CAPN5

Summary: Calcium-dependent cysteine protease activity is the established catalytic function of CAPN5.

Reason: The PAINT inference is consistent with the conserved calpain catalytic apparatus and direct full-length human CAPN5 assays in PMID33811937. C81A and calcium-chelation controls distinguish autoproteolysis from nonspecific breakdown. No target-specific catalytic loss contradicts the ancestral inference. Donor counts are not a measure of IBA validity.

Summary: The InterPro-to-GO activity mapping agrees with direct biochemical characterization.

Reason: CAPN5 retains the calpain catalytic core. Its calcium-dependent activity is supported independently by the full-length assays in PMID33811937; the non-classical domain architecture does not make it an inactive enzyme.

Summary: Retain the historical calcium-dependent protease assignment, corroborated by later direct assays.

Reason: PMID9339374 identifies CAPN5 and CAPN6; its abstract's predicted catalytic inactivity concerns CAPN6. The original CAPN5 activity support cannot be checked beyond the abstract-only cache, so no full-source verification is claimed. Later full-length CAPN5 assays establish the annotated activity (PMID33811937), supporting retention with explicit deference to the original TAS curation rather than a wrong-paralog claim.

Summary: Independent target-specific co-immunoprecipitation supports the CAPN5–STIP1 association as non-core.

Reason: The complete cached BioPlex 2.0 article (PMID28514442) describes affinity purification–mass spectrometry and co-complex associations. Its CAPN5–STIP1 supplementary pair record was not accessible. Independent PMID41597217 explicitly confirms endogenous STIP1 co-immunoprecipitation with full-length WT human CAPN5-3xFLAG and anti-STIP1 immunoblotting in SH-SY5Y cells. This corroborates the association without claiming that the original BioPlex target record was inspected. The confirmatory assays were single experiments; co-complex capture does not demonstrate a direct interface. STIP1 resisted calcium-induced proteolysis under the tested conditions. No chaperone function or more specific CAPN5 molecular function is established, so retain supported generic binding outside the catalytic core.

Summary: Retain the second CAPN5–STIP1 association with independent corroboration and explicit original-source limits.

Reason: The available canonical HTML for PMID33961781 contains the abstract, introduction and part of the discussion. Selected methods in the authenticated alternative XML confirm an AP-MS study, but the exact CAPN5–STIP1 supplementary record was not recovered. Original cell-line specificity and bait direction are therefore not independently asserted. PMID41597217 provides separate target-specific co-IP/immunoblot confirmation using WT full-length human CAPN5 and endogenous STIP1 in SH-SY5Y cells. This supports the generic association as non-core without assigning a direct binding interface, chaperone activity or STIP1 substrate status; STIP1 was resistant to proteolysis in that assay. The original experimental reference and partner identity are preserved.

Summary: Broad cytoplasmic localization is compatible with intracellular soluble and membrane-associated CAPN5.

Reason: The broad PAINT location is consistent with full-length human CAPN5 in cytosolic and membrane fractions in PMID33811937 and PMID37207905. This does not assert exclusive cytosolic localization or exclude nuclear localization in other experimental contexts.

Summary: Focal-adhesion proteomics is relevant, but the CAPN5-specific identification remains unverified.

Reason: PMID21423176 profiles isolated focal adhesions from human HFF1 fibroblasts and explicitly discusses the possibility of cytoskeletal/cytosolic contaminants. The complete cached article was read, but the CAPN5 row in the supplementary proteome was not available. Its calpain-2 mechanistic discussion is not transferred to CAPN5; neither absence nor contamination of CAPN5 is claimed without its target data.

Summary: CAPN5 directly performs proteolysis through its cysteine-protease activity.

Reason: The ancestral process assignment matches direct calcium-dependent cleavage by full-length CAPN5, including its experimentally controlled autoproteolysis (PMID33811937). This broad process does not require assigning an incompletely established physiological substrate pathway.

Summary: Refine the electronic proteolysis summary to directly demonstrated calcium-dependent self proteolysis.

Reason: The InterPro mapping correctly captures broad proteolytic capacity, but it does not itself establish self-cleavage. Independent full-length human CAPN5 experiments in PMID:33811937 supply the more specific evidence: calcium-triggered processing in SH-SY5Y cells depends on catalytic Cys81 and is blocked by EGTA; tagged and untagged constructs were examined. GO:1990092 describes calcium-dependent cleavage of the protein's own peptide bonds without requiring maturation into a more functional product. That outcome fits these assays, whereas GO:0016540 protein autoprocessing inherits a maturation requirement through protein processing that was not established. This is an evidence-backed refinement of one existing broad electronic row, not an inference from the InterPro domain alone, a claim that CAPN5 cannot cleave other substrates, or a change to the preserved broad PAINT proteolysis assertion. No redundant NEW descendant is added. The physiological purpose, cleavage-site identities and cis-versus-trans mechanism remain unresolved.

Proposed replacements: calcium-dependent self proteolysis

Summary: The accessible family-level signaling hypothesis does not resolve the complete historical CAPN5 TAS basis.

Reason: The complete accessible abstract of PMID:9339374 introduces signal transduction as a proposed property of calpains and then reports CAPN5/6 cloning, sequence comparisons and expression. That visible statement alone would not substantiate a CAPN5-specific signaling assignment. However, the original full paper and the curator's supporting passage remain inaccessible: an abstract opening is not proof that the entire TAS basis is visible or speculative. Therefore retain UNDECIDED rather than mark the historical assertion over-annotated on incomplete source access. Later mouse-myoblast intervention/rescue work (PMID:39357823) and human co-IP/in-vitro cleavage studies (PMID:41597217) offer mechanistic leads but do not resolve the particular human signaling response behind this assertion. In contrast, the sibling calcium-dependent protease TAS is independently corroborated by direct human CAPN5 assays (PMID:33811937); its ACCEPT does not verify the unread 1997 passage. Retinal disease alone does not identify the signaling step performed by CAPN5.

Summary: Refine the observed membrane association to plasma membrane without implying extracellular exposure.

Reason: PMID23055945 used permeabilized HEK293T cells expressing tagged human CAPN5 and described wild-type protein near the cell surface. It does not establish external exposure. Official GO defines cell surface as the external part of the wall/plasma membrane; plasma membrane includes associated proteins without prescribing that topology. Later full-length human CAPN5 microscopy colocalizes it with a plasma-membrane marker (PMID37207905), while S-acylation contributes to membrane association (PMID35643222). The localization finding is retained and refined.

Proposed replacements: plasma membrane

Summary: Synaptic localization is supported in a mammalian retinal context and remains an orthology-based human inference.

Reason: GOA transfers the assertion through mouse O08688/ENSMUSP00000048183. PMID27152965 reports mouse photoreceptor-synapse immunolocalization; its separate retinal fractionation experiment used bovine tissue. These data support a context-specific mammalian location, not a directly measured human synapse or a newly established synaptic transmission mechanism. Keep outside the minimal catalytic core.

Summary: Detected in an extracellular-vesicle proteome. The protein is plausible EV content, but this is not where it functions.

Reason: Cytosolic calpain. Cytosolic protein (GO cytosol or cytoplasm annotation); cytosolic proteins, including abundant enzymes and cytoskeletal proteins, are expected and often promiscuous EV contents (MISEV2018 (PMID:30637094) category 2b). True, but not its core location. (Manual re-review for OMICS_EVIDENCE, 2026-10-10; previous action: UNDECIDED.)

Summary: Detected in an extracellular-vesicle proteome. The protein is plausible EV content, but this is not where it functions.

Reason: Cytosolic calpain. Cytosolic protein (GO cytosol or cytoplasm annotation); cytosolic proteins, including abundant enzymes and cytoskeletal proteins, are expected and often promiscuous EV contents (MISEV2018 (PMID:30637094) category 2b). True, but not its core location. (Manual re-review for OMICS_EVIDENCE, 2026-10-10; previous action: UNDECIDED.)