Native epithelial function of human BAIAP2L2/Pinkbar: an evidence audit

Iteration 1 — literature synthesis (no primary data provided).
Scope: distinguish direct BAIAP2L2 evidence from I-BAR-family/paralog inference; evaluate the
unresolved native intestinal/renal epithelial function; recommend GO curation stance.

1. Summary (answer to the research question)

The native epithelial function of BAIAP2L2/Pinkbar in human intestine and kidney remains unresolved.
There is solid direct biochemical and cell-biological evidence that BAIAP2L2 is an epithelial-expressed
I-BAR protein that binds membranes, generates planar (not tubular) membrane sheets, and localizes to
Rab13-positive vesicles and intercellular junctions — i.e., a demonstrable membrane-flattening/scaffold
molecular function. However, the corresponding in-vivo physiological output is not established:
Baiap2l2-knockout mice have normal kidney/colon morphology and normal electrolyte homeostasis under basal
conditions only, and no challenge, aging, or paralog-compound-knockout experiments have been reported.
The single well-established organ-level loss-of-function phenotype is cochlear (stereocilia maintenance /
deafness), which is mechanistically distinct from the epithelial junctional role.

2. Direct BAIAP2L2 evidence vs. paralog/I-BAR inference

2a. DIRECT epithelial evidence (BAIAP2L2 itself)

All direct native-epithelial evidence traces to one foundational paper:

• Pykäläinen et al., 2011, Nat Struct Mol Biol — PMID 21743456; DOI 10.1038/nsmb.2079 (cached/validated abstract).
• Specifically expressed in intestinal epithelial cells.
• Localizes to Rab13-positive vesicles and to the plasma membrane at intercellular junctions.
• I-BAR/IMD domain does not tubulate membranes but promotes planar membrane sheets.
• Structure: relatively flat lipid-binding interface; assembles into sheet-like oligomers (crystals + solution)
  → mechanistic basis for planar deformation.
• Note: phosphoinositide dependence is the expected I-BAR mechanism and is consistent with the flat basic
  lipid interface described here, but the abstract states "lipid-binding interface" rather than naming a specific
  PIP species; a strict phosphoinositide-specificity claim for BAIAP2L2 should be flagged partially inferred.

2b. PARALOG / I-BAR-FAMILY inference (NOT direct native-tissue proof)

• Sudhaharan et al., 2016, J Cell Sci — PMID 27278019 (cached). Rif (Rho-family GTPase) signals through
  I-BAR proteins; explicitly groups IRSp53 (BAIAP2), IRTKS (BAIAP2L1) and Pinkbar (BAIAP2L2) as a morphology
  module. BAIAP2L2-specific downstream biology here is inferred from the family.
• Tholen et al., 2023 — PMID 36520027: BAIAP2L2 binds BAIAP2 and BAIAP2L1 (mass spec) → heteromeric
  I-BAR assemblies plausible; supports redundancy.
• General I-BAR mechanism (negative-curvature/PIP2 binding of IRSp53) is a family property; applying it to
  BAIAP2L2 is inference except where Pinkbar's divergent flat-interface/planar behavior is directly shown (2a).

3. Why knockout mice look normal — and what has NOT been tested

Tholen et al., 2023 (PMID 36520027) (title: "HNF1β-associated cyst development and electrolyte disturbances
are not explained by BAIAP2L2"; journal Pflügers Archiv / Eur J Physiol — DOI uncertain, flag for verification):

• BAIAP2L2 is a direct transcriptional target of HNF1β (luciferase reporter validated).
• Baiap2l2-KO kidney epithelial cells: normal F-actin at cell–cell contacts; normal polarized 3D spheroids with lumen.
• Baiap2l2-KO mice: normal kidney and colon morphology; normal serum and urine electrolytes — but explicitly
  "under physiological conditions."

Interpretation of the null phenotype (unresolved, not disproven):
1. Paralog redundancy — BAIAP2L2 physically binds BAIAP2 and BAIAP2L1; single KO may be buffered.
Untested: compound/double/triple I-BAR knockouts.
2. Challenge-dependence — only basal conditions assayed. Untested: low-Mg²⁺/high-salt diets, ischemia-reperfusion
or nephrotoxic injury, colitis/DSS challenge, infection, or aged cohorts (note the cochlear phenotype is
progressive, becoming overt by 8 months — an epithelial phenotype could likewise be age- or stress-gated).
3. Sub-morphological/functional readouts not probed — junction dynamics, barrier resealing after wounding,
TER/paracellular permeability, Rab13-cargo trafficking kinetics.

4. Candidate epithelial partners, cargo, and structures

No direct epithelial cargo or a specific junctional structure built by BAIAP2L2 has been identified.
The strongest mechanistic hypothesis: a Rab13-coupled, lipid-binding planar scaffold that flattens/stabilizes
membrane at nascent intercellular junctions, potentially in heteromeric I-BAR assemblies.

5. The established (contrasting) cochlear function — keep separate

• Carlton et al., 2021 — PMID 33151556: KO mice lose stereocilia rows 2/3, lose MET current, deaf by 8 months;
  tip localization requires MYO15A + EPS8. (eLife/PLoS-class; DOI flag for verification.)
• Yan et al., 2022 — PMID 34346063: BAIAP2L2 is a "row 2 complex" component (binds EPS8L2, TWF2, CAPZB2, CIB2);
  tip localization requires CIB2.
• Halford et al., 2022 — PMID 35044843: BAIAP2L2–EPS8–MYO15A tripartite complex; row-2 tip retention depends on
  Ca²⁺ influx through transduction channels.
• Ikäheimo et al., 2024 — PMID 39037943: in stereocilia-fusion pathology BAIAP2L2 mislocalizes from tips.

These describe a protrusion-tip actin-scaffold function (negative-curvature geometry), mechanistically distinct
from the planar-junction epithelial activity. BAIAP2L2 thus operates in two different membrane geometries.

6. Supported vs. refuted hypotheses

Supported (direct evidence):
- BAIAP2L2 binds membranes and generates planar sheets via a flat lipid interface + sheet oligomers (PMID 21743456).
- BAIAP2L2 localizes to Rab13 vesicles and epithelial junctions (PMID 21743456).
- BAIAP2L2 is an HNF1β target and binds I-BAR paralogs (PMID 36520027).
- BAIAP2L2 maintains cochlear stereocilia (PMID 33151556, 34346063, 35044843).

Refuted / not supported (basal conditions):
- BAIAP2L2 is required for basal renal/colonic morphology or electrolyte homeostasis — refuted (PMID 36520027).
- BAIAP2L2 explains HNF1β cyst/electrolyte phenotype — refuted (PMID 36520027).

Unresolved (open):
- A specific native epithelial biological process / cargo / junctional structure.
- Challenge- or age-dependent epithelial phenotypes.
- Strict phosphoinositide-species specificity of BAIAP2L2 membrane binding.

7. GO-style curation recommendation

1. Support a membrane-bending/scaffold molecular function. Direct structural + in-vitro data justify MF terms such as
   phosphatidylinositol/phospholipid binding (with the caveat that a specific PIP species is inferred) and a
   membrane deformation / planar membrane-sheet scaffolding activity. This is the best-evidenced, BAIAP2L2-specific claim.
2. Keep epithelial biological-process terms as inferred / non-core. Localization to junctions/Rab13 vesicles is
   direct, but the process (junction assembly, epithelial morphogenesis, barrier function) has no in-vivo support and
   a KO shows no basal phenotype → annotate as inferred/IEA-or-IMP-pending / non-core, not established.
3. Native epithelial function = "unresolved." Do not assert a core epithelial physiological role; the honest curation
   status is candidate/unresolved pending challenge and compound-KO data.
4. Do not transfer stereocilia partners (EPS8/CIB2/MYO15A etc.) to epithelium — these are tissue-specific and would
   be an over-annotation.

8. Limitations & future directions

• Limitation: the entire direct epithelial case rests on a single 2011 paper; independent replication of planar-sheet
  activity and junction localization in human intestinal/renal tissue is lacking.
• DOI caveat: PMIDs are validated against cached PubMed abstracts. DOIs are provided where confidently known
  (Pykäläinen 2011 = 10.1038/nsmb.2079); other DOIs are flagged uncertain and should be verified before publication.
• Priority experiments: (i) compound BAIAP2/BAIAP2L1/BAIAP2L2 knockouts in gut/kidney; (ii) epithelial challenge
  models (injury, colitis, low-Mg/high-salt, aging); (iii) barrier/junction-dynamics and Rab13-cargo trafficking assays;
  (iv) defined phosphoinositide-binding specificity for the BAIAP2L2 I-BAR domain; (v) human intestinal/renal organoid
  localization and knockout phenotyping.

Key citations (PMID / DOI / cache status)

• PMID 21743456 — Pykäläinen 2011, Nat Struct Mol Biol — DOI 10.1038/nsmb.2079 — cached ✔ (core direct evidence)
• PMID 36520027 — Tholen 2023 — DOI uncertain (flag) — cached ✔ (KO phenotype, HNF1β, paralog binding)
• PMID 27278019 — Sudhaharan 2016, J Cell Sci — DOI uncertain (flag) — cached ✔ (I-BAR/Rif family inference)
• PMID 33151556 — Carlton 2021 — DOI uncertain (flag) — cached ✔ (cochlear KO)
• PMID 34346063 — Yan 2022 — DOI uncertain (flag) — cached ✔ (row-2 complex)
• PMID 35044843 — Halford 2022 — DOI uncertain (flag) — cached ✔ (tripartite complex, Ca²⁺-dependent tip retention)
• PMID 39037943 — Ikäheimo 2024 — DOI uncertain (flag) — cached ✔ (stereocilia fusion mislocalization)
• PMID 34695828 — Shakery 2023 — DOI uncertain (flag) — cached ✔ (non-native cancer-cell, correlative)
• PMID 21795389 — Abou-Zeid 2011 / PMID 24377937 — Zahraoui 2014 — cached ✔ (Rab13 junction biology, context)