Differential assembly of RNP granules via activation of distinct dsRNA sensors by adenovirus mutants

This study demonstrates that distinct adenovirus mutants differentially activate dsRNA sensors to drive the assembly of unique ribonucleoprotein condensates—PKR-dependent stress granules or PKR-independent RNase L-dependent bodies—revealing novel pathways linking viral RNA sensing to translational control and granule formation.

The Gist

Imagine your body is a bustling city, and your cells are the individual buildings. When a virus (like a burglar) tries to break in, the city has a sophisticated alarm system to detect the intruder and shut down operations to prevent damage.

This paper is about how the Adenovirus (a specific type of burglar) tries to trick this alarm system, and how the city's security guards (our cells) react differently depending on which part of the virus is missing.

Here is the story of the "Great Cell Granule Mystery," broken down into simple parts:

1. The Alarm System: The "dsRNA" Sensors

Inside our cells, there are special sensors that look for double-stranded RNA (dsRNA). Think of dsRNA as a "Red Flag" or a "Distress Signal." Viruses often create these flags when they are copying themselves.

• PKR: The first guard. When it sees a Red Flag, it sounds a siren and tells the cell to stop all construction work (translation) immediately.
• RNase L: The second guard. When it sees a Red Flag, it pulls out a shredder and destroys the cell's blueprints (RNA) to stop the virus from copying.

2. The Two Burglar Variants

The researchers studied two different versions of the Adenovirus, both of which are missing a piece of their "toolkit."

• The "Missing VA" Virus (∆VA): This virus is missing a specific tool called "VA RNA."

  • What happens: Even though this virus doesn't actually create a visible Red Flag (dsRNA), the PKR guard still gets triggered! It thinks, "Something is wrong!" and shuts down construction.
  • The Result: The cell builds Stress Granules (SGs). Imagine these as emergency bunkers. The cell gathers all its important tools and blueprints into these bunkers to protect them while it waits for the threat to pass.

• The "Missing Splicing" Virus (∆E4): This virus is missing a tool that helps it organize its instructions.

  • What happens: Because it's messy, this virus accidentally creates a huge pile of Red Flags (dsRNA) inside the cell's nucleus.
  • The Result: Both guards (PKR and RNase L) go crazy. The shredder (RNase L) starts destroying blueprints. The cell responds by building RLBs (RNase L-dependent bodies).
  • The Difference: These aren't bunkers; they are trash compactors. They are small, round, and filled with the shredded remains of the blueprints, ready to be hauled away.

3. The Big Discovery: Granules Look Different

The researchers looked closely at these "bunkers" and "trash compactors."

• Stress Granules (SGs) are like a library where books are neatly stacked and protected. They contain specific proteins (like librarians) that keep things organized.
• RLBs are like a recycling plant. They contain different proteins (like trash collectors) and are filled with broken-down pieces.

The study found that the virus-induced granules looked exactly like the ones the cell makes when it's stressed by chemicals. The cell knows exactly which type of "container" to build based on which alarm is ringing.

4. The Twist: The "Ghost" Mechanism

Here is the most surprising part of the story.

Usually, to build the "trash compactor" (RLB), you need the RNase L shredder to be working.

• The Experiment: The scientists removed the shredder (RNase L) from the cells.
• The Expectation: They thought the "trash compactor" wouldn't form because there was no shredder to make the trash.
• The Reality: When the "Missing Splicing" virus (∆E4) infected these cells, the trash compactors still formed!

Even without the shredder, the cell built the RLBs and stopped construction. This means there is a secret, hidden pathway the cell uses. It's like the cell has a "Plan B" alarm that doesn't need the shredder to tell it to build a trash compactor. The virus accidentally triggered this secret alarm by piling up Red Flags in the nucleus.

Summary: The Takeaway

• Different Triggers, Different Responses: The cell has different ways to react to viral trouble. One leads to "bunkers" (Stress Granules) to protect resources, and the other leads to "trash compactors" (RLBs) to destroy the enemy.
• Virus Tricks: The virus tries to hide its tracks, but by missing certain tools, it accidentally sets off different alarms.
• Hidden Backup Plans: The cell is smarter than we thought. Even if we take away its main weapons (like the shredder), it can still figure out how to build the right defense structures using a secret, backup pathway.

In short: This paper shows that our cells are like a highly trained security team that can build different types of emergency shelters depending on exactly what kind of alarm is ringing, and they have secret backup plans that scientists are just beginning to understand.

Technical Summary

1. Problem Statement

Cells employ double-stranded RNA (dsRNA) sensors to detect viral infections, triggering antiviral responses such as translational arrest and RNA decay. Two major pathways involve Protein Kinase RNA-activated (PKR) and the OAS/RNase L system. Activation of these pathways leads to the formation of distinct cytoplasmic ribonucleoprotein (RNP) condensates: Stress Granules (SGs) and RNase L-dependent bodies (RLBs).

While it is known that adenovirus (AdV) mutants lacking Virus-Associated (VA) RNAs ($\Delta$VA) activate PKR, and mutants lacking the E4 region ($\Delta$E4) accumulate nuclear dsRNA, the specific downstream consequences on RNP granule assembly and the precise sensor requirements for these events remain unclear. Specifically, it is unknown whether:

• Different dsRNA sensors drive the formation of distinct granule types (SGs vs. RLBs).
• Virus-induced granules share the same compositional and mechanistic properties as chemically induced granules.
• Granule assembly and translational arrest can occur independently of canonical sensors (PKR, RNase L) and the eIF2$\alpha$ phosphorylation axis.

2. Methodology

The study utilized a combination of virology, cell biology, proteomics, and genetic knockout models:

• Viral Models: Infection of A549, HEK293, HeLa, and U2OS cells with Adenovirus serotype 5 (Ad5) wild-type (WT) and two mutants:
  • $\Delta$VA: Lacks VA RNA I and II (known to activate PKR).
  • $\Delta$E4: Lacks the E4 region (splicing-defective, known to accumulate nuclear dsRNA).
• dsRNA Detection: Used dsRNA-specific monoclonal antibodies (9D5 and J2) for immunofluorescence to visualize dsRNA accumulation.
• Sensor Activation Assays:
  • PKR/RL Pathways: Western blotting for phosphorylated PKR (p-PKR), p-eIF2$\alpha$, p-IRF3, and RT-qPCR for interferon genes.
  • RNase L Activity: Bioanalyzer analysis of ribosomal RNA (18S/28S) integrity to detect degradation.
• Granule Characterization:
  • Immunofluorescence (IF): Staining for markers G3BP1, PABPC1, FXR1, UBAP2L, and others to distinguish SGs from RLBs.
  • Proteomics (APEX2 Proximity Labeling): Used HEK293T cells expressing G3BP1-APEX2-GFP to biotinylate proteins in proximity to granules induced by Sodium Arsenite (SA, for SGs) or poly(I:C) (for RLBs). Mass spectrometry (nLC-MS/MS) identified differential protein enrichment.
• Genetic Knockouts (KO): Utilized A549 and U2OS cell lines with single or double knockouts of PKR, RNase L, OAS1-3, and G3BP1/2 to determine sensor dependencies.
• Functional Assays:
  • Translation: Puromycin incorporation assays to measure active translation.
  • RNA Turnover: Single-molecule FISH (smFISH) for GAPDH mRNA to assess specific transcript degradation.
  • Pharmacological Inhibition: Treatment with Cycloheximide (CHX) to block translation initiation and ISRIB to rescue eIF2$\alpha$-dependent translation.

3. Key Contributions

1. Differential Sensor Activation: Demonstrated that $\Delta$VA mutants activate PKR only (without detectable dsRNA), whereas $\Delta$E4 mutants activate both PKR and OAS3/RNase L (via nuclear dsRNA accumulation). Neither mutant activated the RIG-I-like receptor (RLR) pathway.
2. Granule Specificity: Established that PKR activation alone drives the formation of Stress Granules (SGs), while combined PKR and RNase L activation drives the formation of RLBs.
3. Proteomic Definition: Provided a comparative proteomic map of SGs and RLBs, identifying unique markers (e.g., FAM120A for SGs; LSm14A for RLBs) and confirming that virus-induced granules recapitulate these chemical-stress-induced compositions.
4. Non-Canonical Pathway Discovery: Revealed a novel, PKR/RNase L/eIF2$\alpha$-independent pathway triggered by nuclear dsRNA accumulation that leads to translational arrest and RLB-like granule assembly.

4. Key Results

A. Differential Activation of dsRNA Sensors

• $\Delta$VA Infection: No detectable dsRNA was found, yet PKR and eIF2$\alpha$ were phosphorylated. This led to the formation of large, irregular cytoplasmic granules resembling SGs.
• $\Delta$E4 Infection: Robust nuclear dsRNA accumulation was detected. This activated both PKR and the OAS3/RNase L pathway (evidenced by 18S/28S rRNA degradation). This resulted in the formation of small, spherical granules resembling RLBs, accompanied by nuclear relocalization of PABPC1.
• RLR Pathway: Neither mutant induced IRF3 phosphorylation or significant interferon production, suggesting AdV effectively suppresses or avoids RLR detection.

B. Granule Composition and Morphology

• Morphology: $\Delta$VA-induced granules were large and SG-like; $\Delta$E4-induced granules were small, spherical, and RLB-like.
• Proteomics: APEX2 labeling revealed distinct proteomes.
  • Shared: FXR1, FMRP, Caprin1, UBAP2L.
  • SG-specific: FAM120A, eIF4G1, PRRC2C, Ago2.
  • RLB-specific: LSm14A (significantly enriched).
• Viral Mimicry: Virus-induced granules recruited the same specific markers as their chemically induced counterparts (e.g., $\Delta$E4 granules recruited LSm14A but not eIF4G1).

C. Sensor Dependencies (The "Knockout" Findings)

• $\Delta$VA Granules (SGs): Formation was strictly PKR-dependent. No granules formed in PKR KO cells.
• $\Delta$E4 Granules (RLB-like):
  • In RNase L KO cells, $\Delta$E4 infection still induced granules with RLB morphology (small, spherical) and recruited FXR1/UBAP2L, despite the absence of rRNA degradation.
  • In PKR/RNase L Double KO cells, $\Delta$E4 infection still induced granules. This was a critical finding, proving that granule assembly can occur independently of the two major dsRNA sensors.
  • Note: In Double KO cells, poly(I:C) transfection failed to induce any granules, highlighting that the $\Delta$E4 effect is specific to the viral context/nuclear dsRNA.

D. Translational Arrest Mechanisms

• $\Delta$E4 in Double KO: Despite the lack of PKR, RNase L, and eIF2$\alpha$ phosphorylation, $\Delta$E4 infection caused a strong reduction in puromycin incorporation (translational arrest).
• ISRIB Resistance: Treatment with ISRIB (which bypasses eIF2$\alpha$ inhibition) rescued translation in SA-treated cells but failed to rescue translation in $\Delta$E4-infected Double KO cells.
• Conclusion: $\Delta$E4 infection triggers a non-canonical translational shutoff mechanism that does not rely on eIF2$\alpha$ phosphorylation, PKR, or RNase L.

5. Significance

• Redefining dsRNA Sensing: The study challenges the dogma that cytoplasmic dsRNA sensors (PKR, OAS) are the sole drivers of RNP granule assembly. It identifies a mechanism where nuclear dsRNA accumulation triggers cytoplasmic stress responses via alternative, yet-to-be-defined pathways.
• Granule Plasticity: It demonstrates that the cell can assemble RNP condensates with specific compositions (RLB-like) even in the absence of the canonical enzymatic activity (RNase L cleavage) usually required for their formation.
• Viral Evasion vs. Host Response: The findings suggest that while AdV mutants have evolved to evade certain sensors (like RLRs), the host cell has redundant or alternative mechanisms to detect viral stress (specifically nuclear dsRNA) and shut down translation, potentially as a fail-safe antiviral defense.
• Therapeutic Implications: Understanding these non-canonical pathways could reveal new targets for antiviral therapies or insights into diseases involving RNP granule dysregulation (e.g., neurodegeneration).

In summary, Steinbock et al. provide a comprehensive map of how distinct adenovirus mutants differentially engage host sensors to drive the assembly of specific RNP granules, uncovering a surprising, sensor-independent pathway for translational control and granule formation triggered by nuclear viral RNA.