Auxilin (DnaJ homolog subfamily C member 6, dnajc6) is a J-domain co-chaperone that acts in the terminal, uncoating step of clathrin-mediated membrane traffic. The protein has a three-part architecture, an N-terminal PTEN-like (tensin-type) phosphatase domain followed by a tensin-type C2 domain, a long disordered clathrin-binding region, and a C-terminal J domain carrying the conserved His-Pro-Asp motif. By orthology to bovine and mammalian auxilin 1, it binds assembled clathrin lattices on newly formed clathrin-coated vesicles and, through its J domain, recruits the Hsc70 chaperone (HSPA8) and stimulates its ATP hydrolysis, driving disassembly of the clathrin coat so that vesicles can fuse with their targets and clathrin can be reused. The PTEN-like region is a phosphoinositide-dependent membrane-targeting module rather than an enzyme; the bovine ortholog has no detectable phosphatase activity, and zebrafish auxilin carries the same departure from the PTEN-class P-loop. Zebrafish has two auxilin-family genes, auxilin (dnajc6) and the kinase-containing paralog cyclin G-associated kinase (gak), which diverged before the fish-mammal split. Zebrafish auxilin is expressed maternally and then predominantly in neural tissues, hindbrain and spinal cord neurons and the otic vesicle, whereas gak is broadly expressed. When expressed in Drosophila, zebrafish auxilin rescues the neurogenic and lethal phenotypes of Drosophila auxilin mutants, and this rescue requires its J domain, showing that the Hsc70-dependent uncoating function is conserved. The dnajc6 gene produces many alternatively spliced protein products; this entry is a 974-residue isoform that lacks five N-terminal residues present in the 979-residue canonical product but retains every domain.
Summary: UniRule mapping (UR000414225) fired by the PROSITE tensin-type phosphatase domain (PS51181, residues 109-276), with Cys218 predicted as a phosphocysteine intermediate. The claim is that zebrafish auxilin catalyses protein dephosphorylation. The PTEN-like region of the auxilin 1 ortholog (bovine, same PANTHER subfamily PTHR23172:SF4) was solved crystallographically and tested; it resembles PTEN but has no detectable phosphatase activity, which the authors attribute to a change in the P loop. The zebrafish protein departs from the PTEN-class CX5R P-loop signature in the same way as human auxilin (CSDGR at Cys218 versus human CLDGR), so there is no sequence basis for expecting the zebrafish domain to differ. The PTEN-like region serves instead as a phosphoinositide-dependent membrane-targeting module.
Reason: This is an IEA fold-to-activity mapping that is contradicted by direct biochemical and structural evidence on the auxilin 1 ortholog within the same PANTHER subfamily, and the zebrafish sequence carries the same non-canonical P loop. The previous review marked it as over-annotated on the grounds that activity had not been excluded; the cached full text of the bovine structure paper states that no activity can be detected, so the annotation is demonstrably wrong for this subfamily rather than merely unproven. The human reviewed entry's hedge ("May act as a protein phosphatase") is not positive evidence.
Differences at key residues in the catalytic cleft suggest that the PTEN-like region of auxilin will not have phosphatase activity, however, and indeed, no activity can be detected.
Summary: Same UniRule mapping as GO:0004721, generalised to the parent term. The only hydrolase activity being claimed is the tensin-type phosphatase, which is absent from the auxilin 1 ortholog. Auxilin's characterised molecular role is non-enzymatic, it binds clathrin and recruits and activates Hsc70; the ATP hydrolysis that powers uncoating is performed by Hsc70, not by auxilin.
Reason: Treated consistently with the phosphoprotein phosphatase row from the same rule. The parent term asserts nothing beyond the phosphatase activity, and that activity is contradicted for this subfamily by the bovine auxilin structure and activity assay. No other hydrolase activity is known or predicted for auxilin.
Differences at key residues in the catalytic cleft suggest that the PTEN-like region of auxilin will not have phosphatase activity, however, and indeed, no activity can be detected.
Summary: Mapped from the UniProt subcellular-location line (itself ARBA-transferred). The term is independently supported. For zebrafish auxilin itself, GFP-zAux in HeLa cells is enriched in perinuclear structures that overlap with clathrin and, at higher expression, forms large clathrin-positive aggregates, the behaviour expected of a protein that binds assembled clathrin coats. For the auxilin 1 ortholog, the protein was identified as a coat-associated cofactor that binds assembled clathrin lattices, and the PAINT root node of PTHR23172 (PTN000571055) carries vesicle localisation by IBD. Zebrafish auxilin is the ortholog of mammalian auxilin 1 (Bai et al. 2010 place zAux with human and mouse auxilin rather than with zGAK, and UniProt places it in PTHR23172:SF4 with human DNAJC6).
Reason: This is the location of the core activity. The ARBA-derived subcellular location is not itself used as evidence; the annotation stands on direct co-localisation of zebrafish auxilin with clathrin structures and on the orthologous biochemistry.
IDA PMID:20082716 Disruption of zebrafish cyclin G-associated kinase (GAK) fun...
KEEP AS NON CORE
Summary: ZFIN IDA from Bai et al. 2010 (full text cached). N-terminally GFP-tagged zebrafish auxilin was transiently expressed in HeLa cells; at low expression the signal was "mostly cytosolic and slightly enriched near the perinuclear regions". The construct was shown to be functional (it rescues Drosophila auxilin mutants), so the observation reflects a functional protein, but it was made in a heterologous mammalian cell line by overexpression rather than in zebrafish cells.
Reason: A diffuse cytosolic pool is genuine and expected for a soluble co-chaperone that is recruited onto coated vesicles and released again after each uncoating cycle, so the annotation is correct and as precise as the observation allows. It is not the compartment where auxilin does its work (that is the clathrin-coated vesicle), so it is retained as a non-core localisation. The ZFIN curator's reading of the full text is not disputed.
In cells expressing low levels of GFP-zGAK or GFP-zAux (marked by asterisks), GFP signals were mostly cytosolic and slightly enriched near the perinuclear regions.
These GFP-tagged fusion proteins are functional, as ectopic expression of either GFP-zGAK or GFP-zAux in Drosophila could restore the neurogenic defect and the lethality caused by dAux mutations
IDA PMID:20082716 Disruption of zebrafish cyclin G-associated kinase (GAK) fun...
KEEP AS NON CORE
Summary: ZFIN IDA from the same HeLa-cell experiment. GFP-zAux was "slightly enriched near the perinuclear regions", and these perinuclear structures overlapped with clathrin staining; the authors suggest they represent the trans-Golgi network but did not use a TGN marker.
Reason: The observation is real and correctly captured by the term, and the clathrin overlap makes it biologically sensible (auxilin-family proteins act on clathrin-coated structures at the TGN as well as at the plasma membrane). It is non-core because the perinuclear region is a coarse descriptor of a heterologous overexpression pattern rather than the functional compartment, and the TGN identity was inferred rather than shown, so a MODIFY to a TGN term is not justified from this paper.
ISS PMID:8524399 Role of auxilin in uncoating clathrin-coated vesicles.
NEW
Summary: Proposed. The defining interaction of auxilin is with the Hsc70 chaperone HSPA8 through its C-terminal J domain, whose His-Pro-Asp tripeptide engages Hsp70. Bovine auxilin 1 recruits hsp70c to assembled clathrin in an ATP-dependent manner and loses cofactor activity when the J domain is deleted. The zebrafish J domain (910-974) retains the HPD motif, and in zebrafish auxilin the J domain is required for rescue of Drosophila auxilin mutants.
Reason: The interaction is established on the auxilin 1 ortholog and depends on a domain that is conserved in this sequence and functionally required in vivo for the zebrafish protein. The orthology is supported by PANTHER subfamily SF4 membership and by Bai et al.'s placement of zAux with mammalian auxilins. There is no direct binding assay on the zebrafish protein, hence ISS.
Summary: Proposed. Binding of the auxilin J domain to HSPA8 stimulates the chaperone's ATPase, which is the energy-consuming step of coat disassembly. The curated human auxilin entry describes this cycle, by similarity to bovine auxilin and with experimental support for the co-chaperone role, as successive HSPA8:ATP binding events with ATP hydrolysis stimulation. The zebrafish J domain retains the HPD motif that mediates ATPase stimulation in J-domain proteins.
Reason: This is the standard GO activity for a J-domain co-chaperone and the mechanistic core of auxilin function. It is inferred from the conserved J domain and from the auxilin 1 ortholog; there is no measurement of ATPase stimulation by the zebrafish protein, so it is proposed at ISS level.
Supporting Evidence:
file:human/DNAJC6/DNAJC6-uniprot.txt
binds to the clathrin-auxilin cage, initially at a ratio of one HSPA8 per triskelion leading to ATP hydrolysis stimulation
The primary molecular function is **not enzymatic in itself**, but rather **regulatory and targeting**: auxilin recruits HSC70 to clathrin-coated structures and stimulates HSC70's ATPase activity through its conserved J-domain containing the essential His-Pro-Asp (HPD) motif
IBA file:interpro/panther/PTHR23172/PTHR23172-paint.tsv
NEW
Summary: Proposed. PAINT curators placed clathrin coat disassembly (IBD) at the family root node PTN000571055. Auxilin was identified as the cofactor that, with Hsc70, strips clathrin coats from vesicles, and auxilin performs part of that work (it binds the lattice and drives Hsc70's ATP hydrolysis). Zebrafish auxilin supplies the function of Drosophila auxilin in vivo, in a J-domain-dependent manner, and its overexpression perturbs clathrin distribution in HeLa cells. A more specific vertebrate node, PTN002753975 (Euteleostomi), carries synaptic vesicle uncoating (GO:0016191) and the canonical zebrafish product inherits it by IBA; that term is a descendant of this one and is not proposed separately because the zebrafish evidence does not distinguish synaptic from other clathrin-coated vesicles.
Reason: The participation test is met, since auxilin together with Hsc70 performs the disassembly, and the term is phylogenetically grounded at the family root. This replaces the earlier proposal of clathrin-dependent endocytosis (GO:0072583), which is an ancestor of this term and would be redundant; GO:0072583 is no longer in the GOA set for this isoform (it is present by IBA on the canonical product).
These results suggest that zGAK and zAux can both functionally substitute for dAux.
Core Functions
Hsc70 co-chaperone for clathrin uncoating. Auxilin binds assembled clathrin on newly formed clathrin-coated vesicles and, via its J domain (His-Pro-Asp motif), recruits HSPA8/Hsc70 and stimulates its ATPase to drive ATP-dependent disassembly of the clathrin coat. The function is inherited from the root of the auxilin/GAK family and, for the zebrafish protein, is shown by J-domain-dependent rescue of Drosophila auxilin mutants.
Unlike their full-length counter parts, these J-deletions failed to rescue the extra Elav-positive cell phenotype (Figure 4E &4F) and the lethality, suggesting that, like dAux, the J-domain is critical for the functions of these auxilin family proteins.
Clathrin-lattice recognition. Auxilin binds assembled clathrin with high affinity, which positions the J domain on the coat; the PTEN-like and C2 region adds phosphoinositide-dependent membrane association, so that auxilin is recruited to vesicles after budding.
Q: Should the PTHR23172 PAINT tree be checked for whether the UniRule/PROSITE phosphatase mapping (UR000414225) should exclude the auxilin subfamily SF4, given that the bovine auxilin 1 PTEN-like domain has no detectable activity and the zebrafish and human P loops depart from the PTEN-class signature in the same way?
Q: Does zebrafish auxilin act at synapses, as the vertebrate PAINT node (PTN002753975, synaptic vesicle uncoating) implies? Its neural-restricted expression is consistent with this, but there is no zebrafish synaptic data.
Q: Given that dnajc6 expression is restricted to neural tissues while gak is ubiquitous, what phenotype does dnajc6 loss alone produce in zebrafish neurons, and does gak compensate as it partially does in auxilin knockout mice?
Suggested Experiments
Experiment: Generate a dnajc6 zebrafish mutant and assay synaptic vesicle recycling and clathrin-coated vesicle accumulation in hindbrain and spinal cord neurons by electron microscopy and FM-dye uptake, alone and in combination with gak loss, to test whether the neural-restricted paralog is required for uncoating in vivo.
Experiment: In vitro clathrin-cage disassembly and HSPA8 ATPase stimulation assays with purified zebrafish auxilin (wild type, HPD-to-AAA J-domain mutant, and the J-domain-deleted construct used by Bai et al.) to establish the co-chaperone activity biochemically rather than by rescue.
External Prediction Reviews
These computational predictions are reviewed separately from the GOA annotation set used for this review. The assessments below are from this project and do not constitute official GO annotations or endorsement by GO/UniProt. They are not included in the existing annotation review above.
Prediction method: ProtNLM2 · Version: UniProt 2024_06 pilot
PSEUDOENZYME OVERANNOTATION
Review rationale: The target has the PTEN-like, C2, and J-domain architecture of auxilin/DNAJC6, but the PTEN-like region is a conserved pseudophosphatase rather than an active phosphatase. Structural work on bovine auxilin shows that this region lacks phosphatase activity and instead participates in membrane binding (PMID:20826345), and the focused sequence comparison found the dead C-X3-R P-loop across human, mouse, and zebrafish DNAJC6 instead of PTEN's active C-X5-R loop. The predicted dephosphorylation biological-process term is an over-propagation from a catalytic PTEN-like ancestor: no evidence establishes that dnajc6 itself catalyzes or directly executes a dephosphorylation step.
Supporting Evidence:
file:DANRE/A0A8M9QG43/A0A8M9QG43-uniprot.txt: "ID A0A8M9QG43_DANRE Unreviewed; 974 AA. ... DR InterPro; IPR029023; Tensin_phosphatase. ... DR PANTHER; PTHR23172:SF4; TYROSINE-PROTEIN PHOSPHATASE AUXILIN-RELATED; 1. ... FT DOMAIN 109..276 ... FT /note="Phosphatase tensin-type" ... FT DOMAIN 282..417 ... FT /note="C2 tensin-type" ... FT DOMAIN 910..974 ... FT /note="J""
PMID:20826345: "A change in the structure of the P loop accounts for the lack of phosphatase activity"