The normal Seed8 import supplies ANTXR2/P58335 with 22 original GOA assertions. Preserve all original objects and alternative-product identifiers. A fresh main/branch/open-PR check at 7a664d7f5a2565dc65a77cc8eacccd0e9becfa90 found no existing ANTXR2 or CMG2 review/history or matching review PR. All five original PMID records and six Reactome entries are already cached; no regeneration is needed.
The first two configured research invocations stopped during uv cache/tool initialization. Selecting the existing installed client through the documented override allowed the Falcon/perplexity-lite attempt to run; it failed after 5.772 seconds and generated no report. These are manual notes, not a provider report. The four additional source requests timed out after 45.017 seconds without output in one ordinary isolated batch attempt; this does not mean four individual completed retrieval failures.
The complete cached abstracts of PMID:11683410, PMID:12700348, PMID:15044490, PMID:15243628 and PMID:15326297 were read. The first paper directly reports an ER-localized CMG2-GFP construct and recombinant matrix binding. Its historical construct differs from the full-length cell-surface receptor characterized later. The UniProt record distinguishes isoform 2 ER localization, isoform 1 cell surface, and a predicted secreted isoform 3; this explains apparently different locations without changing the source annotation objects or claiming every isoform shares every location. The short secreted product may undergo nonsense-mediated decay, so its evidence remains predictive.
PMID:12700348(https://pubmed.ncbi.nlm.nih.gov/12700348/) identifies human CMG2 as a second anthrax protective-antigen receptor, separate from ANTXR1/TEM8. PMID:15044490(https://pubmed.ncbi.nlm.nih.gov/15044490/) quantitates direct VWA-domain binding, while PMID:15243628(https://pubmed.ncbi.nlm.nih.gov/15243628/) reports the receptor–ligand structure. These support specific toxin binding rather than unspecified protein binding; receptor binding does not make ANTXR2 the bacterial toxin or a pore-forming subunit.
For PMID:15326297(https://pubmed.ncbi.nlm.nih.gov/15326297/), the normal full-text extraction was read at the complete Materials and Methods and Results/Discussion. The receptor domain physically contacts the toxin; receptor association delays the toxin's membrane insertion until a lower pH. Figure images and deposited coordinates were not inspected. The six cached Reactome entries were read in full; they place ANTXR2 at the plasma membrane and endosome and identify the toxin, rather than ANTXR2, as the pore-forming protein. Their wording does not justify a new pore/channel or proteolysis function for the host receptor.
PMID:28604699(https://pubmed.ncbi.nlm.nih.gov/28604699/) links CMG2 loss to collagen VI accumulation, with human patient and mouse evidence, and reports ligand binding and cellular delivery to degradation. The official abstract, targeted primary tissue/MMP/rescue Results and indexed binding/uptake Results were read. Detailed recombinant construct-species assessment remains pending. Disease rescue alone is not proof that the receptor performs proteolysis.
PMID:36185380(https://pubmed.ncbi.nlm.nih.gov/36185380/) reports failure to detect collagen VI or C5/endotrophin binding to the receptors. The official abstract and indexed primary Results/Discussion were read; the recombinant proteins are murine, expressed in human cell hosts. This is a substantive ligand-specificity disagreement. The independent ANTXR1 reviewer checked its original Methods and confirmed both VWA and full-length murine constructs, with positive toxin-binding controls. No universal collagen-VI receptor assertion or universal negative is warranted before comparing the exact assays.
PMID:35212873(https://pubmed.ncbi.nlm.nih.gov/35212873/) reports growth-factor-dependent endothelial chemotaxis and matrix-dependent receptor effects. The full abstract and indexed rescue Results were read, not the complete Methods. PMID:40777415(https://pubmed.ncbi.nlm.nih.gov/40777415/) is explicitly a preprint reporting direct actin interaction; only its official abstract and publication type have been assessed. Its original assays and construct provenance require examination before proposing a molecular-function addition.
PMID:32428455(https://pubmed.ncbi.nlm.nih.gov/32428455/) was verified against PubMed and the institutional copy of the published article. Targeted Results/Figures 1–6 captions and Methods on cells, constructs and immunoprecipitation were read. The construct is explicitly human P58335-4; native human fibroblasts and RPE1 cells provide additional evidence. Ligand-dependent release of the talin/vinculin/actin association and recruitment of RhoA support transmembrane signaling. The actin association here is mediated through talin/vinculin, not proof of direct actin binding. Mouse Src-family-null cells and zebrafish embryonic experiments remain distinct. No new process term is inferred from the fish spindle-orientation phenotype. Detailed figure images, simulations and supplementary experiments were not independently inspected.
PMID:42409807(https://pubmed.ncbi.nlm.nih.gov/42409807/) was verified against PubMed and the original publisher article. Targeted receptor-switch Results/Figures 4–7 captions and cell/construct Methods were read, rather than the entire article. Human P58335-4 is explicit. The studied receptor maturation and ligand-triggered cytoskeletal release support a signaling receptor role; acyltransferases and the APT2 thioesterase perform the lipid chemistry. CMG2 is their substrate. Cellular collagen-VI handling in this study does not independently resolve the disputed purified ligand-binding specificity.
The additional reading of PMID:35212873 includes the primary ECM-binding ELISA, endothelial adhesion/chemotaxis Results and selected cell/ELISA Methods. Human endothelial models support matrix-dependent effects. The examined fragment did not independently establish the species of every recombinant reagent. Actual normal Source34 records remain pending before quotation and availability closure.
Official AmiGO definitions were checked for GO:0015643 toxic substance binding, GO:0004888 transmembrane signaling receptor activity, GO:0004895 cell adhesion receptor activity and GO:0038024 cargo receptor activity. Toxin binding is a defensible refinement of the three direct protective-antigen interactions. Mere receptor uptake does not establish the simultaneous cargo/coat-adaptor bridge required by the cargo-receptor term. The broader signaling term accommodates the experimentally demonstrated ligand-dependent intracellular switch without asserting universal collagen-VI specificity. No new biological-process assertion is planned.
A bounded local search found no P58335/ANTXR2 entry in gocams/index.tsv or corresponding PTN000412053 assertion in the available local PAINT slices. This is an access limit, not evidence of a faulty ancestral placement. Q6DFX2 is identified as mouse CMG2 in the primary 2020 Methods; complete historical Ensembl donor reconstruction has not been performed.
The independent all-22 prospective consultation passed before authorship. All 22 source assertions and four alternative products are preserved: 16 ACCEPT, five KEEP_AS_NON_CORE and one UNDECIDED, with no NEW annotation. The core is transmembrane signaling receptor activity at the plasma membrane. Existing endosome rows retain their actual curated trafficking evidence; endosome is not added to the core solely from receptor-dependent collagen degradation.
The final protective-antigen decision supersedes the earlier prospective wording that toxic-substance binding might be a useful refinement. PA is the receptor-binding delivery component of anthrax toxin; the isolated ligand's poisonous-substance status was not established. Its three directly supported interactions remain non-core under the user's ActionEnum. Genericity alone is not evidence that a real interaction is incorrect, and a resolved protein interface does not automatically establish a separately meaningful domain-specific-binding activity. Furin cleaves PA, and PA supplies the pore; neither activity belongs to ANTXR2.
The ER assertions retain the historical construct/alternative-product boundary. The predicted soluble isoform remains uncertain because current UniProt flags possible nonsense-mediated decay and endogenous protein secretion has not been verified. No original isoform, qualifier, partner, evidence code or reference field was changed. Missing PAINT-tree reconstruction is not treated as evidence against directly corroborated surface or receptor functions.
The four actual Source34 records are now normal canonical caches imported by the root after exact artifact verification. PMID:28604699, PMID:35212873 and PMID:36185380 have full-text availability; their actual read boundaries remain those recorded above and in the reference reviews. PMID:40777415 is an abstract-only preprint, used as preliminary context without creating an actin-binding or migration assertion. Literal support from the normal records now anchors receptor signaling and the collagen-VI conflict. The 2017 full-length receptor is explicitly human P58335-4; its recombinant VWA ELISA subsection is not independently species-resolved. The 2022 negative experiments use murine receptors, both human and murine C5 ligands, and full-length receptor/NG2 controls. The conflicting specificity result is retained explicitly.
Published PMID:32428455 and PMID:42409807 were read online at the bounded scopes above. Their normal records remain pending Source35 recovery after the one finite failed ordinary batch; their online evidence is presently documented through these notes, without fabricated cache availability or quotations. The authored review remains DRAFT until that source closure, final validation and independent authored review are complete. No canonical cache or provider file was edited.
The authored draft passes schema, ontology-term and reference validation, rendering and its scaffolded codex history check. The three generic-binding advisories are intentional under the explicit user action definitions and independent consultation. A separate integrity check verifies all 22 original source objects, four products, 13 original raw/cache files, 39 exact quote instances and nine PMID titles/availability flags. The two published Source35 normal records remain pending.
The two published signaling sources are now imported as exact normal caches after independent artifact and identity checks. PMID:32428455 remains abstract-only in the cache. Its previously documented institutional-paper Methods and targeted Results provide the human P58335-4 construct context; cache availability is not inflated to reflect that external reading. The cache abstract directly describes the ligand-dependent talin-to-RhoA switch. PMID:42409807 is the final published 2026 article with XML full text, distinct from the separate actin-binding preprint PMID:40777415. Its actual receptor-signaling and surface-transport Results, Figures 4–7 legends, and cell/construct Methods were checked against the imported record. Full-text availability does not claim that every figure, supplement or experiment was read.
The 2026 Methods explicitly identify human P58335-4. Surface biotinylation and Golgi retention after ZDHHC3 knockdown corroborate the plasma-membrane location. The ligand-triggered release of the talin/vinculin/actin association and recruitment of RhoA/other effectors support the existing signaling-receptor function. The acyltransferases and APT2 carry out lipid chemistry, while ANTXR2 is their substrate; PA forms the toxin pore and LF performs proteolysis. No enzyme function or new biological process is assigned to ANTXR2. Human-cell experiments remain separate from mouse tissues and zebrafish developmental phenotypes. The reported collagen-VI handling remains compatible with an unresolved dispute over direct ligand recognition.
The completed authored review retains all 22 original source assertions, four alternative products and the same decisions: 16 ACCEPT, five KEEP_AS_NON_CORE and one UNDECIDED. There are no NEW assertions and one signaling-receptor core at the plasma membrane. The predicted secreted product stays uncertain, the historical ER product and protective-antigen interactions remain non-core, and existing endosome rows retain their curated context without expanding the core location. Exact snippets from the two normal records now support the existing receptor/surface judgments and core. A new scaffolded codex history records this closure; the earlier draft session remains unchanged.
Final source closure passes full schema, ontology-term and reference validation, rendering and new history validation. The same three generic-binding advisories remain intentional under the explicit user ActionEnum. Independent byte checks preserve all 22 source objects, four alternative products and prior cache/raw files; all 56 literal quote instances and 11 PMID title/availability pairs match the unchanged normal records. Final independent authored consultation and its exact pins are retained with the biological closure manifest.
The missing historical PMID12973667 was independently identified through PubMed (DOI10.1086/378781, PMCID PMC1180616). I read the complete abstract, the indexed original paper’s targeted functional-results paragraphs and Figure6 legend. Human JHF/ISH primary fibroblasts showed impaired adhesion to laminin, with no measurable attachment difference for collagens I and IV in that assay. This is patient-cell adhesion evidence, not a new purified-receptor binding experiment; the paper attributes recombinant VWA binding to laminin/collagen IV to the earlier Bell2001 study (PMID11683410). The distinction will be retained alongside the different ligand preparations, signaling assays and negative collagen-VI results in later papers. PMID:12973667(https://pubmed.ncbi.nlm.nih.gov/12973667/), original targeted Results/Figure6. Direct PMC opening failed, but its original indexed Results and figure caption were accessible; this was not a complete-paper read.
One normal fetch for PMID12973667 terminated with exit1 after33.091seconds and produced no cache files. Its exact receipt and logs are in tmp/ANTXR2-followup. No retry or manual publication-cache creation was performed; normal recovery is requested for Source36.
The prospective review is recorded in tmp/ANTXR2-followup/prospective-followup-proposal-v2.json and the independent consultation by amer1. Three protective-antigen IPI rows are refined to the current GO:0015643 label, toxic substance binding. The original restriction that isolated PA must independently poison a cell was too strict: the ontology does not require each component of a multipart toxin to be independently toxic. Direct PA recognition is recognition of anthrax toxin through its receptor-binding component. The structural and affinity studies remain the actual sources, with PMID:12700348 providing independent receptor-dependent toxin-entry context. No binding to isolated LF/EF, pore-forming activity, or physiological antitoxin core is inferred. The earlier prospective NONCORE choice is superseded on this scientific basis, not because genericity alone establishes an erroneous annotation.
One new GO:0006898 process annotation is proposed. ANTXR2 itself supplies extracellular ligand recognition and inducible cytoplasmic adaptor recruitment. In PMID:28604699, the human isoform4 phosphorylation experiment is distinct from endogenous RPE1 beta-arrestin co-IP. The collagen-VI experiment measures clearance at1,3 and6 hours and its bafilomycin sensitivity; it does not directly image the first uptake event. Patient cells and receptor-depleted cells fail this clearance while the IgG control is preserved. The recruitment step, together with human CMG2-dependent anthrax-toxin internalization in PMID:12700348 and receptor-tail switching in PMID:32428455/42409807, establishes a receptor contribution beyond necessity alone. Direct collagen-VI binding remains disputed by PMID:36185380, and that conflict is retained at the finding level.
The current official GO:0006898 definition and parents place this receptor-recognition work under endocytosis and vesicle-mediated transport. Actual original machine-source annotation objects for same-role human LRP2 and CUBN contain this process; CD320 and LRP10 provide additional cargo-receptor MF context. These are pinned local snapshots, not a fresh live QuickGO census: the attempted official QuickGO route was inaccessible. The local GO-CAM index has no ANTXR2/P58335 hit. No ancestor/descendant BP pair is proposed.
The more specifically defined GO:0038024 cargo receptor activity remains withheld. Beta-arrestin recruitment and cargo clearance support receptor participation, but the inspected experiments do not directly establish the simultaneous cargo-to-coat-adaptor bridge specified by this term. The signaling core therefore carries the supported process, with an explicit question about the unresolved bridge. This is also distinct from assigning ANTXR2 the enzymatic work of Src, APT2, lysosomal hydrolases or anthrax lethal factor.
The predicted soluble isoform3 location remains UNDECIDED. Nonsense-mediated decay is possible, and there is no demonstrated endogenous secretion or positive evidence refuting it. The inherited PAINT/source objects remain untouched. Removing repetitive workflow text from row reasons does not claim newly reconstructed ancestry: the earlier node/donor limitations recorded above still apply. The historical ER assignment remains a valid non-core product-specific observation.
The added historical laminin source is awaiting normal Source36 recovery. Its externally read patient-cell adhesion evidence is documented above; no new publication cache or invented cache quote has been created. The review remains DRAFT until this final reference closure.
PMID:12973667 is now present as an exact normal cache after root import. Its complete abstract supports the general patient-fibroblast matrix-interaction finding and supplies the literal quotation in the reference entry. The cache remains abstract-only. The more specific laminin-adhesion Results and Figure 6 legend were read externally at the scope recorded above; they are not presented as recovered full text or as a new purified-receptor binding measurement. No laminin-binding or adhesion annotation is manufactured from this addition.
The completed revision preserves the independently reviewed three toxic-substance-binding refinements and the one receptor-mediated-endocytosis addition, with PMID:12700348 toxin-internalization evidence attached as requested by the independent peer. All 22 original source assertions, four alternative products, remaining decisions and the single core are unchanged by this reference closure. Earlier source-pending entries above are historical. Canonical raw records, publication caches and published history remain unchanged.
Full schema, ontology and reference validation passes without annotation advisories; rendering and the scaffolded follow-up history check pass. The final integrity check preserves all 22 original source objects and four products, verifies 12 PMID titles/availability flags and 63 literal quotation instances, and confirms that reference closure changed no annotation or core. The independent authored review preceded this final reference-only closure; the final closure receipt binds the completed files.