https://orcid.org/0000-0002-7828-8118
ABSTRACT
There is a type of noncanonical autophagy, which is independent of ATG5 (autophagy related 5), also referred to as alternative autophagy. Both canonical and ATG5-independent alternative autophagy require the initiator ULK1 (unc-51 like kinase 1), but how ULK1 regulates these two types of autophagy differently remains unclear. A recent paper from Torii et al. demonstrates that phosphorylation of ULK1 at Ser746 by RIPK3 (receptor interacting serine/threonine kinase 3) is the key difference between these two types of autophagy; this phosphorylation is exclusively found during alternative autophagy.
The molecular mechanism of macroautophagy (hereafter autophagy) has been extensively studied. In canonical autophagy, ATG5 plays a major role as part of the ATG12–ATG5 conjugation system, which is involved in the biogenesis of autophagosomes. In addition to this canonical autophagy, researchers have identified several noncanonical versions, one of which is the ATG5-independent alternative autophagy (hereafter alternative autophagy). There are several differences between canonical autophagy and alternative autophagy: 1. Whereas there are multiple sources for phagophore membrane expansion in canonical autophagy, autophagosomal membranes of alternative autophagy are mainly derived from the trans-Golgi. 2. Although both starvation and genotoxic stresses can induce canonical autophagy, only genotoxic stress induces alternative autophagy. 3. Alternative and canonical autophagy can target different substrates under the same stimulus condition [Citation1–Citation3].
Despite these differences, both types of autophagy are under control of ULK1, a serine/threonine kinase that is a homolog of yeast Atg1. Under normal conditions, MTOR complex 1 (MTORC1) phosphorylates ULK1 at Ser637 and Ser757, modifications that keep ULK1 inactive. In canonical autophagy, stress stimuli will facilitate dephosphorylation of ULK1, allowing translocation of ULK1 to the site of phagophore formation, and activating its kinase activity. Conversely, AMP-activated protein kinase (AMPK) can phosphorylate ULK1 at several different sites [Citation4]. Based on these observations, it has been determined that ULK1 activity is regulated by the status of its multiple phosphorylation sites; however, how ULK1 regulates alternative autophagy was unclear. The newly published paper from Torii et al., elucidates the mechanism of ULK1 activation in this type of noncanonical autophagy, which mainly involves Golgi-localized phospho (p)-ULK1 Ser746 and the kinase RIPK3 that can interact with and phosphorylate ULK1 at this site [Citation5].
Torii et al. first searched for ULK1 phosphorylation sites after using etoposide, which causes the accumulation of DNA strand breaks, to strongly induce alternative autophagy. LC-MS/MS was applied to analyze the immunoprecipitated and digested ULK1 for phosphorylation sites in both untreated atg5 knockout (KO) and etoposide-treated atg5 KO mouse embryonic fibroblasts (MEFs), among which Ser746 showed the most potential. The authors developed a specific antibody against p-ULK1 Ser746, which was used to confirm the phosphorylation on this specific site upon etoposide treatment. This specific antibody was also employed to investigate the subcellular localization of p-ULK1 Ser746 by immunostaining. The result shows that the phosphorylation signal merges completely with a Golgi marker, consistent with the previous result that Golgi membranes are the source of alternative autophagy [Citation2].
To determine the role of ULK1 phosphorylation at Ser746 in alternative autophagy, the authors used correlative light and electron microscopy/CLEM analysis to track the fluorescence of autolysosomes upon alternative autophagy induction. The absence of autolysosomes is observed in MEFs with triple knockouts of Atg5, Ulk1 and Ulk2, which can only be restored by wild-type ULK1 but not by an alanine mutant (S746A), indicating the importance of phosphorylation of ULK1 at Ser746 in alternative autophagy. Next, the authors explored the kinase that is responsible for this specific modification. They found that the target sequence of ULK1 at Ser746 is similar to that of RIPK3 substrates, based on a 2012 report [Citation6]. An immunoprecipitation analysis shows an interaction between endogenous RIPK3 and ULK1 in untreated conditions; etoposide treatment increases the affinity of this binding. The authors present evidence to confirm that RIPK3 is essential for the genotoxic stress-induced phosphorylation of ULK1 at Ser746. Importantly, the necessary role of RIPK3-dependent phosphorylation at ULK1 Ser746 is only observed in alternative autophagy, not in the canonical pathway. As RIPK3 can also phosphorylate MLKL (mixed lineage kinase domain like pseudokinase) and cause membrane rupture, it will result in necroptosis when cells are treated with the pan-caspase inhibitor zVAD (TCZ) [Citation7]. The authors subsequently tested the possible relation between alternative autophagy and necroptosis with regard to RIPK3. It turns out that although RIPK3 can be activated by both etoposide and TCZ treatment, these stimuli induce alternative autophagy or necroptosis, respectively, with no crosstalk, and the mechanisms of RIPK3 activation upon the two treatments are completely different.
A previous report from the same lab shows that etoposide leads to dephosphorylation of ULK1 at Ser637, in a manner dependent on TRP53/p53 and PPM1D (protein phosphatase, Mg2+/Mn2+ dependent 1D) [Citation8]. The authors further determined that this dephosphorylation via the TRP53-PPM1D axis is also required for ULK1 Ser746 phosphorylation and subsequent alternative autophagy. What is more, upon genotoxic stress, TRP53 can regulate RIPK3 at the transcriptional level, and the phosphorylation of ULK1 at Ser746 is blocked in cells lacking TRP53. Therefore, TRP53 plays a dual role with regard to ULK1 activity: together with TRP53, PPM1D dephosphorylates ULK1 at Ser637, while it can modulate RIPK3 expression to promote phosphorylation of ULK1 at Ser746.
The authors applied a close proximity (Duolink) assay and showed that the RIPK3-ULK1 interaction increases upon genotoxic stress; puncta corresponding to RIPK3-ULK1 are maintained in the cytosol. The authors further tested interaction between ULK1 and GOSR1/GS28, a Golgi marker. They found that upon etoposide treatment, the number of wild-type ULK1-GOSR1 signals increases, whereas fewer signals are observed for ULK1S746A-GOSR1, indicating the importance of ULK1 phosphorylation at Ser746 to the translocation of ULK1 from the cytosol to the Golgi. RB1CC1 and ATG13 are two binding partners of ULK1 in canonical autophagy, but their interactions are largely blocked in alternative autophagy. The authors further show that ripk3 knockout and loss of ULK1 Ser746 phosphorylation can increase interactions among ULK1, RB1CC1 and ATG13, whereas RB1CC1 and ATG13 are not found at the Golgi in atg5 KO MEFs after etoposide treatment, indicating that ULK1 Ser746 phosphorylation and Golgi translocation facilitate the dissociation of ULK1 from RB1CC1 and ATG13 in alternative autophagy.
It has been reported that genotoxic stress can induce Golgi morphology alterations, and disturbance of Golgi trafficking can induce alternative autophagy [Citation1]. Therefore, the authors further investigated the association between genotoxic stress and Golgi trafficking, and the role of alternative autophagy in this association. Tracking both an artificial substrate and an endogenous cargo in Golgi trafficking, the authors discovered that RIPK3-dependent ULK1 Ser746 phosphorylation and alternative autophagy can play a vital role in eliminating superfluous undelivered proteins from the Golgi in etoposide-treated MEFs.
In conclusion, RIPK3-dependent phosphorylation of ULK1 Ser746 leads to the disassociation of this protein from components of canonical autophagy, causing the protein to move to the Golgi apparatus, where it participates in ATG5-independent alternative autophagy.
Disclosure statement
No potential conflict of interest was reported by the authors.
Additional information
Funding
References
- Yamaguchi H, Arakawa S, Kanaseki T, et al. Golgi membrane-associated degradation pathway in yeast and mammals. Embo J. 2016;35(18):1991–2007.
- Nishida Y, Arakawa S, Fujitani K, et al. Discovery of Atg5/Atg7-independent alternative macroautophagy. Nature. 2009;461(7264):654–658.
- Honda S, Arakawa S, Nishida Y, et al. Ulk1-mediated Atg5-independent macroautophagy mediates elimination of mitochondria from embryonic reticulocytes. Nat Commun. 2014;5(1):4004.
- Zachari M, Ganley IG, Lane JD. The mammalian ULK1 complex and autophagy initiation. Essays Biochem. 2017;61(6):585–596.
- Torii S, Yamaguchi H, Nakanishi A, et al. Identification of a phosphorylation site on Ulk1 required for genotoxic stress-induced alternative autophagy. Nat Commun. 2020;11(1):1754.
- Wu X, Tian L, Li J, et al. Investigation of receptor interacting protein (RIP3)-dependent protein phosphorylation by quantitative phosphoproteomics. Mol Cell Proteomics. 2012;11(12):1640–1651.
- Sun L, Wang H, Wang Z, et al. Mixed lineage kinase domain-like protein mediates necrosis signaling downstream of RIP3 kinase. Cell. 2012;148(1–2):213–227.
- Torii S, Yoshida T, Arakawa S, et al. Identification of PPM1D as an essential Ulk1 phosphatase for genotoxic stress-induced autophagy. EMBO Rep. 2016;17(11):1552–1564.