Two studies show that the engulfment of certain intracellular membranous structures by vesicles called autophagosomes regulates the structures' degradation in a selective, receptor-protein-mediated manner. See Letters p.354 & p.359
Autophagy — the internal digestion of parts of a cell — is an evolutionarily conserved pathway for recycling damaged or unwanted cellular components. Although the process was initially thought to involve degradation en masse of cytoplasmic contents, evidence has emerged1 to suggest that individual components can be selectively isolated in preparation for autophagy. Two complementary papers papers in this issue2,3 identify specific receptor proteins in yeast and mammalian cells that enable the selective degradation of an organelle called the endoplasmic reticulum (ER).
During autophagy, cup-shaped double-membrane structures called phagophores develop into spherical structures called autophagosomes, which engulf cytoplasmic components destined for autophagic degradation. Autophagosomes fuse with vesicles called lysosomes, which degrade and then recycle the engulfed components (Fig. 1). Selective autophagy of certain bacteria, aggregation-prone proteins and organelles such as mitochondria seems to be mediated by autophagy receptor proteins. These bind to protein components of the phagophore and autophagosome, such as Atg8 in yeast and its equivalent in mammals, LC3, through interactions with LC3-interacting domains (called Atg8-binding pockets and LIR domains in yeast and mammals, respectively)1.
The ER, which directs proteins towards the plasma membrane for secretion or residency, can also be degraded through autophagy (a process called ER-phagy)4. It has not previously been possible to test whether ER-phagy is required for health, although in animal models it seems to protect against α1-antitrypsin deficiency, a condition in which a toxic form of the α1-antitrypsin protein accumulates in the ER and causes liver damage5. But is ER-phagy merely part of bulk autophagy, or does some cellular machinery enable the preferential selection of ER into autophagosomes, as has been previously hypothesized6?
The current studies provide robust support for this hypothesis. The key starting point for both groups was the identification of autophagy receptors that interact with the ER and with autophagosomes, which they achieved by searching for proteins that physically interact with LC3 or Atg8. Khaminets et al.3 (page 354) report that FAM134B, which has an LIR domain and a reticulon domain that binds to and fragments the ER, interacts with LC3 in mammalian cells. In yeast, Mochida et al.2 (page 359) demonstrated that Atg39 and Atg40 (which seems to be a counterpart of FAM134B), interact with Atg8 and the ER. The groups report that loss of Atg39, Atg40 or FAM134B impedes autophagy-dependent degradation of the ER, but does not affect bulk autophagy. Furthermore, mutant forms of the proteins that do not bind to the phagophore and autophagosome are ineffective in ER-phagy.
Mochida et al. showed that Atg40 is associated with the cytoplasmic ER and the ER close to the plasma membrane and seems to be involved in their degradation. By contrast, Atg39 is associated with the ER that surrounds the nucleus. Atg39 regulates degradation of both this nuclear envelope and some nuclear constituents — a previously unidentified form of selective autophagy that the authors call nucleophagy. It will be interesting to determine whether nucleophagy occurs in mammalian cells, and to define when the process occurs, which components are degraded, and why it is required. For instance, nucleophagy might contribute to the breakdown of the nuclear envelope that occurs when cells divide. Nucleophagy could also theoretically be responsible for quality control and repair of damaged nuclear membranes. This could be tested by investigating whether impaired nucleophagy results in the accumulation of malfunctioning pores in the nuclear envelope and compromised segregation of nuclear and cytoplasmic contents.
Khaminets and colleagues provide data suggesting that FAM134B helps to fragment the ER into 'bite-sized' pieces that can be readily incorporated into autophagosomes. This model is analogous with the hypothesis7 that the division of mitochondria enables their own sequestration and degradation through autophagy. The role of FAM134B is of particular interest, because it is mutated in a human disorder of the sensory nerves in which pain perception is impaired, resulting in ulceration of the hands and feet8. Khaminets et al. report that mice lacking FAM134B have similar sensory defects to those seen in the human disorder. In these mice, ER accumulates in the cell bodies of pain-sensing neurons before any morphological abnormalities are apparent, raising the interesting possibility that these neurons may be particularly vulnerable to defective ER-phagy.
The suggestion that ER-phagy might be required for cellular health is supported by Mochida and colleagues' observation that Atg39-deficient yeast are more sensitive than wild-type cells to the stress of nitrogen starvation. However, the possibility cannot be ruled out that sensory abnormalities are mediated by autophagy-independent roles of FAM134B. If the sensory abnormalities observed in mice are directly due to the increase in ER volume, future studies should be able to determine whether the defects are caused by the toxic effects of the uncleared organelles, or whether the increased ER volume affects cellular signalling and neurotransmission.
The two complementary studies2,3 have identified seemingly analogous systems for regulating ER degradation through autophagy, which will lead to a deeper understanding of the process. Goals for the future include understanding the relevance of the ER-phagy machinery to maintenance of the nucleus in mammalian cells, investigating the consequences of impaired or excessive ER-phagy, and defining how the process is regulated. The activities of many autophagy receptors can be modulated by post-translational modifications such as phosphorylation1, and it will be interesting to find out whether such mechanisms are pertinent to ER-phagy. The idea that the activity of receptors involved in ER-phagy can be regulated is supported by Mochida and colleagues' observation that levels of Atg39 and Atg40 increase when cells are treated with the drug rapamycin, which induces some of the signalling changes associated with nitrogen starvation.
Finally, it remains to be seen whether there is any specificity to which parts of the ER and nuclear envelope are cleared by selective autophagy. If there is, what is selected and why? It will be important to discover whether ER-phagy is indiscriminate, or whether the receptors target damaged ER in the way that damaged mitochondria are preferentially degraded by mitophagy and intracellular pathogens by xenophagy9.Footnote 1
Stolz, A., Ernst, A. & Dikic, I. Nature Cell Biol. 16, 495–501 (2014).
Mochida, K. et al. Nature 522, 359–362 (2015).
Khaminets, A. et al. Nature 522, 354–358 (2015).
Bernales, S., Schuck, S. & Walter, P. Autophagy 3, 285–287 (2007).
Hidvegi, T. et al. Science 329, 229–232 (2010).
Kamimoto, T. et al. J. Biol. Chem. 281, 4467–4476 (2006).
Twig, G. et al. EMBO J. 27, 433–446 (2008).
Kurth, I. et al. Nature Genet. 41, 1179–1181 (2009).
Randow, F. & Youle. R. J. Cell Host Microbe 15, 403–411 (2014).
About this article
RETREG1/FAM134B mediated autophagosomal degradation of AMFR/GP78 and OPA1 —a dual organellar turnover mechanism
The EMBO Journal (2020)
Overexpression of family with sequence similarity 134, member B (FAM134B) in colon cancers and its tumor suppressive properties in vitro
Cancer Biology & Therapy (2020)
UFMylation of RPL26 links translocation-associated quality control to endoplasmic reticulum protein homeostasis
Cell Research (2020)
Aging Cell (2018)