Proteolytic dissection of eIF4G reveals the closed-loop mRNP as an architecture for translation repression.

This study demonstrates that while the eIF4G-PABP closed-loop architecture is dispensable for productive translation initiation, it can be actively co-opted through specific proteolytic cleavage to form a dead-end complex that potently represses translation, thereby explaining the distinct translational outcomes of factor depletion versus viral cleavage.

The Gist

The Big Picture: A Factory, a Manager, and a Broken Loop

Imagine a cell as a massive factory that builds proteins (the workers and machines that keep the cell alive). To build these proteins, the factory needs to read instructions (mRNA) and assemble them efficiently.

For decades, scientists believed there was one specific way this factory had to work to be efficient: a "Closed-Loop" system.

• The Analogy: Imagine the instruction manual (mRNA) is a long scroll. To read it efficiently, the factory manager (a protein called eIF4G) grabs the start of the scroll (the 5' cap) and the end of the scroll (the 3' tail) and ties them together in a circle. This "closed loop" was thought to be essential for the factory to run fast and smooth. If you broke the loop, the factory should stop working.

This paper proves that old idea wrong. The researchers found that the factory can actually run perfectly fine even if the scroll is left open and straight (linear). In fact, the "closed loop" isn't a helper for building; it's actually a trap used by viruses to shut the factory down.

---

The Key Discoveries (The Plot Twist)

The researchers used a special "emergency off-switch" (called AID technology) to instantly remove the main manager (eIF4G1) from the factory. Here is what they found:

1. The Factory Didn't Stop (The Backup Manager)

When they removed the main manager, the factory slowed down a bit, but then recovered.

• Why? They discovered a "backup manager" called eIF4G3.
• The Analogy: Think of eIF4G1 as the CEO and eIF4G3 as the Vice President. When the CEO is suddenly fired, the VP steps up. The VP wasn't doing much before, but as soon as the CEO is gone, the VP takes charge and keeps the factory running. Previous studies missed this because they took too long to remove the CEO, allowing the VP to take over before anyone noticed.

2. The "Closed Loop" is Optional

The researchers tested if the factory needed the manager to tie the start and end of the scroll together.

• The Experiment: They created a version of the manager that couldn't grab the end of the scroll (the tail).
• The Result: The factory kept running at full speed!
• The Lesson: You don't need to tie the scroll into a loop to build proteins. The "closed loop" is not required for the machine to work.

3. The Virus Trap (The "Dead-End" Loop)

So, if the loop isn't needed, why do viruses (like the ones that cause the common cold or polio) cut the manager in half?

• The Virus Strategy: Enteroviruses have a special pair of scissors (2A protease) that cuts the manager.
  • Piece A (The C-terminal): This piece helps the virus read its own instructions.
  • Piece B (The N-terminal / 2A-cpN): This is the scary part. This piece can grab the start and end of the scroll and tie them into a loop, BUT it lacks the tools to actually build anything.
• The Trap: The virus uses this piece to tie up all the factory's resources into a "closed loop" that looks perfect but is completely dead. It's like a factory manager who ties all the blueprints into a knot and locks the door, preventing any real work from happening. This is how the virus shuts down the human factory to take over.

4. The Stress Response (Caspase-3)

When a cell is under stress or dying, a different pair of scissors (Caspase-3) cuts the manager.

• The Result: This cut leaves a middle piece that can still build proteins, even without the "loop" capability.
• The Lesson: This suggests that when a cell is stressed, it doesn't just shut down; it switches to a "survival mode" where it builds proteins in a simpler, open-loop way.

5. The Tail's Real Job (Stability, Not Speed)

Finally, they looked at the "tail" of the scroll (PABP).

• The Finding: Removing the tail didn't stop the factory from building; it just made the instruction manuals fall apart and disappear faster.
• The Analogy: The tail isn't the engine that drives the factory; it's the protective cover for the manual. Without it, the manual gets damaged and lost, but as long as the manual exists, the factory can still work without the "loop."

---

Summary: What This Means for Us

1. The "Closed Loop" isn't the engine: We used to think tying the start and end of the mRNA together was the secret to fast protein production. This paper says: Nope. The factory works fine with a straight scroll.
2. Viruses are clever tricksters: Viruses don't just break the factory; they use a "broken" piece of the manager to tie the factory into a useless knot, effectively silencing it.
3. Cells have backups: Our cells have a hidden backup system (eIF4G3) that kicks in when the main system is damaged, which explains why cells are so resilient.
4. Stress changes the rules: When cells are stressed, they switch to a different way of building proteins that doesn't rely on the old "loop" model.

In short: The "closed loop" isn't a requirement for life; it's a feature that can be hijacked by viruses to kill the cell. The cell's ability to keep working without it is what keeps us alive.

Technical Summary

1. Problem Statement

The prevailing model of cap-dependent translation initiation posits that the formation of a "closed-loop" messenger ribonucleoprotein (mRNP) complex is essential for efficient protein synthesis. This model suggests that the 5' cap and 3' poly(A) tail are bridged by interactions between eIF4E (cap-binding), eIF4G (scaffold), and PABP (poly(A)-binding protein), thereby enhancing ribosome recycling and initiation efficiency.

However, direct in vivo validation of this model in mammalian cells has been elusive due to two main limitations:

1. Compensatory Mechanisms: Traditional siRNA-mediated knockdown of eIF4G takes days, allowing cells to activate compensatory pathways or upregulate paralogs, obscuring the acute requirement for eIF4G.
2. Confounding Stability Defects: Depletion of PABP or eIF4G often leads to rapid mRNA destabilization, making it difficult to distinguish between defects in translation initiation and defects in mRNA stability.
3. Viral vs. Cellular Cleavage: It remains unclear why viral proteolytic cleavage of eIF4G (e.g., by enteroviruses) causes near-complete translational shutoff, whereas simple depletion of eIF4G does not.

2. Methodology

The authors employed a combination of acute protein depletion, genetic rescue, and proteolytic fragment analysis to dissect the roles of eIF4G and its interactors:

• Auxin-Inducible Degron (AID) System: The authors generated HeLa and A549 cell lines with endogenous eIF4G1 tagged with an AID-Flag sequence. This allowed for the rapid (within 2–4 hours) and acute depletion of eIF4G1 upon auxin treatment, bypassing compensatory transcriptional responses.
• CRISPR/Cas9 Knockouts:
  • eIF4G3 knockout to test the role of the paralog in buffering translation loss.
  • PABPC4 knockout combined with PABPC1-AID-EGFP tagging to acutely deplete both major cytoplasmic PABP paralogs.
• Tet-On Inducible Rescue Systems: In AID-depleted cells, the authors expressed various exogenous constructs via Doxycycline induction:
  • Full-length eIF4G1.
  • Enteroviral 2A protease cleavage products: N-terminal (2A-cpN) and C-terminal (2A-cpC).
  • Caspase-3 cleavage products: N-terminal (casp3-cpN), Middle (casp3-cpM), and C-terminal (casp3-cpC).
  • Point mutants of eIF4G1 disrupting PABP binding (ΔPABP), eIF4E binding (ΔeIF4E), or both.
• Assays:
  • [35S]-Metabolic Labeling: To measure global nascent protein synthesis rates.
  • Western Blotting & Co-IP: To verify protein levels, degradation kinetics, and protein-protein interactions.
  • Northern Blotting: To assess mRNA stability (Actin vs. Histone H2A) independent of translation.
  • Viral Infection: Infection with Enterovirus A71 (EV-A71) to assess viral protein production in the context of specific eIF4G fragments.

3. Key Contributions and Results

A. eIF4G3 Buffers Basal Translation

• Finding: Acute depletion of eIF4G1 reduced global translation by ~50% at 6 hours, but synthesis recovered by 24 hours despite the continued absence of eIF4G1.
• Mechanism: This recovery was mediated by eIF4G3, a previously underappreciated paralog. In eIF4G3 knockout cells, acute eIF4G1 depletion caused a sustained ~50% reduction in translation with no recovery.
• Significance: This explains why prior siRNA studies (which take days) showed modest defects; the slow depletion allowed eIF4G3 to compensate. eIF4G3 acts as a latent buffer that is mobilized when eIF4G1 is compromised.

B. The Closed-Loop is Dispensable for Bulk Translation

• Finding: The central fragment of eIF4G generated by Caspase-3 cleavage (casp3-cpM), which retains eIF4E and eIF3/eIF4A binding domains but lacks the PABP-binding domain, fully rescued global protein synthesis in the absence of eIF4G1.
• Finding: Full-length eIF4G1 mutants unable to bind PABP (ΔPABP) also fully rescued translation.
• Finding: Acute depletion of both PABPC1 and PABPC4 caused a delayed reduction in protein synthesis (at 24 hours) primarily due to mRNA destabilization, not an acute failure of translation initiation.
• Conclusion: The eIF4G-PABP "closed-loop" architecture is not required for productive bulk translation initiation. Linear mRNPs can support efficient initiation.

C. Enteroviral 2A-cpN is a Dominant Repressor via Closed-Loop Sequestration

• Finding: Unlike caspase fragments, the N-terminal fragment of eIF4G generated by Enteroviral 2A protease (2A-cpN) acted as a potent, dominant translational repressor.
• Mechanism: 2A-cpN retains eIF4E and PABP binding domains but lacks the MIF4G domain required for 43S PIC recruitment. It forms a "dead-end" closed-loop mRNP that sequesters eIF4E and PABP, preventing productive initiation.
• Requirement: The repressive activity of 2A-cpN strictly required simultaneous binding to both eIF4E and PABP. Mutants lacking either interaction failed to repress translation.
• Significance: This resolves the paradox of why viral infection causes total shutoff while simple depletion does not. The virus does not just remove the scaffold; it actively co-opts the closed-loop architecture to silence host translation.

D. Caspase-3 Cleavage Rewires Rather Than Silences

• Finding: Caspase-3 cleavage products (specifically casp3-cpM) support translation even in the absence of PABP interaction.
• Implication: During cellular stress (e.g., sub-lethal apoptosis), eIF4G cleavage may remodel the translational machinery to function in a closed-loop-independent manner, potentially prioritizing survival transcripts or altering termination fidelity, rather than simply shutting down protein synthesis.

4. Significance and Conclusion

This study fundamentally revises the understanding of the "closed-loop" model of translation:

1. Reinterpretation of the Closed-Loop: The eIF4E-eIF4G-PABP bridge is not a universal requirement for cap-dependent translation initiation. Bulk translation proceeds efficiently on linear mRNPs. The closed-loop architecture likely functions primarily in post-initiation processes (e.g., mRNA stability, termination fidelity, and NMD suppression) or for specific subsets of mRNAs, rather than driving the initial recruitment of the 43S complex.
2. Dual Nature of Proteolysis: The paper demonstrates that proteolytic processing of eIF4G can yield diametrically opposite outcomes based on the cleavage site:
   • Caspase-3 cleavage generates fragments that sustain translation (potentially for survival).
   • Enteroviral 2A cleavage generates fragments (2A-cpN) that actively repress translation by hijacking the closed-loop structure.
3. Paralog Redundancy: It establishes eIF4G3 as a critical physiological buffer that maintains basal translation when eIF4G1 is compromised, a mechanism previously masked by the slow kinetics of RNAi.
4. Mechanism of Viral Shutoff: The near-complete translational shutoff observed during enterovirus infection is driven not just by the loss of the eIF4G scaffold, but by the active gain-of-function of the 2A-cpN fragment, which sequesters initiation factors into non-productive closed-loop complexes.

In summary, the authors demonstrate that the closed-loop mRNP is a flexible architecture that can be co-opted for translational silencing by viruses, while productive translation in mammalian cells is robust and does not strictly depend on the eIF4G-PABP bridge.