dnajc6 (A0A8M9QG43) - zebrafish auxilin - notes

Zebrafish auxilin (dnajc6; TrEMBL A0A8M9QG43, 974 aa RefSeq XP_021332835.1). The only
primary study of the zebrafish gene is Bai et al. 2010 PMID:20082716, which is
centred on the paralog gak but characterises zAux in parallel: domain structure
(PTEN-like region, clathrin-binding motif, J domain, no kinase), HeLa-cell
localisation of GFP-zAux, Drosophila rescue with J-domain dependence, and the
neural-restricted expression pattern.

Re-review 2026-09-28

Validation on entry reported one error: the GO:0072583 clathrin-dependent
endocytosis IEA row (GO_REF:0000117, ARBA) is no longer in the refreshed goa.tsv,
although the refreshed uniprot.txt still lists it (DR GO; GO:0072583; P:clathrin-dependent endocytosis; IEA:UniProtKB-ARBA.). The row was therefore
converted to action: NEW (original_reference_id pointing at the uniprot.txt DR
line) rather than deleted, because the zebrafish protein has direct in vivo evidence
for the process PMID:20082716 and the human ortholog carries the term by
IBA and IMP.

Row-level changes:

• GO:0004721 phosphoprotein phosphatase activity (IEA, UniRule UR000414225):
  MARK_AS_OVER_ANNOTATED retained, but the reason is now grounded: the UniRule fires
  on the PROSITE tensin-type phosphatase domain (residues 109-276, predicted
  phosphocysteine Cys218), no auxilin has demonstrated phosphatase activity, human
  UniProt hedges it ["May act as a protein phosphatase and/or a lipid phosphatase."]
  and the only functional test of zAux depended on the J domain PMID:20082716.
• GO:0016787 hydrolase activity (IEA): REMOVE -> MARK_AS_OVER_ANNOTATED, for
  consistency with the phosphatase row from the same rule; the fold is present, the
  activity is unproven, and the ATPase that powers uncoating belongs to Hsc70.
• GO:0030136 clathrin-coated vesicle (IEA SubCell): ACCEPT, now supported by the
  SUBCELLULAR LOCATION line and by zAux/clathrin co-localisation and clathrin
  co-aggregation in HeLa cells PMID:20082716.
• GO:0005829 cytosol (IDA, ZFIN): ACCEPT -> KEEP_AS_NON_CORE. The observation is a
  low-expression GFP-zAux pattern in HeLa cells PMID:20082716; a soluble
  pool is expected for a cycling co-chaperone, but the functional site is the coated
  vesicle and the evidence is heterologous overexpression.
• GO:0048471 perinuclear region of cytoplasm (IDA, ZFIN): ACCEPT -> KEEP_AS_NON_CORE.
  Same experiment; the perinuclear structures overlap clathrin and were "most likely
  representing the TGN" without a marker, so no MODIFY to a TGN term.
• Proposed (NEW) rows kept but re-evidenced with the uniprot.txt SUBUNIT line and
  the paper instead of deep-research paraphrases; evidence_type set to ISS to be
  honest that there is no zebrafish biochemistry: GO:0030544 Hsp70 protein binding
  and GO:0001671 ATPase activator activity (J-domain HPD motif present in the
  sequence, "RKAVLVVHPD"; SUBUNIT "Interacts with HSPA8/HSC70 in an ATP-dependent
  manner; this interaction stimulates the HSPA8's ATPase activity"), and GO:0072318
  clathrin coat disassembly (dAux rescue requiring the J domain).

Also rewritten: description (standalone biology, mentions the gak paralog and
neural expression), core_functions (single co-chaperone entry; cytosol dropped from
locations), references (added the zebrafish and human uniprot.txt files with
findings; reference_review on PMID:20082716 verified against the cached title),
suggested questions/experiments (phosphatase activity, clathrin binding, dnajc6 loss
in zebrafish). Validation ends with zero errors and no warnings.

Follow-up after PR review (2026-10-04)

core_functions now names ATPase activator activity (GO:0001671) as the molecular function instead of Hsp70 protein binding (GO:0030544), matching the core-function prose (auxilin recruits Hsc70 and stimulates its ATPase) and the guidance to prefer an informative activity over a binding term; the binding row remains in existing_annotations as the mechanistic basis. clathrin-dependent endocytosis (GO:0072583) was dropped from directly_involved_in because clathrin coat disassembly (GO:0072318), which remains, is the precise step auxilin performs and the broader term is its ancestor.

Quote cleanup (2026-10-04)

• Replaced the core_functions falcon deep-research quote ('The primary molecular function is not enzymatic in itself ...') with the contiguous source text, restoring the ** emphasis that had been stripped.
• No action, reason or term changed; just validate DANRE A0A8M9QG43 passes with zero errors.

2026-10-21 — Full re-review, independent of the ProtNLM2 predictions

ProtNLM2 output was not opened or used. Evidence: cached literature, UniProt, the
PTHR23172 PAINT file and family review, and a new sequence comparison
(A0A8M9QG43-bioinformatics/compare_isoforms_rereview.py, RESULTS-rereview.md;
no existing bioinformatics files were touched; there were none).

Identity of the benchmark accession

• A0A8M9QG43 is one of 26 unreviewed UniProt entries for zebrafish dnajc6
  (ZFIN:ZDB-GENE-080104-2), one per RefSeq transcript model. The UniProt
  gene-centric canonical is A0A8M9Q6X2 (979 aa, XP_021332834.1, annotation
  score 3). A0A8M9QG43 (974 aa, XP_021332835.1) differs only by a 5-residue
  deletion (GPPSP) after residue 66, N-terminal to every domain; phosphatase-like,
  C2, disordered clathrin-binding and J domains are intact, HPD at 938
  [file:DANRE/A0A8M9QG43/A0A8M9QG43-bioinformatics/RESULTS-rereview.md].
• So this is an alternative-splice isoform, not a fragment. Biology is that of the
  gene; differences in GOA between isoforms are a propagation artefact. The
  canonical A0A8M9Q6X2 receives GO_Central IBAs (clathrin binding, synaptic vesicle
  uncoating, clathrin-dependent endocytosis, cytoplasm, vesicle, postsynaptic
  density) that this isoform does not (QuickGO, 2026-10-21).
• No review exists for A0A8M9Q6X2 anywhere in genes/.

Evolutionary grounding

• PANTHER PTHR23172:SF4 (auxilin) for this entry, the canonical, and human DNAJC6
  (O75061). Bai et al. place zAux with mammalian auxilins rather than zGAK
  PMID:20082716.
• PAINT root PTN000571055 (Eukaryota): IBD clathrin binding, clathrin coat
  disassembly, vesicle, cytoplasm, intracellular membrane-bounded organelle.
  Vertebrate PTN002753975 (Euteleostomi): synaptic vesicle uncoating, postsynaptic
  density. No phosphatase activity at any node, a deliberate omission given the
  PTEN-like fold in every member.
• Family review interpro/panther/PTHR23172/PTHR23172-review.yaml agrees with SF4
  placement (it uses this accession as the SF4 representative) and with the
  non-catalytic reading of the PTEN-like domain. No disagreement.

Changes

• GO:0004721 phosphoprotein phosphatase activity and GO:0016787 hydrolase activity
  (IEA UniRule): MARK_AS_OVER_ANNOTATED -> REMOVE. The bovine auxilin 1 PTEN-like
  region has no detectable activity PMID:20826345,
  attributed to the P loop; zebrafish (CSDGR at C218) and human (CLDGR) share the
  same departure from the PTEN CX5R P loop. Contradicted, not merely unproven.
• GO:0030136 clathrin-coated vesicle: ACCEPT kept, but ARBA SUBCELLULAR LOCATION
  quote replaced by PMID:8524399 and PMID:20082716 (ARBA text is not evidence).
• cytosol, perinuclear region (ZFIN IDA): KEEP_AS_NON_CORE unchanged.
• NEW rows re-grounded: Hsp70 protein binding (ISS, PMID:8524399), ATPase
  activator activity (ISS, curated human UniProt + PMID:8524399), clathrin coat
  disassembly (now IBA-grounded on PTN000571055). ARBA SUBUNIT quotes dropped.
• Added NEW clathrin binding (GO:0030276), root-node IBD inherited by the canonical.
• Dropped the proposed GO:0072583 clathrin-dependent endocytosis row: it is an
  ancestor of GO:0072318 (OLS hierarchicalAncestors), so redundant. Did not propose
  GO:0016191 synaptic vesicle uncoating (descendant) or postsynaptic density; no
  zebrafish synaptic data.
• core_functions: added a clathrin-binding entry; supports switched to primary
  literature and PAINT.
• References added: PMID:8524399, PMID:20826345, PMID:20160091, PAINT tsv,
  RESULTS-rereview.md. None removed; uniprot.txt and deep-research entries carry
  review_notes about their reduced role.

Relative to review next

Canonical product of the same gene, A0A8M9Q6X2 (dnajc6, 979 aa), relation
same_gene_canonical; no review exists. Suggested folder genes/DANRE/dnajc6 (no
dnajc6 folder exists; this folder is keyed by accession). Closest characterised
ortholog human DNAJC6 (O75061) already has a COMPLETE review.

Modules: modules/synaptic_vesicle_endocytosis.yaml and
modules/synaptic_vesicle_cycle.yaml mention auxilin/DNAJC6.

just validate DANRE A0A8M9QG43: valid, no warnings.

2026-10-21 — Characterised comparator (PR #5304 review)

The same-gene relative reviewed in this batch now lives in genes/DANRE/dnajc6__A0A8M9Q6X2/ (renamed from the bare-symbol
folder so it does not shadow this accession's folder). That relative has no experimental data of its own, so the
experimentally characterised comparator for this gene is human DNAJC6 (auxilin), already reviewed in genes/human/DNAJC6/. It sits in the same
PANTHER subfamily (PTHR23172:SF4) as this target.

2026-10-21 — NEW rows withdrawn: already annotated on another product of the same gene

The re-review proposed NEW rows (GO:0030276) on this accession. GOA already carries the same terms on
another product of the same gene, reviewed in genes/DANRE/dnajc6__A0A8M9Q6X2/. GOA propagates gene-level
annotations to one representative product per gene, so adding them to a second product fills a
convention rather than a curation gap. The rows were withdrawn, following the rule applied across the
ProtNLM2 re-review cohorts (first raised in the review of PR #5303). The gene-level function is unchanged.