CLN6 is an ER membrane protein that cooperates with CLN8 during lysosomal enzyme recruitment. The original 54 annotations, two alternative-product objects and 17 reference identity pairs are retained. The draft has no new GO assertion. Positive generic interaction evidence is kept or refined by a verified partner class; a missing native mechanism is not treated as evidence against binding.
Selected actual Methods and Results of PMID:32597833(https://www.jci.org/articles/view/130955) establish CLN6-dependent enzyme recruitment, while the budding assay excludes CLN6 from COPII carriers. The core is consequently descriptive, with ER membrane location. Pairwise association and mutual dependence do not resolve an autonomous cargo-receptor activity or a particular adaptor geometry. No process term is added from deficiency alone.
The corrigendum replaces Figure 2A, 5A and 7A images. Corrected pixels and two supplement pages were inspected in the independent bounded consultation. Results and Methods disagree about the exact loop-deletion interval; the inspected supplement does not settle it. Figure 6 caption/image labels supersede reversed body cross-references. These limitations remain explicit rather than being silently repaired.
The selected normal cache for PMID:15010453(https://doi.org/10.1074/jbc.M400643200) remains an abstract. Authentic author-uploaded original Methods and Results were recovered separately. Human CLN6 expressed in BHK21 cells shows glycosylation at an engineered luminal site and cross-link-dependent dimer formation. Its domain-level lumen assertion is compatible with multipass ER-membrane residence. Patient fibroblasts show impaired degradation of endocytosed arylsulfatase A, while cathepsin D processing is preserved in those experiments. The later recruitment study uses different models and assays; neither source is rewritten to match the other.
Actual publisher text of PMID:11722572(https://febs.onlinelibrary.wiley.com/doi/10.1046/j.0014-2956.2001.02530.x) supplies the pH experiment behind the selected abstract. One control and a CLN6-locus-linked patient fibroblast line were tested before sequence-based CLN6 identification. This supports secondary pH context, not a CLN6 proton pump. Current HPA XML independently corroborates ER localization; original HPA image pixels were not viewed. Secondary early-endosome and raft observations in PMID:17237713(https://doi.org/10.1203/pdr.0b013e31802d8a4a) are retained at their marker-colocalization scope.
The full original CLN6–CRMP-2 and CLN5 studies distinguish a human fragment library bait, tagged-cell associations, and a mouse GST-Cln5 reagent. No returned IntAct record for these older pairs is treated as a negative experiment. The later screens have exact accession/reference joins: HuRI is PMID:32296183, while the WWP2 and EPHB2 pairs belong to PMID:32814053. Viral physical-association records are separated from one colocalization observation. Features documenting mutations, tags and sufficient binding regions are retained in the index.
Partner-class refinements describe what CLN6 binds. They do not transfer protease, kinase, transporter, GTPase or transcription-factor activities to CLN6. TMPRSS2 isoform 2 has a 37-residue N-terminal extension under the current UniProt sequence model; its recorded 38–529 region retains the protease region. EPHB2 isoform 3 retains the receptor kinase domain. The exact RIC3 product is kept broad because current sequence-feature/product labeling is not fully reconciled. SLC30A4 transporter identity is family-supported; its precise substrate flux is not established by the screen.
CLN6 lipid binding in PMID:18317235(https://doi.org/10.1203/PDR.0b013e31816fdc17) is genuine target evidence despite the CLN3-focused title. Actual lysate overlays support the LPA and sulfatide assertions as secondary capacities. They do not establish purified affinity or soluble lipid-carrier chemistry. Current CLN6 isoform 2 replaces the canonical N-terminal 28 residues with 60 residues, yielding a reconstructed 343-residue product; equivalent recruitment or binding has not been demonstrated for both products.
The exact ganglioside and glycosaminoglycan experiments in PMID:16857350 remain inaccessible after bounded primary-source attempts. Its explicit cholesterol accumulation finding is retained separately. The NK-cell proteomic target table in PMID:19946888 remains uninspected. These three source-specific assertions are undecided rather than removed. All four PAINT assertions have valid target self-donor grounding; historical node placement and alignment were not reconstructed, and donor count is not used to judge support.
Normal fetch output and selected exact-main caches are immutable. External original text, source receipts and the bounded consultation are recorded separately in the source index. The Falcon attempt timed out at 90 seconds and the explicit fallback returned an authentication/quota failure; no provider report was fabricated. Manual research supplied the review. The GO-CAM index is checked for the target; no new process is proposed from absence. Short literal primary anchors occur only in the YAML. Their aggregate budget includes these notes and the source index, neither of which repeats those quotations. ROOT will create the genuine history record during publication.
The current ClinGen project instruction explicitly preserves supported generic protein-binding annotations as non-core when a verified narrower class is unavailable. That user instruction takes precedence over the generic annotation-reviewer skill's informational-exclusion recommendation. All 22 positive generic-binding decisions are therefore retained. This policy does not authenticate a missing experiment or turn association into a mechanism. The original history record and the preceding journal entries remain unchanged.
The topology assessment now distinguishes the two source relations. The IBA assertion says is_active_in the ER lumen: CLN6's luminal interface contributes to enzyme recruitment. The IDA assertion says located_in: the original engineered glycosylation experiment exposes a loop, while the protein itself spans the ER membrane. Row 39 is refined to ER membrane, as is the broad membrane NAS row 45. This supersedes the earlier journal's acceptance of a domain-level located_in assertion. Neither original qualifier is rewritten. GO's annotation-relation guidance distinguishes activity location from detected location, and its membrane guidance assigns transmembrane proteins to the membrane compartment.
The source-specific acidification assertions (rows 41 and 42) and protein-catabolism assertion (row 46) are now undecided, replacing their earlier non-core dispositions. In PMID:11722572(https://febs.onlinelibrary.wiley.com/doi/10.1046/j.0014-2956.2001.02530.x), Methods, Table 1 and Discussion show increased pH in one CLN6-locus-linked patient line and explicitly leave a direct or indirect mechanism open. This remains a valid observation. Neither that phenotype nor the later enzyme-recruitment assay identifies a step CLN6 performs in acidification. The historical PAINT node is not reconstructed or alleged to be wrong. The broad catabolism reference PMID:10740217(https://doi.org/10.1002/(SICI)1097-4547(20000415)60:2%3C133::AID-JNR1%3E3.0.CO;2-3) remains abstract-only; direct recruitment and reduced substrate degradation do not settle the boundary between an upstream delivery role and participation in protein breakdown itself. This is not a rule that only pumps or proteases can participate. The distinct regulatory assertion in row 52 retains the measured degradation effect and the demonstrated enzyme-delivery role.
The core now names a proposed lysosomal enzyme recruitment activity at the ER. This is an ontology proposal for the performed cooperative role, not a new GO annotation or a claim that an existing term has been created. PMID:32597833(https://www.jci.org/articles/view/130955), Results/Figures 4–7 and Methods, establishes selective enzyme association and export delay; CLN6 remains in the ER while CLN8 accompanies exported cargo. The current scaffold definition requires holding a complex together, and cargo receptor/adaptor definitions require coat-linked recruitment not established for CLN6. Pairwise association does not distinguish a bridge from cooperative recognition. The proposed term therefore leaves ontology parent placement open. The correction, loop-coordinate discrepancy and Figure 6 caption/body mismatch remain explicit. This follow-up reread the corrected text; it reuses the prior consultation's image inspection without claiming a new image review.
A few cached-primary anchors now make the location and pH observations and the secondary endosome/raft result checkable. Detailed topology claims come from the separately recovered author-uploaded PMID:15010453(https://www.researchgate.net/publication/8671804_Defective_Endoplasmic_Reticulum-resident_Membrane_Protein_CLN6_Affects_Lysosomal_Degradation_of_Endocytosed_Arylsulfatase_A), indexed Methods and Results 398–1138; its normal cache remains an abstract. The source index preserves these external access locators and literal fetched records. It is not cited as biological evidence for itself, and its JSON format is retained.
PMID:28476624(https://doi.org/10.1016/j.bbrc.2017.05.002) remains relevant supporting context: its authentic abstract reports that CLN6 associates with the R120G alphaB-crystallin mutant and reduces aggregation in HeLa cells, with the effect compromised by lysosomal inhibition. Detailed Methods were not accessed. This supports a cellular proteostasis connection without establishing intrinsic folding-chaperone activity. The exact GAG assertion from PMID:16857350(https://doi.org/10.1016/j.bbadis.2006.06.002) remains undecided; its transcript observations cannot substitute for the unavailable full-body assay scope.
All 54 source objects, two products and 22 reference identities are preserved. Five actions change: two localization refinements and three process uncertainties. No GO annotation is added. Primary quotations occur only in the YAML and remain below 25 words per cited source across the published files. The prior source index remains intact apart from the explicit core-scope update and this compact follow-up provenance.
This assessment supersedes the earlier preference to retain the is_active_in ER-lumen review as ACCEPT; the original machine assertion and its qualifier remain unchanged. Formal RO:0002432 is a subproperty of overlaps, with an enables/occurs_in chain, rather than whole-product located_in. Luminal-loop recruitment is therefore formally compatible with the relation. The official GO membrane guidance nevertheless recommends membrane annotation for transmembrane proteins. The refinement to GO:0005789 adopts that representation convention; it does not deny the luminal recruitment interface or allege a wrong historical PAINT node. The official RO ontology and annotation-relation guide distinguish the logical relation from the annotation convention.
Two complete current QuickGO CC queries contain 114 records across human CANX, LMAN1, LMAN2 and CLN8. CANX, LMAN1 and LMAN2 use is_active_in ER membrane, including an LMAN1 record whose returned evidence code is IDA; their official UniProt records describe luminal domains and membrane-spanning segments. CLN8 uses the broader ER term. No is_active_in ER-lumen record occurs in this bounded set. CANX also has located_in lumen Reactome records, so this is evidence of relevant practice, not an absolute exclusion rule. The LMAN1/LMAN2 IBA history was not reconstructed or counted as independent replication. These comparator references describe annotation provenance; they are not newly imported target annotations or claims that their underlying papers were read.
Original PMID:15010453(https://doi.org/10.1074/jbc.M400643200) Methods and Results establish engineered I153S glycosylation in BHK21 cells, glycan sensitivity and ER localization. The selected repository cache remains abstract-only; actual original author-upload text was inspected separately. Original PMID:32597833(https://www.jci.org/articles/view/130955), Results on the second luminal loop and Methods on constructs, preserves positive enzyme recruitment and CLN6 self-association. Loop deletion retains self-association yet loses enzyme association. The existing recruitment core now integrates self-association explicitly, without asserting that its necessity was tested or adding a second core. The unresolved deletion-coordinate discrepancy and published correction scope remain as previously documented.
Only row 38 changes action, ACCEPT to MODIFY; row 39 has a matching explanatory sentence, and the existing core gains the self-association interpretation. Other decisions, source assertions, products, reference identities and previous journal entries remain unchanged. No new biological annotation, quotation, cache or history record is authored here. The compact source index appends the official-source provenance; historical index content remains intact.
The earlier statement about 114 comparator records refers to two complete saved QuickGO CC responses, not to the repository GOA TSV snapshots. The first response, retrieved on 2026-10-11 at 10:59:33.485501 UTC, contains 97 records for CANX, LMAN1 and CLN8; the second, retrieved at 11:00:25.336798 UTC, contains 17 LMAN2 records. Their original hashes, page information and complete literal responses are preserved as unaltered downloads in the reproducible comparison snapshots, with locators in the source index and the exact TSV comparisons. The prior note is preserved as provenance and clarified here.
| Saved record | Gene | GO term/relation | Evidence | Assigned by | Reference recorded by QuickGO |
|---|---|---|---|---|---|
| UniProtKB:P27824!357439916 | CANX | GO:0005789 / is_active_in | IBA / ECO:0000318 | GO_Central | GO_REF:0000033 |
| UniProtKB:P49257!357460019 | LMAN1 | GO:0005789 / is_active_in | IBA / ECO:0000318 | GO_Central | GO_REF:0000033 |
| UniProtKB:P49257!357460020 | LMAN1 | GO:0005789 / is_active_in | IDA / ECO:0000314 | WB | PMID:10559958 |
| UniProtKB:Q12907!327654117 | LMAN2 | GO:0005789 / is_active_in | IBA / ECO:0000318 | GO_Central | GO_REF:0000033 |
Reproducible URLs: CANX/LMAN1/CLN8 CC query and LMAN2 CC query. Repeating a live query may yield newer records; these claims concern the hashed saved responses at the stated times.
The reviewer accurately describes the narrower branch TSV content: CANX lacks is_active_in rows there, and LMAN1 lacks the queried IDA row there. The same TSV bytes occur in the inspected current main. That discrepancy does not erase the actual records in the separate complete query responses. The cause of the differing snapshots is not established here; neither their file dates nor the presence of some other relation rows proves that they have identical coverage. The IDA, WB and PMID:10559958 values above are fields returned by the API. Their presence is not a judgment that the assignment or its biological interpretation is correct. PMID:10559958 was not read as a comparator experiment and is not offered as direct evidence for CLN6. No qualifier is converted into a new functional inference or an asserted historical PAINT placement. The existing membrane refinement remains unchanged.
The current GO:0038024 cargo receptor activity definition specifies joint cargo/coat-adaptor engagement for delivery to transport vesicles. Actual CLN6 experiments establish ER-resident enzyme recruitment and luminal-loop dependence, while CLN8 provides subsequent export; they do not establish the requisite simultaneous CLN6 cargo/coat-adaptor engagement. Naming this existing candidate makes the proposed recruitment term's scope checkable without changing any annotation or claiming that recruitment is absent. The core's document-navigation sentence has been removed; its biology and self-association limits are preserved.
The established JSON evidence index remains a structured source/provenance artifact, not independent evidence for its own claims. Its literal query records and reproducible locators address the checkability concern directly; no conversion or new self-support citation is introduced. That third-pass amendment preserved all original source objects, products, actions, references and primary quotations. The following fourth-pass assessment revises the comparator wording and provenance presentation.
Ordinary validation, unchanged derived status and rendering pass for this third-pass amendment. The 23 advisories comprise 22 supported generic-binding retentions under the explicit campaign exception and one existing source-specific membrane action difference (NAS refinement versus unresolved HDA). None was introduced by the three prose changes. All original source/cache/history bytes and 281 current tool files remain unchanged. The third-pass index embedded the complete comparator payloads; the fourth-pass revision below relocates those same raw responses to reproducible snapshot files.
A fresh direct HTTPS request to the exact combined QuickGO URL returned 97 records and exactly the original 45,011 response bytes (SHA256 a7b92a379413a4a6a44b5322f22c9a1893154d2b7f0d03348dce0790053b4ef2). Both disputed identifiers, UniProtKB:P27824!357439916 and UniProtKB:P49257!357460020, are present in those unaltered returned bytes. A second complete query reproduces all 17 LMAN2 records and its original response hash. The allegation that these entries were invented is therefore not supported by the direct API observations. The differences from repository TSV files remain real; neither record totals nor annotation dates establish their cause. In particular, WB/IDA on the human LMAN1 record is reported only as returned metadata, without endorsing its assignment, interpreting PMID:10559958, or inventing a provenance explanation.
The parameterized comparison script and instructions retrieve every advertised page, preserve raw bytes and response headers, check hit counts and record-ID uniqueness, and produce a structured YAML report. It can also replay saved responses and compare explicitly named common fields with repository TSVs. A separate LMAN2 query with a five-record page size successfully exercises four-page retrieval. The analysis summary records the comparison limits. Future live results may differ; the saved snapshots remain dated observations.
Row 38 retains MODIFY to the ER membrane. Its shorter reason now rests on CLN6's experimentally established membrane topology, luminal recruitment interface and the official GO membrane convention. Comparator queries are corroboration about annotation practice, not experiments establishing CLN6 biology. The source index remains a provenance/navigation artifact and is removed from the scientific reference list because no supported_by entry cites it. The original machine-fetched reference identities, all 54 source assertions, both products, every action and replacement, the core and all primary quotations are unchanged. No existing history record is rewritten.