Version history

{
  "session_id": "sess_SDA-2026-04-06-gap-pubmed-20260406-041428-4c4414ad_task_9aae8fc5",
  "round_number": 1,
  "agent_persona": "persona-theorist",
  "agent_backend": "scidex.core.llm.complete",
  "action": "propose",
  "content": "# Mechanistic Hypotheses: Stress Granule Persistence in Neurodegeneration\n\n## Hypothesis 1: CK2 Hyperphosphorylation Locks G3BP1 in a Hyper-condensed State\n\n**Mechanism:** Casein kinase 2 (CK2)-mediated hyperphosphorylation of G3BP1 at specific serine/threonine residues within its intrinsically disordered region alters the \"tunable switch\" mechanism, converting transient LLPS into irreversible coacervates that nucleate protein aggregation. CK2 activity is upregulated in neurodegeneration (PMID: 28965846), creating a phospho-signature that primes G3BP1 for pathological persistence.\n\n**Target Gene/Protein/Pathway:** CSNK2A1/CSNK2B (CK2 catalytic subunits), G3BP1 phospho-sites (S149, T224)\n\n**Supporting Evidence:** Phosphorylation regulates G3BP1's RNA-binding affinity and phase separation threshold (PMID: 32302571); CK2 phosphorylates numerous RNA granule components (PMID: 26607712); hyperphosphorylation is a hallmark of pathological protein assemblies in tau, TDP-43, and α-synucleinopathies.\n\n**Predicted Experiment:** Phospho-deficient (S149A/T224A) and phospho-mimetic (S149D/T224D) G3BP1 mutants will be engineered in iPSC-derived neurons. Live-cell FRAP and differential centrifugation assays will quantify granule dynamics. CK2 inhibition (CX-4945) will be tested for restoration of physiological granule dynamics in patient-derived cells with TDP-43 or FUS mutations.\n\n**Confidence:** 0.72\n\n---\n\n## Hypothesis 2: Impaired Autophagy Receptor Recruitment Traps G3BP1-Positive Granules\n\n**Mechanism:** The autophagy receptors p62/SQSTM1, OPTN, and NDP52 require specific licensing signals (phosphorylation by TBK1, ubiquitin chains) to recognize and target stress granules for autophagic clearance. In neurodegeneration, TBK1 loss-of-function mutations (linked to ALS/FTD) or granule-specific ubiquitin chain deficiency prevents autophagic recognition, causing persistent granules that coalesce with pathological inclusions.\n\n**Target Gene/Protein/Pathway:** TBK1 signaling axis; p62/UBA domain; G3BP1 ubiquitination (by E3 ligases such as MARCHF7 or HUWE1)\n\n**Supporting Evidence:** TBK1 mutations cause ALS/FTD (PMID: 25188341); p62 colocalizes with stress granules and pathological inclusions (PMID: 24185452); TBK1 phosphorylates p62 to enhance substrate selectivity (PMID: 26242857); G3BP1 granules recruit ubiquitin machinery but may lack proper \"eat-me\" signals.\n\n**Predicted Experiment:** CRISPRi knockdown of TBK1 in neurons will phenocopy stress granule persistence; proximity biotinylation (BioID) will map the ubiquitination landscape of persistent versus reversible granules. A FRET-based biosensor for p62 recruitment to granules will be used to screen for small molecules restoring autophagy receptor licensing.\n\n**Confidence:** 0.78\n\n---\n\n## Hypothesis 3: G3BP1 Serves as a Nucleation Hub for TDP-43 and FUS Pathological Seeding\n\n**Mechanism:** G3BP1-positive stress granules act as \"foci of vulnerability\" where disease-relevant proteins (TDP-43, FUS) undergo conformational conversion to β-sheet-rich aggregates. The phase-separated state of G3BP1 lowers the kinetic barrier for seeded aggregation, and persistent granules become irreversible due to inter-molecular β-sheet propagation between G3BP1, TDP-43, and FUS prion-like domains.\n\n**Target Gene/Protein/Pathway:** G3BP1 condensate surface properties; TDP-43 N-terminal domain; FUS LCD (low-complexity domain)\n\n**Supporting Evidence:** TDP-43 and FUS are recruited to stress granules (PMID: 21981919, 22246329); G3BP1 granules concentrate disease proteins (PMID: 32302571); prion-like domains in FUS/TDP-43 drive LLPS and aggregation; pathological inclusions co-stain with stress granule markers in ALS/FTD.\n\n**Predicted Experiment:** In vitro reconstitution with purified G3BP1, TDP-43, and FUS will test whether G3BP1 condensates nucleate seeded aggregation under conditions mimicking proteostatic stress. Single-molecule fluorescence spectroscopy will quantify the critical concentration for TDP-43 fibrillization within G3BP1 droplets. Cryo-ET will visualize amyloid-like ordered cores in persistent patient-derived granules.\n\n**Confidence:** 0.81\n\n---\n\n## Hypothesis 4: Age-Related Decline in Hsp70 Chaperone Capacity Blocks Stress Granule Reversibility\n\n**Mechanism:** Hsp70/Hsp40 chaperones (HSPA1A, DNAJB proteins) maintain stress granule dynamics by preventing aberrant inter-molecular interactions within the granule proteome. With aging, global chaperone capacity declines, and Hsp70 fails to dissolve granules stalled in the persistent state, allowing liquid-to-solid transition. Overexpression of specific Hsp70 isoforms (HSPA1B) restores granule dynamics by remodeling protein-protein interactions.\n\n**Target Gene/Protein/Pathway:** HSPA1A/HSPA1B (Hsp70), DNAJB6/DNAJB8 (Hsp40), J-domain co-chaperone network\n\n**Supporting Evidence:** Hsp70 regulates stress granule disassembly (PMID: 25437563); Hsp70 supplementation clears pathological protein aggregates in models (PMID: 27474442); aging impairs proteostasis networks (PMID: 22872689); DNAJB6 prevents aberrant phase transitions (PMID: 30392958).\n\n**Predicted Experiment:** Primary neurons from aged mice (18-24 months) will be treated with stress granules inducers (sodium arsenite) to quantify persistence duration versus young neurons. RNAi of HSPA1A/B will phenocopy aging in young neurons. Adenoviral HSPA1B expression will be tested for granule reversibility restoration in aged neurons and patient iPSC-derived motor neurons.\n\n**Confidence:** 0.67\n\n---\n\n## Hypothesis 5: C9orf72 DPR Dipeptides Corrupt G3BP1 Condensate Material Properties\n\n**Mechanism:** Dipeptide repeat proteins (DPRs) translated from C9orf72 hexanucleotide expansions—particularly poly-GA, poly-GR, and poly-PR—bind directly to G3BP1 and alter its phase separation behavior. Positively charged DPRs (GR/PR) engage in aberrant liquid-liquid phase separation with G3BP1's acidic tract, creating hybrid condensates with dramatically increased viscosity and arrested dynamics. Poly-GA seeds G3BP1 granule aggregation via amyloid-like cross-β interactions.\n\n**Target Gene/Protein/Pathway:** C9orf72 expansions producing DPRs; G3BP1 arginine-rich motifs; RNA:G3BP1:DPR ternary complexes\n\n**Supporting Evidence:** C9orf72 mutations are the most common genetic cause of ALS/FTD (PMID: 21944778); DPRs accumulate in patient neurons (PMID: 26637798); G3BP1 granules sequester C9orf72 transcripts and DPRs (PMID: 26326864); arginine-rich DPRs undergo LLPS (PMID: 31439794); poly-GA forms amyloid-like aggregates (PMID: 26951683).\n\n**Predicted Experiment:** Purified G3BP1 will be mixed with synthetic DPRs (poly-GR, poly-PR, poly-GA) to test concentration-dependent effects on G3BP1 LLPS using droplet assays and FRAP. Expression of GFP-G3BP1 with inducible DPR constructs in HeLa cells will track condensate maturation kinetics. Cryo-EM will determine whether DPRs induce G3BP1 amyloid-like order in persistent granules.\n\n**Confidence:** 0.74\n\n---\n\n## Hypothesis 6: Aberrant eIF2α Phosphorylation Creates a Stalled Ribonucleoprotein State\n\n**Mechanism:** In physiological stress, GCN2/PERK-mediated eIF2α phosphorylation triggers translational arrest and stress granule assembly. Resolution requires PP1c-mediated eIF2α dephosphorylation to restart translation and dissolve granules. In neurodegeneration, chronic low-level eIF2α phosphorylation (via PERK hyperactivation from ER stress, or PP1c dysregulation) creates a \"traffic jam\" of stalled translation complexes, preventing the clearance of G3BP1-positive granules.\n\n**Target Gene/Protein/Pathway:** PERK/GCN2 → eIF2α~P axis; PPP1R15B (PP1c regulatory subunit); eIF2B (guanine nucleotide exchange factor)\n\n**Supporting Evidence:** eIF2α phosphorylation is elevated in Alzheimer's, Parkinson's, and ALS (PMID: 25533948, 26142691); PERK haplodeficiency or PP1R15B mutations cause neurodegeneration (PMID: 25239947); eIF2α~P is required for stress granule formation; restoration of eIF2α signaling rescues neurodegeneration models (PMID: 26804002).\n\n**Predicted Experiment:** PERK kinase-dead knock-in mice or AAV-mediated PP1R15B overexpression in disease models will test whether normalization of eIF2α~P flux prevents stress granule persistence. FLIM-FRET biosensors for eIF2α~P levels in individual neurons will correlate phosphorylation kinetics with granule clearance rates. Ribosome profiling will identify translation restart deficits in persistent granule-bearing cells.\n\n**Confidence:** 0.76\n\n---\n\n## Hypothesis 7: mRNA Template Switching Converts Granules from Signaling Hubs to Aggregation Prone\n\n**Mechanism:** Physiological stress granules enrich specific mRNAs (translatome remodeling), while pathological granules accumulate aggregating-prone transcripts (e.g., expanded C9orf72, toxic 3'UTR CGG repeats). G3BP1's RNA-binding selectivity creates granules with distinct RNA \"flavors\" that determine material properties. In neurodegeneration, aberrant RNA templates promote liquid-to-solid transition by providing nucleation cores for amyloidogenic proteins.\n\n**Target Gene/Protein/Pathway:** G3BP1 RNA-binding specificity (RGG domain); structured RNA elements (G-quadruplexes, expanded repeats); RNA helicases (DDX3X, DDX6)\n\n**Supporting Evidence:** G3BP1 selectively enriches mRNA subsets (PMID: 29760419); G-quadruplex structures in stress granule mRNAs (PMID: 28416140); DDX3X and DDX6 regulate granule dynamics via ATPase activity; expanded repeats undergo pathological LLPS (PMID: 31542294); pathological granules have altered RNA composition.\n\n**Predicted Experiment:** STAMP (sequencing of transcribed RNA of membrane protein) or APEX-seq will map the RNA content of reversible versus persistent granules in patient neurons. G3BP1 RRM/RGG mutants with altered RNA specificity will test whether switching RNA content recapitulates persistence. In vitro reconstitution with patient-derived aggregate-prone mRNAs will test whether specific transcripts alter G3BP1 condensate viscosity.\n\n**Confidence:** 0.63\n\n---\n\n## Summary Table\n\n| # | Hypothesis | Primary Target | Confidence |\n|---|------------|----------------|------------|\n| 1 | CK2 hyperphosphorylation | CSNK2A1/B → G3BP1 | 0.72 |\n| 2 | Autophagy receptor impairment | TBK1 → p62/OPTN | 0.78 |\n| 3 | G3BP1-TDP-43 seeding hub | G3BP1 condensate surface | 0.81 |\n| 4 | Hsp70 chaperone decline | HSPA1A/B, DNAJB6 | 0.67 |\n| 5 | C9orf72 DPR corruption | DPRs → G3BP1 viscosity | 0.74 |\n| 6 | eIF2α stalling | PERK/PP1R15B axis | 0.76 |\n| 7 | RNA template switching | G3BP1 RGG, DDX helicases | 0.63 |\n\n---\n\n**Recommended Priority Hypotheses for Therapeutic Development:**\n1. **Hypothesis 3 (G3BP1 seeding hub)** — Highest confidence; directly addresses nucleation of pathological aggregates (TDP-43, FUS), amenable to high-throughput screening of condensate-disrupting compounds.\n2. **Hypothesis 2 (Autophagy receptor)** — Translates well to AAV or small-molecule approaches; TBK1 agonists or p62 activators could restore granule clearance.\n3. **Hypothesis 6 (eIF2α axis)** — ISRIB (eIF2B activator) already in clinical trials; directly modulates granule disassembly kinetics.",
  "tokens_used": "2782",
  "persona_id": "persona-theorist"
}