# ABTB2 notes

## 2026-06-03 Proteostasis PN review pass

- Started from `just fetch-gene human ABTB2`, which seeded three GOA rows: InterPro-derived protein heterodimerization, IntAct/YWHAE protein binding, and HPA nucleoplasm localization.
- Deep research attempt: `just deep-research-falcon human ABTB2 --fallback perplexity-lite` timed out for Falcon after 600 seconds; fallback to `perplexity-lite` failed with a Perplexity API quota error. Manual fallback evidence was written to `ABTB2-deep-research-manual.md` per project instructions.
- PN context places ABTB2 under `Ubiquitin Proteasome System > E3 ubiquitin and UBL ligases > Cul3 substrate receptor > BTB-BACK, variant > ankyrin` and projects `GO:1990756 ubiquitin-like ligase-substrate adaptor activity` [file:projects/PROTEOSTASIS/reports/pn_projection/pn_projected_annotations.tsv "ABTB2		Ubiquitin Proteasome System|E3 ubiquitin and UBL ligases|Cul3 substrate receptor|BTB-BACK, variant|ankyrin"].
- Conservative nomenclature check: older BPOZ-2 papers are useful background for BTB/POZ substrate-adaptor biology, but I did not use them as primary ABTB2 evidence because BPOZ-2 nomenclature overlaps with ABTB1 and needs identifier-level confirmation before being projected to Q8N961. The cached TdT abstract says "BPOZ-2 bound to E3 ligase CUL3 in vitro and in vivo" [PMID:18429817 "BPOZ-2 bound to E3 ligase CUL3 in vitro and in vivo"], but this was not the basis for the ABTB2 `GO:1990756` decision.
- Direct ABTB2 evidence comes from the 2025 PDAC paper. The abstract states that "ABTB2 interacts with tumour necrosis factor receptor-associated protein 1 (TRAP1), promoting its ubiquitin-dependent degradation" [PMID:41322190 "ABTB2 interacts with tumour necrosis factor receptor-associated protein 1 (TRAP1), promoting its ubiquitin-dependent degradation"]. The results section confirms ABTB2-TRAP1 co-IP and TRAP1 ubiquitination [PMID:41322190 "Co-immunoprecipitation using Abs specific to ABTB2 and TRAP1, followed by western blotting, confirmed the specific interaction between ABTB2 and TRAP1"] [PMID:41322190 "multiple mono-ubiquitinated TRAP1 (Figures 8F and S10) in ABTB2-OE cells"]. The discussion adds "We also found that ABTB2 interacts with cullin-3 (Figure S9)" [PMID:41322190 "We also found that ABTB2 interacts with cullin-3 (Figure S9)"].
- Curation conclusion: accept the PN-projected `GO:1990756` for ABTB2 as a new, direct, gene-specific molecular-function annotation, but do not overstate ABTB2 as a catalytic ubiquitin ligase. Mark existing generic binding/heterodimerization rows as over-annotated and keep nucleoplasm as non-core localization.