Receptor-mediated selective autophagy degrades the endoplasmic reticulum and the nucleus

a, Yeast two-hybrid assay. Cells expressing proteins fused with the Gal4 activation domain (GAD) or the Gal4 DNA-binding domain (GBD) were grown on the agar plate with or without histidine (His). Representative results of two technical replicates are shown. Growth without His indicates positive interaction. The P52A R67A mutation in Atg8 weakens its binding to the Atg8-family-interacting motif (AIM)³⁰. b, In the experiments described in Fig. 1b, the protein band intensities were measured, and the relative ratios of co-precipitated proteins to direct targets of immunoprecipitation (coIP (%)) are shown as the mean ± s.d. (n = 3). WT, wild type. c, Cells were treated with rapamycin for 13 h, and the lysates (input) were subjected to co-immunoprecipitation analysis using anti-GFP antibody. The immunoprecipitates (IP) were analysed by immunoblotting. c, Co-immunoprecipitation efficiencies were calculated as described in b, and the results are shown as the mean ± s.d. (n = 3). d, Cells treated with rapamycin for 12 h were subjected to co-immunoprecipitation analysis as described in c. Eight residues in the Asp52–Ser59 region in Atg39 were replaced with Ala in the D52–S59A mutant. Representative results of three technical replicates are shown (c, d). *P < 0.05; **P < 0.01; ***P < 0.001 (unpaired two-tailed Student’s t-test) (b, c, d).

a, Cells were incubated in nitrogen starvation medium, and the processing of the proform (p) of the vacuolar aminopeptidase Ape1 into the mature form (m), which is indicative of activity of the cytoplasm-to-vacuole targeting pathway, was examined by immunoblotting using anti-Ape1 antibodies. mApe1 intensity was divided by the sum of mApe1 and pApe1 intensities, and the relative values are shown (Ape1 maturation (%)). Vacuolar degradation of GFP–Atg8, which correlates with activity of non-selective autophagy, was also investigated. The results were quantified (see Methods), and the relative values (%) are shown. b, c, Mitophagy and pexophagy were examined as described previously³¹. Degradation of GFP-fused proteins was quantified (see Methods), and the relative values (%) are shown. Representative immunoblots of two technical replicates are presented (a–c). d, To assess bulk degradation of the cytoplasm by non-selective autophagy, a Pho8Δ60 alkaline phosphatase (ALP) assay was performed as described previously³². ALP activities are shown as the mean ± s.d. (n = 3). AU, arbitrary units; NS, not significant; ***P < 0.001 (unpaired two-tailed Student’s t-test).

a–e, Cells treated with rapamycin or incubated in nitrogen-starvation medium were subjected to immunoblotting analysis. Pep4 and Ypt7 are required for vacuolar lytic activity and autophagosome-vacuole fusion, respectively. Atg17 is a subunit of the Atg1 kinase complex and required for starvation-induced autophagy¹. Atg19 and Atg34 are autophagy receptors for vacuolar enzymes; Atg32 and Atg36 are those for mitochondria and peroxisomes, respectively. In e, atg39Δ or atg40Δ cells harbouring Atg39–HA or Atg40–HA plasmids, respectively, were analysed. Representative immunoblots of three (a, b, d, e) or four (c) technical replicates are shown. Sec63–GFP degradation (%) was quantified (see Methods), and the results are shown as the mean ± s.d. (n = 3 (b, d, e) or 4 (c)). NS, not significant; *P < 0.05; **P < 0.01; ***P < 0.001 (unpaired two-tailed Student’s t-test) (b–e). The results in c and d showed that ER-phagy requires Atg11 and the AIM-binding ability of Atg8, but not previously characterized autophagic receptors, suggesting the existence of an unknown receptor for this pathway. f, Fluorescence microscopy of cells treated with or without rapamycin. Images are representative of at least 150 cells. Arrowheads, Atg8-and Sec63-positive puncta. Scale bars, 5 μm.

a, Atg39–mCherry or Atg40–mCherry was expressed under the endogenous promoter (endo.) or the CUP1 promoter for overexpression (OE) in the presence of 250 μM CuSO₄. These cells were treated with rapamycin, and degradation of Sec63–GFP was examined by immunoblotting. Pgk1 serves as a loading control. Representative immunoblots of three technical replicates are presented, and the quantification results (see Methods) are shown as the mean ± s.d. (n = 3). *P < 0.05; **P < 0.01 (unpaired two-tailed Student’s t-test). The results for atg1Δ cells suggest that the rapamycin-induced decrease in Atg39–mCherry overexpressed in wild-type cells is due to its autophagic degradation. b, The co-localization of Atg39 or Atg40 with Atg11 in cells treated with rapamycin for 3 h was observed under a fluorescence microscope. Images are representative of at least 100 cells. Scale bars, 2 μm. c–e, Vacuolar degradation of Atg39–GFP and Atg40–GFP was examined by immunoblotting (c, e) and fluorescence microscopy (d). atg39Δ and atg40Δ cells carrying Atg39–GFP and Atg40–GFP plasmids, respectively, were examined in e. Representative immunoblots of three (Atg39–GFP degradation in c, e), or four (Atg40–GFP degradation in c, e) technical replicates are presented, and the quantification results are shown as the mean ± s.d. (n = 3 (Atg39–GFP degradation in c, e) or 4 (Atg40–GFP degradation in c, e)). ***P < 0.001 (unpaired two-tailed Student’s t-test) (c, e). Images are representative of at least 100 cells (d). Scale bars, 5 μm. The results in c–e suggest that vacuolar transport of Atg39 and Atg40 requires their binding to Atg11 and Atg8.

Vacuolar transport of Hmg1–GFP and Rtn1–GFP was examined by fluorescence microscopy. Images are representative of at least 100 cells. Scale bars, 5 μm.

a, pep4Δ cells treated with rapamycin for 13 h were analysed by immunoelectron microscopy using anti-Kar2 antibodies. Images are representative of at least 50 cells. Arrowheads, Kar2 signals (gold particles). These images show the staining of the cortical ER (left) and the perinuclear ER/nuclear envelope (right). b–d, Electron microscopy images (without immunogold labelling) of cells incubated in nitrogen-starvation medium for 12 h (b, d) or treated with rapamycin for 12 h (c). Images are representative of at least 14 cells that contained autophagic bodies. Arrowheads, autophagic bodies containing double-ring structures (b, c) or those containing tubules/sheets (d). Scale bars, 200 nm (a–d).

Atg39 localizes to the perinuclear ER (pnER)/nuclear envelope and acts to load a part of the pnER/nucleus as double-membrane vesicles, which contain intranuclear components, into autophagosomes. Atg40 mainly localizes to highly curved regions in the cortical ER (cER) and cytoplasmic ER (cytoER), and allows sequestration of these ER domains as folded tubules or sheets into autophagosomes. Whereas Atg39 is dispensable for cER/cytoER-phagy, Atg40 also contributes to pnER-phagy/nucleophagy. It is likely that both of these receptors interact with Atg11 to recruit the core Atg proteins that mediate autophagosome formation, and also with Atg8 to promote autophagic sequestration of the ER or a part of the nucleus.

a, Src1–GFP degradation assay for cells before rapamycin treatment. The results for cells treated with rapamycin are shown in Fig. 4a. b, Vacuolar transport of Src1–GFP was examined by fluorescence microscopy. Images are representative of at least 100 cells. Scale bars, 5 μm. c, Autophagic degradation of Nop1–GFP was examined by immunoblotting. Representative results of two (a) or three (c) technical replicates are shown. Nop1–GFP degradation was quantified (see Methods), and the results are shown as the mean ± s.d. (n = 3). NS, not significant; ***P < 0.001 (unpaired two-tailed Student’s t-test). d, pep4Δ cells were treated with rapamycin for 13 h, followed by immunoelectron microscopy using anti-Nop1 antibody. Images are representative of six cells containing Nop1 signals in the nucleolus. Scale bars, 500 nm (left) and 200 nm (right). The data show that the nucleolus is specifically labelled with gold particles. Arrowheads, autophagic bodies in the vacuolar lumen.

a, Cells were examined by electron microscopy. Images are representative of at least 14 cells. b, Cells were subjected to nitrogen starvation, and dead cells were stained with phloxine B. Images are representative of at least 600 cells. DIC, differential interference contrast microscopy images. c, Dead cells were counted (see Methods), and the results are shown as the mean ± s.d. (n = 6). d, Wild-type or atg39Δ cells harbouring Atg39–GFP plasmids or the empty vector were subjected to nitrogen starvation, and dead cells were counted. The results are shown as the mean ± s.d. (n = 3). NS, not significant; *P < 0.05; ***P < 0.001 (unpaired two-tailed Student’s t-test) (c, d). e, f, Cells expressing Rtn1–GFP (e) or Sec63–GFP (f) were subjected to nitrogen starvation, and the fluorescence microscopy images were taken by focusing cell periphery (e) or centre (f). Images are representative of at least 100 (e) or 200 (f) cells. In f, the fluorescence intensity of Sec63–GFP at the periphery of cells marked with asterisks (left) was measured along cyan lines in insets and graphically shown (right). Scale bars, 500 nm (a), 5 μm (b, e, f).

HHpred, a homology search tool based on structure prediction, identified reticulon-related proteins, Yop1 in S. cerevisiae, and REEP4 and REEP5 in Homo sapiens, as proteins similar to Atg40. Red bars, regions structurally similar to Atg40; green bars, reticulon-like domains; orange regions, hydrophobic segments predicted by TMHMM, a membrane protein topology prediction method. LIR, LC3-interacting region.