A missing piece of the puzzle: Atg11 functions as a scaffold to activate Atg1 for selective autophagy

Abstract

The mechanism regulating Atg1 kinase activity for the initiation of selective macroautophagy (hereafter autophagy) under nutrient-rich conditions has been a long-standing question. Canonically in yeast, nutrient starvation or rapamycin treatment repress TOR complex 1 and stimulate the Atg1 complex (including at least Atg1, Atg13, Atg17, Atg29 and Atg31), which allows the recruitment of downstream autophagy-related (Atg) components to the phagophore assembly site (PAS), culminating in phagophore formation, and, subsequently, autophagosome biogenesis. Atg1 also functions under conditions promoting selective autophagy that do not necessarily require nutrient deprivation for induction. However, there has been some debate as to whether Atg1 catalytic activity plays a more important role under conditions of nutrient starvation-induced autophagy (i.e., bulk autophagy) vs. selective autophagy (e.g., the cytoplasm-to-vacuole targeting [Cvt] pathway). A recent paper by Kamber and colleagues investigates the mechanism regulating Atg1 activity during selective autophagy.

Keywords:

• Atg19
• Cvt
• pexophagy
• prApe1
• yeast

Abbreviations

ad	=	autoactivation-dead
as	=	analog-sensitive
Atg	=	autophagy-related
Cvt	=	cytoplasm-to-vacuole targeting
kd	=	kinase-dead
PAS	=	phagophore assembly site
prApe1	=	precursor aminopeptidase I
WT	=	wild type

The study by Kamber and colleagues explores the mechanism by which Atg1 locally promotes selective autophagy under nutrient-rich conditions.¹ Using a cell-free genetic and biochemical approach, the authors sought to determine whether the scaffold protein Atg11 acts as a trigger for Atg1 and selective autophagy activation by binding the Cvt pathway receptor Atg19, and Atg19-bound precursor aminopeptidase I (prApe1). Atg1 is a serine/threonine kinase² that forms a complex with Atg13 and the Atg17-Atg31-Atg29 subcomplex to recruit and activate downstream autophagy components (reviewed in refs. 3, 4). This complex is necessary for the recruitment of additional Atg proteins to the PAS and for autophagosome biogenesis. In a multiple-knockout strain of Saccharomyces cerevisiae, which otherwise lacks the necessary proteins for autophagosome formation,⁵ Atg11 interacts with Atg29 and facilitates recruitment of the Atg17-Atg31-Atg29 complex to the PAS.⁶

Previous work by others has suggested that Atg1 may function in both kinase-dependent (for example, during nutrient-rich selective autophagy such as the Cvt pathway) and -independent (during starvation-induced nonselective autophagy) roles.⁷ In the Cvt pathway, a form of selective, biosynthetic autophagy that occurs under vegetative growth conditions, certain cargo such as prApe1, Ams1 (α-mannosidase), and Ape4 (aspartyl aminopeptidase) are targeted for transport from the cytoplasm to the vacuole (reviewed in ref. 3). Transport of prApe1 along the Cvt pathway requires the receptor Atg19, and the Atg11 scaffold, in addition to Atg1 and Atg13.^8-10

Work by Matsuura et al. demonstrated that Atg1 undergoes autophosphorylation, and that this post-translational modification (and subsequently kinase activity) decreases during starvation-induced autophagy.² Consistent with this, Abeliovich and coworkers showed that Atg1 kinase activity plays a more important role during the Cvt pathway than it does during starvation-induced autophagy.⁷ The conclusion that Atg1 is involved in autophagosome nucleation in a manner that is relatively independent of its kinase activity suggests a key difference in the formation of canonical autophagosomes and selective Cvt vesicles.⁷ Another significant difference between nonselective autophagy and the Cvt pathway is the involvement of the scaffold Atg11, which is required for the Cvt pathway but is not essential for autophagy.⁸

A later study by Kamada and colleagues demonstrated that nonselective autophagy (but not the Cvt pathway) requires an Atg1-Atg13 interaction.¹¹ In contrast to previous reports,^2,7 Atg1 kinase activity was required for (and enhanced by) starvation- or rapamycin-induced autophagy.¹¹ Additional work revealed that Atg1 autophosphorylation of residues Thr226 and Ser230 within the activation loop of the kinase domain is critical for its kinase activity, autophagy and the Cvt pathway.^12,13

The current study by Kamber and coworkers provides a new model to reconcile previous findings regarding the role of Atg1 under nutrient-rich conditions promoting selective autophagy. The authors applied a chemical genetics approach in combination with a cell-free assay system to address the question of Atg1 catalytic functioning during selective autophagy. The cell-free assay system developed by Kamber et al. utilizes an unnatural bulky ATP-derivative that is inaccessible to wild-type (WT) kinases—N⁶-PhEt-ATPγS—to selectively monitor analog-sensitive Atg1 (Atg1^as) activity in cell extracts harboring the genetic background of interest.¹⁴ To generate Atg1^as, the gatekeeper residue within the ATP binding pocket of Atg1 (M102) was replaced with a glycine residue. Atg1 kinase activity was assayed by measuring thiophosphorylation of the substrates of interest by blotting, following the alkylation of cell extracts with the agent paranitrobenzyl mesylate to enable antibody detection during immunoblotting. Thus, only phosphorylation due to Atg1^as kinase activity is measured in this assay as only Atg1^as (and not WT kinases) can accept N⁶-PhEt-ATPγS.

Kamber and colleagues investigated whether receptor-bound targets (such as Atg19) can activate Atg1 using Atg11 as a scaffold, and whether binding to prApe1 activates Atg1 kinase activity under nutrient-rich conditions. To identify interactions between prApe1 and Atg1, FLAG-Atg1 was affinity purified from crude extracts, eluted, and then analyzed by staining, immunoblotting, and quantitative mass spectrometry. Precursor Ape1 coimmunoprecipitates with WT, kinase-dead (kd; D211A mutant), and autoactivation-dead (ad; T226A mutant) Atg1. Atg1 also physically associates with Atg11, Atg13, Atg19, and Atg19-bound prApe1 in agreement with previous studies.^9,11,15 As predicted, in the absence of Atg11, Atg1 and Atg13 form a complex devoid of Atg19-prApe1; whereas without Atg19, Atg1-Atg13 and Atg11 form a complex lacking prApe1. Atg11 binds directly to Atg1, so that in extracts lacking Atg13, Atg1 still interacts with Atg11 (and therefore Atg19-prApe1), although only weakly. These results were further verified by assaying Atg1 kinase activity using myelin basic protein as a substrate, demonstrating that Atg1 catalytic activity requires Atg11 and Atg19. The authors conclude that Atg1 activity under nutrient-rich conditions is controlled by interactions between Atg1-Atg13, Atg11 and Atg19-bound cargo, and that Atg11, Atg19 and prApe1 are not essential during rapamycin-induced autophagy.

The authors also investigated at which stage of selective autophagy (i.e., induction and nucleation, phagophore expansion and closure, etc.), Atg1 kinase activity is required. To assess the role of Atg1 during the formation of Cvt vesicles, the authors first examined the localization of Atg2 under nutrient-rich conditions, as Atg2 localization is dependent on Atg1.¹⁶ In the study by Kamber et al., cells expressing Atg2-GFP but lacking Atg1 or expressing Atg1^kd or Atg1^ad have a significantly lower level of colocalization with mCherry-Ape1, compared to cells expressing WT Atg1. Furthermore, transmission electron microscopy of ypt7Δ cells (that are deficient in autophagosome-vacuole fusion) also expressing Atg1^kd, Atg1^ad, or Atg11ΔRBD (lacking the Atg19 receptor binding domain) display considerably less membrane-enclosed prApe1 complexes than seen in WT cells. Taken together, these results indicate that Atg1 kinase activity is necessary under growing conditions for membrane elongation during the formation of selective Cvt vesicles.

The authors applied the aforementioned cell-free chemical genetics assay to further analyze Atg1 kinase activation by targets of selective autophagy. Using Atg1^as in this system, the authors regenerated the Atg1-Atg13 subcomplex and demonstrated that Atg1^as thiophosphorylates known Atg1 substrates including Atg1, Atg2, Atg9 and Atg13. Additionally, the authors confirmed that Atg1^as thiophosphorylates Vps30/Atg6 (in agreement with data implicating mammalian ULK1 in the phosphorylation of the Vps30 homolog BECN1¹⁷). In vitro Atg1 kinase activity is diminished in cell extracts lacking Atg11, Atg13, or Atg19 in the absence of rapamycin. Pretreatment of cells with rapamycin restores Atg1-mediated phosphorylation to atg11Δ and atg19Δ extracts, but not to those from atg13Δ cells, indicating that Atg13 is necessary for both selective and nonselective autophagy. Taken together, these data show that the cell-free system recapitulates Atg1 activity including known substrate phosphorylation and Atg1 kinase activation by both Atg19-bound cargo and TOR-mediated signals.

Kamber and colleagues use the cell-free Atg1^as system to investigate the mechanism of Atg1 activation by Atg11 and Atg19-bound targets. Atg1 kinase activity appears dependent on Atg11-Atg19 interactions, as kinase activity is minimally recovered in atg11Δ atg19Δ extracts supplemented with increasing amounts of purified FLAG-Atg11, or in atg11Δ extract incubated with purified FLAG-Atg11ΔRBD. However, if purified Atg19-FLAG is added to atg11Δ atg19Δ extracts along with FLAG-Atg11, Atg1 kinase activity is restored. When purified Atg19Δ11BD-FLAG (which cannot bind Atg11¹⁸) is added to the atg11Δ atg19Δ extracts, Atg1 kinase activity again cannot be rescued. In light of these data, Atg11 appears to function as both a scaffolding protein and as a switch to activate Atg1 kinase in response to Atg19-bound targets.

The authors further investigated whether another form of selective autophagy—pexophagy, the selective degradation of peroxisomes by autophagy—could serve as a trigger for Atg1 activation. PEX1, encoding an AAA-ATPase family member necessary for the recycling of Pex5, a peroxisome-targeting signal 1 import receptor, was deleted to induce pexophagy under nutrient-rich conditions.¹⁹ Pexophagy is dependent upon the receptor Atg36 and scaffold Atg11.^19,20 As an alternative approach, Pex1 was eliminated using an auxin inducible-degron system.^19,21 The authors demonstrated that pexophagy increases Atg1 kinase activity in a manner that is dependent on Atg36 and Atg11, but does not require Atg8. A model was proposed in which receptor-bound cargo (such as Atg19-bound prApe1, or Atg36-bound damaged peroxisomes) use Atg11 as a switch to activate Atg1 kinase and drive autophagosome formation, promoting selective autophagy. In this way, selective autophagy receptor-bound targets locally orchestrate Atg1 kinase activation via Atg11 under nutrient-rich conditions despite existing cellular mechanisms to suppress its activity such as through TOR complex 1.

This study provides insight into the role of Atg11 as more than simply a structural scaffold during selective autophagy, but rather implicates it as a dynamic regulator that controls Atg1 kinase activity under conditions typically favoring global inhibition of bulk autophagy. Kamber et al. show that Atg1 kinase activity occurs under nutrient-rich conditions, a process that requires the scaffold Atg11, a selective autophagy receptor, and the sequestered cargo. Additionally, Atg1 catalytic activity requires autoactivation at residue Thr226 and the presence of Atg13. These results support previous reports published by others indicating that Atg1 kinase activity is particularly important during selective autophagy,^7,13 that autoactivation at Thr226 is necessary for its function,^12,13 and that Atg11 is required for optimal Atg1 kinase activity.¹¹

In light of this work, additional questions remain to be answered. In this study, 2 forms of selective autophagy—the Cvt pathway (with Atg19-bound prApe1) and pexophagy (with Atg36-bound peroxisomes)—are reported to require the Atg11 switch for Atg1 catalytic function and selective autophagosome formation. It would be interesting to see if the results are similar when additional Cvt pathway cargo (such as Atg19-bound Ape4 or Ams1) are analyzed, or alternatively, if Atg34-bound Ams1 can also bring about a similar switch during nonselective autophagy.²²

There are obvious physical differences between the canonical autophagosome (300-900 nm) and the smaller Cvt vesicle (140-160 nm). We still do not fully know what determines the upper size limit during the formation of the Cvt vesicle (i.e., why is it smaller than an autophagosome?) and what factor(s) regulates this physical decision. Similarly, what additional components signal to the cellular machinery to form a Cvt vesicle rather than an autophagosome during selective autophagy? Although possible explanations have been proposed,²³ direct evidence for these hypotheses has yet to be provided. Undoubtedly, additional investigation will further elucidate the molecular mechanisms orchestrating Atg1 activity during selective autophagy.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors thank Dr. Aileen Ariosa (University of Michigan) for providing helpful comments.

Funding

This work was supported by NIH grant GM053396 to DJK.