Atg9 Trafficking in Autophagy-Related Pathways

[Autophagy 3:3, 271-274; May/June 2007]; ©2007 Landes Bioscience

Addendum

Atg9 Trafficking in Autophagy-Related Pathways Congcong He Daniel J. Klionsky*

Abstract

Key words cytoplasm to vacuole targeting pathway, mitochondria, pexophagy, protein transport, yeast Addendum to:

.

UT E

IST RIB

Autophagy is a conserved degradative pathway induced by various stress and developmental cues in eukaryotes, in which cytosolic proteins and/or organelles are engulfed into double‑membrane vesicles called autophagosomes, transported to, and degraded in lysosomes/vacuoles. This metabolic process participates in development and a variety of diseases, such as cancer, neurodegeneration, pathogen infection and heart diseases. Autophagy can be selective or nonselective. In the budding yeast Saccharomyces cerevisiae, under starvation conditions, bulk cytosol is randomly sequestered and degraded, and the resulting nutrients reused for essential cellular functions; this starvation‑induced bulk autophagy is considered nonselective. There are also several types of selective autophagy in yeast that target specific cargos. For example, the transport of two vacuolar hydrolases, a‑mannosidase (Ams1) and the precursor form of aminopeptidase I (Ape1 [prApe1]), bypasses the secretory pathway. These hydrolases are synthesized in the cytoplasm and targeted to the vacuole via the cytoplasm to vacuole targeting (Cvt) pathway. The Cvt pathway is morphologically and genetically similar to bulk autophagy, except that it occurs constitutively in normal growing conditions. Another specific route is pexophagy. During pexophagy, excess peroxisomes induced by oleic acid as the sole carbon source are selectively degraded in the vacuole, after cells are shifted back to glucose medium. This pathway also uses double‑membrane vesicles for cargo delivery. In yeast, autophagosomes are thought to emerge from a perivacuolar site called the pre-autophagosomal structure or phagophore assembly site (PAS).1 In autophagic pathways, forming vesicles are likely expanded by addition of new membrane fragments, rather than being generated by budding from the surface of a preexisting organelle or sealing of a piece of continuous membrane; however, the membrane source of autophagosomes/Cvt vesicles and the route and mechanism of lipid transport are not known. Among >25 Atg (autophagy‑related) proteins, Atg9 is the best candidate to help understand these questions. Atg9 is the sole transmembrane protein required in vesicle formation in both bulk and selective autophagy; however, Atg9 is absent from the completed autophagosomes, suggesting that the protein is retrieved on or before vesicle completion. Compared with most other Atg proteins, which localize at the PAS, the subcellular localization pattern of Atg9 is unique. In addition to the PAS, Atg9 also localizes to mitochondria, as well as some unidentified peripheral punctate compartments.2 Atg9 cycles between the PAS and these peripheral sites, suggesting that Atg9 may serve as a membrane carrier. Also, this finding raises the hypothesis that mitochondria act as a membrane source in autophagy. Certain morphological observations support this possibility. For example, during macropexophagy in the methylotrophic yeast Hansenula polymorpha,3 or upon

IEN

Recruitment of Atg9 to the Preautophagosomal Structure by Atg11 is Essential for Selective Autophagy in Budding Yeast

OT D

Previously published online as an Autophagy E-publication: http://www.landesbioscience.com/journals/autophagy/abstract.php?id=3912

ON

Original manuscript submitted: 01/29/06 Manuscript accepted: 01/29/06

.D

*Correspondence to: Daniel J. Klionsky; Life Sciences Institute; University of Michigan; Ann Arbor, Michigan 48109-2216 USA; Tel.: 734.615.6556; Fax: 734.763.6492; Email: [email protected]

CE

Life Sciences Institute; Departments of Molecular, Cellular and Developmental Biology and Biological Chemistry; University of Michigan; Ann Arbor, Michigan USA

The origin of the autophagosomal membrane and the lipid delivery mechanism during autophagy remain unsolved mysteries. Some important hints to these questions come from Atg9, which is the only integral membrane protein required for autophagosome formation and considered a membrane carrier in autophagy‑related pathways. In S. cerevisiae, Atg9 cycles between peripheral sites and the pre-autophagosomal structure/phagophore assembly site (PAS), the nucleating site for formation of the sequestering vesicle. We recently identified a peripheral membrane protein, Atg11, as a binding partner of Atg9, in a yeast two‑hybrid screen. Based on our analysis we propose a model for Atg9 cycling. Our model suggests that a pool of Atg11 mediates the anterograde transport of Atg9 to the PAS along the actin cytoskeleton, and that this delivery process may serve as a membrane shuttle for vesicle assembly during yeast selective autophagy. Here, we discuss the implications of the model and present additional evidence that extends it with regard to membrane trafficking modes during pexophagy.

SC

C. He, H. Song, T. Yorimitsu, I. Monastyrska, W.-L. Yen, J.E. Legakis and D.J. Klionsky

©

20

07

LA

ND

ES

BIO

J Cell Biol 2006; 175:925–35

www.landesbioscience.com

Autophagy

271

Atg9 Trafficking in Autophagy-Related Pathways

invasion of the bacterium Legionella pneumophila into macrophages,4 sequestering membranes or phagosomes are surrounded by and closely associated with mitochondria. Thus, the anterograde transport of Atg9 from the peripheral sites to the PAS may function in delivering lipid and proteins to the site of vesicle formation, and retrieval of Atg9 from the PAS may recycle the protein back to the membrane origin for the next round of delivery.

Atg9 Transport from Mitochondria to the PAS Several components have been shown to be required in Atg9 retrieval from the PAS, including the Atg1‑Atg13 complex, Atg2, Atg18, and the phosphatidylinositol 3‑kinase complex.5 An intriguing question is what factors regulate the anterograde transport of Atg9 to the site of vesicle formation. For this purpose, we performed a yeast two‑hybrid screen among all identified Atg proteins and found several Atg9 interacting partners, including Atg11, Atg23 and Atg27. We found that anterograde transport of Atg9 to the PAS is blocked in atg11D or atg1D atg11D yeast mutants.6 Figure 1. Identification of an Atg9 mutant that specifically loses interaction with Atg11. (A) A point Through random mutagenesis experiments, mutation, H192L, in Atg9 disrupts its interaction with Atg11. An atg9D strain expressing an integrated we obtained a point mutation H192L in Atg11‑GFP fusion (CCH007) was transformed with a plasmid encoding either wild‑type PA‑Atg9 (pCuthe Atg9 N‑terminal domain that disrupts PAAtg9(416); Atg9), PA‑Atg9H192L (pCuPAAtg9H192L(416); H192L), or PA alone (pCuPA(416); PA), drivits interaction with Atg11, but maintains en by the CUP1 promoter. Total lysates and eluted polypeptides following affinity purification (Aff. Pur.) were separated by SDS‑PAGE and detected with anti‑YFP antibody (which detects Atg9 interaction with Atg23 or Atg27 with IgG‑sepharose H192L GFP). (B) The Atg9 mutant is coprecipitated with Atg23. An atg9D strain expressing integrated (Fig. 1). Identification of the specific mutaAtg23‑PA and Atg27‑HA fusions (JLY74) was transformed with a plasmid encoding either wild‑type tion is important for functional studies, Atg9 (pAPG9(416); Atg9) or Atg9H192L (pAtg9H192L(416); H192L) expressed under the endogenous because recent data have suggested that ATG9 promoter. A wild‑type strain expressing integrated Atg23‑PA and Atg27‑HA fusions (JLY68; WT) Atg23 and Atg27 are also involved in Atg9 and an untagged wild‑type strain (SEY6210; ‑) were used as controls. Total lysates and eluted affinity transport from mitochondria to the PAS.7 isolated polypeptides as described above in (A) were detected with anti‑Atg9 antiserum. (C) Atg9H192L (Legakis JE, Yen W‑L, Klionsky DJ, unpub- is coprecipitated with Atg27. An atg9D strain expressing an integrated Atg27‑HA fusion (JLY59) was H192L as in (B), and an untagged lished). The H192L mutation is located transformed with a plasmid encoding either wild‑type Atg9 or Atg9 wild‑type strain was used as a control. Total lysates and eluted polypeptides affinity‑isolated with anti‑HA within the minimal sufficient region for antibody and PA‑sepharose were detected by anti‑Atg9 and anti‑HA antiserum or antibodies. Atg9‑Atg11 interaction (amino acids 159‑255, which map to the central part of the Atg9 N‑terminal arm), assayed by yeast two‑hybrid analysis. on its peripheral compartments in the absence of Atg11, suggesting H192L also locates to the minimal required region for interaction that Atg11 localization to the PAS is not dependent upon the (amino acids 154‑201) reported in a parallel paper.8 Atg9‑Atg11 interaction, whereas Atg9 localization requires interaction Using the Atg9H192L mutant, we found that the Atg9‑Atg11 inter- with Atg11. Taken together, we conclude that Atg11 mediates the action is necessary for the recruitment of Atg9 to the PAS and normal movement of Atg9 to the PAS, but not vice versa. progression of selective autophagy (the Cvt pathway) during vegetaAccordingly, we propose that the anterograde transport of Atg9 tive growth, but is not required in bulk autophagy under starvation may serve as a membrane shuttle for vesicle assembly during selecconditions or with rapamycin treatment (a drug that mimics starva- tive autophagic pathways and that movement from the peripheral tion conditions and induces bulk autophagy). This finding suggests compartments to the PAS is facilitated by Atg11, Atg23 and Atg27 that different mechanisms may be involved in these two fundamental along the actin cytoskeleton (Fig. 2).6 processes, which is consistent with our previous finding that Atg11 is specific for selective autophagy.9 In addition, we demonstrate that the Atg11: A Scaffold at the PAS integrity of the actin cytoskeleton is essential for correct targeting of Atg11 is a peripheral membrane protein, predicted to contain Atg11 to the PAS. We previously reported that actin is also required for proper Atg9 transport to the PAS.10 In the atg9D mutant, Atg11 four coiled‑coil domains that support interactions with multiple localizes to the PAS as in wild‑type cells; in contrast, Atg9 is retained Atg proteins.11 Atg9‑Atg11 interaction is mediated through the 272

Autophagy

2007; Vol. 3 Issue 3


Atg9 N‑terminal cytosolic domain and the second coiled‑coil domain of Atg11. In addition to Atg9, Atg11 also independently interacts with the prApe1 and Ams1 receptor Atg19, and with the Atg1‑Atg13 complex that is involved in vesicle formation. This finding suggests that there may be multiple pools of Atg11 in the cell, recruiting cargos and membrane components to the vesicle forming machinery at the PAS; however, it remains unclear how the Atg9‑Atg11 complex might travel along the cytoskeletal network.

NonMitochondrial Peripheral Pools of Atg9

Figure 2. Artistic representation of the Cvt pathway. Atg11 directs Atg9‑containing membranes from peripheral donor sites (upper right) to the pre-autophagosomal structure/phagophore assembly site. The phagophore membrane then expands to form a Cvt vesicle (center). Completed vesicles fuse with the vacuole, releasing the inner vesicle, termed a Cvt body, into the vacuole lumen (lower right). The figure has been labeled with numbers corresponding to Atg proteins. A capital “V” designates Vps proteins. PE, phosphatidylethanolamine. See the text for details. This figure was modified from one previously published in reference 6, and is reproduced by permission of the American Society for Cell Biology, copyright 2006. Original painting by David S. Goodsell based on the scientific design of Daniel J. Klionsky.

It should be noted that the mitochondrial pool of Atg9 only counts for approximately 40% of the total cellular Atg9. The identity of the other peripheral compartments where Atg9 resides is not known. It is possible that Atg9 is sorted through the secretory pathway and directed to multiple organelles or membranous structures. In wild‑type yeast under steady state conditions, Atg9 does not reside in the ER, endosomes, peroxisomes, vacuole, or plasma membrane;2,12 however, it is possible that localization to the ER is only transient. A recent report from our lab suggests that Atg9 partially cofractionates with a Golgi marker protein by density gradient analysis.7 In addition, Atg27, an Atg9 binding partner that recruits Atg9 to the PAS, localizes to the PAS, mitochondria, and mostly to the Golgi complex. Thus, it is tempting to speculate that Atg9 and Atg27 may colocalize in part at this latter site. A study in mammalian cells suggests that mammalian Atg9 traffics between the trans‑Golgi network (TGN) and

Figure 3. Pexophagy is not disrupted by loss of Atg9‑Atg11 interaction. An atg9D strain expressing an integrated Pex14‑GFP fusion (CCH003) was transformed with a plasmid encoding either wild‑type Atg9 (pAPG9(416)) or Atg9H192L (pAtg9H192L(416)). Cells were grown in medium containing oleic acid as the sole carbon source and shifted to glucose‑rich, nitrogen‑limiting conditions (SD‑N). At the indicated time points, aliquots were taken, and protein extracts were analyzed by western blot using anti‑GFP antibody. Cleavage of Pex14‑GFP to generate free GFP is a measure of vacuole‑dependent peroxisome degradation.10 The positions of Pex14‑GFP and free GFP are indicated. Wild‑type (IRA001), atg1D (IRA002) and atg9D strains expressing the integrated Pex14‑GFP fusion were used as controls. The asterisks mark nonspecific bands.

www.landesbioscience.com

Autophagy

273


endosomes.13 Based on the conservation between mammalian and yeast Atg9, the Golgi complex would be the best candidate as another non-PAS compartment in which Atg9 localizes.

Membrane Delivery During Pexophagy Compared with bulk autophagy and the Cvt pathway, pexophagy is less characterized in S. cerevisiae. Both micro‑ and macropexophagy have been identified in the methylotrophic yeasts.3,14 During micropexophagy, peroxisomes targeted for destruction are engulfed by projections extended directly from the vacuolar membrane; while in macropexophagy, peroxisomes are sequestered by membranes that are of a non-vacuole origin and subsequently released into the vacuole. It is not clear which mechanism(s) S. cerevisiae uses for degradation of peroxisomes. With the Atg9H192L mutant, pexophagy proceeds normally (Fig. 3), which is consistent with the finding from a recent paper that examined a truncated Atg9 mutant lacking N‑terminal amino acids 154‑201 that loses interaction with Atg11.8 This finding suggests that although Atg9 and Atg11 are both required for pexophagy in S. cerevisiae, the interaction between the two proteins does not seem essential for the process in this organism. The result supports a hypothesis that peroxisome uptake may occur through a micropexophagic mode in S. cerevisiae because the membrane source for sequestration would primarily be the vacuole, and dependence of pexophagy on Atg9 transport would be minimal. References 1. Suzuki K, Kirisako T, Kamada Y, Mizushima N, Noda T, Ohsumi Y. The pre-autophagosomal structure organized by concerted functions of APG genes is essential for autophagosome formation. Embo J 2001; 20:5971‑81. 2. Reggiori F, Shintani T, Nair U, Klionsky DJ. Atg9 cycles between mitochondria and the pre-autophagosomal structure in yeasts. Autophagy 2005; 1:101‑9. 3. Kiel JAKW, Veenhuis M. Selective degradation of peroxisomes in the methylotrophic yeast Hansenula polymorpha. In:. Klionsky DJ, ed. Autophagy. Georgetown, Texas: Landes Bioscience, 2004:140‑53. 4. Tilney LG, Harb OS, Connelly PS, Robinson CG, Roy CR. How the parasitic bacterium Legionella pneumophila modifies its phagosome and transforms it into rough ER: Implications for conversion of plasma membrane to the ER membrane. J Cell Sci 2001; 114:4637‑50. 5. Klionsky DJ. The molecular machinery of autophagy: Unanswered questions. J Cell Sci 2005; 118:7‑18. 6. He C, Song H, Yorimitsu T, Monastyrska I, Yen W-L, Legakis JE, Klionsky DJ. Recruitment of Atg9 to the pre-autophagosomal structure by Atg11 is essential for selective autophagy in budding yeast. J Cell Biol 2006; 175:925‑35. 7. Yen W-L, Legakis JE, Nair U, Klionsky DJ. Atg27 is required for autophagy‑dependent cycling of Atg9. Mol Biol Cell 2006; 18:581-93. 8. Chang C-Y, Huang W-P. Atg19 mediates a dual interaction cargo sorting mechanism in selective autophagy. Mol Biol Cell 2006; In press. 9. Kim J, Kamada Y, Stromhaug PE, Guan J, Hefner‑Gravink A, Baba M, Scott SV, Ohsumi Y, Dunn Jr WA, Klionsky DJ. Cvt9/Gsa9 functions in sequestering selective cytosolic cargo destined for the vacuole. J Cell Biol 2001; 153:381‑96. 10. Reggiori F, Monastyrska I, Shintani T, Klionsky DJ. The actin cytoskeleton is required for selective types of autophagy, but not nonspecific autophagy, in the yeast Saccharomyces cerevisiae. Mol Biol Cell 2005; 16:5843‑56. 11. Yorimitsu T, Klionsky DJ. Atg11 links cargo to the vesicle‑forming machinery in the cytoplasm to vacuole targeting pathway. Mol Biol Cell 2005; 16:1593‑605. 12. Kim J, Huang W-P, Stromhaug PE, Klionsky DJ. Convergence of multiple autophagy and cytoplasm to vacuole targeting components to a perivacuolar membrane compartment prior to de novo vesicle formation. J Biol Chem 2002; 277:763‑73. 13. Young ARJ, Chan EYW, Hu XW, Köchl R, Crawshaw SG, High S, Hailey DW, Lippincott‑Schwartz J, Tooze SA. Starvation and ULK1‑dependent cycling of mammalian Atg9 between the TGN and endosomes. J Cell Sci 2006; 119:3888‑900. 14. Habibzadegah-Tari P, Dunn Jr WA. Glucose-induced pexophagy in Pichia pastoris. In: Klionsky DJ, ed. Autophagy. Georgetown, Texas: Landes Bioscience, 2004:126‑37.

274

Autophagy

2007; Vol. 3 Issue 3