AGAP3

AGAP3 is a multi-domain signaling protein that links glutamate receptor activation to downstream small-GTPase and transcriptional outputs in neurons. At excitatory synapses it is a component of the NMDA receptor complex, where it couples receptor activation to Ras/ERK and Arf6 signaling and is required to drive AMPA receptor trafficking during chemically induced LTP (PMID:23904596). A short splice variant, CRAG, carries intrinsic GTPase activity and acts in the nucleus: through its nuclear localization signal and SUMO-interacting motifs it localizes to PML bodies, binds the SRF co-activator ELK1, and activates SRF-dependent c-Fos/AP-1 transcription, an activity absent from the full-length protein (PMID:21832068, PMID:31882856). Consistent with this transcriptional role, CRAG/Centaurin-γ3 knockout mice show blunted kainic acid-induced hippocampal c-fos induction, and forebrain-specific deletion produces an immature dentate gyrus phenotype with hyperactivity, establishing CRAG as required for dentate gyrus granule cell maturation in vivo (PMID:31882856, PMID:33811862). AGAP3 expression is post-transcriptionally controlled by the nuclear speckle splicing factor SRRM2, whose loss causes AGAP3 mis-splicing and protein elevation in both mouse and human iPSC-derived neurons (PMID:42189682). As a recurrent oncogenic event, AGAP3 serves as a 5' fusion partner to BRAF, where it displaces the auto-inhibitory N-terminal region of BRAF to generate a constitutively active kinase that drives MAPK signaling and confers resistance to BRAF and EGFR inhibitors (PMID:31911540, PMID:28539463).

1. 2011 Medium

   Established that the short CRAG splice variant, but not full-length AGAP3, can drive a specific transcriptional program, defining a domain-dependent nuclear function.

   Evidence Reporter assays, siRNA, dominant-negative and deletion mutagenesis in cultured neuronal cells

   PMID:21832068

   Open questions at the time

   • Mechanism of how cytoplasmic vs nuclear isoform partitioning is controlled not defined
   • Direct DNA or co-factor binding by CRAG not yet mapped here
   • Physiological stimulus driving CRAG transcriptional activity not established
2. 2013 Medium

   Placed AGAP3 within the NMDA receptor complex and showed it is the link coupling receptor activation to Ras/ERK and Arf6 signaling and AMPA receptor trafficking during LTP.

   Evidence Reciprocal co-IP for complex membership plus shRNA knockdown with AMPA trafficking and signaling assays in rat primary neurons

   PMID:23904596

   Open questions at the time

   • ArfGAP catalytic activity toward Arf6 not directly demonstrated in this context
   • Direct binding partner within the NMDAR complex not resolved
   • Single lab, neuronal culture only
3. 2017 Medium

   Identified AGAP3 as a recurrent oncogenic 5' BRAF fusion partner conferring drug resistance, extending its biology into cancer signaling.

   Evidence Stable expression of AGAP3-BRAF in BRAFV600E melanoma cells with vemurafenib and MEK inhibitor sensitivity assays

   PMID:28539463

   Open questions at the time

   • Structural basis of N-terminal displacement not shown here
   • Contribution of AGAP3 domains to the fusion phenotype not dissected
   • Resistance demonstrated in cell lines only
4. 2017 Low

   Linked AGAP3 to a microtubule plus-end tracking protein network, suggesting a cytoskeletal-associated role outside neurons.

   Evidence AP-MS with SAINT and reciprocal co-IP in 3T3-L1 adipocytes

   PMID:28550165

   Open questions at the time

   • Interaction membership only; no mechanistic consequence for AGAP3 established
   • +TIP function of AGAP3 not functionally tested
   • Adipocyte system only
5. 2019 Medium

   Defined the molecular basis of CRAG's nuclear transcriptional activity, showing it is a GTPase that activates SRF via ELK1 at PML bodies and contributes to hippocampal c-fos induction in vivo.

   Evidence In vitro GTPase assay, CRAG-ELK1 co-IP, SUMO-interacting-motif mutants, and knockout mouse c-fos measurement

   PMID:31882856

   Open questions at the time

   • GTPase substrate/regulatory cycle of CRAG not characterized
   • How SUMO-interaction directs PML-body recruitment mechanistically unclear
   • Direct vs indirect role of ELK1 binding in SRF activation not separated
6. 2019 Low

   Showed AGAP3 phosphorylation is insulin-regulated, embedding it in an insulin-responsive +TIP network.

   Evidence Targeted quantitative phosphoproteomics in 3T3-L1 adipocytes

   PMID:31018989

   Open questions at the time

   • Single phosphoproteomic dataset with no functional follow-up for AGAP3
   • Responsible kinase and phosphosite consequence unknown
   • Physiological role of phosphorylation untested
7. 2020 Medium

   Demonstrated mechanistically that the AGAP3-BRAF fusion activates BRAF by replacing its auto-inhibitory N-terminus and that the AGAP3 partner sets fusion localization and EGFR-inhibitor resistance.

   Evidence Patient-derived colorectal cancer organoids expressing fusion constructs with drug sensitivity, MAPK signaling, and localization readouts

   PMID:31911540

   Open questions at the time

   • Precise structural mechanism of N-terminal displacement not resolved
   • Which AGAP3 domains drive localization not delineated
   • In vivo tumor relevance beyond organoids not established
8. 2021 Medium

   Established a developmental requirement for CRAG/AGAP3 by linking its loss to failed dentate gyrus neuron maturation and behavioral hyperactivity.

   Evidence Forebrain-specific conditional knockout mice with doublecortin/calbindin immunohistochemistry and open-field testing

   PMID:33811862

   Open questions at the time

   • Molecular pathway linking CRAG loss to maturation arrest not defined
   • Whether the synaptic vs transcriptional function underlies the phenotype unresolved
   • Single lab
9. 2026 Medium

   Identified SRRM2 as an upstream splicing regulator of AGAP3, connecting AGAP3 dosage to a nuclear speckle splicing factor across species.

   Evidence Srrm2+/- mouse splice/transcriptomic analysis and SRRM2-deficient human iPSC-derived neuron splicing analysis

   PMID:42189682

   Open questions at the time

   • Which AGAP3 isoform (e.g. CRAG) is affected by mis-splicing not specified
   • Functional consequence of AGAP3 elevation in neurons not tested
   • Direct vs indirect SRRM2 action on the AGAP3 transcript not distinguished

• The ArfGAP catalytic activity of AGAP3 toward its physiological Arf substrate, and how the synaptic, nuclear-transcriptional, cytoskeletal, and oncogenic-fusion functions are mechanistically integrated, remain unresolved.
• No direct demonstration of Arf6 GAP catalysis by AGAP3 in the corpus
• Structural model of full-length AGAP3 domain regulation absent
• Relationship between isoform usage and each functional context not unified