couscous (Salpingoeca rosetta, F2UJ78, PTSG_07368) — review notes

Automated deep research was unavailable for this gene (no deep-research
provider API keys in the curation environment). These notes are the reviewer's
own reading of the cached primary literature in publications/. No
*-deep-research-*.md file exists for this gene.

Identity

• UniProt F2UJ78 (TrEMBL, unreviewed), 866 aa, ORF PTSG_07368, GenBank EGD77026,
  RefSeq XP_004990866 / XM_004990809.
• UniProt name: "Apple domain-containing protein". Domains: PAN/Apple (PF00024,
  PS50948, aa 18-114), PF11051 Mannosyl_trans3 (two hits), InterPro IPR022751
  Alpha_mannosyltransferase, IPR029044 nucleotide-diphospho-sugar transferase fold.
• PANTHER PTHR31646 "ALPHA-1,2-MANNOSYLTRANSFERASE MNN2" (subfamily SF1). The family
  also contains MNN5 (SF6) - i.e. the fungal Mnn2/Mnn5 alpha-1,2-mannosyltransferase
  clade (UniProt: "Belongs to the MNN1/MNT family" via ARBA).
• N-terminus MARVVGLDVLAPMVLVACSLIAVSVSA is a short hydrophobic segment: SignalP calls
  a signal peptide (1-27); ARBA calls a type II signal anchor. Unresolved whether it
  is cleaved (secreted/luminal) or retained as a membrane anchor.
• Mutant allele couscous^lw1 was originally named "Branched" in Levin et al. 2014.

Primary paper: Wetzel et al. 2018 eLife (PMID:30556809, full text cached)

Phenotype (demonstrated)

• Class C rosette-defect mutant; never forms rosettes, forms amorphous clumps.
  PMID:30556809
• Clumps re-form by aggregation (not division), are shear sensitive; up to 75 cells in
  30 min for Couscous. PMID:30556809
• Mutant cells also stick to wild-type cells. PMID:30556809

Mapping and rescue (demonstrated)

• Single-nucleotide deletion on supercontig 22, frameshift -> early stop.
  PMID:30556809
• Transgenic rescue with couscous-mTFP or mTFP-couscous (not mutant allele or mTFP).
  PMID:30556809

Predicted biochemistry (NOT demonstrated)

• Alpha-mannosyltransferase domain related to yeast/Candida MNN2 (28%/35% identity),
  conserved DXD. PMID:30556809
• No enzymatic assay, no donor/acceptor identified. PMID:30556809
• PAN/Apple domain predicted. PMID:30556809

Localization (tagged overexpression; NOT Golgi)

• Couscous-mWasabi in puncta through cytosol, collar and cell membrane.
  PMID:30556809
• Explicitly not Golgi, possibly ER (untested), with caveats about tag/overexpression.
  PMID:30556809
  PMID:30556809
• Contrast: Jumble (the other predicted GT) does localize to the Golgi-like apical region.

Glycosylation phenotype (demonstrated, mechanism unresolved)

• 21 of 22 lectins unchanged; jacalin basal-pole staining lost in Couscous (and Jumble),
  apical staining normal. PMID:30556809
• Rescue restores jacalin pattern. PMID:30556809
• Authors leave open trafficking vs glycosylation. PMID:30556809

Effect on Rosetteless (demonstrated, downstream)

• Rosetteless localizes basally but is not upregulated/secreted on RIF induction.
  PMID:30556809

Hypotheses (not tested)

• Regulatory glycosylation of adhesion molecules (cadherins); protective glycocalyx
  layer; analogy to yeast MNN2 role in flocculation.

Other literature

• Levin et al. 2014 (PMID:25299189): original screen; the "Branched" mutant (= Couscous)
  forms branched chain colonies. PMID:25299189
• Larson et al. 2020 PNAS (PMID:31896587): frames couscous as an ECM regulator.
  PMID:31896587
• Combredet et al. 2025 Cell Rep (PMID:41037400, abstract only): CRISPR KO of couscous
  (reverse-genetic confirmation, details not in cached abstract); couscous transcript
  regulated by Warts/Yorkie. PMID:41037400
  PMID:41037400

Curation reasoning summary

• MF alpha-1,2-mannosyltransferase: homology-only (TreeGrafter, fungal MNN2 family,
  conserved DXD). Accept as best-supported prediction; flag unverified.
• Golgi terms: family default; directly contradicted (with caveats) by tagged
  localization -> UNDECIDED.
• mannan biosynthetic process: GO definition is plant hemicellulose; fungal cell-wall
  mannan usage does not apply to a choanoflagellate -> REMOVE.
• No NEW process annotations: rosette development / cell adhesion phenotypes are
  necessity evidence, substrate unknown; GO lacks a rosette/clonal colony development
  term (proposed as new term instead).