gbx2

Summary: Phylogenetic inference that Gbx2 binds RNA polymerase II cis-regulatory DNA through its homeodomain. It is grounded in mouse Gbx2 gel-shift data and conserved across the GBX clade.

Reason: Q91907 has an intact metazoan homeodomain (residues 239-298) and sits inside the GBX-2 subfamily (PTHR24334:SF3). There is no evidence of loss of DNA binding. Frog experiments with homeodomain-Engrailed repressor fusions (Gbx2-EnR) that reproduce or rescue Gbx2 function depend on sequence-specific target binding. The PAINT node placement is sound.

Summary: TreeGrafter transfer of the same sequence-specific RNAPII regulatory-region DNA binding activity.

Reason: Consistent with the IBA row and with the homeodomain architecture. Correct at this level.

Summary: Gbx2 is a sequence-specific RNAPII transcription factor. Frog data show it acts as a repressor in early ectoderm and in the pre-placodal region, and as an apparent activator of otic genes.

Reason: This is the core molecular function. Because the sign of regulation depends on context (repressor early, activator for otic onset), the generic RNAPII DNA-binding transcription factor term is the right umbrella. The repressor mode is added separately as NEW GO:0001227. The IBA node is seeded by mouse Gbx2 and zebrafish gbx2 and is phylogenetically sound.

Summary: Combined electronic inference of RNAPII-specific DNA-binding transcription factor activity from the homeobox signature and the PANTHER family.

Reason: Correct, and corroborated by the frog functional data summarized under the IBA row.

Summary: Generic DNA binding from the InterPro homeodomain mapping.

Reason: The statement is true but uninformative. Gbx2 binds RNAPII cis-regulatory DNA in a sequence-specific way, which the gene already carries as GO:0000977. This matches the treatment of the same row in the human GBX2 review.

Proposed replacements: RNA polymerase II transcription regulatory region sequence-specific DNA binding

Summary: Gbx2 acts in the nucleus as a homeodomain transcription factor.

Reason: This follows from its transcription factor activity. Tagged Gbx2 is nuclear in heterologous cells (PMID:17060451, cell culture; mouse/human data in the human GBX2 review). Endogenous frog protein localization has not been measured, but the functional data require nuclear action.

Summary: Electronic nuclear localization, consistent with the IBA row.

Reason: Correct for a homeodomain transcription factor.

Summary: Reactome places the gbx2 gene product in the nucleoplasm in its event "Expression of gbx2 in the neural plate border", which models direct Wnt/TCF activation of the gbx2 promoter.

Reason: Nucleoplasm is a correct, slightly more specific location for a chromatin-acting transcription factor. The Reactome event itself describes regulation of the gbx2 gene (Li et al. 2009), not a Gbx2 activity, so it supports only the location and is not evidence for a Gbx2 process.

Summary: Generic transcription-regulation process from the homeobox signature.

Reason: The more informative RNA polymerase II-specific term is supported by the frog repressor and activator fusion experiments and is already carried by IBA. This matches the human GBX2 review.

Proposed replacements: regulation of transcription by RNA polymerase II

Summary: Gbx2 regulates RNAPII transcription of target genes such as otx2 (repressed) and otic pax8/pax2 (required for onset).

Reason: This is the direct process consequence of the core molecular function, and it is supported in frog by fusion-construct and knockdown experiments.

Summary: TreeGrafter transfer of RNAPII transcription regulation.

Reason: Same as the IBA row.

Summary: Family-level inference, seeded by Drosophila unplugged and mouse Gbx2, that GBX proteins regulate nervous system development.

Reason: Biologically well supported for frog Gbx2, which positions the Otx2/Gbx2 boundary and regulates midbrain-hindbrain boundary and placodal target genes, and as precise as the regulation-of-nervous-system-development branch warrants. It is kept as non-core, matching the human GBX2 review, because it describes a stage- and tissue-specific developmental output downstream of Gbx2's core nuclear RNAPII transcription factor activity, while the specific morphogenetic processes (NEW GO:0030917 midbrain-hindbrain boundary development, NTR neural plate border formation, and NEW GO:0043049 otic placode formation) lie in the developmental process branch rather than being is_a/part_of descendants of GO:0051960.

Summary: TreeGrafter transfer of regulation of nervous system development.

Reason: Same reasoning as the IBA row: biologically supported as a developmental output of Gbx2's transcriptional activity and kept as non-core alongside the specific developmental process annotations (GO:0030917, NTR neural plate border formation, and GO:0043049).

Summary: Gbx2 is the earliest transcription factor of the neural plate border. It is directly activated by Wnt/beta-catenin, it is required for neural crest induction upstream of pax3 and msx1, and it gives the neural folds a posterior identity that separates crest from anterior pre-placodal fate.

Reason: Synthesis across lines of evidence. (1) Expression: from mid-gastrula in posterior dorsolateral ectoderm, including prospective neural crest, with a neural crest enhancer that responds directly to Wnt/beta-catenin (ChIP, transgenesis). (2) Activity: homeodomain RNAPII transcription factor, mainly a repressor in early ectoderm. (3) Loss of function: two morpholinos inhibit neural crest markers and expand the pre-placodal domain, with rescue. Pax3 or Msx1 rescue the Gbx2 knockdown, but Gbx2 does not rescue Pax3/Msx1 loss (full text, via the deep research). (4) Gain of function: Gbx2 turns the pre-placodal domain into neural crest, but only together with Zic1. (5) Conservation: lamprey gbx2 has neural plate border dynamics like the frog gene. Network layer: neural plate border specifier. It acts upstream of the other border genes on the border region as a whole, and it cannot specify crest autonomously because it depends on Zic1. This role is captured by the proposed term 'neural plate border formation' (NTR), which replaces GO:0014029 (proposed for obsoletion: its text describes the border region but it is placed under epithelial to mesenchymal transition). The child GO:0014036 neural crest cell fate specification, used for the crest specifiers (sox10, sox9-a, sox8, snai2), would overstate it. Participation: Gbx2 is a transcription factor acting inside the border cells and repressing six1 there, not a signal or substrate. Comparator check (QuickGO 2026-10-05): the same-layer peers pax3-a and zic1 both carry GO:0014029 by IMP and none carries GO:0014036. No Gbx in human, mouse, rat, chick, zebrafish or Xenopus carries GO:0014029 or its descendants, but mouse Gbx2 carries GO:0001755 neural crest cell migration by IMP. So the family is not systematically excluded, and the frog gap reflects uncurated literature: Q91907 has no experimental annotation at all. Homeolog caveat: the morpholino target copy is not stated in the abstract, and the four X. laevis gbx2 copies are 93-99.7% identical, so the evidence plausibly covers the other copies too.

Summary: Early in development, frog Gbx2 acts as a sequence-specific transcriptional repressor. Obligatory repressor fusions mimic it, obligatory activator fusions do not, and it represses otx2 in the neural plate and the pre-placodal region.

Reason: Two independent frog studies show repressor-mode activity: obligatory activator/repressor fusions of Xgbx2a (Tour 2002), and Gbx2-EnR fusions with knockdown in the pre-placodal region (Steventon 2012). A mechanism is described for the protein family: an eh1-like motif that recruits Groucho/TLE, required for Otx2 repression in medaka (PMID:17060451, cell culture and medaka, so not used as frog evidence). This parallels GO:0001227 on foxd3-a and snai2. Gbx2 can also act as an activator for otic onset, so this row adds to GO:0000981 rather than replacing it. Paralog note: Tour's Xgbx2a is most probably the Q91907 lineage, because Q90YY2 "Gbx2b" differs from Q91907 at about the 13 positions Tour reported between a and b. The two genes "encode functionally homologous proteins" (PMID:11684099).

Summary: Through stage-restricted repression of otx2 and of MHB genes, frog Gbx2 positions the Otx2/Gbx2 boundary where the midbrain-hindbrain boundary (isthmic organizer) forms. Otx2, by contrast, induces the MHB network.

Reason: Synthesis. (1) Expression: the anterior border of gbx2 in the neural plate is at the prospective MHB (rhombomere 1). (2) Activity: it represses otx2 and MHB genes. (3) Perturbation in frog: an inducible wild-type construct and an inducible antimorph show that the gastrula-stage head phenotype tracks otx2 repression. (4) Conservation: mouse Gbx2 carries GO:0030917 and GO:0021555 by IMP (null and conditional mutants; see the human GBX2 review), zebrafish gbx1 carries GO:0021555 by IMP, and frog irx1-a, another frog MHB homeobox gene, carries GO:0030917 by IMP. Amphioxus Gbx and Otx domains also abut at the hindbrain boundary, so the positioning function is chordate-ancestral. The frog-specific evidence is mainly inducible gain of function and antimorph data; the term rests on the combined synthesis with the strong mouse loss-of-function data. Chosen over GO:0060897 neural plate regionalization, which has essentially no curation in GO (one zebrafish row).

Summary: In the posterior pre-placodal region, Gbx2 is required for the onset of the otic markers pax8 and pax2. It restricts otx2 and also gives a positive input to otic specifiers, but it is not needed to form the general pre-placodal region or to maintain otic identity later.

Reason: Splice- and translation-blocking morpholinos (frog) prevent pax8/pax2 onset and reduce otic vesicle size. Early Gbx2-EnR phenocopies this, and Gbx2-EnR blocks otic genes even without Otx2, so the role is more than Otx2 exclusion. Gbx2 alone is not sufficient for otic fate. Participation: a transcription factor acting in otic precursors. Comparators: mouse Gbx2 carries GO:0042472 inner ear morphogenesis (IMP), and zebrafish otic transcription factors (foxi1, dlx3b) carry GO:0043049. This is a second deployment of the same posterior-identity mechanism, so it is treated as non-core relative to the border and MHB roles.