Collagen-DDR signaling pathway module

Native triple-helical collagens in the extracellular matrix bind and activate the discoidin domain receptors (DDRs), receptor tyrosine kinases on the responding cell. Collagen binding to the discoidin domain triggers slow and sustained autophosphorylation of the receptor, which then recruits intracellular signaling proteins that regulate cell adhesion, migration, proliferation, differentiation and remodelling of the surrounding matrix.

Boundary. The module has two steps: a collagen ligand in the extracellular matrix, and a discoidin domain receptor tyrosine kinase on the responding cell. Downstream effectors (SHC1 recruitment to phosphorylated DDR1, NCK, SHP-2/PTPN11, PYK2, SRC-family kinases, ERK1/2 and the induction of matrix metalloproteinases such as MMP1 by DDR2) are not modelled as parts; none is a defining member of the pathway across species. Other collagen receptors are different pathways and are not modelled: the collagen-binding integrins (alpha1beta1, alpha2beta1, alpha10beta1, alpha11beta1), glycoprotein VI and LAIR-1 do not have tyrosine kinase activity and belong under GO:0038065 collagen-activated signaling pathway rather than its tyrosine kinase child GO:0038063. Collagen biosynthesis, secretion and fibril or network assembly are outside the module; the ligand step starts from a native, triple-helical collagen. Taxon scope. The module is scoped to Eumetazoa (NCBITaxon:6072), the taxon placement of the PAINT DDR node PTN002400507 in PTHR24416, which propagates collagen binding and protein tyrosine kinase collagen receptor activity to the DDR clade. DDR homologues are present in cnidarians, nematodes (C. elegans ddr-1 and ddr-2) and insects (Drosophila Ddr), and collagens are older still, so both required parts exist across the clade. The representative members (human DDR1, DDR2 and collagens) are examples, not a species scope. Drosophila Ddr (Q59DZ7) is not used as a representative: the Ddr gene review (genes/DROME/Ddr, Ddr-bioinformatics/RESULTS.md) finds that dipteran Ddr has lost the beta3 lysine (VAVK becomes VAVA), the catalytic-loop asparagine (HRDLATRS) and the DFG aspartate, and that the DDR2 collagen-pocket residues Trp52 and Asp69 are not conserved, so it is a predicted pseudokinase whose collagen binding is untested. Tribolium DDR and C. elegans ddr-2 keep all three catalytic residues (the honeybee predicted model keeps two, lacking the beta3 lysine, and C. elegans ddr-1 one), so the Eumetazoan scope stands, with the dipteran branch a probable loss of the kinase step. Scope criterion, shared with neurotrophin_toll_signaling: a module takes the taxon of the PAINT nodes that ground its steps; where no PAINT node grounds a step, it takes the clade in which the ligand-receptor pairing has been tested. Here the receptor node is Eumetazoan; the collagen ligand role itself is shown only in vertebrates (see the knowledge gap). Which collagens. The ligand step is a variant set by collagen class, following the receptor preferences summarised by Leitinger (PMID:24725424): fibrillar collagens I-III activate both DDR1 and DDR2; basement-membrane collagen IV activates DDR1 but not DDR2; collagen X activates DDR2 much better than DDR1 (PMID:16806867). Collagen VIII is reported to bind DDR1 (cited in PMID:24725424) and is not modelled as a separate variant. Reactome (R-HSA-2327738, R-HSA-2465890) also lists fibrillar collagen V for both receptors; the fibrillar variant's representatives are limited to the types I-III used in the founding binding studies. The ligands are in two PANTHER families (fibrillar and type IV collagens in PTHR24023 COLLAGEN ALPHA; collagen X in PTHR15427:SF48 COLLAGEN ALPHA-1(X) CHAIN), so no family id is asserted for the ligand step as a whole; each variant carries its own family. The Drosophila genome has only three collagen genes (Col4a1, viking and Multiplexin) and no fibrillar collagen. No experiment has shown that Drosophila collagen IV (Col4a1/viking) activates Ddr, so the fly type IV chains are not representatives. The only fly ligand candidate in the literature is Multiplexin (type XV/XVIII homologue, M9ND55 in PTHR24023:SF1098), from a dominant genetic interaction with Ddr in axon ensheathment (PMID:36355066). It is not modelled as a ligand variant: genetic interaction alone does not meet the evidence bar these ligand modules use (the neurotrophin_toll_signaling module leaves out Toll-2 on the same grounds), and a variant would make a satisfiability check count the fly ligand step as filled. It is recorded in the knowledge gap instead. In C. elegans the transmembrane collagen COL-99 acts genetically with ddr-1 and ddr-2 in axon guidance, also without biochemical evidence of binding; it is not modelled. Ligand function term. Each ligand annoton uses GO:0048018 receptor ligand activity, following the GO-CAM signaling receptor guideline that a ligand enables receptor ligand activity or a child. No child of receptor ligand activity fits a matrix collagen (the children are hormone, cytokine, growth factor, chemoattractant/chemorepellent, morphogen and similar activities). The main activity of these collagens is structural (GO:0005201 extracellular matrix structural constituent and related terms); that is not the role they play in this module and is not used for the ligand step. The receptor-side GO:0005518 collagen binding is a binding term carried by the receptors, not a ligand activity. Receptor family and function term. Human DDR1 (PTHR24416:SF333), DDR2 (PTHR24416:SF295) and Drosophila Ddr (PTHR24416:SF580) are in PANTHER family PTHR24416 TYROSINE-PROTEIN KINASE RECEPTOR (UniProt PANTHER cross-references), a broad receptor tyrosine kinase family; the role is restricted to the DDR subfamilies under PTN002400507. The receptor step uses GO:0038062 protein tyrosine kinase collagen receptor activity, the term asserted at PTN002400507 and carried by human DDR1 and DDR2 by IDA (PMID:9659899). Annotation state (QuickGO, queried 2026-11-06). Human: DDR1 (Q08345) and DDR2 (Q16832) enable GO:0038062 (IDA, PMID:9659899; IBA from PTN002400507) and are involved in GO:0038063 (DDR1 IDA PMID:9659899, PMID:21044884 and IMP PMID:12065315; DDR2 IDA PMID:16186108). Of the collagens, only COL4A2 (P08572, IBA from PTN000615596) and COL4A6 (Q14031, IBA from PTN000615545) are annotated to GO:0038063; COL1A1, COL1A2, COL2A1, COL3A1, COL4A1 and COL10A1 are not, and no human collagen has GO:0048018 receptor ligand activity or any descendant. Drosophila: Ddr (Q59DZ7) enables GO:0038062 and GO:0005518 (IBA from PTN002400507) and is involved in GO:0038063 (IEA, GO_REF:0000108 inferred from the MF); Multiplexin (M9ND55), Col4a1 (P08120) and viking (Q9VMV5) have no annotation to GO:0038063 and no receptor ligand activity. No human or fly protein is annotated to GO:0038065 through a collagen. Gene reviews. The human DDR1 and DDR2 reviews have GO:0038062 as their core molecular function. The Drosophila Ddr review does not (see the scope note): it MODIFIES the GO:0038062 IBA to GO:0004888 and gives Ddr's core role as wrapping-glia axon ensheathment (PMID:36355066). Reviews exist for COL1A1, COL1A2, COL2A1, COL3A1, COL4A1 and COL10A1; they keep the Reactome DDR1/DDR2 binding events as ligand context, and the COL1A1 review (updated alongside interpro/panther/PTHR24023/PTHR24023-review.yaml) carries a second core function for receptor ligand activity (GO:0048018) in collagen-activated tyrosine kinase receptor signaling pathway (GO:0038063), whereas COL1A2, COL2A1, COL3A1, COL4A1 and COL10A1 have not yet had a DDR ligand core function added (the GO:0048018 refinement proposed in the COL3A1 review is for its LAIR1 ligand role, not DDR). Receptor preference. The DDR1/DDR2 preference of each collagen class is stated in the variant labels and in this note only; the receptor step is a single annoton for both receptors with one connection, so the pairing (for example collagen IV with DDR1 but not DDR2) is not modelled in machine-readable form.