Abstract
Two principal pathways in eukaryotes, namely the ubiquitin-proteasome system (UPS) and autophagy, mediate selective protein degradation. The UPS typically removes short-lived individual misfolded or regulatory polypeptides that have been tagged with polyubiquitin chains, whereas autophagy eliminates bulkier structures such as large protein complexes, insoluble protein aggregates, organelles, and invading intracellular pathogens. Protein degradation within the UPS is executed by the 26S proteasome, a large multisubunit proteolytic machine whose levels are tightly regulated transcriptionally and during assembly. Our recent studies identified a new mechanism for controlling 26S proteasome abundance through selective autophagy, which we term proteaphagy. This process is separately stimulated by nutrient starvation and proteasome inactivation, the latter occurring independently of ATG1 kinase regulation. Removal of inactive complexes is instead mediated by the proteasomal ubiquitin receptor RPN10, which can simultaneously bind both ubiquitinated proteasomes and lipidated ATG8 lining autophagic membranes.
All cellular organisms, including plants, employ a variety of approaches to control the activity and abundance of their constituent proteins. One such method is selective proteolysis, of which the 2 major pathways in eukaryotes are the ubiquitin-proteasome system (UPS) and autophagy. The UPS involves the attachment of polyubiquitin chains to individual targets via a highly polymorphic conjugation cascade. Proteins tagged in this way are then recognized, unfolded, and degraded by the 26S proteasome. While the UPS and autophagy were long thought to operate independently of one another, it has increasingly become clear that considerable mechanistic overlap exists between the 2 pathways. In particular, the discovery that ubiquitination can target certain cargo to autophagic vesicles via receptor proteins that interact with both ubiquitin and ATG8 has demonstrated a clear role for the ubiquitination machinery in autophagic breakdown.
The UPS plays an essential role in plants by controlling removal of key developmental regulators, facilitating responses to the environment by governing hormone perception and stress responses, and maintaining cellular homeostasis through degradation of damaged or misfolded proteins. As such, the system is tightly controlled and, as the final executer of protein degradation, it is unsurprising that the 26S proteasome is a key focal point of regulation. This can occur at a transcriptional level, during assembly, by post-translational modification and, in mammals and plants, by the incorporation of specific subunit isoforms. However, despite this knowledge, mechanisms for proteasome turnover had remained obscure.
In a recent paper, we described a novel mechanism for proteasome degradation. We initially noticed that proteasome subunit levels in Arabidopsis are elevated in a variety of autophagy-deficient backgrounds, but that this is not accompanied by concomitant increases in transcript levels or total proteasome activity. We therefore speculated that this increase might reflect accumulation of inactive 26S proteasomes normally degraded by autophagy. To confirm autophagic turnover, we tagged subunits of the proteasome core protease and regulatory particle with GFP and, after inducing autophagy by nitrogen starvation, observed GFP-tagged proteasomes accumulating in vacuolar puncta that were also decorated with the autophagic body marker mCherry-ATG8. We confirmed that intact 26S proteasomes are targeted to autophagic bodies by tracking them with epoxomycin-derived fluorescent probes, and again observing punctate structures inside vacuoles.
One striking feature of GFP-tagged autophagy substrates is that, upon delivery to the vacuole, the GFP moiety is relatively resistant to digestion, and hence accumulates. This released free GFP can then be detected via immunoblot, thus providing a facile, semiquantitative measure of autophagic turnover. By quantifying the ratio of free GFP to intact GFP-tagged proteasome subunits, we could directly track proteaphagy. The appearance of free GFP is blocked in autophagy mutants but is dramatically increased upon nitrogen starvation, and is also elevated following proteasome inhibition, achieved either chemically with MG132 and clasto-lactacystin β-lactone, or genetically in mutants that impair proteasome assembly. Interestingly, we found that starvation-induced proteaphagy is dependent on the ATG1 kinase, and hence is likely regulated by upstream nutrient-sensing kinases such as TOR. However, inhibitor-induced proteaphagy is independent of this complex, suggesting that separate routes for proteaphagy exist in Arabidopsis.
The observation that proteasomes can be degraded by autophagy raised a number of questions. For example, how inactive proteasomes are recognized, and the mechanism(s) that deliver them to autophagic vesicles, were unknown. One clue emerged from prior proteomic studies showing that proteasomes become extensively ubiquitinated upon MG132 inhibition. Our current work confirmed this finding and, in addition, showed that proteasome inactivation is also accompanied by increased association of RPN10, a component of the proteasome regulatory lid and one of its major receptors for ubiquitinated substrates. RPN10 is unusual among core proteasome subunits, as it is often present at substoichiometric levels, and can also exist in free form. This increased association is reduced by treatment with a deubiquitinating enzyme, suggesting it results from RPN10 binding ubiquitin moieties present on the proteasome, rather than from increased incorporation into the particle. This made RPN10 an attractive prospect as a receptor for inhibitor-induced proteaphagy.
Autophagic cargo receptors typically recognize not only their cognate targets, but also ATG8, which provides a docking site for these receptors complexed with their targets on the expanding phagophore. We therefore tested whether RPN10 binds ATG8, using a number of interaction assays including yeast 2-hybrid, in planta bimolecular fluorescence complementation, and in vitro pulldowns. In all 3 cases, strong association of RPN10 with ATG8 was detected, with the binding site further localized to the second of 3 ubiquitin-interacting motifs (UIMs) in the C-terminal half of RPN10. Confirmation that RPN10 is involved was shown by the fact that inhibitor-induced proteaphagy is blocked in rpn10–1 plants, which lack the C-terminal portion of RPN10 containing the UIMs. This is consistent with the hypothesis that RPN10 acts as a receptor to tether inactive proteasomes to expanding autophagic membranes.
The discovery of proteaphagy adds yet another example to the ever-expanding repertoire of quality control and recycling processes for which autophagy is responsible. Like proteaphagy, target ubiquitination is required for several other types of selective autophagy, with many cargo receptors containing both ubiquitin-associating (UBA) domains and an ATG8-interacting motif/LC3-interacting region (AIM/LIR). Plants possess a protein with hybrid properties of mammalian SQSTM1/p62 and NBR1, and 2 additional selective autophagy receptors have been described, namely TSPO, which is involved in porphyrin degradation, and ATI1/2, which appear to be involved in turnover of proteins from the endoplasmic reticulum. However, little else is known about plant selective autophagy. While RPN10 has now been identified as a proteaphagy receptor, its existence as a free protein, plus its ability to bind ubiquitin and ATG8, indicates it might play a wider role as a general autophagy receptor for ubiquitinated substrates.
Support for proteaphagy as an evolutionarily conserved mechanism for proteasome clearance has been inferred from proteomic studies, which identified proteasome subunits as abundant contents within purified yeast and mammalian autophagic vesicles. However, proteaphagic mechanisms likely differ between species. Sequence comparisons have revealed that RPN10 has undergone considerable evolutionary divergence among eukaryotes, especially within the C-terminal region containing the UIMs. While all 3 motifs are highly conserved within the plant kingdom, the ATG8-binding UIM is absent in Saccharomyces cerevisiae Rpn10, and has substantially diverged in the human form. Consequently, alternative receptors are likely required to drive proteaphagy outside of plants. Additionally, the E3 ligases involved in proteasome ubiquitination, and the mechanisms by which they identify inactive proteasomes, remain to be determined. As such, further exploration of proteaphagy is likely to be an area of fruitful future study.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Acknowledgments
The authors wish to thank Faqiang Li for critical reading of the manuscript.