sox9-a (Xenopus laevis, UniProt B7ZR65) - review notes

Project context: projects/NEURAL_CREST_ORIGINS.md, Tier 1 (neural crest specifiers).
Homeolog: sox9-b (Q6DFF5). Cross-check: genes/human/SOX9/.

Identity and molecular activity

• SoxE-group HMG-box transcription factor, 477 aa; HMG box 105-173, dimerization
  domain 63-103, transactivation domains TAM (224-308) and TAC (361-477), all by
  similarity to human SOX9 [file:XENLA/sox9-a/sox9-a-uniprot.txt "DNA_BIND 105..173"].
• Activator in vivo in Xenopus: an HMG-box/engrailed-repressor fusion phenocopies
  the morpholino PMID:15464575;
  PMID:15464575.
  This is the best evidence for GO:0001228 (DNA-binding transcription activator
  activity, RNA polymerase II-specific). No Xenopus evidence for a repressor role;
  GO:0000122 (IBA) comes from mammalian/other SoxE members and is kept as non-core.
• SUMOylation at K61 and K365 switches output between crest/melanocyte-promoting and
  otic-promoting activities PMID:16256735;
  [file:XENLA/sox9-a/sox9-a-uniprot.txt "Lys-365 is the major site of sumoylation"].
  The IPI rows with Ube2i (P63282) and Sumo1-a (O57686) reflect Sox9 being a SUMOylation
  substrate, not a binding function of Sox9 itself.
• Nucleus (functional site). Cytoplasm transferred from X. tropicalis Sox9 (Q6F2E7),
  where it is cytoplasmic in previtellogenic oocytes - a storage/oocyte localisation
  [file:XENLA/sox9-a/sox9-a-uniprot.txt "localizes to the cytoplasm of previtellogenic oocytes in the ovary"].

Expression (where/when)

• Maternal, then from mid-gastrula at the lateral edges of the neural plate; persists in
  migrating cranial crest into the pharyngeal arches PMID:11807034.
• Order of onset in Xenopus SoxE: Sox8 precedes Sox9, which precedes Sox10
  PMID:16943273.
• Also otic placode from stage 13/14, eye, genital ridges, notochord, pancreas
  (UniProt tissue specificity).

Loss / gain of function

• MO: loss of NC progenitors with expansion of neural plate; later loss of NC-derived
  cranial skeletal elements PMID:11807034.
• Required for Wnt-mediated NC induction, and needed for specification but not for
  migration PMID:15464575;
  PMID:15464575.
• Sox9 rescues Sox10 MO; SoxE proteins are functionally interchangeable
  PMID:16256735.
• Otic placode: MO loses Pax8/Tbx2; inducible dominant-negative shows placode
  specification, not patterning, requires Sox9
  PMID:15084460.
• Gain of function expands crest/otic markers (deep research summary of Taylor and
  LaBonne 2005; full text not cached here).

Network layer

Sox9 is a neural crest specifier, not a neural plate border specifier and not a
competence/pluripotency factor:
- it acts downstream of Wnt-driven induction (required for it) and its expression follows
Sox8 at the border;
- its loss converts crest to neural plate fate (a fate-specification phenotype);
- it is dispensable for migration once specified;
- it is later reused in the crest-derived chondrogenic lineage.
So the appropriate process term is GO:0014036 neural crest cell fate specification
(a part/descendant of GO:0014029 neural crest formation). The existing IMP rows use
the broader GO:0014029. I MODIFY the Spokony 2002 row (the phenotype it reports is a
specification phenotype) and keep the other two, which I cannot read beyond the abstract,
at the broader term. GO lacks a neural plate border term; not needed for this gene.

Chondrogenesis (core, ancestral)

• Xenopus MO loses neural crest-derived skeletal elements PMID:11807034.
• Deep research (Schock & LaBonne 2020 review): loss of Meckel's cartilage in morphants.
• Conservation: lamprey Sox9 is co-expressed with Col2a1 in the developing skeleton
  PMID:16492784.
• Amphioxus SoxE is expressed in the gill bars (cellular cartilage-like skeleton) - PMID:18369444,
  not cached; mentioned only.

Evolution (for project synthesis)

• Lamprey has three SoxE genes; SoxE1 and SoxE2 are needed for the chondrogenic crest
  PMID:21889937.
  The SoxE duplications in agnathans and gnathostomes were independent
  PMID:21889937,
  so one-to-one Sox9 orthology to lamprey SoxE genes should not be assumed.
• Recruitment of SoxE into the crest is at the cis-regulatory level: amphioxus soxE is
  not co-expressed with AP2 and lacks crest enhancers PMID:22241841.
• Interpretation: the SoxE protein activity (HMG-box activator, chondrogenic program)
  is ancestral; the deployment in the border/crest is a vertebrate novelty.

Homeologs

All experimental (IMP/IDA/IPI) annotations are attached to sox9-a (B7ZR65). The
morpholinos and the mRNA constructs in these papers were described as "Sox9"; whether they
distinguish the L and S homeologs is not stated in the abstracts. Do not transfer by
assertion; note as question.

Pleiotropy: core vs non-core

• Core: Pol II activator activity; NC cell fate specification; chondrocyte
  differentiation/cartilage development (cranial crest-derived).
• Non-core: otic placode formation (real, well-supported but a separate deployment);
  heart development, epithelial morphogenesis, oligodendrocyte differentiation (IBA,
  no Xenopus data); mouse-derived growth-plate, osteoblast, fatty-acid and Wnt terms.
• Sex determination: Xenopus laevis uses DM-W, not a Sox9-centred mechanism; expression in
  genital ridges only. No annotation exists or is proposed.