The enzyme parkin is known to promote disposal of organelles called mitochondria that have suffered damage. The identification of an enzyme that opposes parkin demonstrates how a delicate balance is maintained in the cell. See Article p.370
Cells have a love–hate relationship with mitochondria. As the power plants of cells, these organelles provide the energy required for life, but mitochondrial defects can lead to the production of reactive oxygen species that disrupt crucial cellular functions. Cells therefore use a specialized program, mitophagy, to eliminate damaged mitochondria and so maintain cellular health. Although a mitophagy signalling pathway comprised of two enzymes, PINK1 and parkin, has been identified, it is not clear what factors inhibit the pathway. In this issue, Bingol et al.1 (page 370) report that USP30, a deubiquitinating enzyme, puts the brakes on mitophagy.
In cells with healthy mitochondria, parkin is located in the cytoplasm and is thought to be inactive2,3, whereas PINK1 is associated with mitochondria. Activation of PINK1 in response to mitochondrial damage causes migration of parkin, a ubiquitin ligase, to the outer membrane of the mitochondrion, and its subsequent activation by PINK1 (refs 2, 4, 5). Activated parkin then transfers a small protein called ubiquitin to one or more lysine amino-acid residues on dozens of proteins bound to the mitochondrial outer membrane6,7. Following this ubiquitination process, the ubiquitin tags are recognized by the cell's mitophagy machinery2, leading to mitochondrial degradation. Defects in mitochondrial quality control, brought about by mutations in PINK1 and parkin are the cause of certain neurodegenerative disorders, such as some early-onset familial forms of Parkinson's disease2.
The pathways downstream of ubiquitination at the mitochondrial outer membrane are far from clear, but specific ubiquitinated targets and the total number of ubiquitin modifications on target proteins have been offered as possible factors in the recruitment of the mitophagy machinery to mitochondria2,7. Protein ubiquitination is a reversible modification — indeed, the human genome encodes more than 100 deubiquitinating enzymes. Bingol and colleagues therefore reasoned that inhibiting deubiquitination of parkin targets could be a tool to restore the balance of mitophagy in cells in which the PINK1–parkin pathway is defective (Fig. 1). The authors induced mitophagy experimentally in cells by triggering mitochondrial depolarization, which activates PINK1, and performed a screen for deubiquitinating enzymes that could prevent parkin-dependent mitophagy. This quest resulted in the identification of USP30.
The authors found that USP30 not only prevented the mitophagy machinery from recognizing damaged mitochondria, but also reversed the accumulation of ubiquitin on proteins bound to the mitochondrial outer membrane, indicating that USP30 directly opposes parkin function. In a cell-wide analysis of ubiquitin-tagged proteins, Bingol and co-workers identified 41 targets of parkin ubiquitination that could be deubiquitinated by USP30, including TOM20, a subunit of the mitochondrial translocase enzyme, which is responsible for transport of proteins across the outer mitochondrial membrane. Surprisingly, the 'classic' parkin target protein, mitofusin, was resistant to USP30-driven deubiquitination. An understanding of how USP30 selectively removes ubiquitin from some but not other parkin targets will require further work, but it is conceivable that the proteins that are not targeted by USP30 are simply those that are most efficiently ubiquitinated.
Although the PINK1–parkin pathway is known2 to promote mitophagy in response to chemical or genetic disruption in cell-culture samples, its role in neurons — which are critically affected in Parkinson's disease — has been controversial8. In an experimental tour de force, Bingol et al. examined USP30 and its role in mitophagy in cultured rat neurons and in fruit flies genetically engineered to model Parkinson's disease.
By tracking mitochondria in rat neurons as they underwent mitophagy in 'normal' situations (without the need to artificially activate the pathway), the authors demonstrated that levels of mitophagy were reduced by loss of PINK1 and parkin, and increased by depletion of USP30. Thus, USP30 opposes PINK1–parkin-dependent mitophagy in healthy neurons, in which defective mitochondria probably arise as a result of the oxidative stress that occurs during normal cellular function.
Fruit flies that model Parkinson's disease have defective mitochondria in flight-muscle cells, a reduced ability to climb and reduced levels of the neurotransmitter dopamine9,10. Bingol and colleagues found that these defects were largely reversed when USP30 was removed throughout the animal. To examine the dopamine-producing neurons that are affected by Parkinson's disease, the authors genetically deleted USP30 in these cells specifically, and treated the insects with paraquat, a mitochondrial toxin that elicits Parkinson's-disease-like symptoms in humans, and reduces dopamine levels in fruit flies. Depletion of USP30 in dopamine-producing neurons largely reversed dopamine loss and behavioural defects, and increased survival, indicating that USP30 might actively oppose the PINK1–parkin pathway in cell types affected by Parkinson's disease.
This study may have implications for the treatment of defective PINK1–parkin signalling in Parkinson's disease, and gives us a deeper understanding of this form of mitophagy in general. Inhibitors of USP30 might increase mitochondrial health under conditions in which this system is impaired, for example in patients with mutations in PINK1 or PARKIN. Indeed, Bingol and co-workers found that depletion of USP30 increased mitophagy in human cell lines altered to express a mutant form of PARKIN.
If this activity can be extended to neurons, it is possible that USP30 inhibitors could be beneficial to patients. The development of effective USP14 inhibitors11 suggests that selective targeting of this enzyme class is a possibility. USP30 inhibitors might also improve mitochondrial health in cells with other types of defect, by reducing the threshold for PINK1–parkin-dependent mitochondrial clearance.
Precisely why dopamine-producing neurons are more sensitive to familial mutations in the PINK1–parkin pathway than other cells in the body is unclear. This work raises the interesting possibility that relative levels of USP30 and parkin in various cell types could determine the sensitivity of cells to defects in this pathway. However, a recent report12 suggests that a different family member, USP15, can similarly reverse the PINK1–parkin pathway. Further studies will be needed to understand the relationship between these two pathways.
Finally, Bingol and colleagues provide evidence that ubiquitination of TOM20 is required for mitophagy, providing a new potential link between mitochondria and the mitophagy machinery. Exploring whether TOM20 ubiquitination promotes assembly of the mitophagy machinery on mitochondria may help us to understand what continues to be a central puzzle in the field: the mechanism by which ubiquitinated mitochondria are recognized by autophagosomes, the vesicles that transport damaged mitochondria to be degraded1. Now the race is on to determine whether releasing the parkin brake will benefit patients.