WIPI2 (Q9Y4P8) in Pexophagy, Nucleophagy and Glycophagy — Focused Curation Report

Gene: WIPI2 (human, NCBITaxon:9606) · UniProt: Q9Y4P8 · PANTHER family: PTHR11227 (WIPI/Atg18 PROPPIN family), subfamily SF27
Focus: function_assignment · Hypothesis slug: selective-pexophagy-nucleophagy-and-glycophagy
Terms under review: GO:0000425 pexophagy · GO:0044804 nucleophagy · GO:0061723 glycophagy (all currently IBA)

Executive Judgment

Verdict: Partially supported — retain as non-core, shared-machinery participation; do not treat as WIPI2-specific selective functions.

The seed hypothesis is essentially correct on mechanism. Human WIPI2's only experimentally supported activity is a generic core step of autophagosome biogenesis: it senses PI3P and recruits the ATG12–ATG5–ATG16L1 (E3) complex, allosterically activating LC3/GABARAP lipidation (PMID:24954904, PMID:32437499, PMID:20505359). Pexophagy, glycophagy (both defined by GO as "by macroautophagy") and mammalian macro-nucleophagy all build canonical LC3/GABARAP double-membrane autophagosomes and therefore mechanistically require this WIPI2 step. In that sense WIPI2 "participates" — exactly as it demonstrably does in Salmonella xenophagy, where WIPI2b coats the bacterial membrane and recruits ATG12-5-16L1 (PMID:24954904).

However, three important caveats keep this at partially supported / non-core rather than "supported":

1. No target-specific experiment exists for human WIPI2 in any of the three pathways. All three annotations are phylogenetic inferences (IBA, ECO:0000318, GO_Central, GO_REF:0000033), not IDA/IMP on WIPI2.
2. WIPI2 is not the cargo/selectivity receptor for any of the three: pexophagy uses NBR1/p62 + ubiquitinated PEX5 (+PEX2 E3); mammalian nucleophagy uses a direct LC3–LMNB1 interaction; glycophagy uses STBD1 (CBM20+LIR→GABARAPL1/RB1CC1). The seed's own framing — "may contribute common phagophore/ATG16L1 adaptor machinery without being the selective cargo receptor" — is confirmed.
3. The IBA seeds are loss-of-function "general requirement" experiments on the yeast/fly ortholog acting as core machinery, and the same blanket annotation was propagated identically to all four human paralogs (WIPI1/2/3/4), which have functionally diverged (ATG16L1 arm vs ATG2 arm). IBA cannot resolve this.

Bottom line for the curator: The activity is real but generic. These BP terms are defensible as "required-for/part-of" shared machinery, but they are not evidence that WIPI2 has a peroxisome-, nucleus-, or glycogen-specific function. If the review's convention is to record only pathway-specific or core functions, these three should be flagged non-core / generalizable to macroautophagy (GO:0016236) / autophagosome assembly (GO:0000045), which WIPI2 does hold by direct experiment.

Evidence Matrix

GO Curation Implications (leads — require curator verification)

Recommended framing: WIPI2's core annotations are the PI3P-binding/ATG16L1-adaptor MF and macroautophagy/autophagosome-assembly BP. The three selective-autophagy BP terms are downstream, pathway-context consequences of that same single molecular step, delivered via IBA; they should be labeled non-core / shared-machinery, not removed as wrong (each has a valid within-family experimental seed), and certainly not read as cargo-recognition functions.

Mechanistic Scope

• Immediate molecular function (direct): PI3P-dependent β-propeller effector that recruits/allosterically activates the ATG12–ATG5–ATG16L1 E3 to lipidate LC3/GABARAP onto the phagophore (PMID:24954904, 32437499). Location: phagophore assembly site / omegasome / autophagosome membrane (IDA).
• Downstream / pathway-context (inferred): Because pexophagy and glycophagy are macroautophagy (GO definitions explicitly state "by macroautophagy") and mammalian nucleophagy of lamin B1 is autophagosome-mediated, loss of WIPI2 would impair all of them — but this is a general requirement, identical to losing any core Atg factor, not a peroxisome/nucleus/glycogen-specific activity.
• Not attributable to WIPI2: cargo selection (NBR1/p62/ubiquitin-PEX5; LC3–LMNB1; STBD1–glycogen). These provide the selectivity; WIPI2 provides the shared build step.

Conflicts and Alternatives

• Paralog over-annotation: All four human WIPI paralogs carry identical IBA for the four selective terms. WIPI1/WIPI2 operate in the ATG16L1-recruiting arm; WIPI3(WDR45B)/WIPI4(WDR45) operate with ATG2 in lipid transfer. Blanket IBA cannot distinguish these — a database carry-over artifact, though not necessarily biologically false (all are required for autophagosome formation).
• Organism-specific seeds: The pexophagy/nucleophagy seeds are yeast (piecemeal microautophagy for PMN); glycophagy seed is Drosophila. Extrapolation to human WIPI2 is inference, not observation.
• Competing selectivity mechanisms: For each pathway, the literature attributes selectivity to dedicated receptors, not WIPI2 — so any curation implying WIPI2 confers selective recognition would be an over-statement.
• "General requirement vs direct shared step": The seed asks to distinguish these. The available evidence is loss-of-function requirement (IMP) plus a mechanistic direct step (PI3P→ATG16L1→LC3) that is generic; there is no experiment isolating a WIPI2 step within pexophagy/nucleophagy/glycophagy specifically.

Knowledge Gaps

1. No human WIPI2 loss-of-function in any of the three pathways. Checked QuickGO (0 experimental annotations). Matters because IBA alone is being asked to justify three BP terms. Resolve with WIPI2 KO/KD + peroxisome (PMP70/PEX14), lamin B1, or glycogen turnover readouts.
2. Paralog specificity. Checked annotation blanketing across WIPI1/2/3/4. Matters because curators may want paralog-resolved terms. Resolve with paralog-specific KO rescue in each selective assay.
3. Whether nucleophagy of lamin B1 requires WIPI2 at all vs. proceeding through a non-canonical/WIPI-independent route. Checked: Dou papers implicate LC3 directly but did not test WIPI2. Resolve with WIPI2 dependence test in RAS-induced senescence lamin B1 degradation.
4. Glycophagy dependence on WIPI2 specifically in mammals (Drosophila seed only). Resolve with STBD1-based glycophagy flux assay under WIPI2 depletion.

Discriminating Tests

1. WIPI2 depletion + selective-flux readouts: pexophagy (peroxisome number/PMP70, HaloTag-PTS1 processing), nucleophagy (lamin B1 degradation in oncogene-induced senescence, PMID:26524528 model), glycophagy (glycogen/PAS or 2-NBDG flux, STBD1 reporter). Expectation if "shared machinery only": all three blocked to the same degree as bulk autophagy, with no additional selectivity phenotype.
2. Separation-of-function ATG16L1-binding mutant of WIPI2 (WIPI2b residues that bind ATG16L1 vs PI3P): should block all three pathways coordinately, confirming the single shared step.
3. Paralog swap/rescue: test whether WIPI3/WIPI4 (ATG2 arm) rescue each selective pathway in WIPI2-null cells, to resolve paralog-specific claims.
4. Proximity/interaction proteomics of WIPI2 vs receptors (STBD1, NBR1, LMNB1): absence of direct cargo-receptor complexes would corroborate "no selectivity role."

Curation Leads (require curator verification)

• Candidate action: Keep GO:0000425 / GO:0044804 / GO:0061723 as IBA/non-core, shared-machinery annotations; add an annotation note that WIPI2 acts as generic autophagosome-biogenesis machinery, not a cargo receptor. Do not upgrade to experimental; do not remove outright (each IBA has a valid within-family experimental seed).
• Candidate core terms to foreground instead: MF GO:0032266 (PI3P binding), MF GO:0030674 (protein-macromolecule adaptor activity), BP GO:0000045 (autophagosome assembly), BP GO:0016236 (macroautophagy), CC GO:0000407 (phagophore assembly site) — all direct-evidence supported. Prefer these over "protein binding".
• Candidate references with exact snippets to verify:
• PMID:24954904 — "Atg16L1 directly binds WIPI2b … WIPI2b is a PtdIns(3)P effector upstream of Atg16L1 and is required for LC3 conjugation … WIPI2b is also required for autophagic clearance of pathogenic bacteria."
• PMID:32437499 — "LC3 lipidation is strictly dependent on the recruitment of WIPI2 that in turn depends on PI(3)P … WIPI2 allosterically activate[s] the E3 complex."
• PMID:24265594 — Drosophila core Atg genes (incl. Atg18a) required for autophagic glycogen breakdown (glycophagy IBA seed).
• PMID:26524528 — "The autophagy protein LC3/Atg8 … directly interacts with the nuclear lamina protein lamin B1" (nucleophagy selectivity is LC3–cargo, not WIPI2).
• PMID:39236246 — STBD1 CBM20/LIR mechanism of glycophagy selectivity.
• Suggested question for curator: Does this review's policy annotate core autophagy machinery to each selective-autophagy sub-process (IBA), or restrict such terms to factors with pathway-specific experimental evidence? The answer determines retain-as-non-core vs generalize.
• Suggested experiments: the four discriminating tests above.

Provenance

Machine-readable artifacts: artifacts/WIPI2_evidence_matrix.csv, artifacts/WIPI2_GO_decision_table.csv.

Computed with QuickGO REST (annotation + ontology endpoints) and PANTHER geneinfo, executed in-session (iterations 1–3):
- WIPI2 annotation pull (50 rows) → three target terms are IBA only.
- GO term definition pull → pexophagy/glycophagy "by macroautophagy"; nucleophagy broadly defined (not yeast-PMN-restricted).
- Paralog scan (WIPI1/2/3/4 + yeast Atg18/Atg21/Hsv2) → identical blanket IBA; yeast Atg18 IMP seeds pexophagy/nucleophagy.
- Cross-family experimental-source scan → glycophagy seeded by Drosophila Atg18a (PMID:24265594); human experimental annotations for all three terms are on dedicated receptors/regulators, never WIPI2.

Artifacts

• OpenScientist WIPI2 GO decision table
• OpenScientist WIPI2 evidence matrix
• OpenScientist final report
• OpenScientist final report