|Home | About | Journals | Submit | Contact Us | Français|
Not only is autophagy the major intracellular pathway for degradation and recycling of long-lived proteins and organelles, it is also involved in both the pathogenesis and prevention of many human diseases. Much progress has been made on the identification and characterization of AuTophaGy-related (ATG) genes, in yeast and in mammals. However, our understanding of the molecular mechanisms of autophagy remains quite limited, far from enough to harness autophagy for therapeutic applications. To better understand the molecular mechanisms, we took a unique and novel approach to study autophagy in yeast. We generated a multiple knockout Saccharomyces cerevisiae strain with 24 ATG genes deleted, and determined the minimum requirements for different aspects of autophagy. Our data also provided us with new insights into autophagy, different from those obtained from in vitro analyses. In this addendum, we briefly discuss our findings and consider fields where this multiple knockout strain can be of potential use.
Autophagy is a conserved pathway in eukaryotes for the degradation and recycling of cytosolic components and organelles.1 It is a vesicular process involving the formation of a double-membrane vesicle, an autophagosome, which sequesters portions of the cytoplasm. Autophagy occurs constitutively at a low and basal level, but can be massively induced by environmental cues such as nutrient deprivation, growth factor depletion and pathogen invasion. Beyond its homeostatic function, autophagy has been linked to many human pathologies, including neurodegenerative diseases and cancer. Accordingly, the up-regulation or down-regulation of autophagy in different disease settings may have potential therapeutic applications.2 During autophagy, a membrane of unknown origin, designated the phagophore, sequesters bulk cytoplasm and expands to form a double-membrane vesicle, the autophagosome, which then fuses with the lysosome, or the yeast vacuole. In yeast, this fusion event releases an inner vesicle (termed the autophagic body) into the lumen of the vacuole. In either case, fusion of the autophagosome with the lysosome or vacuole provides access to the degradative hydrolases, allowing degradation of the cargo and recycling of the resulting macromolecules.3
The autophagic pathway has been broken down into several discrete steps: (1) induction; (2) cargo recognition and packaging; (3) nucleation of vesicle formation; (4) vesicle expansion and completion; (5) autophagy-related (Atg) protein retrieval; (6) targeting, docking and fusion of the completed vesicle with the lysosome/vacuole; (7) breakdown of the autophagic body and/or cargo; and (8) efflux of macromolecules.4 31 ATG genes have been identified in autophagy-related pathways in fungi, many of which function in the general vesicle formation process, i.e. nucleation of vesicle formation, and vesicle expansion and completion. But the exact functions of the corresponding gene products are not clear. In addition, many questions remain, such as how these currently “schematic” steps of autophagy are executed and woven together, and in which manner the autophagy machinery is assembled. Previous studies have been confined to single- and double-deletion mutant strains, yielding information about requirements, but not sufficiency of ATG genes in different steps of autophagy. With many such questions in mind, we started constructing a multiple knockout strain by deleting most of the ATG genes involved in autophagy, and we used the strain to reconstitute different steps or aspects of autophagy in vivo. An alternative to this method is in vitro reconstitution. However, with intact cytoskeletal and regulatory networks, our in vivo system is more physiological than in vitro systems. Moreover, compared to in vitro reconstitution, the multiple knockout strain is more suitable for localization studies.
Among the 31 ATG genes, 24 of them were knocked out based on a loxP-Cre system by multiple rounds of deletion and marker rescue.5, 6 Atg31 was not deleted because it was published after we finished constructing the strain and carrying out our initial experiments.7 The remaining ATG genes, ATG15, ATG22, ATG25, ATG26, ATG28 and ATG30, were not deleted because they are not found in Saccharomyces cerevisiae, are only required for peroxisome degradation in other methylotrophic yeasts, or function after autophagosome fusion with the lysosome/vacuole.5 The multiple knockout strain YCY123 is referred to as the MKO strain, and is defective in autophagy-related pathways, but not other vacuolar protein targeting pathways. The strain does not have any obvious unexpected defect other than a slight growth delay.5 Additionally, we constructed a newer MKO strain YCY132 by knocking out ATG31 with a HIS5 marker in strain YCY123, and these two strains show similar phenotypes.
As a proof of principle, we first reconstituted the cargo packaging step in the cytoplasm-to-vacuole targeting (Cvt) pathway, a selective type of autophagy which shares most of the components needed for autophagy, but utilizes additional proteins in cargo recognition and packaging.8–10 We reproduced the temporal order of this packaging step: The adaptor protein Atg11 cannot colocalize with the cargo precursor aminopeptidase I (prApe1) without the receptor Atg19, and the three proteins are necessary and sufficient for this cargo packaging event.5
Previous studies show that the cargo complex comprised of prApe1, Atg19 and Atg11 is a critical component for PAS (phagophore assembly site; a perivacuolar structure presumed to be the nucleating site of autophagosomes, and Cvt vesicles that function in prApe1 import) assembly in vegetative conditions,9 but our reconstitution data suggest that the cargo complex is not sufficient by itself to efficiently target to a perivacuolar site—additional Atg proteins may be required for the delivery of the cargo complex to the PAS.5 With regard to starvation conditions, one study suggests that Atg17 is the initial factor for PAS assembly.11 However, we find that Atg17 is not sufficient for starvation-specific PAS formation; rather, the Atg17-interacting proteins Atg1 and Atg13, along with Atg17 appear to be required and sufficient for the initial PAS assembly during starvation.5 Other factors in the Atg1 complex, such as Atg11 and Atg29, further enhance assembly of the PAS.5
Next, we extended our analyses to two conjugation systems in autophagy: the Atg12–Atg5 and the Atg8–phosphatidylethanolamine (PE) conjugation. Eight conserved Atg proteins are involved in these two separate but related conjugation systems. Atg12, a ubiquitin-like protein, is covalently linked to Atg5 through the serial action of an E1-like activating enzyme, Atg7, and an E2-like conjugating enzyme, Atg10.12, 13 Atg16 interacts with the Atg12–Atg5 conjugate, and is required for maximal efficiency of the conjugation.14 In the other conjugation system, the ubiquitin-like protein Atg8 is conjugated to a lipid, PE.15 The C-terminal arginine residue of Atg8 is first cleaved by the cysteine protease Atg4, exposing a glycine residue for the action of Atg7. The activated Atg8 is then transferred to the E2-like enzyme Atg3, and finally conjugated to PE. The absence of Atg12, Atg5 or Atg16 reduces the Atg8 conjugation efficiency, and Atg4 also affects the equilibrium of Atg8 and Atg8–PE: Atg8 can be cleaved from Atg8–PE through a second cleavage by Atg4, to deconjugate the protein.
Both conjugation systems have been reconstituted in vitro, so the minimum components have been identified for these reactions.16–19 We wanted to know whether the in vitro analyses represented the in vivo situation. By adding back different combinations of Atg proteins to the MKO strain, we find that, consistent with the published data, Atg5, Atg7, Atg10 and Atg12 are necessary and sufficient for Atg12–Atg5 conjugation, and Atg16 facilitates formation of the conjugate.5 We further find that components in the Atg8–PE conjugation system also enhance the conjugation reaction, a connection that was not previously known.5 Even more surprising, we find that even when all eight of the proteins involved in both conjugation systems are expressed, we are not able to detect Atg8–PE.5 However, when Atg4 is removed from the reaction (blocking deconjugation, and thus favoring accumulation of Atg8–PE), and a modified Atg8ΔR is used to bypass the need for the initial cleavage by Atg4, a significant amount of Atg8–PE is detected. The presence of the Atg12–Atg5 conjugate further increases Atg8–PE formation in this setting.5
Thus, in the MKO strain when Atg4 is present, either additional Atg proteins are required for conjugation to be fully efficient, or Atg4 activity is dysregulated, or both. A remaining question is what factor(s) may regulate the conjugation efficiency or Atg4 activity. In addition, if the Atg4 activity is dysregulated, which step of Atg4 cleavage is affected? One possibility is that the initial cleavage of the last arginine of Atg8 does not occur in the MKO strain; in this case, the downstream reactions cannot take place. A second possibility is that Atg4 is overactive, causing deconjugation to occur too fast to accumulate any detectable amount of Atg8–PE. We tested the first possibility by examining the activity of Atg4 in the MKO strain. To detect the initial cleavage, we used a plasmid expressing a C-terminal GFP-tagged Atg8 under its own promoter.20 As a result of the initial cleavage by Atg4, GFP is released together with the last arginine of Atg8 in wild-type cells.20 Thus, the level of free GFP indicates the efficiency of the first cleavage reaction. As a positive control, we examined wild-type cells and verified that only the free GFP band is detected in both growing and nitrogen starvation conditions (Figure 1). As expected, without Atg4, the MKO strain transformed with the empty vector did not process any Atg8-GFP (i.e., there was no free GFP band). In contrast, when Atg4 was expressed from a plasmid, only the free GFP band was detected, just as in wild-type cells either in the presence or absence of nitrogen. These data suggest that the initial cleavage of Atg8 is not blocked in the MKO strain, which argues against the first possibility. An Atg4 construct capable of rapid switching between its active and inactive states may enable us to test the second possibility.
With the MKO strain, we determined minimum requirements for cargo packaging, initial assembly of the starvation-specific PAS, and conjugation of the ubiquitin-like proteins.5 Although we extended previous knowledge of PAS assembly and protein conjugation,5 many questions regarding these steps still remain. The MKO strain will continue to be useful for determining the temporal order of protein arrival at the PAS, and for studying how conjugation and deconjugation are regulated. The strain can also be used to reconstitute other steps or aspects of autophagy, such as Atg protein retrieval.
Besides reconstitution, we can use the MKO strain to study many poorly characterized aspects of autophagy, and gain new information that is difficult to obtain from single- and double-deletion analyses. For example, the MKO strain can be used to determine whether a particular interaction between Atg proteins is direct. As previous data suggest, there are several protein complexes functioning in different steps of autophagy, but there has been no easy way to determine how proteins in a complex interact, and whether sub-complexes exist. The MKO strain expressing various combinations of Atg proteins, will be a useful tool for studying protein interactions and complex formation. The MKO strain will also be valuable for determining minimum requirements for newly discovered aspects of autophagy.
Addendum to: Cao Y, Cheong H, Song H, Klionsky DJ. In vivo reconstitution of autophagy in Saccharomyces cerevisiae. J Cell Biol 2008; 182:703–13.