AVT2 (YEL064C / P39981) review notes

Journal of the AI GO-annotation review for S. cerevisiae AVT2. Provenance is recorded inline
as [PMID:xxxxx "verbatim text"] or [SGD/UniProt ...].

Identity (verified)

• UniProt: P39981 = AVT2_YEAST; systematic name YEL064C; SGD:S000000790.
  480 aa, 53.3 kDa. [UniProt P39981 record; SGD locus S000000790]
• Standard name AVT2 = "Amino acid Vacuolar Transport 2". Protein name in UniProt:
  "Vacuolar amino acid transporter 2".
• Family (multiple concordant sources):
• UniProt SIMILARITY: "Belongs to the amino acid/polyamine transporter 2 family." (ECO:0000305)
• TCDB 2.A.18.6.20 = the Amino Acid/Auxin Permease (AAAP) family.
• Pfam PF01490 (Aa_trans); InterPro IPR013057 (Amino acid transporter, transmembrane domain).
• PANTHER PTHR22950 "AMINO ACID TRANSPORTER", subfamily PTHR22950:SF458
  "SODIUM-COUPLED NEUTRAL AMINO ACID TRANSPORTER 11-RELATED" (i.e. SLC38/SNAT-like clan).
• eggNOG KOG1305; OrthoDB 28208at2759.
  This is the SLC32/SLC36/SLC38 ("amino acid/auxin permease", AAAP) superfamily — the same
  clan Russnak et al. described as "related to the neuronal GABA-glycine vesicular transporters"
  (the vesicular inhibitory amino acid transporter VGAT/SLC32 is in this clan).
• 9 predicted TM helices (UniProt FT TRANSMEM, ECO:0000255): 72-92, 95-115, 145-165,
  214-234, 263-283, 297-317, 338-358, 394-414, 447-467. N-terminal ~48 aa disordered/cytoplasmic
  (REGION 21-48 Disordered, polar-residue biased). So a bona fide polytopic integral membrane
  transporter fold.

NOTE on the "PQ-loop" descriptor in the task brief: the actual UniProt/InterPro/Pfam evidence
assigns AVT2 to the Aa_trans (PF01490 / AAAP / SLC38-SNAT-like) clan, NOT the PQ-loop
(SLC36-PAT/LamB / cystinosin, PF04193) family. I ground the review on the AAAP/Aa_trans
assignment that is actually in the record.

The AVT family (AVT1-7) — what is known, and what is AVT2-specific

The founding paper is Russnak, Konczal & McIntire 2001, J Biol Chem
PMID:11274162.
full_text_available: false in our cache — abstract only. Abstract states:

• "Seven genes in Saccharomyces cerevisiae are predicted to code for membrane-spanning proteins
  (designated AVT1-7) that are related to the neuronal gamma-aminobutyric acid-glycine vesicular
  transporters." PMID:11274162
• "We have now demonstrated that four of these proteins mediate amino acid transport in
  vacuoles." PMID:11274162
• The four with demonstrated activity are named explicitly:
• AVT1 — vacuolar uptake of large neutral amino acids (Tyr, Gln, Asn, Ile, Leu);
  ATP-dependent, abolished by nigericin (pH-gradient driven). PMID:11274162
• AVT3, AVT4 — efflux of Tyr/large neutral amino acids from the vacuole (system-h-like).
  PMID:11274162
• AVT6 — efflux of aspartate and glutamate; ATP-dependent, nigericin-sensitive.
  PMID:11274162
• AVT2, AVT5, AVT7 are NOT among the four with demonstrated transport in that paper — i.e.
  no substrate or direction was demonstrated for AVT2. This is the AVT2-specific gap.

Secondary literature confirms AVT2 remained functionally uncharacterized:
- Regulation-of-amino-acid-transport review (Bianchi/Van Belle/... 2019, MMBR) and other
reviews describe AVT2/AVT5 as members with no observed transport activity. (WebSearch of
the MMBR review text: "No activity has been observed for two of the predicted family members,
Avt2 and Avt5.") — I did not cite this verbatim in the YAML because I could not fetch the full
MMBR text into the publications cache; recorded here as background only.

GO annotations in GOA (9 total; from AVT2-goa.tsv / QuickGO)

MF:
- GO:0003674 molecular_function — ND, GO_REF:0000015 (SGD root placeholder)
- GO:0015179 L-amino acid transmembrane transporter activity — IBA, GO_REF:0000033
BP:
- GO:0003333 amino acid transmembrane transport — IBA, GO_REF:0000033
- GO:1902475 L-alpha-amino acid transmembrane transport — IEA, GO_REF:0000108 (inferred from
the GO:0015179 MF via inter-ontology "logical inference" link)
- GO:0008150 biological_process — ND, GO_REF:0000015 (SGD root placeholder)
CC:
- GO:0005783 endoplasmic reticulum — IDA, PMID:11274162 (SGD; direct assay)
- GO:0005783 endoplasmic reticulum — IBA, GO_REF:0000033 (GO_Central phylogenetic)
- GO:0016020 membrane — IBA, GO_REF:0000033
- GO:0005774 vacuolar membrane — IEA, GO_REF:0000044 (UniProtKB SubCell SL-0271)

Localization: ER (IDA/SGD) vs vacuole membrane (UniProt SubCell / family expectation)

There is a genuine, notable discrepancy:
- SGD IDA (GO:0005783 ER) from PMID:11274162 — direct-assay localization to the ER.
The abstract is about vacuolar transport of the family, but the SGD curator read the full
text and assigned AVT2 specifically to the ER. Per project rules I do NOT overrule an
experimental (IDA) curator call from the abstract-only cache → ACCEPT.
- UniProt SubCell (GO:0005774 vacuole membrane, IEA) — SL-0271, cites the same PMID:11274162
as the source of "Vacuole membrane; Multi-pass membrane protein". So UniProt and SGD read the
same paper differently (family-expected vacuole vs. observed ER). AVT family members are
canonically tonoplast/vacuolar; AVT2 may localize to the ER (possibly ER-retained /
not trafficked to the vacuole), which is itself a hint that AVT2 is atypical. This is a real
knowledge gap (where does AVT2 actually act?).

Domain-based reasoning for the molecular function

• The Aa_trans (PF01490 / AAAP / SLC38-SNAT-like) fold with 9 TM helices strongly supports a
  generic amino acid transmembrane transporter activity — this is domain-defensible.
• BUT no substrate and no direction (uptake vs efflux, and driving force) has been demonstrated
  for AVT2 specifically. Its paralogs split into uptake (AVT1) and efflux (AVT3/4/6) with
  different substrate classes, so paralogy does not pin AVT2's substrate/direction.
• Therefore: keep MF at the generic amino-acid-transporter level (GO:0015179 L-amino acid
  transmembrane transporter activity is a reasonable IBA-supported generic MF; a specific
  substrate/direction would be over-annotation).

Review decisions (summary)

• GO:0015179 L-amino acid transmembrane transporter activity (IBA) — ACCEPT as the
  defensible generic MF (family fold; substrate unknown → do not specialize).
• GO:0003333 amino acid transmembrane transport (IBA) — ACCEPT generic BP.
• GO:1902475 L-alpha-amino acid transmembrane transport (IEA, inter-ontology) — ACCEPT
  (consistent, generic; auto-derived from the MF).
• GO:0005783 ER (IDA, PMID:11274162) — ACCEPT (direct experimental localization; core CC).
• GO:0005783 ER (IBA) — KEEP_AS_NON_CORE (phylogenetic duplicate of the IDA; the IDA is the
  authoritative one).
• GO:0016020 membrane (IBA) — ACCEPT (correct, generic; it is an integral membrane protein).
• GO:0005774 vacuolar membrane (IEA, SubCell) — KEEP_AS_NON_CORE (family-expected/UniProt
  location; conflicts with the ER IDA, so retained as plausible-but-not-core rather than removed —
  cannot overrule either the SubCell mapping or note the unresolved ER-vs-vacuole question).
• GO:0003674 molecular_function (ND) — REMOVE (root placeholder, now superseded by GO:0015179).
• GO:0008150 biological_process (ND) — REMOVE (root placeholder, superseded by GO:0003333).

Knowledge gaps (primary deliverable — this is a dark gene)

1. Transported substrate + direction unknown for AVT2. Family fold ⇒ amino acid transporter,
   but no substrate/direction/driving-force demonstrated (unlike AVT1 uptake; AVT3/4/6 efflux).
2. In-vivo localization unresolved: IDA places it in the ER, but UniProt/family expectation is
   the vacuole membrane — is AVT2 an ER-resident, an ER-retained mis-folded/unassembled member, or
   a vacuolar transporter caught in transit?
3. Physiological role / redundancy with the AVT family: no dedicated phenotype; deletion is
   viable. Is AVT2 redundant with AVT1/3/4/5/6/7 or specialized/silent?
4. No demonstrated activity in the one biochemical survey that assayed the whole family
   (only 4/7 members scored positive; AVT2 was among the 3 negatives).