CTSF is human cathepsin F (HGNC:2531; UniProt Q9UBX1), distinct from its inhibitor cystatin F/CST7. The genuine isolated fetch supplied 21 annotations, and every source field, qualifier and omitted default is preserved. All 21 were manually reviewed: 12 ACCEPT, 3 MODIFY and 6 KEEP_AS_NON_CORE; none are NEW. The default Falcon attempt failed authentication, and the configured Perplexity-lite fallback returned a genuine quota/authentication failure. No synthetic provider report was created. Primary evidence, official records and reproducible source reconstruction therefore supply the review.
Human recombinant CTSF hydrolyzes blocked peptide substrates and protein substrates, supporting cysteine-type endopeptidase activity and proteolysis [PMID:9822672; PMID:10318784; PMID:10748235]. The long official precursor is 484 residues: predicted signal 1–19, activation propeptide 20–270, mature chain 271–484. The current record lists no alternative splice products. The original partial AF071749/AAC78839.1 is 302 residues and maps 183–484; AF071748/AAC78838.1 is 338 residues and maps 147–484. The COS-7 construct begins at the longer partial start; it is not the intact modern precursor. The yeast-processing experiments are kept separate from that imaging construct. Original AF088886/AAD26616.2 is the complete 484-residue sequence PMID:10198209.
The mature 1M6D deposition is 214 residues from human CTSF expressed in Pichia GS115, with N97Q/N108Q/N170Q in mature numbering (N367Q/N378Q/N440Q in the precursor). Its inhibitor-bound structure is not a wild-type kinetic assay or intact-precursor localization experiment PMID:12225749. The cystatin-like N-terminal fold is a structural prediction and regulatory hypothesis, not a separately demonstrated protease-inhibitor activity. No inhibitor MF is proposed for CTSF.
The catalytic/processing consultation by ROOT personally read the recovered 10318784 original PDF and indexed original 15752368 scientific text. Those reads are attributed to ROOT; the owner read the full report and read ledger, reproduced its verifier, and independently assessed the other original sources and all selected abstracts. In 15752368, human CTSF expressed in Pichia was assayed at pH 5.5 with Z-Phe-Arg-AMC. Cystatin F inhibits it with Ki 0.17 ± 0.05 nM. This supports the positive binding assertion, retained as non-core because no justified narrower binding class was established. The protease is the inhibited entity, not the inhibitor PMID:15752368.
The original partial human protein in monkey COS-7 cells produces juxtanuclear vesicles resembling a separately imaged acidic-vesicle pattern. The original explicitly leaves lumen versus cytosolic membrane face unresolved. Its signal-independent trafficking interpretation is superseded by the later full precursor; its enzyme activity remains valid [PMID:9822672; PMID:10318784]. Lysosome is retained holistically with structural and curated mature-chain evidence. The broadly valid vacuole IBA is refined to the human lysosome. Strong LAMP2/CD68 colocalization in 15752368 is for cystatin F, not CTSF; the enzyme/protein antibody limitations are not concealed.
Recombinant human CTSF directly cleaves mouse MHC-II-associated invariant chain toward CLIP at acidic pH. Mouse cathepsin S/L-deficient macrophages and LHVS experiments support the contextual role, but LHVS is not selective CTSF genetics and the paper does not establish an exclusive human requirement [PMID:10748235, original Materials and Antibodies, in-vitro digestion Methods, Figure 6 Results and Discussion]. Reactome events 2130349 and 2130706 explicitly reach catalyst 2130638, physical-entity set 2130682, and hasMember CTSF entity 1181231/Q9UBX1 residues 271–484 in lysosomal lumen. Parent 2132295 contains those events. CTSF performs proteolysis here; it is not simply cargo. The MHC-II process remains non-core.
The complete 15184381 abstract reports secretion of active CTSF from human monocyte-derived macrophages and recombinant apoB-100/LDL cleavage. Full original Methods and Results were not recovered; preserve that access limit. This independent human evidence corroborates the extracellular IBA without treating secretion as a universal dominant location. Biallelic CTSF disease evidence identifies type B Kufs disease/CLN13 but does not alone specify a further performed biological process PMID:23297359.
The original 25037231 Table S1 identifies Q9UBX1/CTSF at one-based row 249 in both the complete MS-data and matrisome sheets, classified as an ECM regulator in liver-related fractions. It is a contextual fraction association, not a structural matrix function PMID:25037231. The original 18570454 Table 2, printed p. 3478, identifies cathepsin F precursor gi6467382 with two matched peptides in fraction 3 from human HB1.F3 neural-stem-cell supernatant. Main Methods/Results and the table pixels were read; fractionation/TEM/LC–MS supports the existing exosome association, not a vesicle-biogenesis function PMID:18570454.
The public PXD000608 MaxQuantOutput archive provides direct target records for the CSF extracellular-vesicle annotation. proteinGroups.txt physical line 1461 (header line 1) identifies Q9UBX1, group id1478, reports seven unique peptides and EV-sample signal. Five separate literal Q9UBX1 peptide entries occur at physical lines 2176, 2418, 3052, 4823 and 7776. One sequence matches current UniProt literally; the other four require isobaric I/L equivalence. The protein-group peptide-ID list differs from those five row IDs in the distributed files. This inconsistency is recorded, not repaired or treated as seven independently reconstructed peptides. No raw-spectrum reanalysis, exact proteoform assignment or main-paper Methods read is claimed PMID:24769233. Both EV annotations remain non-core positive associations.
The genuine current PTHR12411 v19 topology has 4,609 nodes. Target PTN002488050 has an exact 484-residue MSA match and a 22-node path. All four source IBDs are at on-path PTN000274640; no target-path loss record is present. The extracellular IBD has two additional current zebrafish donors; the historical source assertion remains exact. All 53 historical source identifiers were resolved through authentic direct records, family metadata or official locus records. The 59 complete current query pages contain 1,968 records and 107 literal term joins. Current donor evidence does not recreate an unseen historical curator session, and neither small donor sets nor self evidence are grounds for rejection.
The current ARBA00028064 capture has 24 branches and no target match. Its historical source coexists with the valid EC3.4.22.41 and direct human enzyme evidence, so the combined annotation is accepted without falsely claiming current rule reproduction. IPR000668/IPR013128 provide valid broad peptidase mappings that can be refined by target evidence. The current interaction download contains 46 complete target records with verified target sequence ROGIDs, but lacks the old 15752368/O76096 pair; the original inhibition assay supplies positive support. HPA vesicular/ER/plasma-membrane observations are contextual and do not generate additional annotations.
All 12 pre-existing caches remain exact: seven publications, two Reactome entries and three family files. PMID19159218 has a compatible shared source boundary; the different pending PR4576 PTHR12411 cache is preserved separately and is not treated as merged main. Ten additional publication caches and one Reactome cache are genuine normal fetch results. Failed retrieval bodies and provider failures are retained as evidence of limits. The compact source index supplies public locators, target paths and scope judgments; immutable raw captures and computational checks establish integrity separately from biological assessment.
The positive CTSF–CST7 interaction remains KEEP_AS_NON_CORE under the explicit ClinGen binding-retention policy. This project-specific instruction takes precedence over the generic skill's informational-exclusion rule. The decision rests on the original human CTSF inhibition assay, not the policy alone; cystatin F is the inhibitor, CTSF the inhibited protease, and no better-supported binding class was established PMID:15752368.
The active-in vacuole assertion is refined to lysosomal lumen (GO:0043202), using the two explicit mature-CTSF Reactome paths. Its valid ancestral IBD and all original source fields remain unchanged. This further specifies the earlier whole-lysosome refinement; the partial-clone imaging is not presented as proof of the membrane side.
The same catalytic core now includes the already accepted protein-catabolic process (GO:0030163), supported by direct human recombinant apoB-100 degradation. Current GO deliberately separates this term from proteolysis (GO:0006508), since peptide-bond cleavage also covers processing. Both are retained without an ancestor/descendant duplication. The LDL experiment is in vitro and does not measure general intracellular turnover PMID:15184381. No new annotation is proposed.
Reference access flags now distinguish cache format from access to an original. The original 9822672, 10318784, 15752368 and 18570454 texts were available beyond their abstract-only caches; those false full_text_unavailable flags remain accurate. The 17974005 cache contains XML-derived full text, so its flag is corrected to false. This does not claim a CTSF clone-table or functional-assay review; its role remains sequence-resource context. All cached files are unchanged. Original-reader attribution is preserved in the historical notes above and the immutable audit ledger, while public reference and source summaries now state scientific access limits without agent names.
The extracellular-vesicle question now asks which molecular form and activity are present. The PXD000608 numbering inconsistency and the five peptide entries, including four requiring I/L equivalence, remain disclosed in the preceding notes and source index; no seven-peptide reconstruction is claimed. GO_REF method references are marked LOW gene-specific relevance, without changing the soundness of the individual mappings or experimental source assertions.