As (and when) you like it: on-demand phospholipid synthesis drives phagophore expansion during autophagy

ABSTRACT

A key feature of macroautophagy/autophagy is the formation of a transient de novo compartment called the phagophore, which envelops cytoplasmic material, ultimately enclosing it within an autophagosome, allowing it to be targeted for degradation. Schütter et al describe a novel mechanism that spatiotemporally coordinates phospholipid synthesis to drive phagophore expansion and autophagosome formation. These authors show that during starvation, fatty acids (FAs) are channeled into phospholipid synthesis, and the newly synthesized lipids are directed toward autophagosome biogenesis.

Abbreviations: ACS: acyl-CoA synthetase; ER: endoplasmic reticulum; FA: fatty acid; FAS: fatty acid synthetase; MCS: membrane contact sites; PAS: phagophore assembly site

KEYWORDS:

• Autophagosome; fatty acids; lipids
• lysosome
• membrane
• stress
• vacuole

Autophagy is an evolutionarily conserved catabolic process that targets for degradation superfluous cytoplasmic components and damaged organelles that have been sequestered within double-membraned structures called autophagosomes [1]. Autophagy occurs at low basal levels in virtually all cells to perform homeostatic functions and is robustly upregulated during nutrient starvation [2]. Autophagy results in dramatic morphological changes, specifically dynamic membrane remodeling to meet the high demand for membranes required for autophagosome biogenesis [3]. In Saccharomyces cerevisiae, the nucleation of the phagophore begins at the phagophore assembly site (PAS) [4]. Precursor membranes contributed by Atg9-containing and COPII-coated vesicles nucleate at the PAS for the formation of the phagophore, which subsequently expands to engulf cytoplasmic substrates; upon completion, this compartment closes to form an autophagosome [5–8]. Once the autophagosomes fuse with the vacuole, the cytoplasmic components are degraded, and the degradation products generated are released back into the cytosol and subsequently reused [9].

The hierarchical assembly and organization of autophagy-related proteins at the PAS that drive autophagosome formation is very well studied [4]. Additionally, previous studies have proposed several organelles such as the ER, Golgi, endosomes, mitochondria, plasma membrane and lipid droplets as precursor membrane donors for the formation of autophagosomes [10–14]. However, it has been a challenge to identify the stage and to which extent these organelles may supply membranes directly to autophagosome biogenesis. Importantly, the estimated amount of lipids supplied from precursor membranes via Atg9 vesicles is insufficient to maintain autophagosome numbers at the level needed under stress conditions [6,15]. Therefore, the question remains: What are the underlying mechanisms of rapid lipid mobilization and phagophore expansion during autophagosome formation?

A vast majority of cellular lipids are synthesized by ER-resident proteins. Therefore, the ER influences the cellular lipid biomass and enables transfer of lipids to various organelles via membrane contact sites (MCS) [16,17]. Additionally, the ER membrane forms physical contact with autophagosomes [18]. Recently, it was discovered that the conserved Atg2-Atg18 complex tethers phagophores to ER-phagophore MCS [19], and the establishment of this interaction requires Atg9 [20]. Furthermore, in vitro studies show that Atg2 facilitates the transfer of phospholipids between opposing membranes [21], and cells lacking Atg2 fail to form autophagosomes [19]. These studies lend credence to the hypothesis that transfer of phospholipids across ER-phagophore MCS may drive autophagosome formation. In the article highlighted here, Schütter et al test this hypothesis and uncover a novel mechanism wherein rewiring of fatty acid flux toward de novo phospholipid synthesis specifically drives phagophore expansion during autophagy [22].

Lipid flux is regulated by a conserved protein family of long chain acyl-CoA synthetases (ACSs), which activate fatty acids and channel them to various pathways depending on their dynamic subcellular localization [23]. The yeast ACS network is composed of six proteins, and the authors decided to test their potential role in autophagy. The authors used live-cell imagining to determine the spatial organization of the yeast ACS network during autophagy-inducing conditions. They found that Faa1 and Faa4 (members of the ACS family) are predominantly localized to the growing phagophore, and the recruitment of these enzymes occurs downstream of the core autophagy machinery, suggesting a direct role of ACSs in autophagy.

Activation of fatty acids in yeast is performed by two parallel pathways, via the ACSs and through de novo fatty acid synthesis by fatty acid synthetase (FAS). However, in mammalian cells, the fatty acids generated by the FASN (fatty acid synthase) pathway need to be activated by ACSs [24,25]. In order to test for the conserved function of ACSs in autophagy in yeast, the authors inhibited the FAS pathway specifically, using the inhibitor cerulenin. With this approach, they looked at the role of Faa1 in autophagy. To this end, the authors genetically engineered yeast strains that expressed either wild-type (WT) Faa1 or plasma-membrane tethered Faa1 (PM-Faa1). WT Faa1 and PM-Faa1 cells display no difference in autophagy induction or phagophore nucleation; however, PM-Faa1 cells show a significant decrease in auto-phagic flux as measured by GFP-Atg8 processing and fluorescence microscopy analysis upon FAS inhibition. While vacuolar functions appear to remain intact, cells expressing PM-Faa1 fail to accumulate autophagic bodies in hydrolase-deficient (pep4∆) vacuoles. Furthermore, the catalytic activity of Faa1 on autophagic membranes is required to maintain autophagic flux.

To directly test the role of spatial localization of ACSs and how it regulates autophagy, the authors generated synthetic tethers in PM-Faa1 cells, where Faa1 is localized to different organelles; vacuolar membrane, ER membrane and ER exit sites. In cells overexpressing Faa1 at the ER membrane and at ER exit sites, the autophagic flux is partially rescued and shows a strong Faa1 signal next to forming autophagosomes. These results suggest critical functions for Faa1 activity and emphasize the importance of its localization in close proximity to nascent autophagosomes during autophagy.

Schütter et al next examined the molecular function of Faa1 in regulating autophagy by investigating the dynamics of auto-phagosome biogenesis. Using time-lapse imaging they observed that PM-Faa1 cells have multiple Atg8-puncta; however, these cells have a considerable defect in phagophore membrane expansion compared to WT Faa1 cells. Furthermore, the rate of formation of ring-like autophagosomes is significantly reduced in PM-Faa1 compared to WT Faa1 cells. These results indicate that even when phagophores are nucleated, they fail to expand and form autophagosomes in the absence of local FA activation. Furthermore, perturbation of local FA activation significantly delays the maturation of any autophagosomes that do form under autophagy-inducing conditions. Supplementation of oleate to increase the available FA pool drastically increases the rate of phagophore expansion in WT Faa1 cells compared to PM-Faa1 cells, whereas the frequency of phagophore nucleation remains unchanged. These results indicate a critical role for local FA activation by Faa1, downstream of phagophore nucleation, in determining the growth rate of phagophore membranes in a rate-limiting manner during phagophore expansion.

Next, Schütter et al determined the lipid composition of autophagic membranes and found that while PM-Faa1 cells have lower amounts of phospholipids, consistent with a reduced number and size of autophagosomes, the fatty acid profiles of Faa1 and PM-Faa1 cells are similar. Furthermore, they found that Atg8-containing membranes are highly unsaturated, which increases their fluidity, making them conducive to dynamic changes in shape. Taken together, these results suggest that while the rate of autophagosome formation is reduced in PM-Faa1 cells, the autophagic membrane composition is similar to WT Faa1 cells. Therefore, while the localization of Faa1 has no effect on the lipid composition of autophagosomes, its localization on autophagic membranes is important to accelerate phagophore expansion.

The results thus far raise the possibility that accumulation of Faa1 on autophagic membranes is required to direct fatty acids into de novo phospholipid synthesis at the site of phagophore expansion. To directly test this hypothesis, the authors used a combination of isotope labeling and mass-spectrometry to track fatty acid channeling into phospholipid synthesis and subsequent incorporation into autophagic membranes during starvation. Strikingly, they found that the isotopic FA incorporation into phospholipids on Atg8-containing membranes is significantly reduced in PM-Faa1 cells compared to WT Faa1 cells. Furthermore, deletion of genes responsible for committing activated FA to de novo phospholipid synthesis results in defects in autophagy and phagophore dynamics, phenocopying the defects seen in PM-Faa1 cells. Thus, the authors conclusively show that the ACS-mediated channeling of fatty acids into de novo phospholipid synthesis and the utilization of these phospholipids to drive phagophore expansion are critical for autophagosome formation, and this function is an evolutionarily conserved mechanism used to regulate autophagy; heterologous expression of human ACSL4 restores autophagy flux in yeast faa1∆ faa3∆ faa4∆ mutant cells.

In summary, Schütter et al describe the molecular mechanisms underlying phagophore expansion. Upon induction of autophagy, the nucleation of the phagophore requires the hierarchical assembly of core autophagy machinery and the arrival of precursor membranes in the form of Atg9-containing and COPII-coated vesicles [5–8]. Following nucleation, the ACS protein network localizes to the phagophore membrane and activates FAs. Subsequently, the activated FAs are channeled into de novo phospholipid synthesis within neighboring ER. The newly synthesized phospholipids are incorporated into the growing phagophore, enabling its rapid expansion.

Whereas membrane precursors from various organelles in the form of Atg9-containing vesicles play an important role in nucleation of the phagophore, de novo phospholipid synthesis for phagophore expansion might prove to be more biochemically favorable than removal of resident proteins from precursor membranes under the catabolic conditions of autophagy [15]. Furthermore, the ER is an important platform for the synthesis of phospholipids and autophagosome formation [16–18]. Previous studies have shown that phospholipid synthesis enzymes are enriched at ER subdomains proximal to the site of autophagosome biogenesis [26,27]. Furthermore, the transfer of phospholipids synthesized by ER-resident proteins may be facilitated by the Atg2-Atg18 complex, which tethers phagophores to the ER at ER-phagophore MCSs [21]. These studies describe the role for the close association of the ER and autophagosome biogenesis; however, the molecular principle behind this association remained unclear. Now, the work of Schütter et al fills in a missing piece of this puzzle, which describes how FA channeling drives the synthesis of these phospholipids, incorporation into autophagic membranes, and acceleration of phagophore expansion, thus playing an important role in autophagosome biogenesis.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Institute of General Medical Sciences [GM131919].