TEM8–uPA recognition from uPA catalytic activity. The reported receptor-binding/EGFR response supports a ligand-specific receptor activity and contextual positive EGFR regulation. Human HepG2 experiments are distinct from CHO-host or purified-fragment assays. The enzyme assay does not show TEM8 activation of uPA catalysis. The existing GO-CAM gocams/67c10cc400009138/67c10cc400009138-src.yaml explicitly represents ANTXR1 as the plasma-membrane signaling receptor, with positive EGFR regulation and a downstream EGFR activity. Its ANTXR1 activity and adjacent ligand/causal records were inspected; this is not a missing-process argument.
The full cached PMID:21129411(https://pmc.ncbi.nlm.nih.gov/articles/PMC3014418/) was read at construct/cell/antibody Methods and relevant surface-assay Results. Human TEM8 AF279145 and nonpermeabilized external-epitope/surface-biotin assays support surface localization despite antibody epitope masking. GO:0009897 explicitly includes embedded proteins, so an integral transmembrane receptor is not excluded. The full cached PMID:19581412 was read for the surface-labeling proteomics Methods and source context; its ANTXR1-specific supplementary row remains uninspected. Its HDA surface assertion can be retained with explicit curator deference and independent target evidence, without claiming the supplement was read.
All six original Reactome summaries were read in full. PA binding/processing and receptor-complex assembly start at the plasma membrane; uptake reaches endosomal membranes. Furin supplies proteolysis and PA supplies the toxin-translocation channel. These activities are not assigned to ANTXR1. Pathogen-ligand binding and toxin-context endosomal pools are distinguished from its physiological adhesion/cytoskeletal role.
The complete abstract and opening context of PMID:23602711(https://pmc.ncbi.nlm.nih.gov/articles/PMC3644626/) establish the human GAPO association; the complete experimental body has not been read. PMID:38653789(https://pubmed.ncbi.nlm.nih.gov/38653789/) was read at the official abstract and targeted indexed fibroblast Results, not complete Methods. These sources supply disease and perturbation context, not NEW aging, senescence or nuclear-architecture processes.
The PTN000412053 tree/MSA has not been independently reconstructed. Target self-inclusion and donor count are not evidence against PAINT. Ordinary qualifiers remain preserved and do not imply a distinct biological function. The all-31 prospective action map is awaiting independent consultation and actual Source34 cache closure. No new biological-process assertion is planned.
The independent all-31 prospective consultation passed. All original source assertions and six alternative products are retained in the authored draft: 24 ACCEPT, three MODIFY and four KEEP_AS_NON_CORE, with no NEW annotations. One core cell-adhesion receptor function explains the coupled extracellular attachment and cytoplasmic actin-binding mechanism. The uPA/EGFR and pathogen-ligand contexts remain separately described. Five additional normal Source34 records are pending actual artifact inspection and root cache import; their earlier external read scopes above are not claims of normal cache recovery. The final reference list and exact source-specific attachments will be completed after those records are inspected.
The draft passes schema, term and reference validation. Three advisories are intentional: protective-antigen protein binding remains non-core under the explicit user action definitions (a supported interaction is not removed solely for genericity), and two broad receptor terms are accepted in phylogenetic/family rows while their direct experimental rows receive source-specific cell-adhesion or uPA-receptor refinements. The broader inferences are not assigned an experimentally unestablished ancestral ligand specificity. A direct integrity check confirms all 31 original source objects, six alternative products and 37 currently attached verbatim quote instances.
All five additional normal records were recovered in the exact Source34 artifact, inspected at the scopes below, and imported without overwriting any previous cache. The earlier pending statements above describe the draft stage. The final review preserves all 31 source objects, all six alternative products and all decisions: 24 ACCEPT, three MODIFY, four KEEP_AS_NON_CORE, no NEW annotations and one core function.
The actual PMID:19817382 cache was read at the complete Experimental Procedures, Results and Discussion. Purified human beta-actin and human tail 360–420 with Y383A controls support direct binding and bundling of preformed filaments; monomer aggregation is not polymerization. Exact normal-cache support is now attached to the actin-binding, organization and core entries. Historical sv1 names remain unmapped to current isoforms.
The actual PMID:36400786 binding/uptake Results, Figures 3/4 legends and vector, purification, ELISA and internalization Methods refine the earlier species shorthand: AF279145.2 is explicit for the human TEM8 mutant series and collagen-I assays, whereas extracellular AP-fusion purification covers mouse or human proteins. Therefore the unlabelled collagen-VI panels are not all asserted to be human. The collagen-VI finding is marked disputed, with a direct citation to PMID:36185380; this is a conflict link, not a claim that the earlier negative paper chronologically supersedes the later positive one.
The actual PMID:36185380 abstract, principal SPR/cell-binding Results and Figures 1–3 legends, and plasmid, recombinant-protein, SPR and ligand-binding Methods were read. All receptor constructs are murine. Human C5 is used in Figure 2B, mouse C5 is also tested, and murine C5 is used in SPR/2D work. Neither human host cells nor human ligand makes the receptor human. The conditioned-medium collagen-VI experiments include NG2 positive controls. These negative data remain substantive conflicting evidence, with no generalization to every collagen-I assay. No full supplementary-image review is claimed.
PMID:23602711 remains abstract-only in the normal cache and supplies the human GAPO association. For actual PMID:38653789, the abstract, initial WI26/primary-fibroblast perturbation Results with initial Figure 1 context, and cell-culture/transfection Methods were read; other body sections were not fully inspected. Its senescence phenotypes remain contextual findings, without new senescence or nuclear-architecture annotations.
The source-specific receptor refinements, true non-core PA interaction and HDA supplementary-row boundary are unchanged from the independently reviewed draft. No cache, original source field or raw/provider file was rewritten.
Final validation, HTML rendering and scaffolded-history validation passed. The three intentional advisories are unchanged from the draft. Strict integrity checks confirm all 31 original source objects, six alternative products, all decisions and 52 verbatim quote instances; all five imported cache hashes equal the inspected Source34 members.
The original PMID:16762926 Results adjoining Figure 6 explicitly state: “TEM8 was localized at the tip of actin-rich filopodia and was enriched at the membrane of lamellipodia.” The passage was checked in the linked author-uploaded original article and independently checked by a second reviewer. This specific localization evidence replaces the off-topic general actin-coupling snippet on all four IDA/IEA protrusive-membrane rows. The normal PMID cache remains abstract-only and unchanged. PMID:19817382 images an EGFP fusion of the cytoplasmic tail in some experiments; those images alone do not establish localization of a full transmembrane receptor.
The two experimentally established molecular activities are now represented separately: cell adhesion receptor activity and actin filament binding. The latter uses the existing accepted assertion, the purified human tail/human beta-actin experiment and its Y383A control. Actin cytoskeleton organization belongs with this structural activity, while adhesion-dependent spreading stays with the receptor function. No annotation or inferred polymerase activity is added.
The current GO:0005102 definition covers binding to a signaling receptor and does not restrict it to an agonist ligand occupying a ligand-binding site. Its direct child GO:0090722, receptor-receptor interaction, further shows that receptor-complex associations are in scope. The existing LRP6 IPI therefore retains the GO:0005102 refinement at coassociation strength. The original co-IP/domain experiments do not establish a purified binary interface or an agonist site, and the more specific child is not asserted. The independent prospective consultation agreed with this distinction.
All 31 original source assertions and six alternative products remain unchanged, as do the 24 ACCEPT, three MODIFY and four KEEP_AS_NON_CORE decisions. There are now two core functions and no NEW annotations. Earlier one-core statements above describe the previous published version.
The remaining generic protective-antigen interaction is refined to GO:0015643 toxic substance binding. Its current definition concerns binding a poisonous substance and does not require every bound component to be independently toxic. Original PMID:16762926 Figure 1 connects PA attachment to receptor-dependent LF-plus-PA intoxication; the targeted original Methods/Results and figure captions were independently revisited. Thus the receptor recognizes anthrax toxin through its PA moiety. This does not assign toxicity to isolated PA, binding to isolated LF/EF, or toxin-neutralizing activity to TEM8. The original PMID:16762926 IPI source and PA partner P13423 remain unchanged. PMID:11700562 is corroborating external abstract context only and is not claimed as a cached full text.
This corrects the earlier overly restrictive reasoning about isolated PA. One action changes from KEEP_AS_NON_CORE to MODIFY, giving 24 ACCEPT, four MODIFY and three KEEP_AS_NON_CORE. All 31 source assertions, six alternative products and both physiological core functions remain unchanged; no NEW annotation is added. The toxin interaction remains outside those core functions. A separate question records the unresolved functional status of the TEM8-LRP6 association.