ATG14 notes

2026-06-02 review setup

• Started review from fetched UniProt/GOA-derived stub. ATG14 is a Beclin 1-associated autophagy factor and PI3KC3-C1 subunit; the UniProt record states that it is "required for both basal and inducible autophagy", determines localization of the autophagy-specific PI3-kinase complex PI3KC3-C1, enhances PIK3C3 activity in a BECN1-dependent manner, and promotes STX17/SNAP29/VAMP8-dependent autophagosome-endolysosome fusion [UniProt:Q6ZNE5].
• Key early complex evidence: human Atg14 and UVRAG form mutually exclusive Beclin 1/Vps34 complexes; Atg14 localizes to autophagic isolation membranes and Atg14 silencing suppresses autophagosome formation [PMID:18843052 "Atg14 is present on autophagic isolation membranes"; PMID:18843052 "Silencing of human Atg14 in HeLa cells suppresses autophagosome formation"].
• Barkor/Atg14L was identified as an autophagy-specific Beclin 1/class III PI3K factor; depletion reduces autophagy and autophagosome formation, whereas overexpression activates autophagy and increases autophagosomes [PMID:19050071 "Elimination of Barkor expression by RNA interference compromises starvation- and rapamycin-induced lipidation of LC3 and autophagosome formation"; PMID:19050071 "Overexpression of Barkor leads to autophagy activation"].
• Atg14L and UVRAG bind Beclin 1 mutually exclusively; Atg14L localizes at ER, isolation membranes, autophagosomes, and unidentified puncta, and Atg14L knockout caused defective autophagosome formation [PMID:19270696 "Atg14L and UVRAG bind to Beclin 1 in a mutually exclusive manner"; PMID:19270696 "Knockout of Atg14L in mouse ES cells caused a defect in autophagosome formation"].
• ER targeting evidence supports phagophore/ER-membrane localization: ATG14L cysteine repeats are required for ER localization and autophagosome formation, and the BATS domain is important for targeting the PI3K complex to the ER and autophagosome biogenesis PMID:20713597.
• The C-terminal BATS domain targets Barkor/Atg14L to PtdIns(3)P-rich autophagic membranes and is required for PI3KC3 recruitment and autophagy stimulation PMID:21518905.
• Structural and interaction studies place ATG14 as the autophagy-specific organizer of the Beclin 1-VPS34/PIK3C3 complex, supporting class III PI3K complex and protein-membrane adaptor/regulator annotations rather than generic protein binding [PMID:22314358 "Atg14L-containing complex is primarily involved in early stage autophagosome formation"; PMID:25490155 "Architecture and dynamics of the autophagic phosphatidylinositol 3-kinase complex"].
• ATG14 has a later autophagosome fusion role: it binds the STX17-SNAP29 t-SNARE complex on autophagosomes, primes it for VAMP8 interaction, and promotes autophagosome-endolysosome fusion [PMID:25686604 "ATG14 promotes membrane tethering and fusion of autophagosomes to endolysosomes"; UniProt:Q6ZNE5].
• BECN2 also interacts with ATG14 through a metastable coiled-coil to mediate autophagy, consistent with ATG14 acting as a PI3KC3-C1/BECN-family adaptor rather than only as a generic binding protein PMID:28218432.
• Falcon deep research requested for ATG14 as part of the full review process.
• Falcon deep research timed out after the configured 600 second limit and did not produce ATG14-deep-research-falcon.md; final review therefore relies on the cached UniProt, GOA, Reactome, and publication evidence above.

Description cleanup note

The YAML description field was revised to keep it as a standalone biological summary. Project-specific curation framing moved here instead.

• Moved out of the YAML description: the Proteostasis Network framing emphasized ATG14 as an autophagy-initiation/autophagosome-maturation component rather than a generic protein-binding or class I PI3K/AKT signaling factor.

Falcon deep research findings (2026-06-07)

The Falcon (Edison Scientific) report (ATG14-deep-research-falcon.md) was successfully generated (the prior 2026-06-02 timeout note is now superseded). It is consistent with the completed review and adds mostly mechanistic/regulatory detail rather than new core functions. Key points:

• NEW (regulation): ULK1 phosphorylates ATG14L at Ser29, and this phosphorylation is required for VPS34/PI3KC3-C1 activation and PI3P production at autophagosome formation sites; phospho-Ser29 is abolished in ULK1/2 double-knockout or kinase-dead ULK1 cells and restored by WT ULK1 under starvation [PMID:39169022 "the ULK1 palmitoylated enhances the phosphorylation of ATG14L, which is required for activating PI3-Kinase and producing phosphatidylinositol 3-phosphate"; Tabata et al. 2024 Nat Commun, doi:10.1038/s41467-024-51402-w]. This adds an upstream ULK1-kinase->ATG14 site-specific link not previously cited in the review, complementing existing ATG14-dependent BECN1-phosphorylation annotations. The existing review already supports positive/negative regulation of protein phosphorylation involving ATG14; this adds that ATG14 is itself a ULK1 substrate.
• NEW (upstream control, provisional/mechanistic): ULK1 palmitoylation by ZDHHC13 promotes ULK1 translocation to autophagosome formation sites and is required for ULK1->ATG14L Ser29 phosphorylation; 2-bromopalmitate suppresses starvation-induced ATG14 Ser29 phosphorylation [PMID:39169022, Tabata et al. 2024]. This is ULK1/ZDHHC13 biology upstream of ATG14, not a direct ATG14 molecular function - do NOT add as an ATG14 annotation.
• CONFIRMS (complex/architecture): PI3KC3-C1 is a ~360 kDa 1:1:1:1 heterotetramer (VPS34/PIK3C3 : VPS15/PIK3R4 : BECN1 : ATG14) with only ~11% catalytic mass, reinforcing ATG14's non-enzymatic regulatory/targeting role; mutually exclusive with UVRAG-containing PI3KC3-C2 [Chen & Hurley 2025 J Biol Chem, doi:10.1016/j.jbc.2025.110391, PMID:40543587]. Matches existing GO:0035014 / GO:0034271 annotations.
• CONFIRMS (domain/localization): C-terminal BATS domain (~last 80 aa, ~19-aa amphipathic helix) is a membrane-curvature sensor preferring highly curved PI3P/PI(4,5)P2-enriched isolation membranes; N-terminal cysteine-repeat mediates ER targeting; co-localizes with DFCP1/ATG16L1 at omegasomes [Liu et al. 2023 Cells, doi:10.3390/cells PMID:12081132 - review]. Already captured by existing BATS/omegasome/phagophore annotations.
• CONFIRMS (downstream function): ATG14-dependent PI3KC3-C1 activity supports WIPI2/ATG2A recruitment and LC3 lipidation, and complex II cannot functionally substitute [Brier et al. 2019, doi:10.1091/mbc.e18-11-0743, PMID:30540564 (not cached)]. Consistent with existing autophagosome-assembly core function.
• PROVISIONAL (disease, low-confidence): Open Targets lists association-level links (neurodegeneration, COVID-19, dengue, lysosomal storage disease); the report explicitly flags these as association/inference, not causal. No ATG14-targeted clinical trials identified. NOT used to change any annotation.
• Pharmacology context (provisional): VPS34 inhibitors and a reported Beclin1-ATG14L interaction inhibitor are discussed as autophagy-initiation tool compounds [Lee et al. 2024 Biomol Ther, doi:10.4062/biomolther.2024.094 - review]. Tool-compound context only.

Action taken: added the two most citable, ATG14-relevant primary/key sources (Tabata 2024 PMID:39169022; Chen & Hurley 2025 PMID:40543587) to the review references: as statement-only findings (full_text_unavailable), plus one suggested question and one suggested experiment about the ULK1->ATG14 Ser29 axis. Review-only sources (Liu 2023, Lee 2024) and the non-resolved/uncached Brier paper are kept in notes only. No existing annotation action was changed - all Falcon findings either confirm existing calls or are upstream/association-level context.

2026-09-20 full evidence re-review

Reviewed all 95 source rows and the existing Falcon report, including the PAINT IBDs at PTN001784663 (PTHR13664). Human target evidence in the WITH/FROM set is legitimate experimental grounding, not circular support.

• Full PMID:23878393 materially changes the earlier abstract-based decisions. Human HCT116 ATG14 knockout impairs Parkin/valinomycin mitochondrial clearance (Fig. 1G/H): “mitochondrial clearance was greatly reduced in hAtg14 knockout cells”. Restored mitophagy and response to depolarization as contextual functions. ATG14 performs the membrane-building work; cargo recognition need not be its own activity. The same paper measures BECN1 loss and concludes the binding partners contribute to “complex stability”, supporting the stability annotation in addition to phosphorylation.
• The post-transcriptional gene-expression assertion remains UNDECIDED: full Fig. 1 compares p62 protein and transcript levels, but the distinction between a gene-expression regulatory step and downstream protein turnover is not settled. It cannot be rejected merely because the abstract focuses on phosphorylation.
• Full PMID:22797916 demonstrates an ATG14–Snapin mechanism, delayed late endocytic trafficking and receptor degradation, and selective rescue. Restored this trafficking role as core. Early-to-late transport cited to abstract-only PMID:14617358 remains UNDECIDED: absence from an abstract is not misattribution, and late transport is not the same stage.
• Full PMID:21062745 was read from the author-hosted PDF after automated PMC retrieval failed; see ATG14-PMID:21062745-primary-check.md. ATG14 itself is assayed and stimulates VPS34 (Fig. 4C/D). The inhibitor annotation remains REMOVE on that positive target evidence. The source supports PI3P synthesis but does not establish an AKT pathway endpoint.
• Phagocytic-vesicle IEA was traced to mouse Q8CDJ3 and its IDA PMID:22493499 (live QuickGO snapshot in the audit folder). Full text tests SLAMF1 complexes, excluding ATG14 from that specific interaction. This does not exclude every phagosomal pool; no definitive positive ATG14 localization panel was located. Changed the old over-annotation call to UNDECIDED pending adjudication.
• PMID:23332761 explicitly describes ATG14-dependent inhibition of VPS34 phosphorylation and increased BECN1 phosphorylation. Both regulatory annotations represent direct core mechanisms, without making ATG14 a kinase.
• Live GO confirms GO:0097629 and GO:0097632 remain valid extrinsic-component terms, so the prior obsolescence/topology objection is withdrawn. GO:0034045 is now obsolete; GO:7770114 phagophore membrane exists and is used as the replacement. The local validator ontology predates GO:7770114, so the core synthesis retains the correct broader phagophore location GO:0061908, with membrane association explained in prose. No duplicate new-term proposal.
• Falcon corroborates PI3KC3-C1 targeting, BATS curvature sensing and the upstream ULK1–ATG14 Ser29 mechanism. It largely cites reviews for architecture and does not adjudicate the disputed localization or gene-expression assertions. Its upstream ULK1/ZDHHC13 and association-level disease findings do not become new ATG14 process annotations. No OpenScientist duplicate was launched; cache queries for gene/accession/aliases returned no existing report.
• Generic protein-binding removals follow the repository information-content policy; they do not deny the experiments. All original evidence codes, identifiers, source terms and qualifiers are preserved.

2026-09-20 recovered pexophagy report and primary check

Shared PIK3C3 report incorporation: full PMID:21121900 Fig. 4B supports yeast ATG14-containing complex I in pexophagy, consistent with the known human membrane-targeting cofactor role. This gene has no existing pexophagy assertion. No NEW process was added solely by extending the complex-wide report; the established autophagy core and other pending hypotheses are unchanged.

Recovery PR evidence follow-up (2026-09-22)

Correct the obsolete abstract-only statement while retaining the explicit unresolved early-to-late endosomal evidence scope.

Recovery review consistency follow-up (2026-09-22)

Restored readable identifiers in manual prose. Where applicable, reconciled AP3M2 reference notes with the retained contextual claim, removed unrelated PIK3C3 support from unresolved projections, separated PIK3C3 aspect-specific reasons, and documented the surviving/renamed ATG14 membrane term. Source assertion fields and verbatim quotations are unchanged.