ABTB3 Gene Review Notes

2026-06-03 - Proteostasis PN batch review

ABTB3 was newly fetched for the human Proteostasis PN batch. The local GOA set
contains five annotations: membrane, protein heterodimerization activity, PDZ
domain binding, glutamatergic synapse, and nucleoplasm. UniProt also lists
Ensembl-transferred biological-process annotations for exploration behavior,
protein stabilization, and synaptic transmission, glutamatergic, but these were
not present in the fetched GOA TSV.

Deep research status: just deep-research-falcon human ABTB3 --fallback perplexity-lite was run. Falcon timed out after 600 seconds, then
perplexity-lite failed with a Perplexity API quota 401. Per project guidance,
I recorded the evidence synthesis in ABTB3-deep-research-manual.md rather than
creating a fake provider-named deep-research file.

The strongest gene-specific biology comes from mouse Btbd11. Bygrave et al.
identify Btbd11 as an inhibitory interneuron-specific synapse-enriched protein
that is highly conserved and binds Psd-95: PMID:37261953 and
PMID:37261953. The paper directly supports the
orthology-derived human GOA terms for PDZ domain binding and glutamatergic
synapse: PMID:37261953 and PMID:37261953.

The same mouse study supports adding two missing, conservative BP terms by
orthology. For protein stabilization, the relevant evidence is Psd-95 FRAP and
puncta-size data: PMID:37261953 and PMID:37261953. For glutamatergic synaptic
transmission, the direct phenotype is reduced excitatory signaling onto PV
interneurons: PMID:37261953.

Conservative calls:

• Keep GO:0016020 membrane as non-core. UniProt has predicted membrane
  association, but the functional localization is better captured by
  GO:0098978 glutamatergic synapse.
• Mark GO:0046982 protein heterodimerization activity as over-annotated. It is
  inferred from a histone-fold InterPro signature [UniProt:A6QL63 "InterPro;
  IPR009072; Histone-fold."], whereas the direct interaction evidence maps the
  PSD-95 interaction to the C-terminal PDZ-binding motif.
• Keep GO:0005654 nucleoplasm as non-core. It is HPA immunofluorescence-derived
  and does not explain the currently supported synaptic role.
• Do not add GO:0035640 exploration behavior; that is a downstream organismal
  phenotype from mouse knockout behavior, not the core gene-product function.

PN projected context

The Proteostasis PN projection proposes GO:1990756 ubiquitin-like ligase-substrate adaptor activity for ABTB3 from
Ubiquitin Proteasome System|E3 ubiquitin and UBL ligases|Cul3 substrate receptor|BTB-BACK, variant|ankyrin, transmembrane
[file:projects/PROTEOSTASIS/reports/pn_projection/pn_projected_new_to_goa.tsv].
The curated parent mapping says the Cul3 substrate receptor group maps to
GO:1990756, but the audit marks it as requiring manual gene-level review before
gene review changes
[file:projects/PROTEOSTASIS/reports/pn_mapping_audit/current_mapping_scrutiny.tsv].

I did not add GO:1990756 for ABTB3. The evidence found here supports PDZ-domain
binding and synaptic PSD stabilization; it does not show ABTB3-CUL3 binding, CRL3
complex membership, a ubiquitinated substrate, or substrate-adaptor activity.
This is analogous to the KCTD18 exclusion note in the same PN mapping: domain
architecture alone is not enough for a gene-level GO:1990756 assertion.