A modified ubiquitin protein has been identified by three independent studies as the missing link in a cellular quality-control pathway that is implicated in Parkinson's disease. See Letter p.162
Parkinson's disease, a progressive neurodegenerative disorder, has long been hypothesized to be caused by defects in organelles called mitochondria, which power mammalian cells through the production of ATP molecules. An accumulation of dysfunctional mitochondria may lead not only to a cellular energy crisis, but also to excessive production of toxic by-products. Two enzymes implicated in Parkinson's disease, PINK1 and parkin1,2, are thought to be involved in the disposal of defective mitochondria, but how the two proteins interact has been unclear. A trio of studies (by Kane et al.3, writing in the Journal of Cell Biology; by Kazlauskaite et al.4, in the Biochemical Journal; and by Koyano et al.5, on page 162 of this issue) now report that phosphorylated ubiquitin protein is the link between PINK1 and parkin, providing insights into a complex system of parkin regulation.
Kinase enzymes such as PINK1 alter the behaviour of target proteins through the addition of phosphate groups, a process called phosphorylation. PINK1 is imported to mitochondria and, in healthy cells, undergoes rapid degradation6. However, if mitochondria are defective or damaged (for example by exposure to CCCP, a poison that blocks ATP production), PINK1 accumulates, becoming anchored to the outer mitochondrial membrane with its kinase domain exposed to the cytoplasm.
Damaged mitochondria also attract parkin, which is otherwise dispersed throughout the cytoplasm in healthy cells7. Parkin is a ubiquitin ligase, which adds ubiquitin proteins (either singly or in polyubiquitin chains) both to itself through autoubiquitination and to nearby target proteins. Ubiquitinated proteins can serve as a signal to the cell that a cellular compartment should be degraded, which in damaged mitochondria leads to their timely disposal7, a process known as mitophagy.
Mutations in either PINK1 or PARKIN that underlie rare familial forms of Parkinson's disease disrupt mitophagy, implicating this cellular pathway in Parkinson's disease7. Furthermore, PINK1 mutations impede the recruitment of parkin to damaged mitochondria, suggesting that the proteins act in a linear pathway. Consistent with a PINK1–parkin quality-control pathway, mutations in pink1 or parkin in fruit flies cause accumulation of defective mitochondria and cellular degeneration8,9.
Initial models proposed that PINK1 phosphorylates and so activates parkin in damaged mitochondria. Although direct phosphorylation of parkin by PINK1 has been documented10, this modification does not seem to be sufficient for full activation of parkin's ubiquitin-ligase activity3,4,5,10. In search of a functional connection between PINK1 and parkin, three groups undertook cell-wide protein analyses and biochemical studies, and found the missing link between the two — phosphorylated ubiquitin (phospho-ubiquitin).
Each study showed that, in cells in which PINK1 was activated by CCCP treatment, PINK1 phosphorylates ubiquitin at a serine amino-acid residue (serine 65). Strikingly, a corresponding serine-65 residue in a ubiquitin-like domain is the aforementioned target of PINK1 phosphorylation on parkin10. Subsequent analyses by all three groups demonstrated that modified ubiquitin, in turn, induces parkin activity (Fig. 1).
Koyano and co-workers found that modified ubiquitin alone could not fully activate parkin — complete activation required coincident modification of parkin's ubiquitin-like domain as well as of ubiquitin, each at their respective serine-65 residues. A unique aspect of this group's work is their use of a strain of yeast that harbours a mutant form of ubiquitin lacking the serine-65 residue, which cannot be phosphorylated by PINK1. When the authors added human PINK1 and parkin to these cells, they found that parkin was not activated, underscoring the idea of an ordered pathway for mitophagy.
Whereas all three studies implicate phosphorylated ubiquitin as an intermediary in the PINK1–parkin pathway, the role of direct phosphorylation of parkin by PINK1 seems more complex. Koyano and colleagues report that modification of both ubiquitin and parkin at serine-65 is necessary for full activation of parkin in cells. But Kane and colleagues found evidence that modification of ubiquitin alone can activate parkin. This discrepancy is likely to relate to the distinct assays used in the studies, rather than to a biological difference.
Consistent with phospho-ubiquitin's activating role, Kane et al. and Koyano et al. found that it binds directly to parkin. Koyano and colleagues took the studies a step further, demonstrating that phospho-ubiquitin can still be used by parkin as a substrate for ubiquitination and autoubiquitination. But, surprisingly, the group found that parkin could be activated by phospho-ubiquitin that was mutated or modified such that it could not act directly as a substrate in ubiquitination. This implies that phospho-ubiquitin binds to and activates parkin separately from its role as a substrate.
Clues as to how this could be achieved might be gleaned from recent crystallographic analyses of parkin11,12. A phospho-peptide binding pocket has been proposed11 to lie within an inhibitory domain in parkin that, when the protein is inactive, occludes access to its catalytic active site. Kazlauskaite et al. speculate that the active site of parkin could be exposed by conformational changes brought about by the binding of phospho-ubiquitin's phosphate group to this inhibitory domain.
Kane and co-workers' data point to another role for phospho-ubiquitin — recruiting parkin to the outer membrane of damaged mitochondria. A particularly interesting idea is that such recruitment may generate a positive feedback loop, in which recruited parkin would be predicted to ligate additional phospho-ubiquitin to nearby proteins, attracting yet more parkin.
A subset of known parkin substrates, including the proteins mitofusin 2 and Miro, regulate mitochondria13,14, and their ubiquitination by parkin may be required for normal mitophagy. It will be important to determine whether activation by phospho-ubiquitin affects parkin's target selection, the fate of ubiquitinated target proteins, or the structure of polyubiquitin chains formed on targets. Finally, drugs that mimic the effects of phospho-ubiquitin may be candidate therapeutics for inherited and sporadic forms of Parkinson's disease.
Valente, E. M. et al. Science 304, 1158–1160 (2004).
Kitada, T. et al. Nature 392, 605–608 (1998).
Kane, L. A. et al. J. Cell Biol. 205, 143–153 (2014).
Kazlauskaite, A. et al. Biochem. J. 460, 127–139 (2014).
Koyano, F. et al. Nature 510, 162–166 (2014).
Jin, S. M. et al. J. Cell Biol. 191, 933–942 (2010).
Narendra, D. P. et al. PLoS Biol. 8, e1000298 (2010).
Clark, I. E. et al. Nature 441, 1162–1166 (2006).
Park, J. et al. Nature 441, 1157–1161 (2006).
Shiba-Fukushima, K. et al. Sci. Rep. 2, 1002 (2012).
Wauer, T. & Komander, D. EMBO J. 32, 2099–2112 (2013).
Trempe, J. F. et al. Science 340, 1451–1455 (2013).
Poole, A. C. et al. Proc. Natl Acad. Sci. USA 105, 1638–1643 (2008).
Ziviani, E., Tao, R. N. & Whitworth, A. J. Proc. Natl Acad. Sci. USA 107, 5018–5023 (2010).
About this article
δ-Opioid Receptor Activation Attenuates Hypoxia/MPP+-Induced Downregulation of PINK1: a Novel Mechanism of Neuroprotection Against Parkinsonian Injury
Molecular Neurobiology (2019)
Neurobiology of Disease (2015)
Molecular mechanisms underlying PINK1 and Parkin catalyzed ubiquitylation of substrates on damaged mitochondria
Biochimica et Biophysica Acta (BBA) - Molecular Cell Research (2015)
Biosensors and Bioelectronics (2015)