PHF21A

PHF21A (BHC80) is a chromatin-reader scaffold of the LSD1-CoREST/BRAF-HDAC (BHC) repression complex that couples recognition of unmodified chromatin to transcriptional silencing of neuronal genes (PMID:17687328, PMID:15325272). Its C-terminal PHD finger specifically binds unmethylated histone H3K4 (H3K4me0) and is sterically excluded by H3K4 methylation, with a crystal structure defining the basis of this recognition (PMID:17687328). This PHD/H3K4me0 engagement is functionally essential: PHF21A and LSD1 reciprocally depend on one another for chromatin association, and a PHD-finger point mutant that cannot bind H3 fails to restore repression of LSD1 target genes (PMID:17687328). Through the same PHD region PHF21A contacts each of the other BHC subunits, organizing the complex as a scaffold (PMID:15325272), and it stabilizes LSD1 occupancy at specific neuronal targets such as SCN3A, whose de-repression accompanies PHF21A haploinsufficiency in patient cells (PMID:22770980). PHF21A also negatively modulates REST-mediated repression of neurosecretion genes (PMID:19439607) and is required for full cAMP/CRE-driven immediate-early gene induction (PMID:28571721). A neuron-specific microexon isoform (PHF21A-n) loses AT-hook-dependent DNA binding while retaining H3K4me0 recognition, generating a hypomorphic LSD1 complex with reduced nucleosome engagement (PMID:37904995); expression of this neuronal isoform precedes the neuronal LSD1 isoform during differentiation, and through new partners including MYT1-family transcription factors and VIRMA the resulting enzymatically attenuated complex limits excess synapse formation during brain development (PMID:39395799). Disruption of PHF21A causes cognitive impairment, with disease-associated splice/missense variants acting through reduced PHF21A dosage (PMID:22770980, PMID:40622422).

1. 2004 Medium

   Before its molecular role was defined, it was unclear how the BHC repression complex was organized; this established PHF21A as the scaffold that directly contacts every other BHC subunit.

   Evidence Direct pulldown binding assays with BHC components and transcriptional repression assays in cultured cells

   PMID:15325272

   Open questions at the time

   • Isoform-specific HDAC1/HDAC2 binding differences not mechanistically resolved
   • Single-lab binding data without structural mapping of subunit contacts
2. 2006 Medium

   The in vivo requirement for PHF21A was unknown; knockout established it is essential for neonatal survival.

   Evidence Targeted gene knockout in mouse with phenotypic analysis

   PMID:16684532

   Open questions at the time

   • No detectable tissue developmental defect, leaving the lethal cause undefined
   • Does not connect survival phenotype to a specific molecular target
3. 2007 High

   How PHF21A is targeted to chromatin and links to active repression was unresolved; the PHD finger was shown to read H3K4me0 and to be required for LSD1-complex repression.

   Evidence Crystal structure of PHD finger on H3 peptide, in vitro binding, plus RNAi/ChIP epistasis and PHD-mutant rescue

   PMID:17687328

   Open questions at the time

   • Does not address how DNA-binding or non-histone contacts contribute to targeting
   • Genome-wide target spectrum not defined
4. 2009 Medium

   The relationship between PHF21A and the neuronal master repressor REST was unknown; PHF21A was shown to act as a negative modulator of REST-mediated repression at neurosecretion gene promoters.

   Evidence Cell fusion, transfection, ChIP, and immunodepletion in PC12 cells

   PMID:19439607

   Open questions at the time

   • Mechanism of how PHF21A antagonizes rather than reinforces repression at these loci unclear
   • Effect was only partial de-repression in a single cell system
5. 2012 Medium

   Whether human PHF21A dosage controls LSD1 chromatin occupancy at defined genes was untested; haploinsufficiency was shown to reduce LSD1 occupancy and de-repress SCN3A.

   Evidence ChIP in patient-derived lymphoblastoid lines with PHF21A-disrupting balanced translocations

   PMID:22770980

   Open questions at the time

   • Limited to one target gene
   • Causality between SCN3A de-repression and patient phenotype not established
6. 2017 Medium

   PHF21A had been characterized as a repressor; this placed it in active cAMP/CRE signaling by showing it is required for full immediate-early gene induction.

   Evidence RNA-seq, CRE-luciferase reporter, and IEG RT-qPCR after forskolin stimulation in patient-derived cells

   PMID:28571721

   Open questions at the time

   • Molecular basis of a repressor-scaffold promoting CRE induction unexplained
   • Direct CRE-promoter occupancy by PHF21A not shown
7. 2019 Medium

   Functions beyond nuclear repression were unknown; a cytoplasmic splice variant was shown to drive MyD88-p38-TTP signaling and cytokine mRNA stabilization to promote proliferation.

   Evidence Subcellular localization, pathway inhibition, proliferation assays, and xenografts

   PMID:30910347

   Open questions at the time

   • Mechanism connecting isoform localization to pathway activation unresolved
   • Relationship of this cytoplasmic role to chromatin function unclear
8. 2023 Medium

   How a neuron-specific isoform alters complex activity was unknown; the microexon-containing PHF21A-n was shown to lose AT-hook DNA binding while retaining H3K4me0 reading, yielding a hypomorphic demethylase complex.

   Evidence In vitro reconstitution of canonical vs. neuronal PHF21A-LSD1 complexes, AT-hook mutagenesis, DNA binding and splicing analysis (preprint)

   PMID:37904995

   Open questions at the time

   • Preprint not yet peer-reviewed
   • In vivo consequence of reduced nucleosome engagement not established in this study
9. 2024 High

   The developmental purpose of neuronal PHF21A splicing was unknown; asynchronous isoform switching was shown to stepwise deactivate the LSD1-PHF21A complex and to restrain excess synapse formation via new partners.

   Evidence Unbiased proteomics, Phf21a mutant mice, co-IP, and neuron differentiation time-course

   PMID:39395799

   Open questions at the time

   • Functional contribution of MYT1-family and VIRMA interactions to synapse control not dissected
   • How catalytic inactivation translates to limiting synaptogenesis mechanistically unresolved
10. 2025 Medium

    The pathogenic mechanism of a disease-associated variant was ambiguous; characterization showed it reduces neuronal-isoform splicing while preserving DNA binding, implicating dosage rather than binding defect.

    Evidence mRNA splicing analysis and DNA binding assays of the c.1285G>A variant

    PMID:40622422

    Open questions at the time

    • Single-individual variant characterization
    • Direct demonstration of dosage-dependent neuronal phenotype not provided

• It remains unresolved how PHF21A's chromatin-repressor scaffold function, its cytoplasmic signaling isoform, and its non-catalytic neuronal complex partners are mechanistically integrated to control neurodevelopment.
• No structure of the full PHF21A-LSD1 complex on a nucleosome
• Genome-wide target map and partner-specific functions undefined
• Causal chain from variant dosage to cognitive/synaptic phenotype incomplete