A new role for a COPII cargo adaptor in autophagy

ABSTRACT

Endoplasmic reticulum (ER) homeostasis is maintained by the removal of misfolded ER proteins via different quality control pathways. Aggregation-prone proteins, including certain disease-linked proteins, are resistant to conventional ER degradation pathways and require other disposal mechanisms. Reticulophagy is a disposal pathway that uses resident autophagy receptors. How these receptors, which are dispersed throughout the ER network, target a specific ER domain for degradation is unknown. We recently showed in budding yeast, that ER stress upregulates the reticulophagy receptor, triggering its association with the COPII cargo adaptor complex, Sfb3/Lst1-Sec23 (SEC24C-SEC23 in mammals), to discrete sites on the ER. These domains are packaged into phagophores for degradation to prevent the accumulation of protein aggregates in the ER. This unconventional role for Sfb3/Lst1 is conserved in mammals and is independent of its role as a cargo adaptor on the secretory pathway. Our findings may have important therapeutic implications in protein-aggregation linked neurodegenerative disorders.

Protein quality control mechanisms in the ER ensure that only native conformers enter the secretory pathway. The major COPII cargo adaptor, Sec24-Sec23, sorts correctly folded proteins into ER-derived vesicles that traffic to the Golgi. Sec24 has 2 paralogs in budding yeast (Sfb2/Iss1 and Sfb3/Lst1) that also function with Sec23 as cargo adaptors, whereas mammals have 4 SEC24 isoforms (SEC24A, SEC24B, SEC24C and SEC24D). When misfolded proteins accumulate in the ER, the unfolded protein response (UPR) restores ER homeostasis. One way the UPR performs this function is through ER-associated protein degradation (ERAD). ERAD recognizes terminally misfolded proteins and retrotranslocates them into the cytosol where they are degraded. Some proteins, however, cannot be cleared by ERAD. A better understanding of the alternate disposal pathways that degrade this class of proteins is important, as certain aggregation-prone proteins have been linked to neurodegenerative disorders.

Reticulophagy is an alternate disposal pathway that targets ER domains into phagophores and delivers them into vacuoles or lysosomes for degradation. When reticulophagy is induced by nutrient deprivation or TOR inhibitors, ER autophagy receptors recruit Atg8 (yeast) or LC3 (mammals) to discrete sites on the highly dynamic ER network to package ER into phagophores. As autophagy receptors localize throughout the contiguous network of the ER, how specific ER subdomains are recognized and targeted for reticulophagy is unclear. We hypothesized that cytosolic machinery might interact with ER autophagy receptors to mediate this process. COPII coat subunits are potential candidates as they segregate ER domains from the bulk of the network. To ask if COPII subunits are needed for reticulophagy, we analyzed yeast mutants harboring mutations in COPII coat subunits and found the sfb3/lst1∆ mutant to be defective in ER autophagy. Sfb3/Lst1 is specifically required for reticulophagy, and not other autophagy pathways. Furthermore, the role of Sfb3/Lst1 in reticulophagy is independent of its function in the secretory pathway.

In yeast, the cortical and cytoplasmic reticulophagy receptor is Atg40, a reticulon-like ER membrane protein. When reticulophagy is induced with the TOR kinase inhibitor rapamycin, Atg40 expression is upregulated and Atg40-containing foci colocalize with Sfb3/Lst1-Sec23, but not other coat subunits. Sfb3/Lst1 is also the only COPII coat subunit that binds to Atg40. Interestingly, Sfb3/Lst1-Sec23 fails to colocalize with Atg40 and its binding partner Atg8 in the reticulophagy-deficient mutant, lnp1∆. Lnp1 regulates ER network rearrangements by stabilizing the 3-way junctions that form where two tubules fuse. This observation highlights the importance of ER network dynamics in the formation of sites of autophagy on the ER. Consistent with the proposal that Sfb3/Lst1 works with Atg40 to package ER into phagophores, transmission electron microscopy revealed that ER sequestration is severely impaired in sfb3/lst1∆ and atg40∆ mutants.

We found that ER stress, induced by the overexpression of the Z variant of human SERPINA1/alpha-1 antitrypsin (ATZ), upregulates Atg40 levels. Consistent with the notion that aggregation-prone ATZ induces reticulophagy, ATZ aggregates in the ER in the absence of either Sfb3/Lst1 or Atg40. Interestingly, Atg40 expression is modulated by autophagy regulators and not the UPR. In total, these findings imply that TOR-dependent autophagy transcriptional regulators, that control autophagosome abundance, can also restore ER homeostasis in response to certain types of stress. This ER stress response does not depend on the UPR.

The yeast autophagy receptor Atg40 is related to 2 different mammalian reticulophagy receptors, RETREG1/FAM134B and RTN3. Like RTN3, Atg40 localizes to the tubular ER, yet it has a similar domain structure to RETREG1. When reticulophagy is induced in U2OS cells with the TOR inhibitor Torin2, we found that the degradation of ER sheets (marked by RETREG1) and tubules, labeled by RTN3, is dependent on the Sfb3/Lst1 homolog SEC24C. The other SEC24 isoforms are not required for reticulophagy. These studies fit well with published proteomics studies showing an interaction between RTN3 and SEC24C in U2OS cells.

Reticulophagy is also likely to play a pivotal role in neuronal cell health. Mutations in human RETREG1 lead to sensory neuropathies, and the knockout of SEC24C in post mitotic neurons results in cell death. There are additional connections between reticulophagy and neuronal function. For example, we have shown that yeast Lnp1 is required for reticulophagy. Interestingly, the localization of Lnp1 to the ER 3-way junctions is dependent on Sey1/Atlastin, which is also needed for mammalian ER autophagy. Mutations in LNP1 and the reticulophagy receptor ATL3 (atlastin GTPase 3) have been linked to hereditary spastic paraplegias (HSP) and HSP-like neuropathies. HSP is a group of disorders that are associated with spasticity and weakness of the lower limbs.

In summary, our studies have revealed a role for a non-canonical form of the COPII coat in targeting ER domains for autophagy. When reticulophagy is induced by nutrient starvation or aggregation-prone proteins, Sfb3/Lst1-Sec23 in yeast, or SEC24C-SEC23 in mammals, identifies the membrane-bound autophagy receptor on the ER and sequesters the receptor and ER membrane into a phagophore. We have found that ER-reticulophagy sites (ERPHS), lack Sec24, distinguishing them from the ER exit sites (ERES) that bud conventional COPII vesicles routed to the Golgi (Figure 1). In total, our findings indicate that COPII coat subunits mediate the exit of distinct cargoes (secretory cargo or ER domains) from the ER. These 2 trafficking routes are controlled by different stress response pathways. Protein homeostasis in the secretory pathway is regulated by the UPR, while reticulophagy activity is modulated by autophagy regulators.

Figure 1.

Reticulophagy is an alternate disposal pathway for deleterious proteins not degraded by ERAD. Secretory proteins destined to traffic to the Golgi are packaged into COPII vesicles, containing Sec24-Sec23, at ER exit sites (ERES). When unfolded or misfolded proteins accumulate in the ER, the UPR is upregulated and ERAD retrotranslocates misfolded proteins across the membrane into the cytosol where they are degraded by the proteasome. Aggregation-prone proteins that fail to simulate the UPR, or are not degraded by ERAD, are cleared by reticulophagy. ER-reticulophagy sites (ERPHS) lack Sec24. They form on the ER when the Sec24 paralog, Sfb3/Lst1 (SEC24C, in mammals), binds to a reticulophagy receptor. Additionally, autophagy receptors bind to Atg8/LC3, a ubiquitin-like protein essential for autophagosome biogenesis. The formation of ERPHS also depends on Lnp1, an ER shaping protein that resides at and stabilizes the 3-way junctions that arise when two tubules fuse.

These findings have raised important questions that remain to be addressed. For example, although our data indicate that increased levels of reticulophagy receptors drive ERPHS formation, the mechanism by which ERPHS are formed remains to be determined. A possible scenario is that the receptor finds aggregation-prone proteins in the ER, and then concentrates at these sites to drive local coat recruitment. As Atg40, RTN3 and RETREG1 do not have lumenal domains, chaperones and other machinery would be needed to connect the receptor to the aggregation-prone protein. Addressing these questions may have important therapeutic implications for treating HSP and other neurodegenerative diseases.

Funding Statement

S F-N, YC and SP received salary support from the ALPHA-1 foundation [556333] and NIGMS under award numbers R01GM114111, R01GM115422 and R35GM131681 to S F-N; Alpha-1 Foundation; National Institute of General Medical Sciences [R35GM131681]; National Institute of General Medical Sciences [R01GM115422];National Institute of General Medical Sciences [R01GM114111].

Disclosure statement

No potential conflict of interest was reported by the authors.