In situ visualization of autophagy suggests vesicle fusion can contribute to phagophore expansion

Using in situ cryo-electron tomography on S. cerevisiae, this study reveals that cytosolic vesicle fusion contributes to phagophore expansion during later stages of autophagosome biogenesis, challenging the view that lipid transfer by the Atg2-Atg18 complex is the sole mechanism for membrane growth.

The Gist

The Big Picture: The Cell's Recycling Center

Imagine your cell is a bustling city. Every day, this city produces trash—broken machines, old furniture, and protein clumps. To keep the city clean, it has a recycling center called the lysosome.

But you can't just throw big piles of trash into the recycling bin. You need to wrap them up first. In the cell, this wrapping process is called autophagy (literally "self-eating"). The cell builds a special double-layered bubble called an autophagosome to encase the trash before sending it to the recycling center.

The tricky part is building that bubble. It starts as a small, open cup (called a phagophore) that needs to grow huge to swallow the cargo. The question this paper answers is: How does this cup grow so big so fast?

The Main Characters: The Construction Crew

To build this bubble, the cell uses a team of construction workers (proteins). Two of the most important are:

1. Atg2: The "Lipid Truck." It drives back and forth between the cell's main warehouse (the Endoplasmic Reticulum) and the growing bubble, delivering raw materials (lipids/fats) to expand the bubble's walls.
2. Atg9: The "Site Manager." It sits at the edge of the bubble and tells the truck where to drop off the materials.

In a healthy cell, Atg2 and Atg9 hold hands (interact) at the rim of the bubble. This ensures the truck delivers materials exactly where they are needed, and the bubble expands smoothly.

The Experiment: Breaking the Handshake

The scientists wanted to see what happens if the "Site Manager" (Atg9) and the "Lipid Truck" (Atg2) can't hold hands. They created a mutant yeast cell where these two proteins are broken apart (the Atg2-PM4 mutant).

What they expected: The bubble would stop growing because the truck couldn't deliver materials properly.
What they actually saw: The bubble did grow, but it looked weird. The edges (rims) of the bubble became massive and bloated, like a balloon that got stuck halfway through inflating.

The Discovery: The "Delivery Van" Backup Plan

Using a super-powerful microscope (Cryo-Electron Tomography), the scientists took 3D snapshots of these bloated bubbles. They noticed something surprising:

In the mutant cells, there were tiny bubbles (vesicles) floating right next to the giant, bloated rim of the phagophore. In some cases, they saw these tiny bubbles actually fusing (merging) with the rim.

The Analogy:
Imagine you are trying to fill a giant swimming pool (the phagophore) using a garden hose (Atg2).

• Normal Cell: The hose is connected perfectly, and water flows in smoothly.
• Mutant Cell: The hose connection is loose. The water flow is slow and messy.
• The Surprise: Because the hose is struggling, the city sends in delivery vans (the vesicles) to dump buckets of water directly into the pool to help fill it up.

The scientists found that when the main "hose" (Atg2) isn't working efficiently, the cell switches to a backup plan: it fuses extra vesicles directly onto the edge of the bubble to help it expand.

Why This Matters

For a long time, scientists thought the cell only used the "hose" method (lipid transfer) to build these bubbles. They thought the "delivery van" method (vesicle fusion) only happened at the very beginning to start the bubble.

This paper shows that vesicle fusion is a backup strategy the cell uses when the main system is slow or broken. It proves that the cell is flexible and has multiple ways to get the job done.

The Takeaway

• The Problem: The cell needs to build a recycling bubble.
• The Normal Way: A protein truck (Atg2) delivers fats to the edge of the bubble.
• The Glitch: When the truck gets confused (Atg2-PM4 mutant), the bubble gets stuck with a giant, weird rim.
• The Solution: The cell notices the delay and starts fusing extra bubbles (vesicles) directly onto the rim to help it grow.

In short: The cell is like a smart construction crew. If the main conveyor belt breaks, they don't stop building; they just start throwing bricks in by hand (or in this case, fusing extra bubbles) to make sure the recycling center gets built.

Technical Summary

1. Problem Statement

Autophagy is a conserved cellular process where cytosolic cargo is degraded via the lysosome/vacuole. A critical step is autophagosome biogenesis, specifically the expansion of the phagophore (isolation membrane) into a closed double-membrane vesicle.

• Current Understanding: The core machinery involves the Atg9-Atg2-Atg18 complex. Atg2 is known to bridge the endoplasmic reticulum (ER) and the phagophore, transferring lipids unidirectionally to drive expansion. Atg9 acts as a lipid scramblase.
• The Gap: While the biochemical roles of Atg2 and Atg9 are characterized, the spatiotemporal coordination of lipid delivery during the intermediate stages of phagophore expansion remains unclear. Specifically, it is unknown if mechanisms other than direct lipid transfer from the ER (e.g., vesicle fusion) contribute to phagophore growth, particularly when the primary lipid transfer machinery is impaired.

2. Methodology

The authors employed a multi-modal approach combining live-cell fluorescence microscopy with high-resolution structural biology in Saccharomyces cerevisiae (yeast).

• Genetic Models:
  • Atg2-PM4 Mutant: A variant of Atg2 with five C-terminal mutations (QKFST $\to$ AAAAA) that disrupts the interaction between Atg2 and Atg9. This mutant causes decelerated autophagosome biogenesis and mislocalization of Atg2 away from the phagophore rims.
  • $snf7\Delta$ Strain: A deletion of the ESCRT-III component Snf7, which prevents phagophore closure, causing an accumulation of open, expanding phagophores suitable for imaging.
• Live-Cell Fluorescence Microscopy:
  • Used mCherry-Atg8 and GFP-Atg2 (WT vs. PM4) to track autophagosome formation kinetics, puncta counts, and lifetimes.
  • Used Vps8-mScarlet to monitor "Class E compartments" (aberrant endosomes) to ensure the $snf7\Delta$ strain did not introduce artifacts affecting phagophore biogenesis.
• Cryo-Correlative Light and Electron Microscopy (Cryo-CLEM):
  • Cells were vitrified, and regions of interest (Atg9-positive puncta) were identified via cryo-fluorescence.
  • FIB-SEM Milling: Focused Ion Beam milling was used to create thin cellular lamellae (~150–200 nm) for electron tomography.
• In Situ Cryo-Electron Tomography (Cryo-ET):
  • Tomograms were acquired on Talos Arctica and Titan Krios microscopes.
  • Image Processing: Data was processed using WARP, IMOD/AreTomo, and IsoNet (for denoising).
  • Segmentation & Quantification: Membranes were segmented using MemBrain v2 (deep learning). Custom tools (CryoVIA) were used to quantify intermembrane distances, curvature, and rim opening angles.

3. Key Contributions

• Novel Phenotype Identification: The study reveals that disrupting the Atg2-Atg9 interaction (Atg2-PM4) leads to a specific morphological defect: enlarged phagophore rims (intermembrane spacing of 50–200 nm) compared to the wild-type (10–20 nm).
• Discovery of Vesicle Fusion: The authors provide the first in situ visual evidence of cytosolic vesicles (40–200 nm) fusing with the expanding phagophore rims in the Atg2-PM4 mutant.
• Alternative Expansion Mechanism: The paper proposes that when the primary lipid transfer pathway (Atg2-mediated) is impaired, cells compensate by utilizing vesicle fusion to supply lipids for phagophore expansion.

4. Key Results

A. Decelerated Biogenesis in Atg2-PM4

• Live-cell imaging showed that Atg2-PM4 cells have significantly fewer closed autophagosomes (0.2 per cell vs. 0.7 in WT).
• The lifetime of Atg8-positive structures was doubled (22 min vs. 10 min), confirming that the mutation slows down the entire biogenesis process.
• Control experiments confirmed that the $snf7\Delta$ background did not alter the association of phagophores with Class E compartments, validating the model.

B. Altered Phagophore Ultrastructure

• Wild-Type (Atg2-WT): Phagophores exhibited characteristic cup-shaped membranes with closely apposed bilayers (10–20 nm spacing) and protein-mediated membrane contact sites (MCS) with the ER and lipid droplets (bridges ~20–24 nm long, consistent with Atg2/Vps13).
• Atg2-PM4 Mutant:
  • Enlarged Rims: 16 out of 25 observed rims showed a massive expansion in intermembrane spacing (50–200 nm).
  • Morphometric Analysis: Segmentation revealed that the "tip" width of enlarged rims was ~284 Å (vs. ~127 Å in WT), with significantly higher curvature.
  • Contact Sites: Despite the defect, MCS with the ER and lipid droplets were still present, though some tethering densities were longer (29 nm), potentially indicating the involvement of other proteins like Csf1.

C. Vesicle Fusion Events

• Observation: In Atg2-PM4 tomograms, 40–200 nm vesicles were frequently found in close proximity (<200 nm) to phagophore rims. This was not observed in WT cells.
• Fusion Evidence: In one specific tomogram, the authors captured a phagophore with two vesicles attached to opposite rims, connected by a membrane stalk (~100 nm wide). The vesicles appeared to be fusing with the phagophore rim rather than the flat backbone.
• Vesicle Identity: The vesicles lacked recognizable protein coats (clathrin, COPII, COPI) and were larger than canonical COPII vesicles (60–80 nm). Their origin remains unidentified but is distinct from Class E compartments.

D. Quantitative Correlation

• In WT cells, rim opening angle correlated positively with tip width (consistent with growth).
• In Atg2-PM4 "enlarged" rims, this correlation was lost, suggesting a structural deviation from the standard growth trajectory, likely due to the physical addition of vesicle membranes.

5. Significance and Conclusion

• Mechanistic Insight: The study challenges the view that phagophore expansion relies solely on lipid transfer from the ER via Atg2. It demonstrates that vesicle fusion is a viable, compensatory mechanism for membrane expansion, particularly when the primary lipid transfer machinery is compromised.
• Physiological Relevance: The observation that vesicles fuse specifically with the high-curvature rims (rather than the flat sheet) suggests a geometric preference for fusion events during autophagosome biogenesis.
• Broader Implications: This finding adds a new layer to the understanding of autophagy, suggesting a plasticity in lipid supply mechanisms. It implies that cells possess redundant pathways (Atg2-mediated transfer vs. vesicle fusion) to ensure autophagosome formation even under stress or mutation.
• Future Directions: The study highlights the need to identify the specific origin and cargo of these vesicles and the molecular signals triggering their fusion with the phagophore.

In summary, this paper utilizes advanced in situ cryo-ET to visualize a "backup" mechanism for autophagosome expansion, revealing that when the Atg2-Atg9 lipid transfer axis is disrupted, cells resort to fusing cytosolic vesicles with the phagophore rim to sustain growth.