AKIP1 / BCA3 (Q9NQ31) — review notes

2026-10-03. PAINT no-IBA backlog. Provider: affinage (gates clear, good recall).

Shape

27 rows over 4 distinct terms; 23 are GO:0005515 IPI. Unlike AIRIM, the partners split three
informative ways:

AKIP1 is not an AKAP despite its name. It binds the PKA catalytic subunit's N-terminal A helix,
not the RII regulatory subunit:
PMID:15630084

The PDZ rows have a sequence basis. AKIP1 ends …F-P-V, a class II PDZ-binding motif
(Φ-x-Φ-COOH). Caveat recorded: the fragmentomics paper never names AKIP1 in its text (data are
supplementary), and these are peptide–domain affinities, not cellular interactions.

An unanswered question worth flagging: the same C-terminus binds both the PKA catalytic subunit
and PDZ domains. Competitive?

The InterPro mapping chose the wrong NF-κB branch

GO:1901222 regulation of non-canonical NF-κB (IEA, InterPro IPR033214). But the mechanism is
PKA phosphorylation of p65 Ser-276 — the canonical subunit:
PMID:20562110

Non-canonical NF-κB runs through NIK, p100 processing and RelB. InterPro's own description of
IPR033214 talks about p65 colocalization and never mentions the non-canonical pathway, so the
branch isn't supported even by the mapping's source. MODIFY → GO:0043122 regulation of
canonical NF-κB signal transduction — deliberately direction-neutral, since InterPro notes splice
variants that repress or enhance.

Recorded, not added

Mitochondrial localization: PMID:23319652
— shown in rat myocytes, curated for no species (mouse Akip1 has only nuclear terms), and the
paper itself notes PMID:23319652.
Transfer isn't safe → knowledge gap, not NEW.

Rac1 (PMID:17227220) and β-catenin (PMID:30936461) binding: single studies each → questions.

Process

• Four papers (20562110, 23319652, 17227220) were sitting in this worktree as untracked
  old-format files (reference_id: header) from an earlier session. Re-fetched with
  force=True rather than committing stale-format records.
• Quotes checked whitespace-collapsed and otherwise character-exact: 14 distinct, 0 failures.

Round 2 (review feedback)

1. The NF-κB MODIFY rested on uncited InterPro paraphrases. It now quotes PMID:23319652, which
   states the isoform dependence directly: AKIP1b recruits SIRT1 and represses; AKIP1a enhances.
   (The reviewer pointed to line 101; the statement is a few sentences on — I checked rather than
   quoting the line number.)
2. So GO:0043123 positive regulation, not the neutral GO:0043122. UniProt's canonical isoform
   (Q9NQ31-1) is AKIP1a, which enhances. The repression belongs to AKIP1b, a different product.
3. PMID:18178962 was missing — the paper UniProt cites with ECO:0000269 for exactly this function,
   cited by no GO annotation. AKIP1 binds p65 directly, PKAc is in the AKIP1·p65 complex, knockdown
   abolishes the response.

That third point changed the core function. AKIP1 binds both the kinase and its substrate, so it
gets GO:0035591 signaling adaptor activity as NEW and as the core MF — the same treatment as AKAP7's
PKA-to-channel bridge — with GO:0034236 kept as a second core function. No IPI partner accession,
because the abstract doesn't say which species the co-IP used.

Not added: a nuclear-retention process term. Both results behind it are overexpression, and whether
AKIP1 anchors or blocks export isn't shown. Recorded as a question.