smad3b notes (Danio rerio, SMAD family member 3b; UniProt Q8AY16)

2026-09-28 — session log (DANRE_DUPLICATION batch 3, random sample)

Deep research: not available for this gene (Edison/Falcon returned 402 Payment Required;
the OpenAI key is invalid). Not attempted, per instructions. No -deep-research-*.md file
exists. The literature search was done by hand (Europe PMC; same queries as in
../smad3a/smad3a-notes.md).

Accession: Q8AY16 (TrEMBL, 423 aa) holds all 19 GOA rows. Its only experimental row is the
Tob1a protein-binding IPI (PMID:16890162). Paralog: smad3a (Q8AY15). Human ortholog SMAD3.
The sequence analysis is shared with smad3a: ../smad3a/smad3a-bioinformatics/RESULTS.md.

TGD origin

See smad3a notes. PANTHER TGD_tree; synteny and phylogeny in PMID:27703851
PMID:27703851.

Protein

• smad3b vs human SMAD3 93.0%, vs gar SMAD3 93.9%, vs smad3a 94.1%.
• Differences concentrate in the linker (82.3% identical to human, vs 92.7% for smad3a); 2-aa
  linker deletion; S418 (CK1 phosphosite in human) is N. Zn site, K40/K41, T179/S204/S208/S213
  and the SSVS motif are conserved.
• Faster evolution and linker positive selection in teleost smad3b
  PMID:27703851
  PMID:27703851
  (These positions are in flounder numbering; I did not map them to zebrafish.)

Expression

• First described, with expression and overexpression activity, by Pogoda and Meyer 2002
  PMID:12112463
• ZFIN rows from that paper: maternal; gastrula margin; lateral mesoderm; tail bud; brain
  regions (diencephalon, epithalamus, pretectum, tegmentum, midbrain, hindbrain, rhombomeres);
  retina; spinal cord; somites (smad3a-bioinformatics/output.txt).
• Whole-embryo expression, strongest in eyes and tail at somitogenesis
  PMID:21159776
• Not circadian PMID:29940038
• Bgee top entities: heart, muscle tissue, pharyngeal gill, eye, liver.

Function

• Overexpression induces mesoderm and organizer genes
  PMID:12112463
  PMID:12112463
• Weaker than smad3a on myf5 PMID:21159776
• caSmad3b raises cardiomyocyte proliferation (31%, vs 70% for caSmad3a)
  PMID:29196619
• DN-Smad3b used as a general Smad2/3 blocker in neural induction and cell-competition work
  PMID:19580801
  PMID:39690179
  These dominant-negative results show what Smad2/3 signalling does, not what smad3b specifically does.
• Knockdown alters neural differentiation (as for smad3a) PMID:25286120
• Double knockout with smad3a: see smad3a notes (PMID:42584512, PMID:40066353).

Annotation decisions

• All IBA/IEA rows reviewed with the same actions as the matching smad3a rows (conserved protein,
  same PAINT nodes). Protein binding (Tob1a IPI) removed as uninformative on both copies.
• No smad3b-specific experimental process annotation exists in GOA. None was added: the
  overexpression and dominant-negative data act on Smad2/3 signalling generally.