ARHGAP44

ARHGAP44 (RICH2) is a Rho-family GTPase-activating protein that couples membrane geometry to actin remodeling, restraining Rac1, Cdc42, and RhoA signaling across neuronal and epithelial contexts (PMID:24352656, PMID:25498153, PMID:28527113, PMID:28642683). Its N-BAR domain senses inward plasma membrane curvature generated by acto-myosin contractile forces, recruiting the protein to nascent filopodia sites where its GAP domain drives local Rac-GTP hydrolysis to suppress filopodia initiation in developing neurons (PMID:25498153). As a Rac1-specific GAP it controls dendritic spine morphogenesis, with loss-of-function disinhibiting synaptic Rac1 in vivo and altering spine number and receptor composition (PMID:24352656, PMID:26969129); at postsynaptic densities it forms an LTP-regulated complex with Shank3 that governs endosomal recycling and GluA1 AMPA-receptor exocytosis during spine enlargement (PMID:23739967, PMID:36630934). In polarized epithelial cells ARHGAP44 links CD317/tetherin to the apical actin cytoskeleton through EBP50 and ezrin, maintaining the apical actin network and microvilli (PMID:19273615). Its catalytic (R291A-sensitive) GAP activity toward Cdc42 restrains cell spreading and migration and is transcriptionally suppressed by mutant p53 (PMID:28527113).

1. 2009 High

   Established the first cellular role for RICH2 as a cytoskeletal adaptor, answering how the membrane protein CD317/tetherin is anchored to the apical actin network in polarized epithelia.

   Evidence siRNA knockdown, co-immunoprecipitation, and confocal imaging in polarized epithelial cells

   PMID:19273615

   Open questions at the time

   • Did not test GAP activity or a specific Rho GTPase in this epithelial context
   • Direct binding stoichiometry within the EBP50/ezrin complex not resolved
2. 2013 High

   Defined RICH2 as a Rac1-specific GAP acting upstream of Rac1 to control dendritic spine morphology, answering which GTPase mediates its synaptic function.

   Evidence Knockdown/overexpression with Rac1-inhibitor (EHT 1864) epistasis rescue, mEPSC recording, and immunofluorescence in cultured hippocampal neurons

   PMID:24352656

   Open questions at the time

   • Did not address regulation of recruitment to spines
   • Did not distinguish Rac1 from other potential substrates biochemically
3. 2013 High

   Identified the Shank3-RICH2 complex as an LTP-regulated machine controlling AMPA receptor recycling, connecting GAP function to synaptic plasticity output.

   Evidence Proteomic screen, BRET in live cells, siRNA and interfering peptide disruption, GluA1 exocytosis assay in hippocampal neurons

   PMID:23739967

   Open questions at the time

   • Molecular basis of LTP-induced interaction increase not defined
   • Link between GAP catalysis and receptor exocytosis not directly dissected
4. 2014 High

   Revealed the mechanistic logic of curvature-sensing, showing the N-BAR domain reads acto-myosin-generated membrane curvature to target GAP activity and suppress filopodia initiation.

   Evidence Live-cell imaging, dominant-negative/constitutively-active constructs, BAR-domain curvature binding assay, Rac-GTP pull-down in neurons

   PMID:25498153

   Open questions at the time

   • In vivo relevance of curvature sensing not tested
   • Whether the same mechanism operates at spines or epithelial surfaces unknown
5. 2016 Medium

   Confirmed RICH2 as an in vivo negative regulator of synaptic Rac1, linking its loss to structural and receptor-composition synaptic phenotypes.

   Evidence RICH2 knockout mouse with immunohistochemistry, electron microscopy, Rac1-GTP pull-down, and behavioral testing

   PMID:26969129

   Open questions at the time

   • Single lab
   • Behavioral consequences not mechanistically tied to specific spine changes
6. 2017 Medium

   Extended the substrate repertoire to Cdc42 and RhoA and placed RICH2 in a tumor-suppressive transcriptional circuit downstream of mutant p53.

   Evidence RNA-seq/RT-qPCR, Cdc42-GTP and RhoA-GTP pull-downs, WT vs GAP-dead R291A overexpression, cell spreading/migration assays, and amygdala KO analysis

   PMID:28527113 PMID:28642683

   Open questions at the time

   • Region- and context-specificity of substrate choice (Rac1 vs Cdc42 vs RhoA) unexplained
   • Single lab for each substrate claim
7. 2019 Medium

   Documented physical association with Cdc42, Rac1, and β-catenin in carcinoma cells and a Cdc42-dependent anti-tumor, anti-filopodia function in vivo.

   Evidence Co-immunoprecipitation, filopodia assay, stable overexpression, and xenograft in hepatocellular carcinoma cells

   PMID:31136984

   Open questions at the time

   • β-catenin interaction functional significance not established
   • GAP-dependence of tumor suppression not formally tested here
8. 2023 Medium

   Anchored RICH2 biochemically to the postsynaptic density in vivo and noted developmental nuclear accumulation, refining its subcellular localization.

   Evidence Subcellular fractionation, immunoblotting, and immunohistochemistry of mouse brain across regions and developmental stages

   PMID:36630934

   Open questions at the time

   • Functional role of nuclear pool not addressed
   • Single lab
9. 2023 Low

   Reported a non-canonical effect of RICH2 overexpression on mitochondrial dynamics and MAPK/ERK/HIF-1 signaling in glioma.

   Evidence Fluorescence microscopy, qRT-PCR, Western blot, calcium imaging, and glioma mouse model (overexpression only)

   PMID:37926169

   Open questions at the time

   • Overexpression only, no loss-of-function
   • Direct molecular link between RICH2 and mitochondrial machinery not dissected
   • Not independently replicated

• How ARHGAP44 selects among Rac1, Cdc42, and RhoA in different cellular and tissue contexts, and how curvature sensing, scaffold binding, and transcriptional control are integrated, remains unresolved.
• No structural model of substrate selectivity
• Mechanism coupling N-BAR recruitment to specific GTPase choice unknown
• No unified model across neuronal, epithelial, and tumor contexts