The collagen-recognition experiments use defined receptor fragments and ligands. In PMID:17703188, Methods, “Sample preparation”, the human DDR2 discoidin domain comprises residues 26–186 and is expressed in Pichia. The collagen preparation is described as “collagen type II from bovine nasal septum”. Results, the collagen-binding subsection, and Figure 1 report a monomeric-domain dissociation constant of 2.5 × 10^-5 M, equivalent to 25 micromolar. The source is the original article (DOI:10.1038/sj.emboj.7601833). The two MINT records IM-11002-1 and IM-11002-2 map the ligand to human P13942 and one curated comment reports 2.5 micromolar. Neither the human partner mapping nor that decimal discrepancy overrides the original bovine collagen-II assay and value. NMR transferred cross-saturation identifies a domain contact surface; it does not assay an intact native human receptor–COL11A2 complex.
PMID:20004161, Methods “Protein Expression Vectors”, Results and Figure 6, distinguishes human DS 26–190, soluble ectodomain 22–398 and full-length wild-type/W52A constructs. The crystal contains a synthetic collagen-like triple helix with norleucine replacing methionine. Figure 6A tests collagen-I binding; Figure 6B tests collagen-triggered phosphotyrosine in HEK293 cells. The latter readout is not a site-by-site assignment of kinase performer. The separate PMID:16186108 abstract distinguishes Src activation-loop phosphorylation from DDR2 cis-autophosphorylation. The 2009 structural original was recovered; the full 1997 PMID:9659899 and 2005 PMID:16186108 originals remained inaccessible. Their available abstracts and curator assertions are retained honestly.
PMID:21423176 is source-specific evidence for focal-adhesion-associated DDR2. The original NHLBI author-hosted Supplementary Table 1 contains this literal row, under columns Gene Name, Common Protein Name, Detectable List and Reproducible List:
| Gene Name | Common Protein Name | Detectable List | Reproducible List |
|---|---|---|---|
| DDR2 | Discoidin domain-containing receptor 2 | V | V |
The table reports detection and reproducibility in the HFF1 adhesion proteome. Its DDR2 row is the 173rd zero-based HTML table row, including the title and header rows. Original Methods and Results describe hypotonic isolation/trituration of human HFF1 adhesion material and proteomic reproducibility criteria. This establishes detection in the preparation, not direct focal-adhesion structural activity, a specific interaction interface, or absence of associated cytoskeletal material. Failed PMC supplement downloads returned challenge HTML; the actual source used here is the linked author table.
PMID:20564243, Methods “Cell culture” and “DDR2 knockdown”, Results on osteogenic differentiation and mineralization, combines human MG63 marker assays with mouse C2C12/C3H10T1/2/MC3T3-E1 experiments. The mineralization readouts are from mouse cultures; the p38/Runx2 mechanism is not a direct DNA-binding function of DDR2. The publisher-indexed original at DOI:10.1002/jbmr.159 was read despite direct HTML/PDF access failures. The preserved abstract-only publication cache is not silently expanded with those original-only details.
PMID:19900459, Methods “Plasmid Constructs” and “Matrix Mineralization”, and Results/Figure 7, tests an engineered mouse DDR2 Met1–Ile400 V5/His ectodomain in MC3T3-E1 cultures. It changes fibrillogenesis and mineral deposition. The original article explicitly does not establish endogenous DDR2 ectodomain shedding. This experiment therefore informs contextual annotation but does not create a natural soluble human isoform.
PMID:11375938, Results/Figures 4–6 and Methods, uses Ddr2-null mice and mutant fibroblast rescue. Reduced growth-plate proliferation, rather than increased apoptosis, accompanies shortened bones; wild-type receptor rescues fibroblast proliferation whereas kinase-dead receptor does not. Original article. PMID:23409069, Methods and Results on DDR2 suppression, uses rat HSC-T6 cells and an alcoholic-liver-disease rat model; the proliferation, activation and apoptosis effects are retained within those contexts. Original article.
The binding papers are evaluated at the actual assay level. PMID:22939624 uses tagged-cell LUMIER/co-immunoprecipitation; PMID:25036637 supplies the CDC37 LUMIER association; PMID:33961781 supplies the DDR1-bait BioPlex association; PMID:35384245 uses MAC-tagged RTKs, pervanadate, AP-MS/BioID and selected co-immunoprecipitations. The accompanying evidence guide gives literal EBI/IMEx records and partner identities. Partner-class binding refinements do not imply purified direct interfaces, physiological ligand-dependent complexes, or phosphorylation/activation of the partner by DDR2.