AIGR Deep Research Report — S. pombe nce101 (C6Y4B6): "non-classical protein export" function

Gene: nce101 / SPAC12G12.17 (SCHPO) · UniProt: C6Y4B6 (NCE1_SCHPO) · Family: NCE101 (Pfam PF11654 / InterPro IPR024242 / PANTHER PTHR28011)
Seed hypothesis: S. pombe nce101 directly participates in signal-sequence-independent non-classical protein export, rather than merely inheriting an ambiguous family-level annotation from the S. cerevisiae NCE101 screen.

Executive Judgment

Verdict: REFUTED / OVER-ANNOTATED (with respect to "direct participation").

The evidence points to the opposite of the seed hypothesis. Far from having direct, gene-product-specific evidence, S. pombe nce101 carries no experimental functional data of any kind, and every "non-classical export / exocytosis / protein secretion" annotation is electronically or phylogenetically transferred. This is precisely the "inherited ambiguous family-level annotation" scenario the hypothesis was framed to exclude.

Key pillars:
1. No direct evidence in fission yeast. UniProt C6Y4B6 is a 58-aa single-TM peptide at Protein Existence level 2 (transcript only). Its FUNCTION comment is 100% "By similarity" (ECO:0000250); GO terms are ISO (exocytosis), IBA (protein secretion), and IEA (membrane).
2. PomBase itself calls it uncharacterized — and formally annotates it "unknown." Characterisation status = "conserved unknown"; product name is the family label "non-classical export protein family Nce101." QuickGO confirms PomBase curates all three GO roots as No Data (ND, GO_REF:0000015): molecular_function (GO:0003674) = ND, biological_process (GO:0008150) = ND, cellular_component (GO:0005575) = ND. The only non-root functional terms are GO:0009306 "protein secretion" via IEA (InterPro IPR024242) + IBA (GO_Central, from the S. cerevisiae NCE101 node S000003742) and GO:0016020 membrane via IEA. No experimental (EXP/IDA/IMP/IGI) annotation exists. The only phenotypes are generic chemical-genomic drug/stress sensitivities.
3. The founder annotation is weak and ambiguous. S. cerevisiae NCE101 (Q02820, PE = 3 "inferred from homology") rests on a single 1996 galectin-1 over-expression screen (PMID 8655575) and is explicitly hedged: "may be part of the export machinery or may also be a substrate."
4. The "NCE" family is a phenotype label, not a molecular-function family. Sibling screen hits were reassigned to unrelated functions: NCE103 = β-carbonic anhydrase (an enzyme; PMID 25109265) and NCE102 = plasma-membrane sphingolipid sensor / eisosome-MCC microdomain protein (PMID 35758748) — neither a dedicated export-machinery component.

1. Both reference databases call it "unknown function." SGD's curated description of the founder (Verified ORF YJL205C) is "Protein of unknown function; … SWAT-GFP and mCherry fusion proteins localize to the cytosol" — and PomBase calls the S. pombe gene "conserved unknown." The experimental cytosolic localization further conflicts with the transferred "single-pass membrane protein" CC (a TMHMM prediction), so even the location term is uncertain.

Most important caveat: "Refuted" here means the direct-participation claim is unsupported and the annotation is over-strong; it does not prove nce101 is uninvolved in secretion. It is a genuine, fungi-conserved small peptide of currently unknown molecular function (Verified ORF, not dubious). Absence of evidence ≠ evidence of a different function. Note the location is itself ambiguous: predicted single-TM vs experimentally cytosolic tag fusions.

Evidence Matrix

GO Curation Implications (leads — require curator verification)

The three functional GO annotations on C6Y4B6 all trace, ultimately, to transferred/ambiguous evidence:

• No MF term is justified by current evidence beyond, at most, generic membrane localization. Do not assert an export-machinery molecular activity.
• Recommend the curated position mirror PomBase's "conserved unknown": a fungi-conserved small single-pass membrane protein of unknown molecular function, with any secretion/export terms explicitly marked as homology-transferred and non-core.
• Avoid defaulting to "protein binding" — no interaction evidence exists for the S. pombe product.

Mechanistic Scope

Immediate molecular function being tested: whether the 58-aa nce101 peptide is itself a component (or dedicated substrate) of a signal-sequence-independent export apparatus.

• Direct gene-product activity: unknown. The only hard biochemical/structural fact is a single predicted transmembrane helix (aa 10–27) → integral membrane peptide.
• Downstream / inferred: the "export" descriptor derives from an over-expression phenotype on a heterologous reporter (galectin-1) in a different species (S. cerevisiae), then propagated by orthology. This is a pathway-level phenotype, not a demonstrated molecular activity of the endogenous peptide.
• The S. pombe deletion phenotypes (drug/stress sensitivities) are pleiotropic and do not localize the protein to an export step.

Conflicts and Alternatives

• Family carry-over / mis-labeling (strongest alternative): NCE103 (carbonic anhydrase) and NCE102 (sphingolipid sensor) demonstrate that "NCE" screen membership routinely does not correspond to a direct export function. nce101 is the least-characterized member.
• "Substrate vs machinery" ambiguity: even the founder UniProt statement cannot decide whether the peptide is machinery or cargo — a fundamental unresolved distinction that undercuts any "directly participates in the export machinery" claim.
• Species differences: all S. pombe evidence is transferred; there is no fission-yeast export assay for nce101.
• Over-expression artifact: the founding phenotype is an over-expression/heterologous-reporter effect, a known source of gain-of-function artifacts that need not reflect endogenous function.

Knowledge Gaps

Discriminating Tests

1. Endogenous localization: C-terminal tag of S. pombe nce101 at native locus; ask PM/eisosome vs ER/Golgi vs cytosolic membranes.
2. Leaderless-secretion functional test: measure unconventional secretion of an established cargo (e.g., Acb1, or heterologous galectin-1/GAPDH) in wild-type vs Δnce101 and vs over-expression — replicating the founder assay in the actual organism.
3. Interactome: AP-MS / proximity labeling to test physical association with any secretion apparatus vs membrane-microdomain proteins (parallels the NCE102 outcome).
4. AlphaFold/structural check: compare the modeled fold/topology against NCE102 (tetraspanin-like) to assess whether nce101 shares a microdomain-protein architecture rather than a transporter fold.
5. Cross-species annotation audit: confirm no experimental anchor exists anywhere in PF11654 before propagating export terms.

Curation Leads (all require curator verification)

• Action change: Treat the "non-classical protein export / exocytosis" function as non-core, homology-transferred, and over-strong for C6Y4B6. Align the review with PomBase's "conserved unknown" characterisation.
• GO leads:
• Retain GO:0016020 membrane (CC, IEA) — supported by a predicted TM helix (aa 10–27).
• Flag/weaken GO:0006887 exocytosis (ISO) and GO:0009306 protein secretion (IBA) as non-experimental transfers; consider removing the more specific "exocytosis" term. Do not add an MF export/transporter term.
• Candidate references to cite for the review:
• PMID 8655575 (founder screen; snippet: "A screen for components of the export machinery has identified genes that are involved in nonclassical export.") — note it is a heterologous over-expression screen and the UniProt function is hedged "machinery or substrate."
• PMID 35758748 (NCE102 = sphingolipid sensor) and PMID 25109265 (NCE103 = carbonic anhydrase) — precedents that "NCE" ≠ molecular export function.
• Suggested curator questions:
• Is there any experimental (non-IBA/ISO/IEA) evidence for S. pombe nce101 function? (Current answer: no.)
• Should family-name-derived BP terms be demoted to "unknown molecular function" pending an organism-specific assay?
• Suggested experiment (highest value): Δnce101 leaderless-cargo secretion assay in S. pombe (Discriminating Test #2).

Provenance

All programmatic results above were retrieved live during this run via the UniProt REST API (C6Y4B6, Q02820, Q12207 NCE102, PF11654 family search), the PomBase gene API (SPAC12G12.17), and the SGD backend locus API (YJL205C), plus PubMed abstracts (PMIDs 8655575, 35758748, 25109265). No local repository bioinformatics files were used. No result was fabricated; where evidence is absent it is stated as absent.

Iteration 2 additions: SGD confirms YJL205C is a Verified (not dubious) ORF but "protein of unknown function" with cytosolic GFP/mCherry localization; NCE101 (PF11654) and NCE102 (PF01284, MARVEL) are confirmed unrelated families, ruling out paralog confusion.

Iteration 3 additions: QuickGO API confirms PomBase curates C6Y4B6's MF/BP/CC roots as ND (unknown; GO_REF:0000015) and that "protein secretion" is present only as IEA + IBA transfers (IBA source = S. cerevisiae NCE101 node S000003742) — the definitive database-level demonstration that the export annotation is inherited, not direct. Computed provenance saved as artifacts/nce101_evidence_matrix.csv and artifacts/nce101_go_decision_table.csv.

Artifacts

• OpenScientist nce101 evidence matrix
• OpenScientist nce101 go decision table
• OpenScientist final report
• OpenScientist final report