DDR2 senses extracellular triple-helical collagen through its discoidin domain and transmits that signal through an intracellular tyrosine kinase. Collagen recognition, Src-dependent activation-loop phosphorylation and subsequent DDR2 autophosphorylation are related but distinct steps. A single collagen-receptor core captures this mechanism; ATP binding and collagen recognition are components rather than independent physiological cores. [PMID:9659899; PMID:16186108; PMID:17703188; PMID:20004161]
The official Q16832 record contains one 855-residue precursor, with a predicted signal peptide at 1–21 and mature chain at 22–855. Extracellular 22–399, transmembrane 400–421 and cytoplasmic 422–855 features describe its architecture. These cleavage/topology coordinates are prediction-supported, not newly measured processing sites. No alternative products were seeded. Isolated discoidin domains, Fc ectodomains, V5/His fragments, kinase-dead constructs and disease variants are experimental states rather than demonstrated natural splice products. The mouse secreted ectodomain experiment does not establish an endogenous human shed receptor. [PMID:17703188; PMID:20004161; PMID:19900459]
The original collagen-II experiment and current human-mapped interaction records differ in ligand species/type and reported affinity. The source objects are preserved, while the original bovine ligand and 25 micromolar value govern the scientific interpretation. Exact source locators and the independently recovered focal-adhesion table are in DDR2 primary-source notes. [PMID:17703188; PMID:21423176]
HSP90AB1, CDC37, DDR1 and NTRK3 associations support binding-class refinements. The target is a client or associated receptor, not thereby a chaperone, neurotrophin receptor, or kinase acting on each partner. C-terminal tags, cellular co-complex capture, proximity labeling and pervanadate stimulation limit claims about native interfaces and physiological activation. [PMID:22939624; PMID:25036637; PMID:33961781; PMID:35384245]
Plasma-membrane localization is central to collagen sensing. The HPA textual record supports that location, with a separate uncertain actin-filament signal; no microscopy images were independently reinterpreted. The actual human HFF1 focal-adhesion proteome contains DDR2 in its reproducible set, supporting a contextual localization without a new structural-adaptor claim. Rat nephron apical staining is retained as a tissue-specific ortholog inference. [PMID:20004161; PMID:21423176; PMID:15218324]
Mouse knockout and rescue experiments connect DDR2 to chondrocyte and fibroblast proliferation, endochondral growth and matrix-dependent migration. They distinguish impaired Matrigel migration from unchanged simple motility. Those outputs, as well as osteogenic differentiation and matrix remodeling, remain contextual consequences of receptor signaling. Human MG63 marker knockdown is distinguished from mouse mineralization/p38 experiments. Activating human disease variants implicate DDR2 in connective-tissue remodeling but did not increase every downstream signaling marker in the examined fibroblasts. [PMID:11375938; PMID:11723120; PMID:20564243; PMID:30449416]
Rat donor annotations describe cardiac fibroblast, vascular smooth-muscle, nephron and hepatic-stellate contexts. Expression responses to stretch, TGF-beta, hypoxia or angiotensin do not make DDR2 the primary sensor for each stimulus. Downstream survival and cell-cycle responses are not constitutive molecular activities. The inaccessible rat vascular proliferation experiment remains unresolved rather than being replaced by evidence from fibroblasts. [PMID:16087782; PMID:18664364; PMID:26674152; PMID:32412792; PMID:23409069]
The phylogenetic assessment uses actual PTHR24416 topology, exact DDR2 alignment, positive IBDs and literal experimental donor joins. Target self-inclusion is legitimate descendant evidence. The flattened tree-display NOT strings are not substituted for the actual GAF relationship/loss records. The inspected lineage carries the positive assertions and no on-path IKR/IRD/NOT loss. PI3K/Akt remains a context-dependent inferred output, compatible with the absence of increased AKT in a particular variant-fibroblast experiment. PMID:30449416
No new GO assertion is proposed. The collagen-receptor mechanism is already represented in the seeded annotation set; the two unresolved rows concern source-specific adhesion and vascular smooth-muscle proliferation. The current production GO-CAM index has no literal DDR2/Q16832 match, which is not itself a reason to manufacture an annotation.
Research provenance: the genuine default Falcon attempt failed authentication; the configured fallback also failed its quota/authentication check. No provider text was fabricated. This biological synthesis comes from the cited caches, actual selected originals, official records and independently reviewed source reconstructions. Publication caches remain unchanged. The binding decisions follow the campaign's supported-class/generic-positive policy; none relies on an activation requirement absent from the binding term definition.
The earlier description of all four partner classes as refinements is narrowed for CDC37. The DDR2–CDC37 associations remain positive: LUMIER in PMID:25036637, and both AP-MS and proximity labeling in PMID:35384245. CDC37 binds kinase clients and recruits HSP90, but the inspected study distinguishes co-chaperone recognition of flexible domains from HSP70 recognition of unfolded polypeptides. This does not establish the complete protein-folding-chaperone class required by GO:0051087 for CDC37 itself. The two source-specific CDC37 annotations therefore remain generic binding, retained as non-core under the project supported-interaction policy; their interactions are not rejected. The HSP90AB1 and receptor-kinase binding refinements are unchanged. [PMID:25036637, Results kinase-client specificity and Discussion; PMID:35384245, RTK interactome]
The collagen-binding and phosphorylation assertions from PMID:20004161 now each point directly to that publication without repeating the existing quotation. The biological conclusions, original source fields and single integrated collagen-receptor core are unchanged.
The rat cyclic-stretch study reports DDR2 mRNA/protein induction and TGF-beta-dependent regulation in vascular smooth-muscle cells (PMID:16087782). The corresponding rat donor annotations use IEP. Both response-term definitions allow changes in gene expression, and GO manual-annotation guidance, section “Annotating from gene or protein expression studies,” permits response-to-stimulus IEP in appropriate contexts while cautioning that expression changes do not invariably establish functional contribution. The former non-core decisions for GO:0035994 and GO:0071560 are therefore superseded by UNDECIDED: the detailed original experiments and the biological basis for transferring these context-dependent responses to human DDR2 remain unverified. The full original remains unavailable. This preserves the positive rat observations and does not assert that the unseen paper lacks participation assays, that IEP is intrinsically invalid, or that absence of a direct human assay proves loss of the response.
NTRK3 remains a supported receptor-tyrosine-kinase binding partner. PMID:35384245 reports an AP-MS screen in pervanadate-treated HEK293-derived cells and separate tagged RTK–RTK co-immunoprecipitation; the literal DDR2–NTRK3 record points to Figure EV1E, lane 18. These cellular association assays and the independently established NTRK3 receptor class support GO:0030971 without proving a purified interface, native colocalization, or receptor cross-activation. The two CDC37 associations remain positive and non-core. CDC37 is an established kinase-client co-chaperone; whether it independently meets the stricter nascent/unfolded-polypeptide folding-chaperone definition remains a curator question, rather than a claim that CDC37 never assists folding (PMID:25036637; PMID:35384245).
For the monomeric human DDR2 discoidin domain binding bovine collagen II, the original SPR Results give 2.5 × 10^-5 M, equivalent to 25 micromolar (PMID:17703188; Results, “Collagen-binding properties of the DDR2-DS domain,” Figure 1). The MINT record instead carries 2.5 micromolar. This numerical discrepancy is now explicit in the reference assessment and an expert question; the original measurement takes precedence. The separate human COL11A2 accession in the imported interaction record remains untouched and is not substituted for the actual bovine collagen-II ligand.
No additional SRC- or SHC1-binding NEW row is introduced solely to enumerate downstream partners already represented in collagen-receptor signaling. Broad location terms remain handled by their existing specificity judgments; this finite follow-up makes no automatic cross-aspect action changes. The contextual positive regulation of kinase activity remains supported by the measured p38 phosphorylation/perturbation results without claiming direct DDR2 phosphorylation of p38 (PMID:20004161; PMID:20564243).