ATP6V0A2 is the human V-ATPase 116 kDa a2 subunit, a V0-sector membrane subunit of the V-ATPase. UniProt describes it as a "Subunit of the V0 complex of vacuolar(H+)-ATPase" and states that V-ATPase acidifies intracellular compartments [file:human/ATP6V0A2/ATP6V0A2-uniprot.txt "Subunit of the V0 complex of vacuolar(H+)-ATPase"; "V-ATPase is responsible for acidifying and maintaining the pH of intracellular compartments"].
The strongest direct ATP6V0A2 literature support is endosomal. Hurtado-Lorenzo et al. reported that "The a2-isoform is targeted to early endosomes" and that it "interacts with ARNO in an intra-endosomal acidification-dependent manner"; they also found that "Inhibition of endosomal acidification abrogates protein trafficking" [PMID:16415858 "The a2-isoform is targeted to early endosomes"; "interacts with ARNO in an intra-endosomal acidification-dependent manner"; "Inhibition of endosomal acidification abrogates protein trafficking"]. This supports endosome membrane, proton transport/acidification, and pH-sensing trafficking context. It does not justify retaining generic protein binding as an informative MF annotation.
The human V-ATPase structural paper supports the general complex architecture and the endosome/lysosome acidification role: V-ATPases are "ATP-driven proton pumps" with a membrane Vo complex for proton transfer, and organellar V-ATPases establish pH homeostasis of endosomes and lysosomes [PMID:33065002 "V-ATPases are ATP-driven proton pumps"; "homeostasis of endosomes and lysosomes"]. The 2020 V-ATPase review similarly states that V-ATPases are the "primary source of organellar acidification" PMID:32001091.
The PN projection has two ATP6V0A2 candidate additions: GO:0007042 lysosomal lumen acidification and GO:0046610 lysosomal proton-transporting V-type ATPase, V0 domain [file:projects/PROTEOSTASIS/reports/pn_projection/pn_projected_annotations.tsv "ATP6V0A2"; "GO:0007042"; "GO:0046610"]. The mapping audit requires manual gene-level review for lysosomal acidification propagation [file:projects/PROTEOSTASIS/reports/pn_mapping_audit/current_mapping_scrutiny.tsv "manual_gene_level_review_required_before_gene_review_change"]. I accepted both PN-projected additions because they are supported by gene-level V0 subunit identity plus lysosomal membrane evidence, but I did not promote the broader PN autophagy/nutrient-sensing context into GO beyond these terms.
Lysosomal membrane support comes from lysosomal membrane proteomics that found "17 polypeptides comprising or associated with the vacuolar adenosine triphosphatase" in purified placental lysosomal membranes PMID:17897319. This is enough to support lysosomal membrane and the PN lysosomal V0-domain addition when combined with ATP6V0A2 V0-subunit identity.
The HIF/iron paper found ATP6V0A2 among V-ATPase subunits enriched in a HIF1A reporter screen and concluded that V-ATPase disruption causes "intracellular iron depletion" [PMID:28296633 "ATP6AP1, ATP6V1A, ATP6V1G1, ATP6V0A2 and ATP6V0D1"; "disrupting the V-ATPase results in intracellular iron depletion"]. This supports keeping intracellular iron ion homeostasis as non-core, but the broader oxygen-response term is over-annotated relative to the mechanism.
The macroautophagy and immune-response annotations should be treated conservatively. The lipofuscin paper says "both the autophagosomes and the lysosomal system are not mandatory" for lipofuscin formation PMID:22982048, which is weak support for ATP6V0A2 as a macroautophagy regulator. The old J6B7/T-cell cDNA paper frames the clone as having a "putative immune regulatory role" PMID:2247090, but this does not support a modern core ATP6V0A2 immune-response annotation.
This re-review preserves all 30 original machine-supplied assertions, their order, qualifiers and references. Decisions are 17 ACCEPT, eight KEEP_AS_NON_CORE and five UNDECIDED. One additional IDA molecular-function proposal records PI4P binding, giving 31 reviewed rows and two core functions. Two legacy NEW proposals for lysosomal lumen acidification and lysosomal V0 membership are withdrawn because the added lysosomal specificity was not established. No new biological-process annotation is proposed.
The a2 subunit contributes proton translocation in the assembled V0/V1 pump; it does not independently hydrolyze ATP. The synthesis emphasizes Golgi and early-endosome membranes. The human structure in PMID:33065002 contains a1, supplying conserved architecture rather than direct a2 lysosomal evidence. Its short normal cache lacks Results/Methods despite a full-text flag. The existing experimental lysosomal-membrane assertion remains UNDECIDED pending its precise target table. The synthesized Golgi/endosomal core uses GO:0033176, the general V-type complex. GO:0016471 explicitly requires vacuolar-membrane context in its definition and part_of parent; original machine-supplied terms remain unchanged.
PMID:16415858 distinguishes a2–ARNO/PSCD2 association from subunit-c–Arf6 binding. The independent core consultation examined vesicle co-IP and recombinant-tail evidence; the tail donor was unresolved. No GEF activity is asserted. Mouse acrosome and macrophage perinuclear studies remain model-specific context. Mouse macrophage a2 exclusion from bacterial phagosomes does not establish a universal human negation. PMID:28296633 supports human a2-dependent HIF effects under aerobic culture, but a true oxygen step-up assay and a2-specific iron rescue remain unresolved. Accessible PMID:22982048 text names an a1 perturbation; incomplete access cannot invalidate the a2 annotation. PMID:2247090 describes a real mixed-lymphocyte assay whose human specificity remains uncertain.
PMID:18157129 (doi:10.1038/ng.2007.45) is an abstract-only normal cache. Human ATP6V0A2 disease variants cause glycosylation defects and impaired patient-fibroblast Golgi trafficking. This establishes disease and organelle context without independently demonstrating lumen pH.
PMID:38396846 (doi:10.3390/ijms25042170) tests purified human a2 residues 1–400 and PI4P-containing liposomes. Mutations near residues 237–240 reduce binding; circular dichroism controls gross secondary structure. Human HEK293 fragment targeting, membrane fractionation and acute Golgi PI4P depletion connect that binding to localization. These results justify NEW GO:0070273 and a separate lipid-recognition core. The paper does not establish affinity, exclusive lipid preference, full-length pump activation or full-length rescue. The full-length targeting mechanism is an inference from a defined human fragment. The isolated K238_V238.del Methods spelling is inconsistent with the paper's usual K237_V238del designation. Original Results 2.1–2.3 and Methods 4.1–4.5 were read from the exact normal XML cache; images and supplements were not independently reanalyzed. The official GO term and its phosphatidylinositol-phosphate-binding parent were checked; this MF is distinct from the existing pump, ATPase-binding and protein-binding terms.
PMID:39680136 (doi:10.1007/s00018-024-05506-7) uses calibrated GalT-RpHLuorin2 in immortalized mouse MEFs, with mean Golgi pH 5.80 in wildtype, 6.52 in knockout and 6.25 in R755Q cells. Ionophore and bafilomycin controls support interpretation. A separate experiment expresses human ATP6V0A2-V5 in HeLa cells and finds TGN46 co-localization. Species and assays are kept separate: no human pH IDA row is added. The article's consent boilerplate does not override its explicitly documented human construct and cell line. Selected Results, Fig. 1/3 text, reporter, cell-culture and viral-expression Methods were read; whole-paper or raw-data reanalysis is not claimed.
All three new caches came unchanged from the normal source pipeline. Reference identities were independently checked against primary records before marking them VERIFIED. Other reference assessments remain UNVERIFIED where their target-specific support is unresolved. Generic binding stays non-core where supported; limited functional specificity alone does not establish a false interaction under the user-supplied ActionEnum. No complete PAINT tree or supplemental proteomics dataset was inspected. Earlier project-specific notes above remain historical and are superseded by this assessment.
The normal source pipeline has now supplied unchanged abstract-only caches for PMID:11872743 and PMID:19549681. Both are included as top-level references with full_text_unavailable true. The former abstract directly describes mouse a2 acrosomal staining. The latter abstract emphasizes a3; the a2 perinuclear/phagosome distinction remains supported by the separately documented primary Results/Fig. 1 web reading, not by an invented cached passage. Neither source is presented as a direct human assay. This reference completion changes no annotation, action, core function, description, or literal supporting quote.
The donor citations are now attached directly to their existing rows: PMID:11872743 to acrosomal localization, and PMID:19549681 to the perinuclear and phagocytic-vesicle assessments. The former normal abstract explicitly names mouse a2 in acrosomes. The latter normal abstract foregrounds a3; the a2 comparison remains based on the previously documented original Results/Figure 1 reading at https://doi.org/10.1242/jcs.050443 and is not represented by an invented cache quotation.
PMID:18157129 is linked to the first core as corroborating human Golgi-function evidence: affected-fibroblast trafficking and patient glycosylation support the biological context, without independently demonstrating proton flux or Golgi pH. PMID:16415858 is marked VERIFIED for its identity and bounded a2–ARNO/endosome claims. Its normal cache remains abstract-only; the earlier selected original-paper read does not establish the recombinant tail donor species. The verified generic association is retained as non-core under the user ActionEnum definition. No specific GEF, scaffold or adaptor activity is inferred solely from this association.
The complete official indexed abstract of PMID:17295899 (https://pubmed.ncbi.nlm.nih.gov/17295899/; DOI https://doi.org/10.1111/j.1600-0897.2006.00463.x) was read on 2026-09-29. Female human PBMC/JEG-3 co-cultures exposed to a2NTD showed increased IL-1beta secretion and altered IL-1-receptor gene expression. The measurements were ELISA and gene arrays. The abstract describes a released N-terminal product, not a splice isoform. Direct PubMed rendering was empty and the publisher full text was inaccessible; exact recombinant residues/species, purification and endotoxin controls, complete Methods, figures and supplements were not independently read. Thus the earlier reason cannot imply that no human immune-cell experiment exists. This is contextual support, not resolution of the original mouse J6B7 TAS evidence or proof of a general immune function of the intact pump. Row 30 remains UNDECIDED. The normal PMID:17295899 cache is absent; no source file, formal reference object or cache quotation has been manufactured.
The official GO:0006955 definition and parents were inspected at https://amigo.geneontology.org/amigo/term/GO:0006955. It is an immune-system response to a potential threat; term breadth alone does not refute a contextual immune role. A target-table question now records the exact missing PMID:17897319 lysosomal-proteome evidence. Existing immutable PN notes remain historical. All original source assertions, actions, the existing PI4P addition, products and both core mechanisms are preserved.
The applied follow-up passed focused validation (two existing advisories for generic binding and the uncited manual research report), rendering, and validation of the newly scaffolded history record. The independent consultation covered these finite changes and preserved all 31 annotation source objects; no new whole-repository validation pass or complete-paper reassessment is claimed.
The normal just fetch-pmid 17295899 attempt failed with a name-resolution error and produced no cache. Its actual failure is retained in the task receipts. The earlier bounded abstract reading above remains a research note; the specific uncached PBMC/JEG-3 and a2NTD findings are removed from the machine-readable immune-response reason. That original TAS row stays UNDECIDED based on the unresolved transfer and assay specificity of the cached historical J6B7 source. No quotation or reference object is fabricated.
Eight concise, cache-verbatim findings now document early-endosome targeting, patient Golgi-trafficking defects, family-level proton pumping, historical lymphocyte suppression, macroautophagic uptake, human a2 CRISPR/HIF context, the purified human 1–400 PI4P-binding fragment and the human HeLa localization experiment. The mouse fibroblast pH assay remains separate. Sixteen existing rows were checked for explicit source links; four gain links, while the HIF-context and historical immune rows receive bounded direct anchors. Complete already-read abstracts and selected exact primary passages were checked again; the broader concatenated search output was truncated and is not represented as a new whole-paper reading.
The 21 old findings and 72 old quotes are not restored mechanically: evidence still has to support the revised claim, respect donor/construct limits and avoid unnecessary repeated extracts. All 31 source objects and actions, both cores, the established fragment PI4P activity, original caches and prior history remain unchanged. The verified a2–ARNO association remains non-core under the user ActionEnum; no unsupported replacement MF is invented. The proposed result has 20 literal snippets including eight reference findings. Validation and publication of this second proposal have not yet occurred.
The preceding section records the earlier proposal state. Independent finite scientific review passed, and the reviewed proposal was applied. Focused validation passed with two advisories: the retained verified generic-binding association and the unlinked manual research document. Rendering passed. A new history record was scaffolded for this response; earlier histories and cached sources are preserved. PR publication and review of the new head remain separate steps; no new repository-wide validation pass is claimed.
The retained GO:0005515 decisions implement the user's explicit instruction for this ClinGen task: keep a supported, biologically correct interaction as KEEP_AS_NON_CORE when no defensible, more specific replacement has been established, rather than remove it solely for lacking functional information. This deliberately departs from the annotation-reviewer skill's informational-exclusion recommendation. That recommendation can remove an uninformative annotation without declaring the interaction false; the latest review is correct about that distinction. Earlier references here to the supplied ActionEnum are incomplete as an explanation of authority: the basis is the explicit instruction for this task, not the enum alone, a repository-wide policy change, an exemption granted by the agent, or an observed maintainer sign-off. No external approval comment is claimed.
The one retained generic row is the a2–ARNO/PSCD2 association from PMID:16415858. The existing reason and source-access record distinguish the a2–ARNO association from Arf6 binding to subunit c. The normal cache remains abstract-only; the earlier selected primary-paper consultation did not resolve the recombinant a2-tail donor species. The open question about a more specific molecular role is retained, not treated as proof of a GEF, scaffold or adaptor activity. Thus this is an intentional retention under the task instruction despite the repository's informational-exclusion criterion, not an assertion that the criterion requires biological falsity.
All 31 review objects (30 original source assertions and one existing NEW PI4P-binding proposal), both core functions, all 20 existing evidence excerpts and the prior source records remain unchanged. The unavailable PMID:17295899 remains outside the review YAML; its prior failed retrieval is not represented as successful access.
This addendum clarifies the authority and exclusion criterion for the existing decisions. It adds no primary-source reading, biological assertion or new annotation adjudication. Earlier journal entries and validation results remain historical; this clarification does not claim reviewer approval or PR completion.