New regulators of PRKN-independent mitophagy

ABSTRACT

Mitophagy, a type of selective autophagy targeting damaged or superfluous mitochondria, is critical to maintain cell homeostasis. Besides the well-characterized PRKN-dependent mitophagy, PRKN-independent mitophagy also plays significant physiological roles. In a recent study, researchers from Anne Simonsen’s lab discovered two lipid binding kinases, GAK and PRKCD, as positive regulators of PRKN-independent mitophagy. The researchers further investigated how these two proteins regulate mitophagy and demonstrated their roles in vivo. Focusing on the less known PRKN-independent mitophagy regulators, these findings shed light on understanding the mechanism of mitophagy and its relation to diseases.

KEYWORDS:

• Autophagy
• GAK
• lipid-binding kinases
• PRKCD
• PRKN

Mitochondria are essential to cell metabolism and physiology, but they have to be removed at some point during development in some cells or when they get damaged. Mitophagy, a type of selective autophagy, is critical for the clearance of superfluous or defective mitochondria. Different types of mitophagy have been characterized in mammalian cells [1] and the best-characterized form is mediated by PINK1 and the E3 ligase PRKN; in response to membrane potential dissipation, PRKN is recruited and activated by PINK1 on the outer mitochondrial membrane (OMM) and ubiquitinates several OMM proteins. In turn, this causes the recruitment of mitophagy receptors, which subsequently connect mitochondria to the phagophore [1]. In addition to this pathway, mitophagy can be PRKN independent. Some E3 ligases other than PRKN can ubiquitinate OMM proteins [1], and some OMM proteins, such as BNIP3L/NIX, BNIP3 and FUNDC1 can directly bind to LC3 and function as receptors [2].

PRKN-independent mitophagy plays important roles in physiology and in response to stress. For instance, BNIP3L-mediated mitophagy is crucial for programmed mitochondrial clearance during erythroid cell maturation [3]. Additionally, PRKN-independent mitophagy is triggered in response to hypoxia, which depends on the higher expression level of BNIP3 and BNIP3L induced by HIF1A (hypoxia inducible factor 1 subunit alpha) [4,5]. However, many questions remain open regarding the regulation of PRKN-independent mitophagy [6], especially with regard to lipids and lipid-binding proteins [7]. In the paper highlighted here, Munson et al. identified GAK and PRKCD, two lipid-binding kinases, as positive regulators of PRKN-independent mitophagy and further demonstrated the different mechanisms of these two proteins in mitophagy regulation and their roles in vivo [8].

To specifically examine PRKN-independent autophagy, the authors used U2OS cells, which have a low PRKN expression that is not sufficient to induce mitophagy in response to the loss of mitochondrial membrane potential [9]. In addition, the iron chelator deferiprone (DFP) and dimethyloxallyl glycine (DMOG) were used; these two chemicals stabilize HIF1A, thus replicating hypoxia-induced mitophagy [10,11]. In addition, unlike CCCP, which depolarizes the mitochondrial membrane, DFP treatment maintains the mitochondrial membrane potential [10,12] so that the PINK1-PRKN pathway is less likely to be induced.

To start with, a mitophagy reporter, EGFP-mCherry fused with a mitochondrial localization sequence, is expressed in U2OS cells (hereafter named IMLS cells) to measure mitophagy. Yellow mitochondrial structures corresponding to overlapping green and red fluorescent signals can be seen under normal conditions but after DFP treatment red puncta appear indicating delivery to the lysosome and quenching of the GFP signal. Formation of the red puncta is compromised by ULK1 knockdown or bafilomycin A₁ treatment, suggesting DFP-induced mitophagy. In line with this, DFP treatment results in a reduction in both CS (citrate synthase) activity and in the abundance of mitochondrial proteins. Using IMLS cells, the authors conducted a screen of 197 lipid-binding proteins expressed in U2OS cells, from which they found that knocking down GAK or the PRKC/PKC family member PRKCD inhibits the DFP-induced red puncta formation, suggesting that these two proteins positively regulate DFP-induced mitophagy.

Because these two lipid-binding proteins are serine-threonine protein kinases, the authors first investigated whether their kinase activities are required for DFP-induced mitophagy regulation. Both GAK inhibitor SGC-GAK-1 (GAKi) and a PRKC inhibitor, sotrastaurin (hereafter termed PKCi), impair the DFP-induced red puncta formation in IMLS cells and inhibit loss of mitochondrial proteins. However, kinase activity inhibition does not affect the red puncta formation when PRKN-overexpressing IMLS cells are treated with CCCP. Additionally, GAKi or PKCi treatment does not lead to a significant change in LC3B-II protein level or puncta number after starvation. These results indicate that the kinase activities of GAK and PRKCD are only required for an efficient PRKN-independent mitophagy, but not for PRKN-dependent mitophagy or starvation-induced nonselective autophagy.

To determine how these two proteins regulate mitophagy, the authors first checked the HIF1A pathway and found that the expression level and phosphorylation status of BNIP3 and BNIP3L following DFP treatment are not altered by GAKi or PKCi. Next, the authors focused on the recruitment of core autophagy machinery components. When mitophagy is induced by DFP, early autophagy proteins such as ATG13, ULK1, LC3B and WIPI2 are recruited to mitochondria. Inhibiting PRKC kinase activity results in fewer ATG13 and ULK1 puncta, suggesting that PRKCD is required for the formation of the ULK1 complex. In contrast, blocking GAK kinase activity does not affect the puncta number of early ATG proteins; however, less LC3 signal colocalizes with mitochondria in cells treated with GAKi, which implies deficient cargo loading to the phagophore. The authors further found that the cells treated with GAKi have stacked mitochondrial layers. This morphological alteration, however, may not be the cause of the deficient cargo loading because artificial fragmentation of mitochondria via CCCP treatment cannot rescue the loss of mitophagy following GAKi treatment. In addition, large autolysosomes and an increasing number of lysosomal structures are found in the GAKi-treated cells, and in line with this, more TFEB is found the in the nucleus. However, because GAKi does not affect nonselective autophagy or PRKN-dependent mitophagy, the researchers suggest that the deficiency in DFP-induced mitophagy caused by GAKi treatment is more likely to result from inefficient cargo loading instead of the lysosomal defect.

The authors further investigate the roles of GAK and PRKCD in mitophagy in vivo, using C. elegans and D. rerio, respectively. In C. elegans expressing mtRosella GFP-DsRed in body-wall muscle cells, knocking down gakh-1, one of the two orthologs of GAK results in a significant decrease in DsRed-only puncta, indicating an impaired mitophagy. Zebrafish has two PRKCD orthologs, prkcda and prkcdb, and both have high expression in the hindbrain. In zebrafish expressing EGFP-mCherry-fused Cox8, the authors find fewer red puncta in the hindbrain in prkcd double-knockout (DKO) fish under both basal and DMOG-induced conditions. DKO fish also show a reduced swimming trend in the dark phase, which is possibly related to the motor neurons in the hindbrain as a consequence of impaired mitophagy. From the phenotypes in C. elegans and D. rerio, the researchers demonstrate that both GAK and PRKCD are important for the regulation of mitophagy in vivo.

To summarize, in this study, the authors determined that GAK and PRKCD, two lipid-binding kinases, positively regulate PRKN-independent mitophagy. However, questions remain regarding a more detailed mechanism. For instance, how GAK inhibition leads to abnormal mitochondria morphology and deficient cargo loading are not clear. Additionally, whether the lipid-binding domain of PRKCD is important for mitophagy regulation, as it is dispensable for mitochondrial localization, needs further investigation. Even though these and other questions remain, the discovery that these two proteins are not necessary for PINK1-PRKN-mediated mitophagy and starvation-induced autophagy suggests a fundamentally different machinery of PRKN-independent mitophagy. The role of these proteins in regulating basal mitophagy in C. elegans and D. rerio also highlight their conserved functions in vivo. More importantly, an increasing number of studies indicate the contribution of PRKN-independent mitophagy to neurodegenerative diseases [13,14]. Because GAK has been identified as a gene associated with Parkinson disease from genome-wide association studies [15], and knocking out the orthologs of PRKCD in zebrafish leads to a phenotype possibly related to motor neurons, the discovery of these two proteins may shed light on not only the understanding of the mechanism of PRKN-independent mitophagy but also its relation with neurodegenerative disorders.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Institutes of Health [GM131919].