The completed Falcon report identifies VAPA as an ER receptor/tether rather than a lipid carrier. Independent primary inspection confirms this: PMID:33124732(https://pubmed.ncbi.nlm.nih.gov/33124732/) supplies structures, FFAT peptide-binding experiments and contact-site reconstitution; PMID:24209621(https://pubmed.ncbi.nlm.nih.gov/24209621/) distinguishes the VAPA anchor from the lipid-transfer domain in OSBP. Lipid-transport BP annotations can therefore be mechanistically appropriate without claiming intrinsic lipid-transporter MF.
PMID:18713837(https://pubmed.ncbi.nlm.nih.gov/18713837/) tests VAPA overexpression, ER cargo transport and microtubule association, including rescue by FFAT peptide. PMID:10523508(https://pubmed.ncbi.nlm.nih.gov/10523508/) supports peripheral VAP-33/occludin localization. These sources support the core and several ancillary annotations. Protein-folding, early membrane-fusion, NF-kappaB-screen and opposing viral-direction annotations require more detailed source-specific resolution and remain UNDECIDED. IPI interaction rows are retained as non-core observations rather than being converted automatically into core FFAT interactions.
The human/horse alignment preserves the MSP domain and membrane anchor and identifies a horse linker insertion corresponding to the human alternative-splicing region. This explains why the shared MSP-domain name cannot transfer the nematode sperm-filament mechanism. See the horse prediction review for the exact original text and assessment.
Context: ER_PM_TETHERING_OBSOLETION project (GO:0061817 ER-PM tethering obsoleted; MF replacement GO:0160214). VAPA carries no GO:0061817 annotation in GOA, so nothing needs repair.
goId=GO:0160214&goUsage=exact, IEA rows filtered out; re-run 2026-09-27, 1003 total hits, 12 non-IEA). The 12 rows:No human ORP (OSBPL3/5/8) or ESYT has it experimentally. So GO curators have applied it to VAP orthologs in fission yeast, which conflicts with a strict reading of the "via membrane lipid binding" clause.
- GO-CAM check (gocams/):
- gocams/682fbcd000003765 ("Regulation of focal adhesion assembly: OSBPL3 transports PI4P and PC between PM and ER (Human)", production) is curated from PMID:32234213 itself (ECO:0000314). It types VAPA (Q9P0L0) as GO:0043495 protein-membrane adaptor activity occurring in GO:0005789 ER membrane. GO:0140268 ER-PM contact site appears in the model on partner activity nodes (IQSEC1, OSBPL3), not on VAPA. A curator who modelled this exact contact from this exact paper reached the same conclusion as this review. This is the strongest support for not adding GO:0160214.
- gocams/68fac5ed00002250 ("Phosphatidylinositol and phosphatidic acid transport by PITPNM1 in ER-PM MCS", production, UniProt-contributed) has human VAPB (O95292) as a GO:0160214 activity node, with ISS evidence (ECO:0000250, GO_REF:0000024, with MGI:1928744) on its enabled_by and occurs_in GO:0140268 associations. So the VAPB precedent is a GO-CAM activity node as well as a GAF row. This sharpens the question, since VAPA and VAPB are the same kind of FFAT receptor.
- CC asymmetry in VAPA GOA: VAPA is_active_in GO:0140284 ER-endosome membrane contact site (IDA PMID:41741634) and GO:0160258 ER-trans-Golgi network membrane contact site (IDA PMID:34688657, PMID:39106189), but has no GO:0140268 ER-PM contact site row. The MF annotations have the same shape: VAPA has GO:0170016 ER-endosome tether activity (IDA), whose definition allows bridging "by interacting with an endosome protein", and no ER-PM term, whose definition requires lipid binding. GOA's GO:0043495 IDA row for VAPA also cites PMID:32234213, the same paper used here.
- Decision: do not add GO:0160214 as NEW for VAPA. Its mechanism (FFAT receptor on the ER side) is already captured by the core GO:0043495 protein-membrane adaptor activity; the term definition names PM lipid binding that VAPA does not perform; and the PomBase/VAPB precedent is recorded as a suggested question rather than propagated by us.
Scope: full re-review of all 182 existing annotations, description and core functions for Q9P0L0. ProtNLM2 output was not consulted.
Identity: Q9P0L0 is the reviewed canonical human VAPA entry (249 aa, isoform 1). It has the MSP domain (Pfam PF00635), the coiled coil and the C-terminal tail anchor. It is not a fragment or alternative gene model. Isoform 2 (VSP_038648) is a linker-region splice variant; no annotation in GOA is isoform-specific.
Evolutionary grounding: PANTHER PTHR10809:SF155 (VESICLE-ASSOCIATED MEMBRANE PROTEIN-ASSOCIATED PROTEIN A), together with mouse Vapa Q9WV55, rat Vapa Q9Z270, bovine and orangutan VAPA. The paralog VAPB/VAPC (O95292) is in SF12. The family review interpro/panther/PTHR10809/PTHR10809-review.yaml places VAPA in SF155 as well, so there is no disagreement. The IBA rows (GO:0043495, ER membrane, plasma membrane, neuron projection development) come from the PAINT family node. GO:0043495 and ER membrane agree with direct human evidence. The plasma-membrane and neurite rows are kept as non-core.
Main changes and reasons:
- GO:0005515 protein binding (100 rows) migrated to the generic protein-binding policy. 11 rows whose source shows FFAT-dependent binding by the partner are MODIFY to GO:0033149 FFAT motif binding: CERT PMID:16895911, protrudin PMID:19289470, STARD3 PMID:24105263, OSBP/ORP2 [PMID:24209621, PMID:25681634, PMID:20178991], VPS13A PMID:30741634, WDR44 PMID:32344433 and phospho-FFAT partners PMID:33124732. The other 89 rows are REMOVE: interactome screens, viral proteins, VAP family oligomers (already covered by GO:0042803/GO:0046982), and non-FFAT partners. Removing them does not mean the interactions are false.
- The three NOT rows from PMID:18713837 (ER organization, protein folding in ER, ER exit site) are now ACCEPT with summaries that describe the negation. Before, the text read as if it described a positive role.
- COPII-coated vesicle budding is now MARK_AS_OVER_ANNOTATED. The evidence is that VAPA overexpression inhibits budding, which is dominant interference and not participation PMID:18713837. ER-to-Golgi vesicle transport is now KEEP_AS_NON_CORE.
- Host-mediated suppression of viral genome replication is REMOVE. The cited abstract assigns suppression to VAP-C, a VAPB splice product, and says VAPA overexpression enhanced HCV replication PMID:19515777. The activation term is kept as non-core. Viral release remains UNDECIDED because there is no release assay in the cached text.
- Membrane fusion (TAS, PMID:9657962) is MARK_AS_OVER_ANNOTATED. It was speculation based on VAMP binding ("suggesting possible roles in vesicle fusion"). VAPA is a tether, not a fusion factor.
- NF-kappaB (HMP, PMID:12761501) is MARK_AS_OVER_ANNOTATED. It is an overexpression reporter screen hit.
- Ceramide metabolic process, sphingomyelin biosynthesis, cholesterol homeostasis and regulation of cholesterol transport are downgraded from ACCEPT to KEEP_AS_NON_CORE. VAPA is the ER-side FFAT receptor and catalyses none of these steps. The contact-site lipid transport terms (ceramide, sterol, cholesterol and phospholipid transport) remain ACCEPT under the scaffold/tether case of the participation test, because the transfer depends on the positioning that VAPA provides.
- Core functions: added a third core function, ER-endosome tether activity (GO:0170016) at ER-endosome contact sites PMID:24105263, and directly_involved_in processes for the adaptor function. The description now says that VAPA has no intrinsic lipid-transfer activity.
Relative for follow-up: human VAPB (O95292), the nearest well-characterised paralog (PTHR10809:SF12). Mouse Vapa (SF155 ortholog) has only 3 experimental annotations, so it adds little beyond human VAPA. No VAPB review exists in the repo.
Modules: none in modules/ reference VAPA (only deep-research artifacts mention it). GO-CAMs: gocams/682fbcd000003765 (ORP3/IQSEC1 focal adhesion; VAPA as GO:0043495) and gocams/682fbcd000003781 (OSBPL1A cholesterol transport).