Curvature Sensing and Membrane Remodeling of the VPS37A N-terminal Domain during Autophagy

Background

Macroautophagy is responsible for degrading and recycling cellular components in eukaryotic cells. All of the recycling cargo of the cell is collected within a compartment called the autophagosome. The autophagosome starts off as a small vesicle, which expands into a cup-shaped membrane [1]. Its formation is complete when the two tips of the membrane are joined, resulting in a double-membrane compartment enclosing the cargo to be recycled. ESCRT proteins have recently been shown to play a role in this final step of autophagosome formation [2, 3]. The ESCRT machinery is involved in many membrane remodelling processes within the cell, including endosomal sorting, cell division and virus budding [4]. In an earlier study, the researchers behind this preprint reported that the ESCRT-I protein Vps37A recruits other ESCRT components to the forming autophagosome and is required for its closure. [3]. They now present molecular details of the interaction of Vps37A with membranes, offering a possible mechanism for the specific recruitment of Vps37A to the site of autophagosome closure.

---

Key findings

The UEVL domain of Vps37A is specific for autophagosome closure

In their previous work, the research group presenting this pre-print, has shown that the N-terminal of Vps37A is required for its localisation to the autophagosome [3]. This region was predicted to contain a ubiquitin E2 variant-like (UEVL) domain. Intriguingly, this domain is absent from the paralogues Vps37B-D, and no notable sequence homology to known protein structures could be found. The authors therefore decided to determine the structure of Vps37A UEVL domain by nuclear magnetic resonance (NMR). Although the domain proved to be structurally similar to other UEV domains, further experiments showed that it does not perform the same functions. Knock out of Vps37A in human cells led to accumulation of autophagy target proteins. When the wild-type (WT) protein was re-introduced, it could rescue this defect, but a chimera of Vps37A with a UEV domain from TSG101, another ESCRT-I protein, could not. The authors later showed that unlike the N-terminus of Vps37A, the TSG101 domain cannot bind to membranes, and while TSG101 UEV bound ubiquitin, Vps37A UEVL binding to ubiquitin could not be detected by NMR spectroscopy. These results show that the UEVL domain-containing N-terminus of Vps37A has a unique role in autophagosome closure, distinct from the canonical ubiquitin-binding function of other UEV domains.

 

The Vps37A N-terminus interacts with highly curved membranes containing negatively charged lipids

The authors wanted to test, if Vps37A, like its yeast homologue, also interacts with membranes [5]. For this, they incubated purified Vps37A^1-148 , which includes the UEVL domain, with liposomes of different sizes and used flotation assays to check for binding.  Vps37A^1-148 readily interacted with small, highly curved, liposomes, but binding drastically decreased when incubated with larger liposomes. As high membrane curvature leads to packing defects, the authors investigated whether this is required for Vps37A binding. Indeed, when the content of the packing defect-inducing lipid PE was reduced, binding of Vps37A^1-148 to liposomes decreased. Moreover, negatively charged lipid head groups were indispensable: when DOPG was depleted from liposomes, no Vps37A^1-148 was bound. These results are intriguing as they suggest a mechanism by which Vps37A may localise to the highly curved tips of the phagophore inside cells.

Two hydrophobic motifs represent the membrane-binding regions of the Vps37A N-terminus and their absence reduces autophagic flux

By comparing NMR spectra of the Vps37A N-terminus in aqueous solution and in bicelles, the authors could identify residues likely involved in membrane interactions. They found two conserved clusters of bulky hydrophobic amino acids. Loss of these hydrophobic motifs by either deletion or mutagenesis led to decreased binding to liposomes. Their ablation also caused reduced localisation of Vps37A to phagophores in vivo and led to an arrest of autophagosome biogenesis. This resulted in diminished autophagic flux and accumulation of autophagy targets in cells.

The Vps37A N-terminus can remodel liposomes to higher-order structures

The authors measured liposome size by dynamic light scattering (DLS) and found that when Vps37A^1-148 was added to the liposomes, their average size increased. Importantly, the effect was only seen under conditions when Vps37A^1-148 binds to membranes (high curvature and negative charge), and was more pronounced when the hydrophobic membrane binding motifs were intact. The authors demonstrated that the DLS observations correspond to membrane remodelling by using negative stain electron microscopy. Upon addition of the WT Vps37A^1-148, liposomes visibly increased in size and rosette-like liposome clusters started to form. These remodelling events seem to be accompanied by bilayer destabilisation, as Mn²⁺ leakage from liposomes could be measured in an NMR-based assay upon protein addition.

 

---

What I like about the preprint

I enjoyed reading this preprint because it considers principles I often think about in my own research. I like the idea that curvature sensing plays a role in how proteins are targeted to the right destinations within cells. This is something that also applies at the highly curved cristae membranes in mitochondria, which is the cellular compartment of interest to my own work. I also liked the versatile use of the NMR method to address different problems within this study. What really sparked my curiosity in this preprint was the way in which the authors followed up on their previous paper in order to dig deeper into the molecular details of Vps37A’s recruitment to the phagophore membrane. This not only helps to uncover principles of protein-membrane interactions in general, but may also be important should this step of autophagy become a medical target.

 

Future directions and questions for the authors

1. You mention the need to identify interaction partners of Vps37A on the phagophore membrane. Is it known whether the ESCRT-I complex is pre-assembled in the cytosol or whether it assembles sequentially on the membrane? Did you check the localisation of other ESCRT-I subunits in the VPS37A membrane binding mutant?
2. To perhaps make the data more physiological, are you able to also purify the full-length Vps37A? It would be interesting to determine whether full-length Vps37A can also bind to the membrane, or if some conformational changes – induced for example by binding partners (other ESCRT-I members?) – are needed for membrane binding.
3. Finally, in Figure 3b/c, it looks like the mutation of the C-terminal hydrophobic motif rescues the deletion of the N-terminal hydrophobic motif. Do you have an explanation for this?

 

---

References:

1. Nakatogawa, H., Mechanisms governing autophagosome biogenesis. Nature Reviews Molecular Cell Biology, 2020. 21(8): p. 439-458.
2. Takahashi, Y., et al., An autophagy assay reveals the ESCRT-III component CHMP2A as a regulator of phagophore closure. Nature Communications, 2018. 9.
3. Takahashi, Y., et al., VPS37A directs ESCRT recruitment for phagophore closure. Journal of Cell Biology, 2019. 218(10): p. 3336-3354.
4. Vietri, M., M. Radulovic, and H. Stenmark, The many functions of ESCRTs. Nature Reviews Molecular Cell Biology, 2020. 21(1): p. 25-42.
5. Kostelansky, M.S., et al., Molecular architecture and functional model of the complete yeast ESCRT-I heterotetramer. Cell, 2007. 129(3): p. 485-98.

 

 

 

doi: https://doi.org/10.1242/prelights.33321

Read preprint (No Ratings Yet)