Menu

Close

The autophagic membrane tether ATG2A transfers lipids between membranes

Shintaro Maeda, Chinatsu Otomo, Takanori Otomo

Preprint posted on February 20, 2019 https://www.biorxiv.org/content/10.1101/555441v1

Article now published in eLife at http://dx.doi.org/10.7554/eLife.45777

Fattening up for autophagy: lipid transfer by ATG2A could enable expansion of the phagophore

Selected by Sandra Malmgren Hill

Context
Autophagy is an essential process for degradation and recycling of material where cytosolic material, including whole organelles, are enclosed in a double-membraned vesicle and fused to the lysosome for proteolytic digestion. The location of autophagosome formation is determined by a local enrichment of the lipid signaling molecule phosphatidyinositol-3-phosphate (PI(3)P) (Bento et al., 2016). This lipid signal recruits PI(3)P-binding proteins to the site, and a cascade of protein recruitment allows for membrane curvature, and formation of a cup-shaped membrane precursor called the phagophore. The phagophore expands and encloses its cargo to form the mature autophagosome, and the autophagosome fuses with the lysosome for the proteolytic degradation of autophagosome content. While a lot is known about the molecular mechanisms inducing autophagy, and the key effectors mediating autophagosome formation, the exact location of autophagosome formation and the source of phagophore membrane material is debated (Tooze and Yoshimori, 2010). The ER, mitochondrial contact sites and the recycling endosome have been suggested to act as sites for phagophore formation and the phagophore has been suggested to grow due to de novo lipid synthesis, and/or donation from various membrane sources such as the ER, endosomes and lipid droplets (Carlsson and Simonsen, 2015).

In this preprint, the authors look into the process of phagophore expansion, and show that the protein ATG2A, which is recruited to the phagophore together with the PI(3)P-binding WIPI proteins (Kotani et al., 2018), acts as a lipid transferring protein. The authors suggest that this function could allow it to transfer lipids from the ER to the growing phagophore and thereby mediate phagophore expansion.

 

Major findings
Atg2 is a protein consisting of two membrane-binding domains: the N-terminal corein domain and the C-terminal CAD domain (Chowdhury et al., 2018). Yeast Atg2 has been shown to bind and tether liposomes, and it is suggested that, at least in yeast, Atg2 could tether the phagophore to the ER to allow membrane extension (Kotani et al., 2018). In this preprint by Maeda et al, the authors utilize in vitro applications to show that the mammalian homologue ATG2A not only tethers liposomal membranes (Chowdhury et al., 2018), but also transfers lipids between them.

The authors use small and large unilamellar vesicles (SUVs and LUVs respectively) to represent high versus low curvature membranes, and show that ATG2A can extract lipids efficiently from LUVs and dissociate from the membrane with those lipids. The reverse action of lipid transfer can also be observed, where ATG2A bound to a lipid can unload this lipid and insert it into the membrane of LUVs. Using an energy-transfer based assay to study the kinetics of lipid transport, the authors show that ATG2A facilitates lipid transfer also between high curvature SUVs.

Previous data showed that ATG2A tethering of the lower curvature membrane LUVs could only be achieved upon binding of ATG2A and WIPI4 and required PI(3)P present on at least one of the two vesicles to be tethered. Data in this preprint confirm the necessity of WIPI-linked tethering for lipid transfer, and show that ATG2A can mediate lipid transfer between a PI(3)P-containing membrane and a PI(3)P-free membrane when WIPI4 or WIPI1 is added. The authors also show that the lipid transfer is bidirectional: When PI(3)P is incorporated in the acceptor membrane instead of the donor membrane, causing a shift in the orientation of ATG2A with the C-terminal CAD tip towards the PI(3)P bound WIPI protein, lipid transfer is observed at the same rate as when the ATG2A has its C-terminal bound to PI(3)P-WIPI on the donor membrane.

A direct interaction between ATG2A and WIPI1 could not be observed, instead the authors present data illustrating that ATG2A and WIPI proteins cooperatively associate with PI(3)P containing membranes.

Based on this data, the authors present a bridge mechanism for how ATG2A transfer lipids between membranes (Figure 1). In phagophore extension, hypothesizing that the phagophore is formed from PI(3)P-positive structures on the ER, ATG2A forms a bridge together with WIPI1 or WIPI4 to provide lipid transfer from the ER to the highly curved ends of the expanding phagophore (Fig. 1A). The bridge mechanism also holds true for ATG2A lipid transfer between other high curvature membranes (Fig 1B), where ATG2A can be bound in either orientation. For lipid transfer between low curvature membranes, the presence of PI(3)P and WIPI proteins will guide the direction of ATG2A (Fig 1C), although it is currently unknown if and how this affects the direction of lipid transfer.

Figure 1: A bridge model explains the mode of ATG2A-mediated lipid transfer. A) Lipid transfer by ATG2A from ER to the phagophore could represent a mechanism for phagophore expansion. ATG2A can also transfer lipids between high curvature membranes (SUVs) independently of WIPI proteins (B), but requires the presence of these proteins for lipid transfer between high curvature membranes (LUVs) (C). Figure from the preprint by Maeda et al., made available under a CC-BY-NC-ND 4.0 international license.

 

Why I choose this preprint
I think the question of where and how autophagosomes are formed is very interesting, and it is a field of study that is rapidly evolving with a lot of new data and theories being published continuously. The preprint highlighted in this preLight utilizes clever in vitro experiments to test their hypothesis on lipid transfer for membrane expansion, and I look forward to following the progress on how these findings can be correlated with in vivo observations.

 

Open Questions 

  • This preprint provides convincing evidence for lipid transfer activity of ATG2A in vitro, but the in vivo physiological role of this mechanism remains to be investigated. What are the lipids transferred by ATG2A, and what are the donor membranes? The authors present a model where phagophore expansion is mediated by transfer of lipids from the ER, but is it possible that ATG2A could also mediate lipid transfer from other membrane sources, such as Atg9-vesicles which Atg2 have been reported to interact with (Gomez-Sanchez et al., 2018)?
  • The autophagy deficient Atg210-12D mutant retains its interaction with WIPI homologue Atg18, and localizes to the site of phagophore formation but cannot mediate expansion (Kotani et al., 2018). Could the deleted region be important for the lipid transferring activity of Atg2?
  • What shifts the equilibrium to favor transfer in a certain direction? The authors’ model of phagophore expansion requires a unidirectional transport of lipids, yet the in vitro data shows that lipid transport is bidirectional, regardless of the orientation of ATG2A. Thus, there must be additional regulatory factors involved to affect the binding affinity of ATG2A N-term versus C-term and regulate the lipid transfer efficiency.
  • Do the WIPI proteins (WIPI1 vs WIPI4) have different roles in ATG2A mediated lipid transfer or are their functions redundant? The authors present data on WIPI1 self-oligomerisation affecting vesicle clustering, which could indicate a role in regulation of membrane curvature. If WIPIs have different properties of self-oligomerisation, is it possible that WIPIs affect different lipid transfers to regulate curvature and shape the autophagosome?

 

References

Bento, C.F., Renna, M., Ghislat, G., Puri, C., Ashkenazi, A., Vicinanza, M., Menzies, F.M., and Rubinsztein, D.C. (2016). Mammalian Autophagy: How Does It Work? Annu Rev Biochem 85, 685-713.

Carlsson, S.R., and Simonsen, A. (2015). Membrane dynamics in autophagosome biogenesis. J Cell Sci 128, 193-205.

Chowdhury, S., Otomo, C., Leitner, A., Ohashi, K., Aebersold, R., Lander, G.C., and Otomo, T. (2018). Insights into autophagosome biogenesis from structural and biochemical analyses of the ATG2A-WIPI4 complex. Proc Natl Acad Sci U S A 115, E9792-E9801.

Gomez-Sanchez, R., Rose, J., Guimaraes, R., Mari, M., Papinski, D., Rieter, E., Geerts, W.J., Hardenberg, R., Kraft, C., Ungermann, C., et al. (2018). Atg9 establishes Atg2-dependent contact sites between the endoplasmic reticulum and phagophores. The Journal of cell biology 217, 2743-2763.

Kotani, T., Kirisako, H., Koizumi, M., Ohsumi, Y., and Nakatogawa, H. (2018). The Atg2-Atg18 complex tethers pre-autophagosomal membranes to the endoplasmic reticulum for autophagosome formation. Proc Natl Acad Sci U S A 115, 10363-10368.

Tooze, S.A., and Yoshimori, T. (2010). The origin of the autophagosomal membrane. Nature cell biology 12, 831-835.

Tags: autophagy, lipid transfer, membrane dynamics, phospholipids

Posted on: 12th April 2019 , updated on: 24th April 2019

Read preprint (No Ratings Yet)




  • Author's response

    Takanori Otomo shared

    • This preprint provides convincing evidence for lipid transfer activity of ATG2A in vitro, but the in vivo physiological role of this mechanism remains to be investigated. What are the lipids transferred by ATG2A, and what are the donor membranes? The authors present a model where phagophore expansion is mediated by transfer of lipids from the ER, but is it possible that ATG2A could also mediate lipid transfer from other membrane sources, such as Atg9-vesicles which Atg2 have been reported to interact with (Gomez-Sanchez et al., 2018)?

    These are excellent questions. Since we did not investigate the lipid specificity in our preprint, at this point we can only refer to others’ studies: Kumar et al., JCB 2018; Osawa et al., NSMB 2019; Valverde et al., JCB 2019. These studies have collectively shown that Vps13, a homolog of Atg2, and Atg2 both can bind to various kinds of glycerolipids, inferring that these proteins would transfer those glycerolipids.

    Based on the intimate spatiotemporal relationships between the ER and the phagophore, we prefer to think the ER as the primary donor, although our in vitro data alone cannot exclude vesicles, such as Atg9 vesicles or COPII vesicles, as a potential lipid/membrane source. But any models based on lipid transfer from vesicles would encounter various issues. For instance, how small vesicles could provide a large number of lipids to enable phagophore expansion without losing their structural integrity would be difficult to answer.

     

    • The autophagy deficient Atg210-12D mutant retains its interaction with WIPI homologue Atg18, and localizes to the site of phagophore formation but cannot mediate expansion (Kotani et al., 2018). Could the deleted region be important for the lipid transferring activity of Atg2?

    Osawa et al. have provided answers to this question.

     

    • What shifts the equilibrium to favor transfer in a certain direction? The authors’ model of phagophore expansion requires a unidirectional transport of lipids, yet the in vitro data shows that lipid transport is bidirectional, regardless of the orientation of ATG2A. Thus, there must be additional regulatory factors involved to affect the binding affinity of ATG2A N-term versus C-term and regulate the lipid transfer efficiency.

    This is one of the most interesting implications that have emerged from ATG2 studies, but we currently don’t have good answers to offer. It’s a great opportunity for all in the field to figure this out.

     

    • Do the WIPI proteins (WIPI1 vs WIPI4) have different roles in ATG2A mediated lipid transfer or are their functions redundant? The authors present data on WIPI1 self-oligomerisation affecting vesicle clustering, which could indicate a role in regulation of membrane curvature. If WIPIs have different properties of self-oligomerisation, is it possible that WIPIs affect different lipid transfers to regulate curvature and shape the autophagosome?

    Sure, the rate of ATG2-mediated lipid transfer must be affected by various parameters including membrane curvature. But it is difficult to know from our data whether WIPI1 changes membrane curvature and we didn’t focus on this issue in this preprint. If it does so indeed as yeast Atg18 does (Gopaldass et al., EMBO J, 2017), WIPI1-driven changes of membrane curvature could certainly affect the efficiency of ATG2-mediated lipid transfer. We are hoping that our continued biochemical and physical studies will define the precise roles of different WIPIs in the future.

    Have your say

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Sign up to customise the site to your preferences and to receive alerts

    Register here
    Close