COL4A4 provides the alpha4 chain of the alpha3/alpha4/alpha5 basement-membrane collagen network. Its isolated human NC1 domain also has experimentally demonstrated integrin-binding activity. Basement-membrane structure is its established primary role. The domain activity is retained as a context-specific experimental finding; it does not establish that intact COL4A4 constitutively presents the same ligand surface.
Human P53420 has 1,690 residues, signal sequence 1–38, mature chain 39–1690 and currently annotated NC1 domain 1465–1690. No alternative-product comment was present in the authenticated official record. The mature chain participates in the alpha3/alpha4/alpha5 collagen IV network. The original human sequence paper uses its own NC1 domain-length convention; its complete abstract is available, not full Methods. PMID:7523402(https://pubmed.ncbi.nlm.nih.gov/7523402/)
The complete positive NC1-hexamer abstract reports alpha3/alpha4/alpha5 composition, including alpha4–alpha4 crosslinked dimers and association supporting a shared protomer. This is not an alpha4 homotrimer. Publisher attempts failed; a successful Elsevier HTTP response contained metadata only. Exact full Methods/species/construct scope remain uninspected, and the selected shared abstract is preserved. PMID:12193605(https://pubmed.ncbi.nlm.nih.gov/12193605/)
Actual human NC1 Methods/Results show that isolated alpha4 was monomeric in solution and crystallized in a noncanonical dodecamer; the authors question its physiological relevance. An equimolar alpha3/alpha4/alpha5 mixture yielded alpha5 crystals, while SPR measurements concerned alpha3/alpha5. The structural contacts inform the matrix contribution but do not demonstrate native alpha345 reconstitution or an autonomous alpha4 adaptor mechanism. Row 36 is refined to the existing collagen structural term, without introducing a scaffold function. The official GO:0030020 comment permits extracellular-matrix collagen proteins; the type IV network is not described as fibrillar collagen or as a directly measured isolated-chain tensile assay. PMID:30443360(https://pubmed.ncbi.nlm.nih.gov/30443360/)
The publisher corrigendum corrects an author name and affiliation only. Its operative text does not alter the experimental conclusions.
The 2007 clinical abstract explicitly includes COL4A4 as well as COL4A3 in Cypriot families with thin basement membranes and renal outcomes. It leaves variant predisposition versus modifier explanations open; no full original methods were recovered. The 2009 original publisher PDF was genuinely recovered and its Methods/Results/Discussion read: 45 biopsy-confirmed patients and 143 controls, sequencing, segregation and clinical comparisons. Proposed variant pathogenicity is qualified and functional validation is still requested by the authors. Neither study is described as a direct mechanical measurement or complete developmental time course. The existing structural and glomerular-basement-membrane assertions are retained/refined through the independently established physical network role, with these source boundaries explicit. PMID:17942953(https://pubmed.ncbi.nlm.nih.gov/17942953/), PMID:19675380(https://pubmed.ncbi.nlm.nih.gov/19675380/)
The source-specific collagen fibril-organization annotation remains UNDECIDED. The current GO:0030199 definition concerns processes determining fibril size and arrangement, not only molecules that themselves form fibrils. The available review abstract addresses fibril assembly generally and does not establish a COL4A4-performed contribution. The original full text remains unavailable after the documented publisher/repository failures. Network-forming identity alone is neither evidence of fibril organization nor evidence against an influence on fibrils. Independently established type IV complex membership is retained. PMID:29853175(https://pubmed.ncbi.nlm.nih.gov/29853175/) The older merged COL4A1 REMOVE/MISCITED assessment relied on a collagen-class exclusion that the available source cannot establish. That assessment has been reopened: the current COL4A1 review also retains this assertion as UNDECIDED and its source link as UNVERIFIED. The comparison therefore does not justify changing the COL4A4 action to a confident rejection. COL4A1 is not an alpha3/alpha4/alpha5 protomer chain, so that proposed shared-protomer premise is also inapplicable.
The original 2012 Methods and Results directly demonstrate binding of recombinant human NC1 to alphaVbeta3 by SPR, with EDTA-dependent dissociation and partly antibody-sensitive melanoma adhesion. The study amplified its sequence-checked NC1(1–229) construct from human kidney mRNA, expressed tagged protein in bacteria and tested it against immobilized integrin. The mammalian expression constructs were separate preparations. The insert is not assumed to equal the current 226-residue UniProt NC1 interval. This supports the retained integrin-binding annotation for the isolated domain. It does not establish that the fragment activity is a primary conserved role: naturally released receptor-active human NC1, its physiological concentration and its contribution under normal endogenous expression remain unresolved. The formal functional synthesis therefore retains the established structural role; the positive NC1 and synthetic QS-13 findings remain separate experimental activities below. Public source locators: Methods, NC1 cloning and purification; SPR assay, Results and Figure 7. PMID:22539938(https://pubmed.ncbi.nlm.nih.gov/22539938/)
Synthetic QS-13 exactly matches P53420 residues 1678–1690; the SPR derivative adds a biotin-linked terminal lysine. It binds alpha5beta1, unlike the full-NC1 alphaVbeta3 experiment. HUVEC migration/pseudotubes and mouse Matrigel vascularization decrease, while HUVEC proliferation does not. Direct binding, receptor-sensitive adhesion and the angiogenic endpoints are distinct assays; docking does not solve the receptor complex. These positive pharmacological findings remain visible, but natural release/activity and comparable full-NC1 antiangiogenesis are not established. No NEW GO:0016525 is proposed. PMID:32850867(https://pubmed.ncbi.nlm.nih.gov/32850867/)
The 2007 paper tests other synthetic 19–20-residue collagen-derived peptides in HUVEC response and migration assays, with in-vivo testing left for the future. They are not equated with QS-13 or full NC1. Selected Table 1 sequence cells are absent and were not independently recovered. PMID:17239819(https://pubmed.ncbi.nlm.nih.gov/17239819/)
Integrin binding describes the directly measured domain–receptor interaction. Its more specific child additionally places that binding within cell-matrix adhesion; the retained generic annotation records the directly measured binding. Its core-status assessment is separate and does not negate the positive coated-domain adhesion assay. The positive synthetic-peptide experiments do not establish an endogenous angiogenesis-regulatory role. Earlier comparator records for COL4A2 and COL4A3 have mixed evidence types and concern different fragment preparations, so they do not establish the missing endogenous QS-13 premise.
The original basal-lamina antibody abstract explicitly identifies alpha4 localization. Its species is not stated there; the recovered PMC XML contains a PDF-only notice rather than full Methods. Positive basement-membrane localization is retained without claiming a newly read human tissue experiment. Basement membrane is extracellular matrix, not a lipid membrane. PMID:2211832(https://pubmed.ncbi.nlm.nih.gov/2211832/)
The original 25037231 human colon/liver ECM-enrichment Methods and exact publisher workbook were inspected. COL4A4/P53420 occurs at row 118 of both the complete MS dataset and Matrisome sheets. The source establishes protein detection and collagen classification, not a cancer mechanism or force measurement. The two workbook rows are separate views of the same source data, not independent experiments. PMID:25037231(https://pubmed.ncbi.nlm.nih.gov/25037231/)
There are 55 resolved IBA source joins and 282 literal current same-term record occurrences. Evidence types vary and these are current corroborating records, not a reconstruction of historic PAINT IBD placement. Four original ancestral-node identifiers remain explicitly uninspected; target self-evidence is legitimate descendant grounding. Three electronic-source joins separately authenticate mouse Q9QZR9 and its current structural/complex/basement-membrane records. No donor count or presumed circularity is used to overturn an IBA.
All 17 distinct Reactome sources were read and traced to P53420 through 344 actual current paths. Three ER events resolve the precursor through triple-helix formation, association and secretion and are retained as biosynthetic non-core location. Extracellular degradation places the chain as a substrate; peroxidasin performs sulfilimine chemistry. Family-set or complex membership does not prove each isolated alpha4 receptor interface. The tetrastatin event is supported independently by the original direct-binding source. R-NUL4084497 currently reports Homo sapiens; its literal identifier and separate source assertion are retained. No underlying primary-paper read is claimed solely from reading a Reactome summary.
The source-evidence index records publication identifiers, source-access limits, exact target-table rows, donor records and Reactome paths. External full-text access is distinguished from abstract-only cached material. The 344 verified target paths describe the recovered event context rather than every possible graph path.
Biological questions concern natural fragment release and concentration, receptor specificity, physiological alpha3/alpha4/alpha5 assembly and any direct effect on fibril organization. Processing does not make COL4A4 the enzyme that cleaves it.