CLN8 has two experimentally distinguished activities. It is a COPII cargo receptor for selected lysosomal enzymes and an acyl-CoA-dependent lysophospholipid acyltransferase. The 25 original annotations, their source fields and the absence of alternative products are preserved. Two molecular functions are added; no new biological-process annotation is proposed.
The selected full Methods and Results of PMID:30397314(https://doi.org/10.1038/s41556-018-0228-7) establish human CLN8 recruitment of newly synthesized lysosomal enzymes and interaction of its cytosolic tail with Sec24A/C. Export-motif mutations disrupt the coat interaction and Golgi trafficking; a luminal-loop deletion separates cargo recruitment from localization and recycling. Human-cell knockout/rescue and metabolic labeling support delivery of selected enzymes. Mouse lysosome measurements provide complementary evidence, rather than replacing the human assays.
PMID:32597833(https://doi.org/10.1172/JCI130955) distinguishes CLN6 recruitment at the ER from CLN8 export in COPII vesicles. Its budding assay combines HeLa membranes and rat liver cytosol. The proteins cooperate in enzyme recruitment, but CLN8 export does not require CLN6. The published corrigendum, PMID:41392992, replaces Figure 2A, 5A and 7A panels. A bounded independent reading inspected those corrections and the relevant supplement; the original selected normal cache remains unchanged. The original body also reverses the Figure 6A/B references relative to the legends, and its CLN6 deletion interval differs between Results and Methods. Neither inconsistency is silently carried into the CLN8 description.
Original PMID:10861296(https://doi.org/10.1093/hmg/9.11.1691) remains abstract-only in the normal cache. Its authentic abstract directly reports human CLN8 ER/ERGIC localization and retrieval-motif effects. The author's 2004 thesis separately describes those original experiments and labels additional unpublished observations. The thesis is not represented as recovered full journal text. Later human experiments corroborate the location independently. The broad membrane annotations are refined to the ER membrane without asserting exclusive residence.
The actual Methods and Results of PMID:39970228(https://doi.org/10.1126/sciadv.adr3723) show purified human CLN8-HA converting an LPG substrate and acyl-CoA into BMP, with radiolabel/TLC and diagnostic LC-MS/MS detection. Separate purification batches and substrate/enzyme controls support the activity. Translation of the explicit synthesized coding sequence matches all 286 residues of reviewed Q9UBY8; the C-terminal HA vector tag is additional. This is a sequence check of the tested construct, not an assertion about alternative products absent from the seed.
The LPG reagent has a racemic glycerol headgroup. The demonstrated BMP product does not by itself establish exclusive R,S or mature S,S stereochemistry. The complete route and inversion steps discussed in the paper remain mechanistic proposals. CLN8 concentration was not determined because residual HA peptide interfered with quantification; TLCD1 concentrations, kinetic constants, reaction mechanism and catalytic-mutant assays are not assigned to CLN8. GO:0071617 fits the observed transfer from acyl-CoA to a lysophospholipid. Its LPA- and LPE-specific children do not describe the CLN8 LPG-headgroup reaction.
The original PMID:16086686(https://doi.org/10.1111/j.1471-4159.2005.03376.x) abstract positively reports phospholipid and sphingolipid changes in two human EPMR brain samples. Full original Methods were not recovered. Broad phospholipid metabolism is retained with independent direct catalytic support, while the specific ceramide and cholesterol source interpretations remain unresolved. The PMID:30453012(https://doi.org/10.1016/j.bbadis.2018.11.011) abstract reports reduced ceramide in patient fibroblasts and rescue of phosphorylation changes by PP2A inhibition; the tested ER-to-Golgi ceramide conversion is normal. These observations provide cellular context without making CLN8 a phosphatase or a lipid transporter.
The full PMID:18317235(https://doi.org/10.1203/PDR.0b013e31816fdc17) explicitly reports CLN8-positive ceramide and GalCer overlays with normal-cell lysate. A CLN3-centered title does not invalidate the CLN8 result. Both direct and inherited ceramide-binding assertions are retained as non-core, with the distinction between lysate association and purified affinity or transport made explicit.
PMID:12151215(https://doi.org/10.1016/S0968-0004(02)02154-0) remains an inaccessible full family-analysis paper. Its authentic abstract proposes lipid-related functions. Later CLN8 chemistry supports broad lipid synthesis, but does not resolve its older lipid-transport, ceramide-synthesis or negative-proteolysis assertions. Those source-specific claims remain undecided. The positive later cargo-protection result is preserved: CLN8-deficient cells clear retained lysosomal enzymes more rapidly, and ERAD inhibition increases their abundance without restoring trafficking. This is not silently substituted for the original NAS linkage or treated as a demonstrated general protease-inhibitor activity.
The actual original PMID:10508524(https://doi.org/10.1038/13868) establishes pathogenic human variants and the mouse mnd mutation, with progressive neurological deterioration. Cloning from fetal-brain material is not a demonstration of a developmental step performed by CLN8. The direct developmental assertion is marked over-annotated at this source scope, while the historical PAINT developmental placement remains unresolved. This does not assert that a separate developmental role is absent.
The authentic PMID:10740217(https://doi.org/10.1002/(SICI)1097-4547(20000415)60:2%3C133::AID-JNR1%3E3.0.CO;2-3) abstract labels cathepsin D as CLN8 under historical nomenclature and proposes common lysosomal catabolism roles. Its full text remains unavailable. The reference assessment is therefore unverified, with the observed nomenclature issue explicit. The modern CLN8 receptor and lipid enzyme are not shown to perform protein breakdown; lysosomal cargo enzymes do that work.
Current PAINT target and donor records were retrieved without reconstructing the historical tree. Positive target judgments and unresolved ancestral-placement assessments are distinct. A target in its own WITH/FROM list is legitimate experimental grounding, not circularity. Official identities authenticate mouse Cln8 and Tlcd3b and zebrafish tlcd3ba. A direct WormBase stable-identifier mapping remains inaccessible; failed queries are not treated as zero biological evidence.
The original CLN3 association is retained under the project's policy for experimentally supported generic binding. Actual original coimmunoprecipitation and antibody-control Methods were read; an empty current IntAct target response does not refute them. The MINT target response contains an unrelated FZD7 screen record and is not used to invent a CLN3 join.
All normal fetched variants are archived. Existing selected main caches are unchanged, including abstract-only records where authentic external full text was read separately. Genuine Falcon research timed out after 90 seconds and the configured Perplexity fallback returned HTTP 401 quota failure; no provider-labelled report was authored. The review instead rests on the documented manual primary reads and bounded consultations. The current GO-CAM index had no match for the exact CLN8 identifiers queried. This is a limited index check, not evidence of biological absence.
The two proposed molecular functions have no typed is_a ancestor or descendant overlap with the original annotations or one another. The receptor and catalyst are distinct activities, so they are represented as two coordinated core functions. No additional process is created merely from loss-of-function necessity or shared pathway membership.
Normal validation, calculated-status checking and HTML rendering pass. The two advisory warnings have explicit dispositions: supported CLN3 binding remains non-core under project policy, and the direct developmental source receives an over-annotation judgment while the unreconstructed PAINT source remains unresolved. These evidence-specific decisions preserve every original source object. The calculated review status remains DRAFT because six assertions remain undecided. Project completion requires publication, review and merge.
The supported CLN3 interaction remains non-core under the explicit ClinGen project curation instruction. This task-specific user instruction supersedes the skill's informational-exclusion default; it neither asserts that generic binding is a core function nor validates inaccessible interactions. The original co-immunoprecipitation and control experiments supply the biological evidence.
The three externally recovered papers retain their genuine full-text access assessments while their selected abstract-only caches remain untouched. Selected original passages provide short transcriptions with public publisher URLs, exact Methods/Results headings or figure-panel locators and retrieved-file hashes. They document the preserum control, fetal-brain cloning material and CLN8-positive ceramide overlay; they are primary transcriptions rather than an authored summary cited as evidence.
The thesis mentioned above is Liina Lonka (2004), Neuronal ceroid lipofuscinosis CLN8 – from gene to protein, University of Helsinki dissertation, public repository PDF, pp. 53–55, sections 2.2–2.3. The PDF SHA256 is 0dbcc5225fbbd7207a97830dfed05d1a8a474921de5338c38d7155bc844a64ae. Its study-II summaries are distinguished from explicitly unpublished observations and from the unavailable full journal article. No annotation depends solely on the thesis: the authentic PMID:10861296 abstract and later PMID:30397314 experiments independently establish the ER/ERGIC itinerary.
The official GO:0160121 reaction uses two LPG molecules as acyl donor/acceptor and produces BMP plus glycerophosphoglycerol. CLN8 instead transfers acyl groups from acyl-CoA; GO:0071617 remains the appropriate measured activity. This donor-chemistry distinction is independent of the unresolved product stereochemistry. Core 2 now uses GO:2001312, lysobisphosphatidic acid biosynthetic process, because the purified human construct performs BMP formation, corroborated by cellular loss/rescue. This core-only refinement does not create a redundant BP annotation or claim that CLN8 performs the proposed stereochemical inversion pathway. The two original NEW molecular-function assertions, all 25 source assertions and their actions remain unchanged.
Follow-up normal validation, calculated-status checking and rendering pass on the unchanged DRAFT YAML. There are three advisories: retained supported generic binding, source-specific developmental actions, and the core-only GO:2001312 process absent as an exact annotation row. The latter is intentionally not added as NEW because existing phospholipid-metabolism coverage already includes its ancestor; the core records the more precise performed step without a redundant assertion.