HMG20A

HMG20A is a chromatin-regulatory HMG-box protein that operates as a context-dependent subunit of histone-modifying complexes to control epithelial-to-mesenchymal transition (EMT) and lineage differentiation programs (PMID:25639869, PMID:36709316). It functions as a mutually exclusive subunit of the LSD1-CoREST histone demethylase complex, replacing its paralog HMG20B through a coiled-coil interaction with BHC80, and is recruited with SNAI1 and SMAD4 to epithelial gene promoters where it promotes LSD1-mediated H3K4 demethylation to repress epithelial genes and drive TGF-β-triggered EMT (PMID:25639869). Structurally, HMG20A combines an N-terminal intrinsically disordered transactivation domain, an HMG box that binds four-way-junction DNA with higher affinity than linear DNA, and a long coiled-coil domain that mediates assembly with the histone reader PHF14; the PHF14–HMG20A complex engages TEAD1 to modulate Hippo signaling and is likewise required for EMT (PMID:36124662). Beyond LSD1-CoREST, HMG20A associates with NuRD and the H2A.Z-binding protein PWWP2A, localizes to open promoters and enhancers, and is required for neural crest and cardiomyocyte differentiation in Xenopus and mouse ESC models (PMID:36709316). The same chromatin machinery is deployed in differentiated cells: HMG20A occupies the Pax4 promoter and is required for beta-cell functional maturity and glucose-stimulated insulin secretion (PMID:29449530), and acts through the LSD1-CoREST axis to sustain reactive astrogliosis in hypothalamic astrocytes (PMID:34093866). HMG20A also participates in a PHF5A–PHF14–HMG20A–RAI1 RNA polymerase-associated subcomplex with KMT2A in pancreatic cancer stem cells (PMID:37709746). Independently of its host chromatin role, HMG20A is co-opted during vaccinia virus infection, translocating to viral factories and binding viral genomic DNA via its HMG box, an interaction with the host range factor CP77 that is essential for viral replication (PMID:16840350).

1. 2000 Medium

   Establishing HMG20A's basic identity as an HMG-box protein defined it as a sequence-specific/structure-specific DNA-binding factor and bounded its homology to other proteins.

   Evidence cDNA sequencing, genomic mapping, and expression analysis identifying an HMG box most similar to yeast NHP6A

   PMID:10773667

   Open questions at the time

   • No complex membership or functional pathway identified
   • DNA-binding specificity not biochemically characterized at this stage
2. 2006 High

   The first direct functional engagement showed HMG20A is a DNA-binding host factor that vaccinia virus must neutralize, demonstrating its HMG box binds genomic DNA in cells.

   Evidence Yeast two-hybrid, GST pulldown, deletion mapping, imaging, viral genome ChIP, and vaccinia growth assay in CHO-K1 cells

   PMID:16840350

   Open questions at the time

   • Endogenous (non-viral) chromatin targets not yet defined
   • Relationship to host transcriptional complexes unaddressed
3. 2010 Medium

   Linking HMG20A to REST-responsive gene activation and beta-dystrobrevin binding implicated it in neuronal differentiation and chromatin dynamics.

   Evidence Yeast two-hybrid, reciprocal binding assays with kinetics, and luciferase reporter assays

   PMID:20530487

   Open questions at the time

   • Mechanism by which beta-dystrobrevin modulates chromatin not resolved
   • Direct genomic targets in neurons not mapped
4. 2015 High

   Defining HMG20A as a mutually exclusive LSD1-CoREST subunit established the central mechanistic model: it recruits LSD1 via BHC80 to demethylate H3K4 and repress epithelial genes, driving SNAI1/SMAD4-dependent EMT.

   Evidence Reciprocal Co-IP, siRNA knockdown, ChIP, transcriptomics, and migration/invasion assays

   PMID:25639869

   Open questions at the time

   • Determinants selecting HMG20A versus HMG20B in a given context unknown
   • Direct enzymatic role of HMG20A in demethylation versus scaffolding not separated
5. 2018 Medium

   Tissue-specific roles emerged: HMG20A directly occupies the Pax4 promoter and is required for beta-cell maturity and insulin secretion, linking it to T2DM risk via a 3' UTR SNP.

   Evidence siRNA knockdown, ChIP, qRT-PCR, luciferase reporter, and glucose-stimulated insulin secretion in islets and INS-1E cells

   PMID:29449530

   Open questions at the time

   • Whether beta-cell function uses the LSD1-CoREST complex not directly tested here
   • Causal contribution of the rs7119 SNP in vivo not established
6. 2021 Medium

   An adult-brain role was defined showing HMG20A drives reactive astrogliosis gene programs and astrocyte bioenergetics through the LSD1-CoREST axis.

   Evidence siRNA knockdown, RNA-seq, Seahorse metabolic assay, conditioned-medium neuronal viability assay, and in vivo ORY1001 (LSD1 inhibitor) rescue in mice

   PMID:34093866

   Open questions at the time

   • Direct chromatin targets in astrocytes not mapped
   • Mechanism linking gene-expression changes to mitochondrial bioenergetics unresolved
7. 2022 High

   Structural and biochemical dissection resolved HMG20A's domain architecture and a new complex: a coiled-coil-mediated PHF14–HMG20A module that binds TEAD1 to modulate Hippo signaling and is required for EMT, with the HMG box preferring four-way-junction DNA.

   Evidence Proteomics, Co-IP, deletion analysis, AlphaFold2 modeling, four-way-junction vs linear DNA binding assays, transcriptomics, and invasion assays

   PMID:36124662

   Open questions at the time

   • How PHF14–HMG20A and LSD1-CoREST associations are coordinated unknown
   • In vivo significance of four-way-junction DNA preference not established
8. 2022 Medium

   HMG20A was assigned a metabolic differentiation role, acting with LSD1 to deposit H3K4me2 at the MEF2C promoter and restrain adipogenesis.

   Evidence Gain-/loss-of-function, RNA-seq, ChIP-PCR, and adipogenic differentiation assays

   PMID:36142473

   Open questions at the time

   • Apparent activating role at MEF2C contrasts with repressive LSD1 demethylase activity and is not mechanistically reconciled
   • Direct versus indirect H3K4me2 effects not separated
9. 2023 High

   Broader complex membership and developmental requirement were established: HMG20A partners with NuRD and PWWP2A, sits at open promoters/enhancers, and is required for neural crest and cardiomyocyte differentiation.

   Evidence Proteomics, ATAC-seq, ChIP-seq, RNA-seq, Xenopus morpholino knockdown, and mESC differentiation assays across two model systems

   PMID:36709316

   Open questions at the time

   • Functional contribution of NuRD versus LSD1-CoREST associations in development not dissected
   • Direct PWWP2A–HMG20A interaction mode unresolved
10. 2023 Medium

    A KMT2A-containing subcomplex placed HMG20A in pancreatic cancer stem cell maintenance, defining a PHF5A–PHF14–HMG20A–RAI1 RNA polymerase-associated module.

    Evidence Co-IP, proteomics, KMT2A-WDR5 inhibitor treatment, and in vivo xenograft

    PMID:37709746

    Open questions at the time

    • Direct role of HMG20A within the inhibitor's mechanism not isolated
    • Chromatin targets of this subcomplex not mapped
11. 2023 Low

    HMG20A was linked to cardiac fibrosis as a downstream effector of HAX1 activating TGF-β1/Smads signaling.

    Evidence Co-IP, luciferase reporter (HAX1 3' UTR), Western blot, siRNA knockdown, and in vivo myocardial infarction model

    PMID:37352750

    Open questions at the time

    • Co-IP for HAX1–HMG20A without reciprocal validation; direct HMG20A–Smads link not characterized
    • Whether the effect requires HMG20A's chromatin complexes untested

• What governs the choice among HMG20A's multiple complexes (LSD1-CoREST, NuRD, PHF14–TEAD1, PHF5A–PHF14–KMT2A–RAI1) and whether it acts as an activator or repressor in a given cell type remains unresolved.
• No unified model reconciling repressive (H3K4 demethylation) and activating (H3K4me2 deposition) outcomes
• Cell-type-specific determinants of complex assembly unknown
• No high-resolution structure of full-length HMG20A within an assembled complex