ILT1 (YDR090C / Q03193) — curation notes

Journal-style notes for the AI GO-annotation review of the S. cerevisiae dark gene ILT1.
Standard name ILT1 = "Ionic Liquid Tolerance 1"; systematic name YDR090C; UniProt Q03193;
SGD S000002497. SGD qualifier: Verified ORF.

Identity / provenance

• ILT1 is the SGD standard name assigned in 2018 by Higgins et al.
  PMID:30045857.
• SGD one-line description: "Protein of unknown function; deletion confers sensitivity to cationic
  compounds; green fluorescent protein (GFP)-fusion protein localizes to the plasma membrane."
• UniProt name: "Uncharacterized membrane protein YDR090C". Evidence level PE 1 (protein level).

Sequence / domain reasoning (inline, from UniProt Q03193 record)

• 310 aa multi-pass membrane protein. UniProt models 6 TRANSMEM helices (topology from
  PubMed:12524434 / PubMed:16847258 C-terminal reporter-fusion topology studies); Higgins 2018
  cite a 7-transmembrane-helix prediction PMID:30045857. Either way it is a polytopic ~6–7 TM integral membrane protein.
• PQ-loop repeats: UniProt annotates two PQ-loop domains (residues 5..69 and 138..194). Pfam
  PF04193 (PQ-loop), InterPro IPR006603 (PQ-loop_rpt), SMART SM00679 (CTNS, ×2). The PQ-loop /
  "MtN3/saliva" fold is a duplicated 3-TM unit typical of SWEET-like and PQ-loop transporters/
  sensors (e.g. cystinosin/CTNS, PQLC2, SWEETs, lysosomal amino-acid transporters).
• InterPro family IPR051415 = LAAT-1 (Lysosomal/Vacuolar Amino Acid Transporter 1).
• PANTHER: family PTHR16201 ("SEVEN TRANSMEMBRANE PROTEIN 1-RELATED" / LAAT-1);
  subfamily PTHR16201:SF37 = "PQ-LOOP REPEAT-CONTAINING PROTEIN".

Family / orthology reasoning — the key nuance for the IBA annotations

The GOA file carries four IBA (GO_Central / PAINT) annotations propagated from PTHR16201:
- GO:0005774 vacuolar membrane (C)
- GO:0015174 basic amino acid transmembrane transporter activity (F)
- GO:0034488 basic amino acid transmembrane export from vacuole (P)
- GO:0080144 intracellular amino acid homeostasis (P)

Inspecting the PANTHER PAINT table (interpro/panther/PTHR16201/PTHR16201-paint.tsv), these IBD
seeds are the characterized members of other subfamilies, not ILT1's own clade:
- SGD:S000005452 = YPQ1 (YOL092W) — vacuolar membrane lysine exporter
- SGD:S000002760 = YPQ2 (YDR352W) — vacuolar membrane arginine exporter
- SGD:S000000351 = RTC2/YPQ3 (YBR147W) — putative vacuolar cationic-amino-acid transporter
- WB:WBGene00021546 = C. elegans LAAT-1 (lysosomal cationic amino acid transporter)
- UniProtKB:Q6ZP29 = human PQLC2/LAAT1 (lysosomal cationic amino acid transporter)
- PomBase:SPAC17C9.10 (S. pombe)

ILT1 is NOT a PAINT seed (it is uncharacterized). It sits in subfamily SF37, whose only two
members are ILT1 (Q03193) and the S. pombe SPAC2E12.03c (Q10227) — both uncharacterized
(PTHR16201-entries.csv). So the specific "vacuolar basic amino acid export" function attached to
ILT1 by IBA is inherited from the distantly related YPQ/PQLC2 subfamilies, i.e. it is a
family-level (paralog) inference, not evidence that ILT1 itself does this.

Two concrete conflicts with the family-level IBA inference:
1. Location. ILT1 is experimentally at the plasma membrane, not the vacuolar membrane:
- Huh et al. GFP localization (IDA:SGD; UniProt SUBCELLULAR LOCATION "Cell membrane
{ECO:0000269|PubMed:14562095}").
- Higgins et al. Ilt1-GFP PMID:30045857. So the IBA GO:0005774 vacuolar membrane term conflicts
with two independent experimental plasma-membrane localizations.
2. Process. The only characterized phenotype is cationic-toxin tolerance at the cell
surface, not intracellular/vacuolar amino-acid homeostasis.

Therefore the IBA amino-acid-transport terms are best treated as over-annotations / uncertain
paralog inferences for ILT1: the fold (PQ-loop) is consistent with a small-molecule
transporter/sensor, but the substrate (basic amino acids) and site (vacuole) are not
supported for ILT1 and are in fact contradicted (location).

Experimental knowledge (KNOWN) — all from Higgins et al. 2018 (PMID:30045857, full text)

• Deletion phenotype: ydr090cΔ shows reduced growth at a subtoxic 31 mM [C2C1im]Cl (imidazolium
  ionic liquid) PMID:30045857. Basis for the standard-name
  assignment ILT1.
• Complementation / gain of function: plasmid ILT1 rescues the ilt1Δ growth defect in
  [C2C1im]Cl in the BY (S288c-derived) background PMID:30045857; extra copies increase tolerance.
• Broader cationic-toxin tolerance: ilt1Δ is sensitive to the cationic dye crystal violet
  (CV) and to [C4C1im]Cl PMID:30045857.
• Localization: Ilt1-GFP at the plasma membrane PMID:30045857.
• Independent of SGE1: ILT1 acts independently of the SGE1 efflux pump PMID:30045857.
• Strain-specific effect: no ILT1 requirement in the 378 background (which carries the tolerant
  SGE1PLL allele) PMID:30045857.
• PTMs (UniProt large-scale): Phospho-Ser229 (PubMed:17330950, 19779198); predicted N-glyco at
  N251, N259.

NOT known / knowledge gaps

• Molecular function is unknown. No transport activity, substrate, or mechanism has been
  measured for Ilt1. It is plausibly a PQ-loop transporter/permease or a transport-associated /
  regulatory membrane protein, but there is no direct biochemical assay. The IBA MF term (basic
  amino acid transporter) is a family inference from a different subfamily and is not supported.
• The relevant substrate/transported species (if any) is unknown — whether Ilt1 exports the
  ionic-liquid cation, imports something protective, or acts indirectly (e.g. modulating another
  transporter or membrane property) has not been determined.
• The direct biological role beyond cationic-toxin tolerance is unknown. Ionic liquids are
  synthetic xenobiotics; the native physiological function of ILT1 (its selective advantage in
  the wild) is not established. No growth phenotype other than cationic-compound sensitivity is
  documented.
• No physical interactors / pathway are functionally established (BioGRID lists high-throughput
  interactions only).

Curation plan (summary)

• Location: keep plasma membrane (IDA, PMID:30045857) as core CC; keep generic "membrane" IEA.
• The IBA vacuolar membrane (GO:0005774) location conflicts with two experimental PM
  localizations → MARK_AS_OVER_ANNOTATED / REMOVE-leaning; treat as non-core paralog inference.
• IBA MF (basic amino acid transporter, GO:0015174) and IBA BP (vacuolar basic amino acid export
  GO:0034488; intracellular amino acid homeostasis GO:0080144) are unsupported paralog inferences
  for ILT1's subfamily → MARK_AS_OVER_ANNOTATED (uncertain; do not assert; document in
  knowledge_gaps).
• ND MF/BP (GO:0003674 / GO:0008150): experimental data now support at least a BP
  (cationic-compound / xenobiotic tolerance) → the BP-ND is superseded; MF remains genuinely
  unknown.
• core_functions: minimal — CC plasma membrane; BP response to xenobiotic/toxic cationic compound
  (defensible from deletion phenotype). MF left as a knowledge gap.

Falcon deep-research cross-check (2026-07-05)

Ran just deep-research-falcon yeast ILT1 (Edison; 24.7 min). The report
(ILT1-deep-research-falcon.md) independently reaches the same core conclusion: YDR090C/ILT1 is a
PQ-loop family member that is distinct from the characterized Ypq1-3 vacuolar transporters, and
any transport/substrate assignment is "inference rather than direct evidence". It adds useful
family context (3+1+3 alternating-access fold from cystinosin structures; PQLC2/LAAT-1 lysosomal
cationic-amino-acid exporters; Ypq2 nitrogen-regulated vacuolar arginine export) — all consistent
with my paralog-over-annotation reasoning for the IBA terms. Two caveats: (1) the report leans
toward a "vacuolar transporter" prediction based purely on homology, which conflicts with the two
experimental plasma-membrane localizations for Ilt1 — I weight the direct evidence higher; (2) the
report did NOT surface the Higgins et al. 2018 primary experimental paper and instead mentions an
unverifiable "Reed et al. 2019 Yarrowia lipolytica" study for the ILT1 name — likely a spurious
citation. I have NOT used that claim; the review's functional/phenotypic claims are all anchored to
the verified Higgins et al. 2018 full text (PMID:30045857).

Full-gene rereview, 2026-09-20

All ten input rows were assessed; nine GOA source rows and all source fields are preserved. The four disputed IBAs (vacuolar membrane, basic-amino-acid export from vacuole, intracellular amino-acid homeostasis and basic-amino-acid transport activity) are restored to ACCEPT, with explicit inherited scope. Broad membrane becomes ACCEPT because being less specific than plasma membrane does not make it non-core. The original ND placeholders are preserved as provenance of absent direct experimental assignments, not negation of independently inferred functions.

The actual PTHR16201 tree identifies Q03193/YDR090C as leaf PTN000415976, in SF37 and below the four-term ancestor PTN001044753. Different characterized subfamilies do not invalidate this ancestral inference. No loss assertion occurs on the inspected target lineage. Exact tree path, response hash, accession mapping and current target/ancestral GAF rows are preserved in ILT1-paint-lineage.json and explained in the lineage note.

The dedicated primary study PMID:30045857(https://doi.org/10.1534/genetics.118.301161) was read in full. Chromosomal GFP fusions retained the tested growth phenotype and localize to the plasma membrane in YPD. That establishes a plasma-membrane pool, without testing or excluding a separate vacuolar pool under other conditions. Deletion, complementation and gene-dosage experiments establish strain-dependent cationic-toxin tolerance; the authors explicitly leave direct export by Ilt1 unresolved. A different studied role and absent transport assay therefore do not refute inherited amino-acid transport/homeostasis. No new OpenScientist request is needed without a concrete conflicting target datum.

The authored NEW response-to-xenobiotic-stimulus row was withdrawn under the user's participation requirement: necessity, rescue and dosage do not identify the step Ilt1 performs in the response. Its evidence remains in the description and questions; the inferred amino-acid homeostasis role provides an independently supported core. No matching GO-CAM target entry was found. The direct tolerance mechanism and relation to native amino-acid transport remain experimental questions.

The complete Falcon report was incorporated critically. Its comparative PQ-loop/cystinosin/Ypq/PQLC2 architecture and transport synthesis is useful, but it missed the 2018 ILT1 paper, claimed no target localization existed, and offered an inaccurate naming account. Primary target microscopy corrects that omission; it does not itself overturn inherited vacuolar functions. Exact cache searches found no OpenScientist report for ILT1/YDR090C/Q03193. Description, core and gaps distinguish the directly observed PM pool from inferred vacuolar transport, while keeping the exact target substrate and tolerance mechanism open.