C18orf32 (Q8TCD1) notes

2026-10-08 review session (MICROPROTEINS Tier 2)

Gene: C18orf32, UPF0729 protein, 76 aa, PE1 (protein level), HGNC protein-coding, MANE
NM_001035005.4. Canonical-ORF small protein (not an alternative-ORF peptide), so the
standard genes/human/C18orf32/ folder is correct. Domain: Pfam DUF4512 (PF14975),
PANTHER PTHR13456 (UPF0729). N-terminal hydrophobic region (1-37), C-terminal disordered
region (46-76).

Literature found (PubMed esearch "C18orf32": 5 hits; plus GPI screen)

• PMID:12761501 (Matsuda 2003, abstract only): genome-scale overexpression screen of
  150,000 cDNAs in HEK293 with an NF-kB luciferase reporter; source of the "putative
  NF-kappa-B-activating protein" name and the HMP row. No follow-up on C18orf32 in the
  abstract PMID:12761501.
• PMID:23864651 (Huang 2013): MYTH screen with GLP-1R bait; C18orf32 not named in the
  cached text (likely in a supplementary interactor table). Source of IPI GLP1R row.
• PMID:29275994 (Bersuker 2018, full text): APEX2 proximity labelling of LD proteome.
  Key points:
• ER under basal conditions, redistributes to LDs with oleate PMID:29275994.
• Endogenous protein on LDs, antibody validated by KO PMID:29275994.
• N-terminal hydrophobic domain necessary and sufficient for ER/LD targeting.
• Constitutively degraded by gp78 (AMFR)/derlin-1 ERAD, VCP- and proteasome-dependent;
  DERL1 and AMFR co-purify PMID:29275994. These are the IPI rows: they reflect
  C18orf32 being an ERAD substrate, not an MF of C18orf32.
• KO had no LD phenotype PMID:29275994; minor lipidomic changes.
• PMID:29255114 (Liu 2018, JCB, full text; found via PubMed): haploid screen for factors
  needed for GPI-inositol deacylation by PGAP1 identified C18orf32 among 7 genes; KO gives
  partial PIPLC resistance PMID:29255114. Molecular function unknown PMID:29255114. PGAP1
  expression/stability/localization unchanged in KO cells (Fig 3).
• PMID:35107634 (Salian 2022, abstract only): homozygous c.90dupC (p.Phe31Leufs*3) in two
  siblings with neurodevelopmental disorder (GPIBD25); KO HEK293 PIPLC-resistance rescued by
  WT but not mutant cDNA; mutant protein in ER and nuclear aggregates. Source of the EXP ER
  row (abstract says the mutant was in ER; WT localization presumably in full text).
• PMID:41351175 (Kumar 2025, Cell Commun Signal, full text): single-group study claiming
  C18orf32 on SEC22B-positive secretory autophagosomes, binding SEC22B via its C-terminus,
  and required for LD secretion; liver knockdown causes steatosis. Interesting but
  unreplicated; much of the binding evidence is docking/in vitro affinity; Fig 1E is a GO
  enrichment of interactors. Not used for NEW annotations.
• PMID:41218492 (chimeric RNA RPL17-C18orf32) and PMID:37438770 (blood transcriptome),
  PMID:27833855 (KLHL23 readthrough): not about the C18orf32 protein's function; not cited.

Assessment

• Locations ER and LD: well supported by two independent groups (IDA, EXP); ACCEPT all
  five CC rows.
• Protein binding x3: REMOVE (bare protein binding). DERL1/AMFR binding reflects the
  protein being an ERAD client; GLP1R is a MYTH hit with a hydrophobic segment.
• NF-kB HMP: overexpression reporter hit from a genome-scale screen; no follow-up in
  23 years; MARK_AS_OVER_ANNOTATED.
• GPI anchor: the KO phenotype (partial PIPLC resistance; rescue) shows necessity for
  efficient PGAP1-mediated inositol deacylation, but nothing says what C18orf32 does.
  PGAP1 performs the deacylation. Per the participation test, no NEW BP is proposed;
  raised as a question. No MF term fits.

2026-11-05 ORGANELLES ER batch 2 (endoplasmic reticulum rows)

Rows: GO:0005783 EXP (PMID:35107634), IDA (PMID:29275994), IEA (GO_REF:0000044, SL-0095). No ER sub-compartment row existed.

Checks:
- IEA source: UniProt "SUBCELLULAR LOCATION: Endoplasmic reticulum {ECO:0000269|PubMed:29275994, ECO:0000269|PubMed:35107634}" [file:human/C18orf32/C18orf32-uniprot.txt], so the keyword is curated from the two papers behind the experimental rows.
- PMID:29275994 (full text): "In the absence of oleate, GFP-tagged c18orf32 was present in the ER (Figure S7A)."; the N-terminal region is required: "In contrast, the c18orf32(Δ1-37)-GFP mutant was diffusely distributed throughout the cell and was absent from LDs (Figure 6I)."; endogenous protein in "ER-enriched pellet fractions"; ERAD by gp78/derlin-1 ("Affinity purification revealed that c18orf32-S interacts with both derlin-1 and gp78").
- PMID:41351175 (full text): N-terminal helix, "The hydrophobic segment (residues 7–15) likely inserted into the lipid droplet’s core".
- No signal peptide, no TM feature in UniProt (REGION 1..37 "Necessary for its localzation to the endoplasmic reticulum and lipid droplets").

Outcome: all three root rows stay ACCEPT (constitutive ER residence, not transit), each with a dated ORGANELLES sentence. NEW GO:0005789 ER membrane (IC, PMID:29275994) added: no signal peptide plus an N-terminal hydrophobic anchor that is necessary and sufficient for targeting and that partitions onto the LD monolayer (cytosolic-leaflet insertion), the endogenous protein in the membrane pellet only, and membrane ERAD machinery. Judgement call: the anchor may be monotopic (amphipathic/hairpin) rather than spanning; it is still a membrane-embedded pool, not a recruited cytosolic one. core_functions unchanged.