PTPN4

PTPN4 (PTPMEG) is a cytoplasmic FERM–PDZ–phosphatase protein tyrosine phosphatase that operates at the intersection of synaptic plasticity, immune signaling, and cell survival (PMID:10748123, PMID:17953619, PMID:25425441). Its catalytic activity is held in an autoinhibited state by an intramolecular interaction in which the PDZ domain suppresses the phosphatase domain in cis; the active two-domain protein adopts a compact solution conformation, and binding of a PDZ ligand releases this autoinhibition to restore catalytic competence (PMID:25158884). This regulation depends on the inter-domain linker, where a conserved hydrophobic patch mediates the interdomain communication required for both autoinhibition and ligand-triggered activation (PMID:28801650). A range of C-terminal PDZ-binding motifs engage the PTPN4 PDZ domain and can disrupt autoinhibition, including glutamate receptor subunits GluRdelta2 and GluN2A, the p38γ MAP kinase C-terminus, and viral proteins such as rabies virus G protein and HPV16 E6, with structures defining the binding determinants (PMID:10748123, PMID:22000519, PMID:27246854, PMID:35089587). Once activated, PTPN4 dephosphorylates substrates spanning distinct pathways: it dephosphorylates the p38γ activation loop (PMID:27246854), the TLR4 adaptor TRAM (TICAM2) to suppress TRIF-dependent IRF3 activation and IFN-β production (PMID:25425441), and STAT3 at Tyr705 to restrain STAT3 transcriptional activity, a function lost in colorectal cancer to accelerate tumor growth (PMID:31025789). In the nervous system, PTPN4 is required for cerebellar long-term depression at parallel fiber–Purkinje cell synapses and for motor learning, and its FERM domain directs localization to dendritic spines (PMID:17953619, PMID:30238967). PTPN4 activity is also controlled at the protein level: calpain-mediated proteolytic cleavage in the PEST-containing intermediate domain activates the phosphatase several-fold and occurs during platelet activation (PMID:8910369), while the E3 ubiquitin ligases MARCH8 and NSPA drive its ubiquitin–proteasomal degradation (PMID:33158444, PMID:37747937).

1. 1996 High

   Established that PTPN4 catalytic activity is post-translationally controlled by proteolysis, defining a regulatory mode beyond constitutive phosphatase function and linking it to physiological agonist signaling.

   Evidence In vitro protease treatment and phosphatase assays on recombinant protein, plus calpain-inhibitor experiments in agonist-stimulated platelets

   PMID:8910369

   Open questions at the time

   • The physiological substrates dephosphorylated by the calpain-activated form in platelets were not identified
   • Whether proteolytic activation operates in non-platelet contexts was not tested
2. 1996 Medium

   Showed PTPN4 overexpression suppresses proliferation and anchorage-independent growth, providing early evidence for a growth-suppressive role with both catalytic and non-catalytic components.

   Evidence Stable COS-7 lines expressing wild-type or catalytically dead PTPN4; growth curves and soft-agar colony assays

   PMID:8917530

   Open questions at the time

   • The substrate(s) mediating growth suppression were not defined
   • Overexpression in a single cell line does not establish endogenous role
3. 2000 High

   Identified the PDZ domain as the device that targets PTPN4 to glutamate receptor subunits, placing the phosphatase physically at synaptic receptor complexes.

   Evidence Yeast two-hybrid, reciprocal Co-IP in cultured cells and brain tissue, with domain mapping; cell-based phosphorylation assay with catalytic mutant

   PMID:10748123

   Open questions at the time

   • The functional consequence of enhanced GluN2A phosphorylation was not mechanistically resolved
   • Direct receptor dephosphorylation in vivo was not demonstrated
4. 2006 High

   Demonstrated through the Drosophila ortholog that distinct PTPN4 domains have separable roles in neuronal axon projection, with catalytic activity and the FERM domain controlling different processes.

   Evidence Drosophila loss-of-function genetics with domain-specific rescue and neuroanatomical analysis of mushroom body lobes

   PMID:17138662

   Open questions at the time

   • Substrates underlying axon-projection phenotypes were not identified
   • Direct extrapolation to vertebrate PTPN4 not established
5. 2007 High

   Established a definitive in vivo neural function: PTPN4 is required for cerebellar synaptic plasticity and motor learning.

   Evidence PTPN4-knockout mouse behavioral testing (rotarod, eyeblink conditioning) and electrophysiology of PF–PC LTD

   PMID:17953619

   Open questions at the time

   • The synaptic substrate dephosphorylated to enable LTD was not identified
   • Molecular link between phosphatase activity and LTD machinery left open
6. 2008 Medium

   Tested a candidate immune function (TCR ζ ITAM dephosphorylation) and found in vitro substrate engagement but no in vivo requirement, revealing functional redundancy among PTPN4 and its paralogs in T cells.

   Evidence Substrate-trapping Co-IP and NF-κB reporter assays contrasted with PTPN4 single, double and triple KO mouse TCR/cytokine analyses

   PMID:18614237 PMID:19107198

   Open questions at the time

   • Redundant phosphatase(s) compensating in vivo not identified
   • In vitro ITAM trapping does not establish physiological substrate
7. 2011 High

   Provided the structural basis of PDZ-ligand recognition and showed that peptide ligands targeting the PDZ domain trigger glioblastoma cell death, linking ligand binding to pro-apoptotic outcomes.

   Evidence Crystal structures of PTPN4-PDZ/peptide complexes (GluN2A, rabies G) with affinity measurements and cell-death assays correlating killing with affinity

   PMID:22000519

   Open questions at the time

   • The downstream signaling between PDZ occupancy and cell death was not mapped
   • Endogenous physiological ligand competition not addressed
8. 2014 High

   Resolved the core regulatory mechanism: the PDZ domain autoinhibits the phosphatase in cis and PDZ-ligand binding relieves this inhibition.

   Evidence AUC, SAXS, NMR and kinetic phosphatase assays on the two-domain construct

   PMID:25158884

   Open questions at the time

   • Structural snapshot of the autoinhibited interdomain interface not directly visualized
   • Whether all PDZ ligands activate to equivalent extents not quantified here
9. 2015 Medium

   Connected the activated phosphatase to innate immune control by identifying TRAM as a substrate whose dephosphorylation restrains TLR4/TRIF-driven IFN-β.

   Evidence Co-IP, phosphorylation assays, TRAM localization, and IRF3/IFN-β readouts

   PMID:25425441

   Open questions at the time

   • The TRAM phosphosite(s) targeted were not pinpointed
   • Single-lab pathway placement without in vivo confirmation
10. 2016 High

    Linked PDZ-ligand activation to substrate turnover by showing p38γ both binds the PDZ domain to relieve autoinhibition and is itself dephosphorylated on its activation loop.

    Evidence Crystal structure of the PTPN4-PDZ/p38γ complex with kinetic phosphatase assays and cell-death readouts

    PMID:27246854

    Open questions at the time

    • Cellular consequences of p38γ dephosphorylation by PTPN4 not fully traced
    • Whether p38γ acts as activator, substrate, or both in vivo not resolved
11. 2017 Medium

    Pinpointed the inter-domain linker, and a conserved hydrophobic patch within it, as the structural element required for both autoinhibition and ligand-mediated activation.

    Evidence Site-directed mutagenesis of linker residues with kinetic assays and sequence analysis

    PMID:28801650

    Open questions at the time

    • Atomic-resolution view of how the linker couples PDZ occupancy to the active site not obtained
    • Single-lab mutagenesis data
12. 2018 Medium

    Assigned the FERM domain a localization function in mammalian neurons, showing a disease-associated variant abolishes dendritic spine targeting without affecting expression.

    Evidence Transfection of wild-type vs p.Leu72Ser PTPN4 in mouse hippocampal neurons with spine imaging

    PMID:30238967

    Open questions at the time

    • The FERM-domain binding partner mediating spine targeting was not identified
    • Single variant in a single study
13. 2019 Medium

    Identified STAT3-Tyr705 as a PTPN4 substrate and established a tumor-suppressive role in colorectal cancer through STAT3 restraint.

    Evidence Co-IP, phosphorylation assay, STAT3 reporter, overexpression/knockdown, and xenograft model

    PMID:31025789

    Open questions at the time

    • Whether STAT3 dephosphorylation requires PDZ-ligand activation was not addressed
    • Single-lab tumor model
14. 2020 Medium

    Showed PTPN4 protein levels are controlled by E3-ligase-driven degradation (NSPA), coupling PTPN4 abundance to synaptic receptor phosphorylation and postsynaptic density composition.

    Evidence Cell-based ubiquitination assay, synaptic fractionation, and KO mouse phosphorylation analysis

    PMID:33158444

    Open questions at the time

    • The GluN2B substrate link was inferred by correlation, not direct dephosphorylation
    • Direct NSPA–PTPN4 ubiquitination site not mapped
15. 2021 High

    Defined the structural basis for viral hijacking of PTPN4 by showing HPV E6 oncoproteins from multiple genotypes engage the PDZ domain.

    Evidence X-ray crystallography of the PTPN4-PDZ/HPV16 E6 complex and ITC across HPV genotypes

    PMID:35089587

    Open questions at the time

    • The functional consequence of E6 binding on PTPN4 activity or stability was not determined
    • Cellular impact in HPV-infected cells not tested
16. 2023 Medium

    Extended degradative control to MARCH8 and reinforced the PTPN4–STAT3 tumor-suppressor axis in pancreatic cancer.

    Evidence Ubiquitination assays, Co-IP, overexpression/knockdown and xenograft functional assays

    PMID:37747937

    Open questions at the time

    • MARCH8 ubiquitination site on PTPN4 not mapped
    • Relationship between MARCH8- and NSPA-mediated degradation not reconciled
17. 2025 Low

    Proposed a nuclear pool of PTPN4 imported by KPNA5 that suppresses STAT3 signaling in ovarian cancer.

    Evidence Co-IP, nuclear fractionation, overexpression functional assays, and xenograft model

    PMID:40145330

    Open questions at the time

    • Nuclear localization inferred from functional assays with limited NLS mechanistic detail in a single paper
    • Whether nuclear vs cytoplasmic STAT3 dephosphorylation predominates not resolved

• How the diverse activating PDZ ligands, proteolytic activation, and E3-ligase-driven degradation are integrated to set PTPN4 activity toward specific substrates in a given cell type remains unresolved.
• No unified model linking conformational activation to substrate selection across tissues
• Atomic structure of the full-length autoinhibited enzyme lacking
• Physiological hierarchy among synaptic, immune, and STAT3 substrates undefined