ATG16L1

UniProt ID: Q676U5
Organism: Homo sapiens
Review Status: INITIALIZED
πŸ“ Provide Detailed Feedback

Gene Description

ATG16L1 is the targeting subunit of the E3-like machinery that conjugates ATG8-family proteins to membrane phospholipids. It has no catalytic activity of its own: it binds the ATG12-ATG5 conjugate, homodimerizes through a coiled coil, and oligomerizes into 400 kDa and 800 kDa assemblies that determine where ATG3 delivers an activated ATG8 protein. What ATG16L1 supplies is placement. Two separable membrane-binding regions and a set of protein interactions decide which bilayer gets modified: an N-terminal region required for lipidation under all conditions, a beta-isoform-specific C-terminal region, RB1CC1 and the WIPI proteins that read the PtdIns3P mark on the phagophore, and RAB33A and RAB33B, which bind the ATG16L1 homodimer through a noncanonical Rab-effector interface and recruit the complex to phagophores. The protein serves two mechanistically distinct pathways through different domains. In canonical macroautophagy it drives ATG8 lipidation on the double-membrane phagophore and so controls its elongation. In non-canonical autophagy, often called CASM, its WD40 C-terminal domain directs lipidation onto single, pre-existing endolysosomal and endosomal membranes - a route dispensable for canonical autophagy but required for LC3-associated phagocytosis, for antigen presentation in dendritic cells, and for surviving influenza A infection in mice. ATG16L1 also links autophagy to innate immune signalling through NOD1, NOD2, TLR2, TMEM59, MEFV and IRGM, and the T300A coding variant is a well-known Crohn's disease risk allele.

Proposed New Ontology Terms

phagophore membrane

Definition: The membrane system of a phagophore, comprising its two closely apposed lipid bilayers and the continuous rim that joins them, which expands by lipid acquisition and closes to form the double membrane of an autophagosome.

Justification: The four GO:0034045 assertions on this gene, and the GO:0000421 autophagosome membrane annotation, all describe the same thing: the ATG12-ATG5-ATG16L1 complex acting on the elongating phagophore bilayer. GO has autophagosome membrane and omegasome membrane but no phagophore membrane, so such annotations split between a term for the finished organelle and GO:0034045, which asserts via bounding_layer_of a bounding membrane the PAS does not have. See GO issue #29437 recommendation R3 and projects/CONDENSATES/GO_0034045-annotation-audit.md. Same term proposed in the ATG9A and WIPI2 reviews.

Parent term: phagophore

endoplasmic reticulum-phagophore membrane contact site

Definition: An organelle membrane contact site formed at the junction of the endoplasmic reticulum membrane and the phagophore or isolation membrane, across which lipid is delivered to the expanding phagophore and which is resolved before autophagosome closure.

Justification: PMID:28890335 shows in mammalian and C. elegans cells that isolation membranes contact the ER, that VMP1 controls those contacts through SERCA, and that losing VMP1 locks isolation membranes onto the ER and blocks autophagosome formation. GOA recorded this as GO:0120095 vacuole-isolation membrane contact site, a term defined for the junction of the vacuolar membrane with the isolation membrane - the yeast VICS - in an organism that has no vacuole. The mis-assignment happens because GO:0120095 is the ontology's only isolation-membrane contact site term. The organelle membrane contact site branch already has ER-vacuole, ER-endosome, ER-lysosome, ER-plasma membrane, ER-trans-Golgi and ER-lipid droplet children; the ER-phagophore contact, which is where ATG2-mediated lipid transfer occurs, is missing. See modules/phagophore_organelle_contact_site.yaml.

Parent term: organelle membrane contact site

protein lipidation at single membranes

Definition: The conjugation of an ATG8-family protein to a phospholipid of a single, pre-existing membrane such as an endosome, lysosome, phagosome or plasma membrane, without formation of a double-membrane autophagosome and without sequestration of cytosolic cargo. Also known as non-canonical autophagy or conjugation of ATG8 to single membranes (CASM).

Justification: Two IDA experiments on this gene, PMID:29317426 and PMID:33586810, both characterise this process and both were annotated to GO:0016237 microautophagy, whose definition requires cargo ingestion by direct invagination of the compartment membrane. CASM involves no invagination and delivers no cargo by itself; it is a lipidation event on an existing bilayer, and it is separable from canonical autophagy at the domain level, needing the WD40 C-terminal domain of ATG16L1 that canonical autophagy does not. A search of the autophagy branch returns macroautophagy, microautophagy, chaperone-mediated autophagy and the cargo-specific children, but no term for this process, which is why curators reached for the nearest wrong one. LC3-associated phagocytosis, the best-known instance, likewise has no GO term.

Parent term: C-terminal protein lipidation

Existing Annotations Review

GO Term Evidence Action Reason
GO:0000045 autophagosome assembly
IBA
GO_REF:0000033
ACCEPT
Summary: Autophagosome assembly, the process ATG16L1's conjugation activity drives.
Reason: ATG16L1 positions the ATG12-ATG5 conjugate on the phagophore, where the complex acts as the E3-like factor for ATG8-family lipidation and thereby drives membrane elongation. Supported by IMP as well as IBA across the family. Core.
GO:0034045 phagophore assembly site membrane
IBA
GO_REF:0000033
MODIFY
Summary: Phagophore assembly site membrane. Four assertions describing ATG16L1 recruitment to the isolation membrane, not to a membrane bounding the assembly site.
Reason: GO:0034045 asserts via bounding_layer_of that a membrane bounds the phagophore assembly site. The PAS is a liquid-like protein condensate with no bounding bilayer, and the term carries phagophore and isolation membrane as related synonyms of the distinct term GO:0061908 - see GO issue #29437 and projects/CONDENSATES/GO_0034045-annotation-audit.md. The underlying experiments here are unambiguous about which sense is meant: PMID:23392225 is titled for targeting of Atg16L1 to the isolation membrane, and PMID:32960676 shows RAB33B recruiting the complex to the phagophore. Moved to GO:0061908, the existing term those synonyms belong to; the precise destination is the phagophore membrane proposed under proposed_new_terms. One of the four assertions carries UniProtKB-SubCell:SL-0221, so this is also an SL propagation.
Propagation Review
Root cause: TERM SCOPING PROBLEM
Failure modes: GRANULARITY MISMATCH
Sources checked:
MGI:MGI:1924290 Β· Atg16l1 (mouse) SUPPORTS TRANSFER
Direct one-to-one ortholog. The localization transfers cleanly; the defect is in the destination term, not the inference.
UniProtKB:Q676U5 Β· ATG16L1 (the target's own experimental evidence) SUPPORTS TRANSFER
The target appears among its own WITH/FROM sources because its experimental annotation was one of the descendant evidences behind the curator's IBD. That strengthens the transfer rather than making it circular, and it means the term issue flagged here is purely one of destination, with the underlying localization experimentally grounded on this gene.
MGI:MGI:1920933 Β· Atg16l2 (mouse) SUPPORTS TRANSFER
Atg16l2 is a paralog rather than an ortholog, and its inclusion means the curator placed the IBD ancestral to both ATG16 paralogs - a positive judgment that the localization predates the duplication, not a dilution of the evidence. Nothing here bears on the term-scoping objection, which is the whole of the issue with this annotation.
PANTHER:PTN000459085 Β· ATG16 family node SUPPORTS TRANSFER
The node grouping is sound; all members act at autophagic membranes.
Proposed replacements: phagophore
Supporting Evidence:
PMID:23392225
Atg16L1 deletion mutants, which lack the FIP200-interacting domain, are defective in proper membrane targeting.
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore
GO:0000421 autophagosome membrane
IBA
GO_REF:0000033
ACCEPT
Summary: Autophagosome membrane, the site of the canonical conjugation reaction.
Reason: This is where the ATG12-ATG5-ATG16L1 complex works during canonical autophagy. Strictly the reaction happens on the elongating phagophore rather than on a completed autophagosome, which is the same phagophore/autophagosome-membrane conflation that the proposed phagophore membrane term is meant to resolve. Accepted as the best available term for a core location.
GO:0034274 Atg12-Atg5-Atg16 complex
IBA
GO_REF:0000033
ACCEPT
Summary: Atg12-Atg5-Atg16 complex. The complex ATG16L1 names and organizes.
Reason: ATG16L1 homodimerizes and oligomerizes the ATG12-ATG5 conjugate into the 400 kDa and 800 kDa assemblies that carry out ATG8-family lipidation. Supported by IDA, IPI and IBA. Core.
GO:0043495 protein-membrane adaptor activity
IBA
GO_REF:0000033
ACCEPT
Summary: Protein-membrane adaptor activity. The molecular function ATG16L1 contributes to the complex: it puts the conjugation machinery on the right membrane.
Reason: ATG16L1 has two separable membrane-binding regions, an N-terminal one required for lipidation under all conditions and a beta-isoform-specific C-terminal one required only for lipidation on perturbed endosomes. ATG12-ATG5 alone cannot bind membrane; adding Atg16 restores it. That is exactly the GO:0043495 definition - bringing a protein complex together with a membrane - and it is what distinguishes ATG16L1's contribution from the catalytic chemistry supplied by ATG3 and ATG7. Core.
Supporting Evidence:
PMID:30778222
the N-terminal membrane-binding region is essential for LC3B lipidation under all conditions tested
GO:0000045 autophagosome assembly
IEA
GO_REF:0000120
ACCEPT
Summary: Autophagosome assembly, the process ATG16L1's conjugation activity drives.
Reason: ATG16L1 positions the ATG12-ATG5 conjugate on the phagophore, where the complex acts as the E3-like factor for ATG8-family lipidation and thereby drives membrane elongation. Supported by IMP as well as IBA across the family. Core.
GO:0005737 cytoplasm
IEA
GO_REF:0000120
MARK AS OVER ANNOTATED
Summary: Cytoplasm, the broadest possible location for a protein with six specific ones.
Reason: ATG16L1 is annotated to autophagosome membrane, lysosomal membrane, endosome membrane, endolysosome membrane, the phagophore assembly site and the cytosol. The bare cytoplasm parent adds nothing to any of them. The IEA copy of it carries UniProtKB-SubCell:SL-0086, so this is also an instance of the SL project's pattern A; see projects/SL.md.
GO:0005765 lysosomal membrane
IEA
GO_REF:0000044
ACCEPT
Summary: Lysosomal membrane. Core location for the non-canonical, WD40-dependent arm of ATG16L1 function.
Reason: ATG16L1 has two separable jobs, and this is the location of the second one. Its WD40 C-terminal domain is required for ATG8-family lipidation onto single endolysosomal membranes and dispensable for canonical autophagy, which makes this a genuine site of action rather than a trafficking waypoint. Core.
Supporting Evidence:
PMID:29317426
In contrast, non-canonical autophagy is a pathway through which LC3 is lipidated and inserted into single membranes, particularly endolysosomal vacuoles during cell engulfment events such as LC3-associated phagocytosis.
GO:0010008 endosome membrane
IEA
GO_REF:0000044
ACCEPT
Summary: Endosome membrane, a demonstrated site of non-canonical ATG8 lipidation.
Reason: Lipidation onto perturbed endosomes requires the beta-isoform-specific C-terminal membrane-binding region of ATG16L1 and is independent of VPS34, so this is a mechanistically distinct site of action rather than an incidental localization. Core, as part of the non-canonical arm.
Supporting Evidence:
PMID:30778222
Thus, we propose that distinct membrane binding by the two isoforms of ATG16L1 mechanistically distinguish lipidation of LC3 and GABARAP proteins at distinct cellular locations and/or under different cellular conditions.
IEA
GO_REF:0000108
MARK AS OVER ANNOTATED
Summary: Axon, inferred from a biological process that was itself inferred from a mouse ortholog.
Reason: GO_REF:0000108 assigns a cellular component automatically from an existing process annotation - here GO:0098930 axonal transport, which is itself only an Ensembl projection from mouse Atg16l1. Two automatic inferences chained together, with no observation of human ATG16L1 in an axon anywhere in the chain. The same objection applies to the process annotation it derives from.
GO:0034045 phagophore assembly site membrane
IEA
GO_REF:0000120
MODIFY
Summary: Phagophore assembly site membrane. Four assertions describing ATG16L1 recruitment to the isolation membrane, not to a membrane bounding the assembly site.
Reason: GO:0034045 asserts via bounding_layer_of that a membrane bounds the phagophore assembly site. The PAS is a liquid-like protein condensate with no bounding bilayer, and the term carries phagophore and isolation membrane as related synonyms of the distinct term GO:0061908 - see GO issue #29437 and projects/CONDENSATES/GO_0034045-annotation-audit.md. The underlying experiments here are unambiguous about which sense is meant: PMID:23392225 is titled for targeting of Atg16L1 to the isolation membrane, and PMID:32960676 shows RAB33B recruiting the complex to the phagophore. Moved to GO:0061908, the existing term those synonyms belong to; the precise destination is the phagophore membrane proposed under proposed_new_terms. One of the four assertions carries UniProtKB-SubCell:SL-0221, so this is also an SL propagation.
Proposed replacements: phagophore
Supporting Evidence:
PMID:23392225
Atg16L1 deletion mutants, which lack the FIP200-interacting domain, are defective in proper membrane targeting.
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore
GO:0034497 protein localization to phagophore assembly site
IEA
GO_REF:0000117
ACCEPT
Summary: Protein localization to phagophore assembly site, supported by IMP.
Reason: ATG16L1 brings ATG12-ATG5 to the phagophore, and RAB33B binding is one route by which that recruitment happens. The process counterpart of the accepted GO:0043495 adaptor activity. Core.
GO:0042802 identical protein binding
IEA
GO_REF:0000117
KEEP AS NON CORE
Summary: Identical protein binding. Unusually informative for this term, since ATG16L1 homodimerization is structurally required.
Reason: Both IPI experiments name UniProtKB:Q676U5, ATG16L1 itself, and the homodimer is real and functional - two RAB33B molecules bind independently to one ATG16L1 homodimer, and the coiled-coil-mediated dimer is the unit that assembles into the larger ATG12-ATG5-ATG16L1 complexes. The term still describes an interaction rather than an activity, and the functional consequence is already captured by GO:0034274, so this is kept as genuine but non-core.
GO:0005515 protein binding
IPI
PMID:20562859
Network organization of the human autophagy system.
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:23202584
Structure of the human ATG12~ATG5 conjugate required for LC3...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:23262492
Interaction between FIP200 and ATG16L1 distinguishes ULK1 co...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:23376921
TMEM59 defines a novel ATG16L1-binding motif that promotes l...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:23392225
FIP200 regulates targeting of Atg16L1 to the isolation membr...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:32296183
A reference map of the human binary protein interactome.
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:33961781
Dual proteome-scale networks reveal cell-specific remodeling...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:40205054
Multimodal cell maps as a foundation for structural and func...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0042802 identical protein binding
IPI
PMID:20562859
Network organization of the human autophagy system.
KEEP AS NON CORE
Summary: Identical protein binding. Unusually informative for this term, since ATG16L1 homodimerization is structurally required.
Reason: Both IPI experiments name UniProtKB:Q676U5, ATG16L1 itself, and the homodimer is real and functional - two RAB33B molecules bind independently to one ATG16L1 homodimer, and the coiled-coil-mediated dimer is the unit that assembles into the larger ATG12-ATG5-ATG16L1 complexes. The term still describes an interaction rather than an activity, and the functional consequence is already captured by GO:0034274, so this is kept as genuine but non-core.
GO:0042802 identical protein binding
IPI
PMID:23392225
FIP200 regulates targeting of Atg16L1 to the isolation membr...
KEEP AS NON CORE
Summary: Identical protein binding. Unusually informative for this term, since ATG16L1 homodimerization is structurally required.
Reason: Both IPI experiments name UniProtKB:Q676U5, ATG16L1 itself, and the homodimer is real and functional - two RAB33B molecules bind independently to one ATG16L1 homodimer, and the coiled-coil-mediated dimer is the unit that assembles into the larger ATG12-ATG5-ATG16L1 complexes. The term still describes an interaction rather than an activity, and the functional consequence is already captured by GO:0034274, so this is kept as genuine but non-core.
GO:0000421 autophagosome membrane
IEA
GO_REF:0000120
ACCEPT
Summary: Autophagosome membrane, the site of the canonical conjugation reaction.
Reason: This is where the ATG12-ATG5-ATG16L1 complex works during canonical autophagy. Strictly the reaction happens on the elongating phagophore rather than on a completed autophagosome, which is the same phagophore/autophagosome-membrane conflation that the proposed phagophore membrane term is meant to resolve. Accepted as the best available term for a core location.
GO:0005776 autophagosome
IEA
GO_REF:0000107
KEEP AS NON CORE
Summary: Autophagosome, the parent of the autophagosome membrane term already supported.
Reason: Less informative than GO:0000421, which this gene carries from direct evidence. Both assertions here are transfers from the mouse entry rather than independent observations. Retained as non-core.
GO:0005930 axoneme
IEA
GO_REF:0000107
MARK AS OVER ANNOTATED
Summary: Axoneme, asserted twice from the same mouse entry by two different automatic routes.
Reason: Both the Ensembl-compara projection and the ISS trace to UniProtKB:Q8C0J2, mouse Atg16l1, so they are one claim counted twice rather than two lines of evidence. A ciliary or flagellar axoneme localization is a strong and tissue-restricted claim for a ubiquitously expressed component of the ATG8 conjugation machinery, and nothing in the human record supports it. The related sperm midpiece annotation at least has an independent anchor in the SPATA33 interaction on this same entry; this one does not.
GO:0010508 positive regulation of autophagy
IEA
GO_REF:0000107
MARK AS OVER ANNOTATED
Summary: Positive regulation of autophagy, a role conflation: ATG16L1 is machinery, not a regulator.
Reason: ATG16L1 executes a step of autophagy - it is the targeting subunit of the ATG8 conjugation machinery - rather than modulating the rate at which autophagy proceeds. Annotating a core component to positive regulation of the process it constitutes blurs the machinery/regulator distinction that the regulation branch exists to preserve, and it is doubly weak here because the only support is an ortholog projection from the mouse entry. The executed process is already captured by GO:0016236 and GO:0000045.
GO:0016236 macroautophagy
IEA
GO_REF:0000107
ACCEPT
Summary: Macroautophagy, the canonical pathway ATG16L1 serves.
Reason: Supported by two independent IDA experiments and by Reactome. Core.
GO:0016237 microautophagy
IEA
GO_REF:0000120
MODIFY
Summary: Microautophagy. Both supporting experiments are about ATG8 conjugation to single endolysosomal membranes, which is not what GO:0016237 means.
Reason: GO:0016237 is defined as cytosolic components being ingested by late endosomes, lysosomes or vacuoles through direct invagination of the compartment membrane, without prior sequestration into an autophagosome. That is a cargo-uptake mechanism. What PMID:29317426 and PMID:33586810 actually show is that the WD40 domain of ATG16L1 directs lipidation of ATG8-family proteins onto single, already-existing endolysosomal membranes - the process usually called non-canonical autophagy or CASM, which involves no invagination and delivers no cargo by itself. The curators appear to have reached for the nearest available term because GO has none for CASM: a search of the autophagy branch returns macroautophagy, microautophagy, chaperone-mediated autophagy and the cargo-specific children, but nothing for conjugation of ATG8 to single membranes. Moved to GO:0006501 C-terminal protein lipidation, which both papers support directly and which this gene already carries from one of them. A dedicated term is proposed under proposed_new_terms.
Proposed replacements: C-terminal protein lipidation
Supporting Evidence:
PMID:29317426
In contrast, non-canonical autophagy is a pathway through which LC3 is lipidated and inserted into single membranes, particularly endolysosomal vacuoles during cell engulfment events such as LC3-associated phagocytosis.
GO:0034274 Atg12-Atg5-Atg16 complex
IEA
GO_REF:0000120
ACCEPT
Summary: Atg12-Atg5-Atg16 complex. The complex ATG16L1 names and organizes.
Reason: ATG16L1 homodimerizes and oligomerizes the ATG12-ATG5 conjugate into the 400 kDa and 800 kDa assemblies that carry out ATG8-family lipidation. Supported by IDA, IPI and IBA. Core.
GO:0051607 defense response to virus
IEA
GO_REF:0000107
KEEP AS NON CORE
Summary: Defense response to virus. An ortholog projection that happens to be well founded, but a downstream physiological role.
Reason: The IEA is a compara projection from mouse Atg16l1, and in this instance the mouse evidence is strong and specific: mice lacking the WD and linker domains fail to target ATG8-family proteins to single membranes and die of influenza A infection. The claim is therefore defensible, but it is an organism-level consequence of the non-canonical lipidation activity rather than a function of the protein. Non-core.
Supporting Evidence:
PMID:33586810
ATG16L1-dependent targeting of LC3 to single-membrane, non-autophagosome compartments - referred to as non-canonical autophagy - protects mice from lethal IAV infection.
GO:0097225 sperm midpiece
IEA
GO_REF:0000107
KEEP AS NON CORE
Summary: Sperm midpiece, projected from the mouse entry and weakly corroborated by the SPATA33 interaction.
Reason: As with the axoneme annotation, the Ensembl projection and the ISS both trace to UniProtKB:Q8C0J2 and are one claim, not two. Unlike the axoneme annotation this one has an independent thread on the human record: ATG16L1 interacts through its WD repeats with SPATA33, a spermatogenesis-associated protein. That makes a sperm-restricted localization plausible enough to keep, but it is a tissue-specific context rather than a core location.
GO:0098930 axonal transport
IEA
GO_REF:0000107
MARK AS OVER ANNOTATED
Summary: Axonal transport, an Ensembl-compara projection with no supporting observation on this gene.
Reason: Autophagosomes do undergo retrograde transport in axons, so a neuronal role for the autophagy machinery is not implausible in principle. But this annotation asserts participation in axonal transport as a process, on the strength of an automatic ortholog projection alone, and it is then used as the basis for a second automatic inference to GO:0030424 axon. An unsupported claim that seeds a further claim is worth flagging in both places.
GO:0098978 glutamatergic synapse
IEA
GO_REF:0000107
MARK AS OVER ANNOTATED
Summary: Glutamatergic synapse with an is_active_in qualifier, from an ortholog projection.
Reason: is_active_in asserts that this is a site where the gene product carries out its molecular function, which is a considerably stronger statement than located_in and not one an Ensembl-compara projection can support. Synapse-type terms of this kind normally rest on dedicated synaptic curation with direct evidence. Flagged as over-annotated on the strength of the qualifier as much as the term.
GO:0016236 macroautophagy
TAS
Reactome:R-HSA-1632852
ACCEPT
Summary: Macroautophagy, the canonical pathway ATG16L1 serves.
Reason: Supported by two independent IDA experiments and by Reactome. Core.
GO:0005515 protein binding
IPI
PMID:39284914
ER-phagy restrains inflammatory responses through its recept...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005829 cytosol
IDA
GO_REF:0000052
KEEP AS NON CORE
Summary: Cytosol, the soluble pool of a peripherally membrane-associated complex.
Reason: The ATG12-ATG5-ATG16L1 complex is soluble until it is recruited to a membrane, so a cytosolic pool is real, but the functionally informative locations are the membranes where lipidation happens. Non-core.
GO:0005737 cytoplasm
EXP
PMID:23392225
FIP200 regulates targeting of Atg16L1 to the isolation membr...
MARK AS OVER ANNOTATED
Summary: Cytoplasm, the broadest possible location for a protein with six specific ones.
Reason: ATG16L1 is annotated to autophagosome membrane, lysosomal membrane, endosome membrane, endolysosome membrane, the phagophore assembly site and the cytosol. The bare cytoplasm parent adds nothing to any of them. The IEA copy of it carries UniProtKB-SubCell:SL-0086, so this is also an instance of the SL project's pattern A; see projects/SL.md.
GO:0005765 lysosomal membrane
EXP
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
ACCEPT
Summary: Lysosomal membrane. Core location for the non-canonical, WD40-dependent arm of ATG16L1 function.
Reason: ATG16L1 has two separable jobs, and this is the location of the second one. Its WD40 C-terminal domain is required for ATG8-family lipidation onto single endolysosomal membranes and dispensable for canonical autophagy, which makes this a genuine site of action rather than a trafficking waypoint. Core.
Supporting Evidence:
PMID:29317426
In contrast, non-canonical autophagy is a pathway through which LC3 is lipidated and inserted into single membranes, particularly endolysosomal vacuoles during cell engulfment events such as LC3-associated phagocytosis.
GO:0010008 endosome membrane
EXP
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
ACCEPT
Summary: Endosome membrane, a demonstrated site of non-canonical ATG8 lipidation.
Reason: Lipidation onto perturbed endosomes requires the beta-isoform-specific C-terminal membrane-binding region of ATG16L1 and is independent of VPS34, so this is a mechanistically distinct site of action rather than an incidental localization. Core, as part of the non-canonical arm.
Supporting Evidence:
PMID:30778222
Thus, we propose that distinct membrane binding by the two isoforms of ATG16L1 mechanistically distinguish lipidation of LC3 and GABARAP proteins at distinct cellular locations and/or under different cellular conditions.
GO:0034045 phagophore assembly site membrane
EXP
PMID:23392225
FIP200 regulates targeting of Atg16L1 to the isolation membr...
MODIFY
Summary: Phagophore assembly site membrane. Four assertions describing ATG16L1 recruitment to the isolation membrane, not to a membrane bounding the assembly site.
Reason: GO:0034045 asserts via bounding_layer_of that a membrane bounds the phagophore assembly site. The PAS is a liquid-like protein condensate with no bounding bilayer, and the term carries phagophore and isolation membrane as related synonyms of the distinct term GO:0061908 - see GO issue #29437 and projects/CONDENSATES/GO_0034045-annotation-audit.md. The underlying experiments here are unambiguous about which sense is meant: PMID:23392225 is titled for targeting of Atg16L1 to the isolation membrane, and PMID:32960676 shows RAB33B recruiting the complex to the phagophore. Moved to GO:0061908, the existing term those synonyms belong to; the precise destination is the phagophore membrane proposed under proposed_new_terms. One of the four assertions carries UniProtKB-SubCell:SL-0221, so this is also an SL propagation.
Proposed replacements: phagophore
Supporting Evidence:
PMID:23392225
Atg16L1 deletion mutants, which lack the FIP200-interacting domain, are defective in proper membrane targeting.
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore
GO:0034274 Atg12-Atg5-Atg16 complex
IPI
PMID:23202584
Structure of the human ATG12~ATG5 conjugate required for LC3...
ACCEPT
Summary: Atg12-Atg5-Atg16 complex. The complex ATG16L1 names and organizes.
Reason: ATG16L1 homodimerizes and oligomerizes the ATG12-ATG5 conjugate into the 400 kDa and 800 kDa assemblies that carry out ATG8-family lipidation. Supported by IDA, IPI and IBA. Core.
GO:0000045 autophagosome assembly
IMP
PMID:30778222
Distinct functions of ATG16L1 isoforms in membrane binding a...
ACCEPT
Summary: Autophagosome assembly, the process ATG16L1's conjugation activity drives.
Reason: ATG16L1 positions the ATG12-ATG5 conjugate on the phagophore, where the complex acts as the E3-like factor for ATG8-family lipidation and thereby drives membrane elongation. Supported by IMP as well as IBA across the family. Core.
GO:0043495 protein-membrane adaptor activity
IMP
PMID:30778222
Distinct functions of ATG16L1 isoforms in membrane binding a...
ACCEPT
Summary: Protein-membrane adaptor activity. The molecular function ATG16L1 contributes to the complex: it puts the conjugation machinery on the right membrane.
Reason: ATG16L1 has two separable membrane-binding regions, an N-terminal one required for lipidation under all conditions and a beta-isoform-specific C-terminal one required only for lipidation on perturbed endosomes. ATG12-ATG5 alone cannot bind membrane; adding Atg16 restores it. That is exactly the GO:0043495 definition - bringing a protein complex together with a membrane - and it is what distinguishes ATG16L1's contribution from the catalytic chemistry supplied by ATG3 and ATG7. Core.
Supporting Evidence:
PMID:30778222
the N-terminal membrane-binding region is essential for LC3B lipidation under all conditions tested
GO:0005515 protein binding
IPI
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore via a ...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0034045 phagophore assembly site membrane
IDA
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore via a ...
MODIFY
Summary: Phagophore assembly site membrane. Four assertions describing ATG16L1 recruitment to the isolation membrane, not to a membrane bounding the assembly site.
Reason: GO:0034045 asserts via bounding_layer_of that a membrane bounds the phagophore assembly site. The PAS is a liquid-like protein condensate with no bounding bilayer, and the term carries phagophore and isolation membrane as related synonyms of the distinct term GO:0061908 - see GO issue #29437 and projects/CONDENSATES/GO_0034045-annotation-audit.md. The underlying experiments here are unambiguous about which sense is meant: PMID:23392225 is titled for targeting of Atg16L1 to the isolation membrane, and PMID:32960676 shows RAB33B recruiting the complex to the phagophore. Moved to GO:0061908, the existing term those synonyms belong to; the precise destination is the phagophore membrane proposed under proposed_new_terms. One of the four assertions carries UniProtKB-SubCell:SL-0221, so this is also an SL propagation.
Proposed replacements: phagophore
Supporting Evidence:
PMID:23392225
Atg16L1 deletion mutants, which lack the FIP200-interacting domain, are defective in proper membrane targeting.
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore
GO:0034497 protein localization to phagophore assembly site
IMP
PMID:32960676
RAB33B recruits the ATG16L1 complex to the phagophore via a ...
ACCEPT
Summary: Protein localization to phagophore assembly site, supported by IMP.
Reason: ATG16L1 brings ATG12-ATG5 to the phagophore, and RAB33B binding is one route by which that recruitment happens. The process counterpart of the accepted GO:0043495 adaptor activity. Core.
GO:0019787 ubiquitin-like protein transferase activity
TAS
Reactome:R-HSA-5678490
MODIFY
Summary: Ubiquitin-like protein transferase activity. Correct in kind but two levels too general; the specific term exists.
Reason: GO:0019776 Atg8-family ligase activity is defined as catalysis of the covalent attachment of an Atg8-family modifier to phosphatidylethanolamine or phosphatidylserine on a membrane. That is precisely what the ATG12-ATG5-ATG16L1 complex does, and the phosphatidylserine clause even covers the non-canonical arm that the WD40 domain supports. The Reactome reaction behind this annotation is literally titled for LC3 transfer from ATG3 to PE, so nothing is lost by using the specific term. Note that ATG16L1 is the targeting subunit rather than the catalytic one, so contributes_to on the complex-level activity would be more accurate than enables; that is a qualifier question separate from the term choice.
Proposed replacements: Atg8-family ligase activity
GO:0005515 protein binding
IPI
PMID:24191030
Structural basis of ATG3 recognition by the autophagic ubiqu...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:33087875
SPATA33 is an autophagy mediator for cargo selectivity in ge...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0097225 sperm midpiece
ISS
GO_REF:0000024
KEEP AS NON CORE
Summary: Sperm midpiece, projected from the mouse entry and weakly corroborated by the SPATA33 interaction.
Reason: As with the axoneme annotation, the Ensembl projection and the ISS both trace to UniProtKB:Q8C0J2 and are one claim, not two. Unlike the axoneme annotation this one has an independent thread on the human record: ATG16L1 interacts through its WD repeats with SPATA33, a spermatogenesis-associated protein. That makes a sperm-restricted localization plausible enough to keep, but it is a tissue-specific context rather than a core location.
GO:0000421 autophagosome membrane
IDA
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
ACCEPT
Summary: Autophagosome membrane, the site of the canonical conjugation reaction.
Reason: This is where the ATG12-ATG5-ATG16L1 complex works during canonical autophagy. Strictly the reaction happens on the elongating phagophore rather than on a completed autophagosome, which is the same phagophore/autophagosome-membrane conflation that the proposed phagophore membrane term is meant to resolve. Accepted as the best available term for a core location.
GO:0006501 C-terminal protein lipidation
IDA
PMID:33586810
Non-canonical autophagy functions of ATG16L1 in epithelial c...
ACCEPT
Summary: C-terminal protein lipidation. The reaction ATG16L1 exists to promote, stated as a process.
Reason: ATG8-family proteins are activated by exposure of a C-terminal glycine and then conjugated through it to a membrane phospholipid. As the targeting subunit of the E3-like ATG12-ATG5-ATG16L1 complex, ATG16L1 determines where that happens. Core, and the correct home for the observations currently sitting on GO:0016237 microautophagy.
Supporting Evidence:
PMID:33586810
ATG16L1-dependent targeting of LC3 to single-membrane, non-autophagosome compartments - referred to as non-canonical autophagy - protects mice from lethal IAV infection.
GO:0016236 macroautophagy
IDA
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
ACCEPT
Summary: Macroautophagy, the canonical pathway ATG16L1 serves.
Reason: Supported by two independent IDA experiments and by Reactome. Core.
GO:0016236 macroautophagy
IDA
PMID:33586810
Non-canonical autophagy functions of ATG16L1 in epithelial c...
ACCEPT
Summary: Macroautophagy, the canonical pathway ATG16L1 serves.
Reason: Supported by two independent IDA experiments and by Reactome. Core.
GO:0016237 microautophagy
IDA
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
MODIFY
Summary: Microautophagy. Both supporting experiments are about ATG8 conjugation to single endolysosomal membranes, which is not what GO:0016237 means.
Reason: GO:0016237 is defined as cytosolic components being ingested by late endosomes, lysosomes or vacuoles through direct invagination of the compartment membrane, without prior sequestration into an autophagosome. That is a cargo-uptake mechanism. What PMID:29317426 and PMID:33586810 actually show is that the WD40 domain of ATG16L1 directs lipidation of ATG8-family proteins onto single, already-existing endolysosomal membranes - the process usually called non-canonical autophagy or CASM, which involves no invagination and delivers no cargo by itself. The curators appear to have reached for the nearest available term because GO has none for CASM: a search of the autophagy branch returns macroautophagy, microautophagy, chaperone-mediated autophagy and the cargo-specific children, but nothing for conjugation of ATG8 to single membranes. Moved to GO:0006501 C-terminal protein lipidation, which both papers support directly and which this gene already carries from one of them. A dedicated term is proposed under proposed_new_terms.
Proposed replacements: C-terminal protein lipidation
Supporting Evidence:
PMID:29317426
In contrast, non-canonical autophagy is a pathway through which LC3 is lipidated and inserted into single membranes, particularly endolysosomal vacuoles during cell engulfment events such as LC3-associated phagocytosis.
GO:0016237 microautophagy
IDA
PMID:33586810
Non-canonical autophagy functions of ATG16L1 in epithelial c...
MODIFY
Summary: Microautophagy. Both supporting experiments are about ATG8 conjugation to single endolysosomal membranes, which is not what GO:0016237 means.
Reason: GO:0016237 is defined as cytosolic components being ingested by late endosomes, lysosomes or vacuoles through direct invagination of the compartment membrane, without prior sequestration into an autophagosome. That is a cargo-uptake mechanism. What PMID:29317426 and PMID:33586810 actually show is that the WD40 domain of ATG16L1 directs lipidation of ATG8-family proteins onto single, already-existing endolysosomal membranes - the process usually called non-canonical autophagy or CASM, which involves no invagination and delivers no cargo by itself. The curators appear to have reached for the nearest available term because GO has none for CASM: a search of the autophagy branch returns macroautophagy, microautophagy, chaperone-mediated autophagy and the cargo-specific children, but nothing for conjugation of ATG8 to single membranes. Moved to GO:0006501 C-terminal protein lipidation, which both papers support directly and which this gene already carries from one of them. A dedicated term is proposed under proposed_new_terms.
Proposed replacements: C-terminal protein lipidation
Supporting Evidence:
PMID:29317426
In contrast, non-canonical autophagy is a pathway through which LC3 is lipidated and inserted into single membranes, particularly endolysosomal vacuoles during cell engulfment events such as LC3-associated phagocytosis.
GO:0034274 Atg12-Atg5-Atg16 complex
IDA
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
ACCEPT
Summary: Atg12-Atg5-Atg16 complex. The complex ATG16L1 names and organizes.
Reason: ATG16L1 homodimerizes and oligomerizes the ATG12-ATG5 conjugate into the 400 kDa and 800 kDa assemblies that carry out ATG8-family lipidation. Supported by IDA, IPI and IBA. Core.
GO:0034274 Atg12-Atg5-Atg16 complex
IDA
PMID:33586810
Non-canonical autophagy functions of ATG16L1 in epithelial c...
ACCEPT
Summary: Atg12-Atg5-Atg16 complex. The complex ATG16L1 names and organizes.
Reason: ATG16L1 homodimerizes and oligomerizes the ATG12-ATG5 conjugate into the 400 kDa and 800 kDa assemblies that carry out ATG8-family lipidation. Supported by IDA, IPI and IBA. Core.
GO:0036020 endolysosome membrane
IDA
PMID:29317426
The WD40 domain of ATG16L1 is required for itsΒ non-canonical...
ACCEPT
Summary: Endolysosome membrane, the site of WD40-dependent non-canonical lipidation.
Reason: is_active_in is the right qualifier: this is where the WD40 domain does its work, not merely where the protein is found. Core, as part of the non-canonical arm.
GO:0005515 protein binding
IPI
PMID:28890335
The ER-Localized Transmembrane Protein EPG-3/VMP1 Regulates ...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0120095 vacuole-isolation membrane contact site
IDA
PMID:28890335
The ER-Localized Transmembrane Protein EPG-3/VMP1 Regulates ...
MODIFY
Summary: Vacuole-isolation membrane contact site. The supporting paper describes ER-isolation membrane contacts in mammalian cells; the vacuole is the wrong organelle.
Reason: GO:0120095 is defined as a contact site at the junction of the vacuolar membrane and the isolation membrane, with synonym vacuole-phagophore contact site - the yeast VICS, which in budding yeast is organized by Vac8. PMID:28890335 is about mammalian and C. elegans cells and is explicit that the contacts in question are between the isolation membrane and the endoplasmic reticulum, controlled by the ER-resident protein VMP1 through SERCA. Mammalian cells have no vacuole. The reason the annotation landed here is that GO has no term for the structure the paper describes: GO:0120095 is the only isolation-membrane contact site term in the ontology, and the organelle membrane contact site branch has ER-vacuole, ER-endosome, ER-lysosome, ER-plasma membrane, ER-Golgi and ER-lipid droplet children but no ER-phagophore child. Generalized to the parent GO:0044232, which is true as stated, and the missing child is proposed under proposed_new_terms. This is the same gap that modules/phagophore_organelle_contact_site.yaml is built around.
Proposed replacements: organelle membrane contact site
Supporting Evidence:
PMID:28890335
During autophagosome formation in mammalian cells, isolation membranes
PMID:28890335
Loss of VMP1 causes stable association of IMs with the ER, thus blocking autophagosome formation.
GO:0051020 GTPase binding
IPI
PMID:25891078
IRGM governs the core autophagy machinery to conduct antimic...
KEEP AS NON CORE
Summary: GTPase binding, annotated from the IRGM interaction. The functionally central GTPase partners are the RAB33 proteins, and they are not what this annotation records.
Reason: IRGM is an immunity-related GTPase and the interaction is real, so the term is correct, but it captures an antimicrobial-defence context rather than the core mechanism. The GTPase interaction that matters mechanistically is with RAB33A and RAB33B, which recruit the ATG16L1 complex to phagophores through a noncanonical Rab-effector interface; those are currently buried in the GO:0005515 protein binding pile and would justify GO:0031267 small GTPase binding on their own evidence. Non-core as annotated.
Supporting Evidence:
PMID:25891078
IRGM governs the core autophagy machinery to conduct antimicrobial defense.
GO:0005515 protein binding
IPI
PMID:28561066
WIPI3 and WIPI4 Ξ²-propellers are scaffolds for LKB1-AMPK-TSC...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005515 protein binding
IPI
PMID:26347139
TRIM-mediated precision autophagy targets cytoplasmic regula...
MARK AS OVER ANNOTATED
Summary: Bare protein binding from twenty-eight IPI experiments; uninformative as a function statement, and it is where several of this gene's most important interactions are buried.
Reason: Per project curation guidance, protein binding says nothing about what the gene product does. The problem is acute here because ATG16L1 is a hub and the pile hides mechanistically distinct partners: ATG5 and ATG12 form the conjugate it carries; RB1CC1 targets it to the isolation membrane; WIPI1 and WIPI2 deliver it there from the PtdIns3P mark; RAB33A and RAB33B recruit the complex to phagophores through a noncanonical Rab interaction; ATG3 hands over the ATG8 protein; MAP1LC3B, MAP1LC3C and GABARAPL2 are its substrates; TLR2, NOD2, TMEM59, MEFV, SPATA33 and UBAC2 connect it to signalling and cargo contexts. The substrate and Rab interactions in particular support the accepted GO:0019776 and GO:0034274 annotations and would be better captured there.
GO:0005829 cytosol
TAS
Reactome:R-HSA-5678490
KEEP AS NON CORE
Summary: Cytosol, the soluble pool of a peripherally membrane-associated complex.
Reason: The ATG12-ATG5-ATG16L1 complex is soluble until it is recruited to a membrane, so a cytosolic pool is real, but the functionally informative locations are the membranes where lipidation happens. Non-core.
GO:0005829 cytosol
TAS
Reactome:R-HSA-5679255
KEEP AS NON CORE
Summary: Cytosol, the soluble pool of a peripherally membrane-associated complex.
Reason: The ATG12-ATG5-ATG16L1 complex is soluble until it is recruited to a membrane, so a cytosolic pool is real, but the functionally informative locations are the membranes where lipidation happens. Non-core.
GO:0005829 cytosol
TAS
Reactome:R-HSA-5682012
KEEP AS NON CORE
Summary: Cytosol, the soluble pool of a peripherally membrane-associated complex.
Reason: The ATG12-ATG5-ATG16L1 complex is soluble until it is recruited to a membrane, so a cytosolic pool is real, but the functionally informative locations are the membranes where lipidation happens. Non-core.
GO:0005930 axoneme
ISS
GO_REF:0000024
MARK AS OVER ANNOTATED
Summary: Axoneme, asserted twice from the same mouse entry by two different automatic routes.
Reason: Both the Ensembl-compara projection and the ISS trace to UniProtKB:Q8C0J2, mouse Atg16l1, so they are one claim counted twice rather than two lines of evidence. A ciliary or flagellar axoneme localization is a strong and tissue-restricted claim for a ubiquitously expressed component of the ATG8 conjugation machinery, and nothing in the human record supports it. The related sperm midpiece annotation at least has an independent anchor in the SPATA33 interaction on this same entry; this one does not.
GO:0000045 autophagosome assembly
NAS
PMID:15620219
Cloning and analysis of human Apg16L.
ACCEPT
Summary: Autophagosome assembly, the process ATG16L1's conjugation activity drives.
Reason: ATG16L1 positions the ATG12-ATG5 conjugate on the phagophore, where the complex acts as the E3-like factor for ATG8-family lipidation and thereby drives membrane elongation. Supported by IMP as well as IBA across the family. Core.
GO:0005776 autophagosome
ISS
PMID:15620219
Cloning and analysis of human Apg16L.
KEEP AS NON CORE
Summary: Autophagosome, the parent of the autophagosome membrane term already supported.
Reason: Less informative than GO:0000421, which this gene carries from direct evidence. Both assertions here are transfers from the mouse entry rather than independent observations. Retained as non-core.

Core Functions

As the targeting subunit of the ATG12-ATG5-ATG16L1 complex, promotes covalent attachment of ATG8-family proteins to membrane phospholipid, the E3-like step of the ATG8 conjugation cascade. ATG16L1 supplies membrane placement and complex oligomerization rather than catalysis, so the activity belongs to the complex it completes.

Supporting Evidence:
  • PMID:30778222
    Thus, we propose that distinct membrane binding by the two isoforms of ATG16L1 mechanistically distinguish lipidation of LC3 and GABARAP proteins at distinct cellular locations and/or under different cellular conditions.

Brings the ATG12-ATG5 conjugate to a membrane. ATG12-ATG5 alone does not bind membrane; ATG16L1 supplies two separable membrane-binding regions, an N-terminal one required for lipidation under all conditions and a beta-isoform-specific C-terminal one required only on perturbed endosomes, and this is what selects which bilayer is modified.

Supporting Evidence:
  • PMID:30778222
    the N-terminal membrane-binding region is essential for LC3B lipidation under all conditions tested

Through its WD40 C-terminal domain, directs lipidation of ATG8-family proteins onto single, pre-existing endolysosomal and endosomal membranes - the non-canonical or CASM arm. This is domain-separable from canonical autophagy, which does not need the WD40 domain, and it is the activity that underlies LC3-associated phagocytosis and the protective response to influenza A virus in mice.

Supporting Evidence:
  • PMID:29317426
    In contrast, non-canonical autophagy is a pathway through which LC3 is lipidated and inserted into single membranes, particularly endolysosomal vacuoles during cell engulfment events such as LC3-associated phagocytosis.
  • PMID:33586810
    ATG16L1-dependent targeting of LC3 to single-membrane, non-autophagosome compartments - referred to as non-canonical autophagy - protects mice from lethal IAV infection.

References

Loading supporting content…

Download this section (compressed HTML)

Suggested Questions for Experts

Q: ATG16L1 is annotated with enables on the complex-level ligase activity, but it is the targeting subunit and the chemistry is done by ATG3 and ATG7. Should the conjugation annotations on ATG16L1, ATG5 and ATG12 all be contributes_to, with enables reserved for the catalytic subunits?

Q: Do the axoneme, sperm midpiece, axon, axonal transport and glutamatergic synapse annotations rest on any observation of human ATG16L1, or are all five downstream of the single mouse entry UniProtKB:Q8C0J2? Four of them are automatic, and one of those is inferred from another.

Q: Does the ontology need a CASM branch, and if so where should it attach - under protein lipidation, as proposed here, or as a sibling of macroautophagy under autophagy? The field is explicit that these are not bona fide autophagic processes, which argues against the autophagy branch.

Q: The T300A Crohn's risk variant increases caspase cleavage of ATG16L1. Which of this gene's two functional arms does it impair, and is the disease association better explained by loss of canonical autophagy or of WD40-dependent lipidation at single membranes?

Suggested Experiments

Experiment: Separation-of-function rescue in ATG16L1-null cells comparing full-length protein, a WD40-deleted mutant and an N-terminal membrane-binding mutant across canonical starvation-induced autophagy, LC3-associated phagocytosis, monensin-induced endosomal lipidation and Salmonella xenophagy, to establish which annotations belong to which arm and give GO a basis for splitting them.

Experiment: Direct test of the human localizations projected from mouse: immunostain endogenous ATG16L1 in human spermatozoa and in cultured human neurons, with knockout controls, to determine whether the axoneme, sperm midpiece, axon and synapse annotations are supported outside the mouse entry.

Experiment: Proximity labelling from ATG16L1 in cells with and without VMP1, scoring ER contact-site proteins, to test whether the contact recorded as GO:0120095 is an ER-phagophore contact as PMID:28890335 argues and to supply the evidence a new GO term would need.

πŸ“„ View Raw YAML

Loading supporting content…

Download this section (compressed HTML)