Proteostasis Review Batch 8 — Gene Selection

Date: 2026-06-13
Branch: claude/proteostasis-gene-review-3zv0vt

Context

Batches 1–7 are complete (pr-1217, 2026-06-03, 2026-06-06, 2026-06-07,
2026-06-07b, 2026-06-07c, 2026-06-11). Batch 7 covered the ER
proteostasis branch (SRP/translocon, EMC/GET insertion, glycoprotein-folding
QC, ERAD ubiquitin machinery). Batch 8 moves into the Ubiquitin–Proteasome
System (UPS) branch, selecting 50 unreviewed PN candidates
(ok_for_propagation_to_go scope, no prior *-ai-review.yaml).

Theme: Cullin–RING ligase (CRL) substrate-recognition & assembly modules

The UPS branch of the PN workbook is intentionally domain-heavy and inclusive,
built around the E3 ubiquitin/UBL ligase scaffold. This batch walks the largest
and most systematically organized CRL substrate-receptor family — the F-box
proteins of the SCF (SKP1–CUL1–F-box / CRL1) complex — plus the CRL4 core
and CRL assembly/regulation machinery. It directly exercises the UPS-branch
E3 ubiquitin and UBL ligases -> CRL family part of the PN taxonomy.

F-box proteins are grouped by their substrate-recognition module:

• FBXL — F-box + leucine-rich repeats
• FBXW — F-box + WD40 repeats
• FBXO — F-box "other" (diverse/no canonical second domain)

Selected genes (50)

F-box / leucine-rich-repeat receptors — FBXL (14)

1. FBXL3 — SCF receptor for CRY1/CRY2 (circadian clock)
2. FBXL5 — iron-sensing receptor for IRP2/IREB2 (FBXL5–SKP1 hemerythrin)
3. FBXL6
4. FBXL7 — receptor for AURKA, survivin
5. FBXL8
6. FBXL12 — receptor for CDKN1B/p27, ALDH3A1
7. FBXL13
8. FBXL14
9. FBXL15
10. FBXL16
11. FBXL17 — receptor for BACH1; SCF dimerization QC
12. FBXL18
13. FBXL20 (SCRAPPER) — synaptic RIM1 turnover
14. FBXL22

F-box / WD40-repeat receptors — FBXW (8)

1. FBXW2
2. FBXW4 (dactylin)
3. FBXW5 — receptor for SASS6, TSC2, EPS8
4. FBXW7 (CDC4) — receptor for MYC, cyclin E, NOTCH, JUN (tumor suppressor)
5. FBXW8 — CRL7 receptor (with CUL7)
6. FBXW9
7. FBXW10
8. FBXW12

F-box / "other" receptors — FBXO (22)

1. FBXO2 — ER glycoprotein lectin F-box (N-glycan recognition, ERAD)
2. FBXO5 (EMI1) — APC/C inhibitor (not a canonical SCF receptor)
3. FBXO6 — ER glycoprotein lectin F-box (ERAD)
4. FBXO7 (PARK15) — PINK1/Parkin mitophagy; PI31 proteasome regulator
5. FBXO8
6. FBXO10
7. FBXO15
8. FBXO16
9. FBXO17
10. FBXO21
11. FBXO22 — receptor for KDM4A, PTEN, BACH1
12. FBXO25
13. FBXO27 — ER glycoprotein lectin F-box
14. FBXO30
15. FBXO33
16. FBXO34
17. FBXO36
18. FBXO39
19. FBXO40
20. FBXO41
21. FBXO43 (EMI2/ERP1) — APC/C inhibitor, meiotic CSF
22. FBXO47

CRL4 core & CRL assembly/regulation (6)

1. DDB1 — CRL4 adaptor (DDB1–CUL4 scaffold for DCAF receptors)
2. DDB2 — CRL4 substrate receptor (UV-damaged DNA recognition)
3. DDA1 — DET1/DDB1-associated, CRL4 stabilizer
4. DTL (CDT2) — CRL4 substrate receptor (CDT1, p21, SET8 degradation)
5. CAND2 — CAND1 paralog; cullin–RBX sequestration / CRL exchange factor
6. GLMN (glomulin) — RBX1-binding CRL assembly regulator (FBXW7/CUL7 control)

Method

Same as batches 5–7: fetch-gene scaffolding (uniprot + goa + seeded
-ai-review.yaml + cached publications + PANTHER family), per-annotation review
against GO guidelines with verbatim supported_by, populated
description/core_functions/suggested_questions/suggested_experiments,
reference reference_review adjudication, then uv run linkml-validate +
uv run ai-gene-review validate.

Curation watch-points

• F-box = substrate receptor, not the catalytic core. The catalytic RING
  activity of an SCF complex lives on RBX1/RBX2, not the F-box protein. Prefer
  GO:1990756 ubiquitin-like ligase-substrate adaptor activity /
  GO:0061630 ubiquitin protein ligase activity in the SCF context plus
  SCF-dependent proteasomal ubiquitin-dependent protein catabolic process
  (GO:0031146) and part_of the SCF ubiquitin ligase complex (GO:0019005).
  Do not treat bare protein binding (GO:0005515) as a core function —
  replace/supplement with the specific substrate-recognition role.
• SKP1 binding is the defining F-box feature. GO:0019904/SKP1 binding or
  F-box domain binding annotations are mechanistically central and should be
  kept (ACCEPT/non-core), but the core statement is substrate-specific
  ubiquitination.
• Lectin F-box subfamily (FBXO2/FBXO6/FBXO27, and FBXO17/FBXO44). These bind
  N-linked high-mannose glycans on misfolded glycoproteins and feed ERAD — use
  carbohydrate/high-mannose oligosaccharide binding + ERAD/SCF-dependent… catabolic process, a genuine proteostasis-core function.
• Non-canonical "F-box" members. FBXO5/EMI1 and FBXO43/EMI2 are APC/C
  inhibitors, not SCF substrate receptors despite the F-box motif — their core
  biology is cell-cycle/meiotic APC/C regulation; treat SCF/ubiquitin-ligase
  propagations skeptically.
• Tumor-suppressor / well-studied receptors (FBXW7, FBXL5, FBXL3,
  FBXO7, FBXO22) have rich experimental annotations; ground the core
  function in IDA/IMP/IPI literature, keep the many IEA/IBA terms as ACCEPT or
  non-core appropriately.
• Domain-based / poorly characterized paralogs. Many FBXL/FBXO/FBXW members
  rest largely on IBA/IEA family propagation with little direct evidence; use
  KEEP_AS_NON_CORE/UNDECIDED for substrate-specific claims that are not
  supported, and never REMOVE an experimental annotation whose full text was
  not read.
• CRL4 core vs receptor. DDB1 is the adaptor/scaffold (not a receptor);
  DDB2 and DTL/CDT2 are DCAF substrate receptors. DDA1 stabilizes CRL4.
  CAND2 (CAND1 paralog) and GLMN are regulators of CRL assembly, not
  ligases — avoid assigning them ubiquitin-ligase catalytic activity.