snai2 (Slug; Q91924) — Xenopus laevis — review notes

Project: NEURAL_CREST_ORIGINS (Tier 1, neural crest specifier). Reviewed 2026-10-01.

Identity

• Q91924 is snai2.L (Xenbase XB-GENE-6255814, chr 6L); 266 aa, N-terminal SNAG repression
  domain (1-20) and five C2H2 zinc fingers (126-262). The S homeolog snai2.S is TrEMBL
  Q90VW5 (also 266 aa) and carries no experimental GO annotations. Most frog loss-of-function
  reagents (antisense RNA, morpholinos, dominant-negative zinc-finger constructs) would not
  distinguish L from S.
• Distinct from snai1 (P19382, Tier 3). Lamprey has a single snail gene "which has affinities
  to both snai1 and snai2" PMID:39060477.
• UniProt: "Probable transcriptional repressor. Acts downstream of snai1 in the specification
  of the neural crest and neural crest migration." [file:XENLA/snai2/snai2-uniprot.txt]

Expression (where/when)

• Xslu is expressed in the prospective crest after the convergence-extension movements, later
  than Xsna, and then in pre- and post-migratory cranial and trunk crest PMID:7720581.
• Induced by noggin + bFGF in animal caps PMID:7720581; promoter has a LEF/beta-catenin element (Vallin 2001, PMID:11402039,
  not cached; via UniProt INDUCTION).
• York 2024: in Xenopus snai2 behaves like a definitive crest factor (monotonic increase from
  blastula to late neurula), unlike lamprey PMID:39060477. Contrast with snai1, which is
  co-expressed with pluripotency factors in blastula animal cells PMID:25931449.

Molecular activity

• Transcriptional repressor: an Engrailed-repressor fusion mimics XSlug, a Gal4-activator fusion
  inhibits it PMID:10772801.
• Direct target: E-cadherin. Flag-Snail2 ChIP at late neurula PMID:25617436; Snail2 recruits PRC2 PMID:25617436. EZH2 binds Snail2 directly PMID:25617436.
• Corepressors: Ajuba LIM proteins (Ajuba, LIMD1, WTIP) bind the SNAG domain PMID:18331720 (abstract-only in cache; UniProt records frog
  Snai2 SNAG-LIM interactions).
• Elp3 co-IP (the GOA protein binding IPI): mElp3 pulled down xSlug, and xSlug rescues Elp3
  morphants PMID:27189455.
  The paper's model is Elp3 stabilising Snail1 protein; for Snai2 this is an upstream regulator
  of Snai2 protein, not a Snai2 activity. Generic protein binding -> REMOVE.

Gain of function

• XSlug overexpression expands crest markers and melanocytes PMID:10772801.
• But Slug alone is not sufficient to induce crest in naive ectoderm; it needs Wnt/FGF
  PMID:9609823; Snail2 + Wnt
  is sufficient PMID:25931449. Snail (snai1), by contrast, induces Slug and other markers; Slug cannot
  PMID:12490555.
• Snail2 expands border and crest markers but this needs EZH2 PMID:25617436.

Loss of function

• Stage-dependent: early inhibition prevents crest precursor formation, later inhibition blocks
  migration PMID:10772801.
• Antisense slug: migration and derivatives (cartilage) lost; rescued by Slug or Snail
  PMID:10452849.
• Slug MO abolishes Sox9 and raises apoptosis; mesoderm markers reduced PMID:17205110. The NF-kB/apoptosis loop is a separate (mesoderm/survival) role; not core.
• Network layer: Snail2 knockdown does not reduce border genes PMID:25617436. So snai2 is
  downstream of the border specifiers, and downstream of snai1 PMID:12490555.

Layer placement

NC specifier (and later EMT/delamination effector), not a border specifier and not a blastula
competence factor. Evidence: onset after snai1 and inside the crest domain; border genes
unaffected by its loss; insufficient alone, sufficient with Wnt; repressor activity; direct
repression of E-cadherin via PRC2 during NC EMT; required early for precursors and later for
migration.

Process terms: GO:0014036 NC cell fate specification (core), GO:0001755 NC cell migration (core).
GO:0036032 NC cell delamination was considered (Snai2 recruits PRC2 to the E-cad promoter) but
WITHHELD: E-cad RNA rises only ~1.5-fold in Snai2 morphant crest, migrating Xenopus cranial crest
retain E-cadherin and need it (Huang 2016, cited in the updated deep research
[file:XENLA/snai2/snai2-deep-research-falcon.md "E-cadherin depletion impaired their migration and
protrusions, and other tested classical cadherins did not replace it."]), and no Snai2-specific
delamination assay exists in frog. Raised as a suggested question. GO:0014029 TAS row (Id2 cardiac crest ablation paper, which
does not appear to study Slug) -> MODIFY to GO:0014036 per the project convention (sox10, sox9-a).

Evolution

• Amphioxus Snail is the one NC specifier transiently expressed at the border PMID:18562679. So Snail border expression predates vertebrates; the
  Snai1/Snai2 split and snai2's specifier-specific behaviour are vertebrate (gnathostome-level,
  given a single lamprey snail).
• Lamprey snail is in the conserved NC-GRN PMID:17765683.

Open issues

• snai2.S (Q90VW5) carries none of the experimental annotations.
• Snai1/Snai2 redundancy: which roles are snai2-specific?
• Apoptosis/NF-kB and mesoderm roles (PMID:17205110) not annotated; possible non-core.
• No GO neural plate border term; snai2 is NOT a border gene, so it is unaffected by that decision.

Update after deep-research rewrite (2026-10-01)

The falcon report was regenerated mid-review. New points: two X. laevis Slug pseudoalleles
(L/S) with identical zinc fingers (Vallin 2001); LEF/beta-catenin site in the snai2 promoter
(upstream input, not a Snai2 activity); Ppa F-box degradation of Slug; Bcl-xL rescue indicates a
survival component in Slug MO phenotypes [file:XENLA/snai2/snai2-deep-research-falcon.md "it
should not be annotated as secreted or as a cell-surface receptor."]. Quotes re-checked against
the new file.