Mutations in the PARKIN gene (chromosome 6q25–27) were first identified in 1998 as a cause for autosomal-recessive juvenile parkinsonism (AR-JP) (Kitada et al. 1998). By now hundreds of PARKIN mutations including single amino acid changes and deletions/duplications have been identified and account for more than 50% of cases with familial recessive Parkinson’s disease (PD) (Corti et al. 2011). In addition to their causative role in PD, mutations in PARKIN have also been found in numerous cancer tissues and a tumor suppressive role has been described for the protein (Xu et al. 2014). PARKIN is located on chromosome 6 and is one of the largest genes in the human genome spanning more than 1.38 Mb. PARKIN orthologs are found in all animal species. Human PARKIN has 12 exons that encode for a 465 amino acid protein with an N-terminal ubiquitin (Ub)-like (UBL) domain and four C-terminal zinc finger motifs. The encoded protein was initially described as a classical really-interesting-new-gene (RING)-type E3 Ub ligase (Shimura et al. 2000) that facilitates covalent attachment of the small modifier Ub to lysine residues of substrate proteins. However, PARKIN has recently been recognized as a RING-in-between-RING (RBR)-type protein, a new class of E3 Ub ligases that utilize a distinct catalytic mechanism (Wenzel et al. 2011). Due to the numerous unrelated substrate proteins that have been proposed as a target for the E3 Ub ligase PARKIN, its functions have been linked to several subcompartments of the cell mostly related to stress response (Walden and Martinez-Torres 2012). In fact, PARKIN appears to be broadly cytoprotective and has been shown to defend cells and animal models against various environmental and genetic insults (Feany and Pallanck 2003).
In 2003, mitochondrial pathologies had been first described in a Drosophila knockout model that exhibited reduced lifespan, locomotor defects, and male sterility (Greene et al. 2003). While deletion of PARKIN in mice resulted only in subtle phenotypes, mitochondrial dysfunctions were observed as well, yet without overt pathological changes in ultrastructure (Kahle and Haass 2004). However, none of those knockout animals showed selective loss of dopaminergic neurons as seen in PD patients. In 2006, two studies showed that the mitochondrial localized PTEN-induced kinase 1 (PINK1), another gene mutated in recessive PD, and PARKIN genetically interact in a linear pathway, with PARKIN being downstream of PINK1 (Pallanck and Greenamyre 2006). While PARKIN was thought to be a cytosolic E3 Ub ligase with numerous unrelated substrates, in 2008 PARKIN was reported for the first time to translocate to specifically damaged mitochondria (Narendra et al. 2008). In 2010, PINK1 and PARKIN were then shown to physically associate and functionally cooperate in order to identify, label, and target damaged mitochondria to selective degradation via autophagy (mitophagy) (Geisler et al. 2010). This pathway not only linked two PD genes but also both major cellular dysfunctions implicated in disease pathogenesis (Springer and Kahle 2011). The last few years have provided enormous insights into the structure and function of PARKIN. PINK1 was found to phosphorylate the UBL domain of the E3 Ub ligase PARKIN and the small modifier protein Ub at a conserved serine-65 residue (Durcan and Fon 2015). Both phosphorylations fully activate PARKIN’s latent E3 Ub ligase activity through release of its autoinhibited conformation and recruitment to damaged mitochondria. By now, it has been well established that PINK1 and PARKIN mediate a multifaceted mitochondrial quality control to regulate organelle morphology and transport, function and metabolism, as well as biogenesis and degradation (Roberts et al. 2016).
PARKIN and Human Disease
PARKIN in Parkinson’s Disease
PARKIN in Cancer
In addition to its neuroprotective functions, accumulating evidences not only from cell-based studies but also from animal models and from human patients suggest a tumor suppressive role for PARKIN (Xu et al. 2014). While the exact association between loss of PARKIN and cancer susceptibility remains unclear, alterations of its expression and functions are highly prevalent especially in glioma, lung, breast, colon, and ovarian cancer. Yet, it is important to note that the PARKIN gene is prone to mutations and deletions due to its very large size and the chromosomal fragile site FRA6E (6q26). Nevertheless, numerous mutations, deletions, and copy number variants as well as promotor hypermethylation, and aberrant mRNA and protein expression have been detected in various cancer cell lines and primary tumors. In contrast to familial PD where PARKIN mutations occur in the germline, the majority of PARKIN variants identified from numerous cancer tissues are somatic (Fig. 1a, bottom). Only very recently, germline mutations have been reported in patients with glioma, lung, and ovarian cancer. In addition, mutations in PARKIN were heterozygous in the majority of cases, and only few were homozygous. Together with a frequently observed underexpression, this may suggest a haploinsufficient mode for PARKIN in cancer. It is intriguing that mutations found in PD and in cancer can occur on the very same residue and sometimes even result in the same amino acid exchange (e.g., R275W and A398T). However, the structural and functional consequences of most mutations found in cancer tissues are not clear and must be determined. Likewise, the biological effects remain to be identified, as it is uncertain if mitochondrial quality control or other functions of PARKIN are affected in cancer.
(In-)Activation of PARKIN by Mutations and Posttranslational Modifications
Detailed structural and functional characterization of the individual missense mutations has already led to a better understanding of PARKIN regulation (Fiesel et al. 2015b). It is worth mentioning that while all pathogenic PD mutations result in loss of PARKIN function, the molecular mechanisms are quite distinct and can disrupt mitochondrial quality control at different steps of a complex, sequential process (Geisler et al. 2010). First, several mutations that affect structurally important residues including the Zn2+coordinating residues often result simply in misfolding and aggregation of Parkin. Second, mutations can interfere with its activation and translocation, disrupt E3 ubiquitin ligase activity through direct or indirect effects that prevent association with coenzymes, recognition/binding to substrates, or Ub transfer. Third, since PARKIN can autoubiquitylate itself and thereby induce its own degradation, mutations that constitutively enhance its activation/activity (i.e., through reduced autoinhibition) can also result in loss of its functions.
In addition to inactivation by mutations, several posttranslational modifications of PARKIN have been shown to result in decreased enzymatic activity and/or solubility of the protein (Fig. 1b, top). Most phosphorylations that have been identified so far are thought to negatively regulate PARKIN functions and are mediated by several different kinases including casein kinase 1 (CK1), protein kinase A (PKA), protein kinase C (PKC), cyclin-dependent kinase 5 (cdk5), and the nonreceptor tyrosine kinase c-Abl. Moreover, modifications of its many cysteine residues by oxidation, dopamine-adducts, and nitrosylation have been shown to inactivate PARKIN and have been detected upon stress conditions in cells and animals as well as in human postmortem brain samples of PD patients. Thus, PARKIN appears highly sensitive towards stress conditions that occur in cells and tissues with age and disease. In contrast, only very few modifications such as sulfhydration of some PARKIN cysteine residues have been reported to enhance its E3 Ub ligase activity. Most importantly, phosphorylation of serine-65 in the UBL domain by the upstream PD-related mitochondrial kinase PINK1 has been shown to trigger release of autoinhibition and is crucial to activate the enzymatic functions of PARKIN (Caulfield et al. 2014).
Enzymatic Functions of PARKIN
E3 Ubiquitin Ligase Mechanism of PARKIN
PARKIN has been initially identified as a classical RING-type E3 Ub ligase (Shimura et al. 2000). However, more recently PARKIN has been recognized as a member of the RBR-type family that forms a distinct class of E3 Ub ligases which operate by a homologous-to-the-E6-AP-carboxyl-terminus (HECT)/RING hybrid mechanism, utilizing features of both of these types of enzymes (Wenzel et al. 2011). While classical RING-type E3 ligases bind an Ub-charged E2 enzyme with their canonical RING finger domain, they do not physically receive the Ub moiety but catalyze the direct transfer from the E2 onto a lysine residue of a substrate protein. In this case, the respective E2 coenzyme directs lysine linkage between Ub moieties in the forming poly-Ub chain. In contrast, HECT-type E3 ligases form an intermediate Ub thioester on their active site cysteine (i.e. they get charged with Ub themselves), before final transfer to a substrate. In that case, poly-Ub chain formation and lysine linkages are solely dependent on the respective E3 enzyme. RBR-type Ub ligases such as PARKIN utilize a hybrid mechanism in which they bind the E2 coenzyme with their canonical RING finger (RING1), for transthiolation of their active center and final ubiquitylation of the substrate protein. In addition, RBR-type E3 ligases appear to feature an allosteric Ub-binding site in RING2 near their active site that may bind the Ub moiety of the charged E2 enzyme to facilitate closer proximity between the catalytic sites of E2 and E3 enzymes for efficient transthiolation.
Regulators, Cofactors, and Substrates of PARKIN
Since the initial discovery of PARKIN, numerous protein regulators, cofactors, and substrates with diverse biological functions have been described (Walden and Martinez-Torres 2012). Until today, the exact enzymatic functions and thus biological consequences remain poorly understood. PARKIN has been shown to catalyze the formation of various Ub modifications ranging from (multi-) monoubiquitylation to polyubiquitin chains with different lysine linkages and thus distinct topologies. These various ubiquitylations serve different degradative and nondegradative functions. In addition to protein substrates, PARKIN can autoubiquitylate itself and thereby regulate its own levels and functions. PARKIN can operate as a single molecule or as part of a multiprotein E3 Ub ligase complex. However, it is important to note that many of the earlier studies were based on (co-)overexpression of PARKIN, Ub variants, cofactors, and substrates.
During mitophagy, PARKIN is regulated by several distinct E2 enzymes that either positively or negatively affect its enzymatic functions (Fiesel et al. 2014). For instance, members of the UBE2D family (UbcH5) as well as UBE2L3 (UbcH7) can charge the catalytic center of PARKIN with Ub and thereby regulate activation and translocation to damaged mitochondria. Others such as the K63-linkage specific UBE2N complex (UbcH13/Uev1A) appear to regulate downstream functions in a sequential process, as they were involved in mitochondrial clustering that precedes autophagic clearance. In addition, UBE2R1 was shown to negatively regulate PARKIN upstream of its translocation to damaged mitochondria. PARKIN enzyme functions and mitochondrial quality control are also regulated by deubiquitylation enzymes (DUBs) that remove the conjugated Ub from substrate proteins (Durcan and Fon 2015). For instance, while ataxin-3 binds and directs the E2 coenzyme UbcH7 away from PARKIN thereby impeding its enzymatic activity and autoubiquitylation, USP8 removes inhibitory K6-linked Ub chains directly from the UBL domain of PARKIN and thus has a positive regulatory effect. In contrast, USP15 and USP30 were both shown to deubiquitylate PARKIN substrates during mitophagy, thereby opposing its enzymatic activity.
It is noteworthy that different types of poly-Ub chains linked through K6, K11, K27, K48, or K63 of Ub have been identified during mitophagy as have been hundreds of mitochondrial and cytosolic substrate proteins (Sarraf et al. 2013). While the exact enzymatic complex required for formation and the biological roles of these different Ub linkages remain unclear, it is likely that their conjugation to individual substrate proteins may serve very specific functions. Structural diversity of substrates and the binding and activation by PINK1-phosphorylated Ub raise the possibility that once activated, PARKIN acts as a rather promiscuous E3 Ub ligase with limited substrate specificity. Alternatively, PARKIN may act as an Ub-chain elongating E4 enzyme that requires priming of substrates with (phospho-) Ub or may exert different catalytic activities as an E3/E4 enzyme combination. In this respect, it will be important to elucidate the exact enzymatic functions of PARKIN as autoubiquitylations, generation of free poly-Ub chains as well as of substrates, has been detected.
Activation of PARKIN
While the kinase PINK1 is constantly imported into healthy mitochondria where it is cleaved and degraded, PINK1 accumulates on the outer mitochondrial membrane upon mitochondrial damage. Here it dimerizes and gains enzymatic activity towards the conserved serine-65 residues of Ub and the UBL domain of PARKIN (Fig. 1b, bottom). PINK1-phosphorylated Ub binds to PARKIN at a phospho-binding site (PBS) in RING1 to activate its E3 Ub ligase functions and also serves as its receptor on damaged mitochondria. In addition, PINK1-mediated phosphorylation of PARKIN in the UBL domain enhances its enzymatic functions and both phosphorylation events are thought to fully activate the E3 Ub ligase activity of PARKIN (Durcan and Fon 2015). Both phospho-Ub binding (Fig. 2b) and phosphorylation of PARKIN itself (Fig. 2c) can induce major structural changes that result in release of autoinhibition and allow association of the Ub-charged E2 coenzyme, thereby unleashing the enzymatic activity of PARKIN.
Biological Functions of PARKIN
PARKIN has been linked to numerous other biological processes including, but not limited to, protein aggregation and clearance of unfolded/misfolded proteins, synaptic functions and neurotransmission, vesicular dynamics and endocytosis, DNA damage repair and cell cycle regulation, and cytoskeletal stability and cell survival signaling. Given the clear-cut genetic and functional links between PINK1 and PARKIN, research in the past few years has focused on their roles in mitophagy (mitochondrial degradation of damaged mitochondria by autophagy). By now it has become clear that PINK1 and PARKIN regulate many more aspects of mitochondrial biology, probably depending on the nature, level, and extent of mitochondrial damage. While clearance of terminally damaged whole organelles via the autophagy-lysosome system may be the last resort, specific roles for mitochondrial-membrane-associated degradation and Ub-proteasome-dependent clearance of individual mitochondrial proteins have been shown. For instance, degradation of Mitofusin 1 and 2 prevents fusion with healthy mitochondria, while degradation of Miro prevents their further transport, both of which ensures segregation of damaged organelles from the mitochondrial network. In addition, certain proteins of either of the two mitochondrial membranes and the matrix can be engulfed into specialized mitochondrial-derived vesicles (MDVs). While different cargos, trafficking routes and destinations of those are just emerging, MDVs provide additional means to selectively remove individual damaged components (Roberts et al. 2016).
However, so far the best-characterized biological function of PARKIN is linked to mitophagy. Once localized to damaged mitochondria and activated, PINK1 and PARKIN jointly catalyze the formation of phosphorylated poly-Ub chains on mitochondrial substrates in a feed-forward mechanisms. Mass spectrometry has shown that the numerous substrates are decorated with poly-Ub chains of various linkages and thus topologies (Sarraf et al. 2013). Since those are known to trigger the activation of different downstream pathways, it is likely that distinct enzymatic activities of PARKIN (and PINK1) regulate the diverse fates of individual mitochondrial components and/or whole organelles. Whether proteasomal degradation, formation of MDVs, or autophagic clearance, Ub labeled mitochondria have been shown to recruit numerous proteins such as HDAC6 and VCP/p97 as well as the adaptors p62/SQSTM1, NBR1, NDP52, OPTN, and TAXBP1 to decode the respective Ub signals. The kinase TBK1 has further been shown to phosphorylate those dual adaptor proteins to enhance their affinities towards Ub and the autophagosomal membrane protein LC3. In fact, phosphorylated poly-Ub chains (pSer65-Ub) have been described as the “mitophagy tag” that mediates autophagosomal engulfment of damaged mitochondria and end-stage clearance of entire organelles in lysosomes. pSer65-Ub appears to be a highly sensitive marker for mitochondrial damage and its (patho-)physiological role has recently been validated in cells, primary neurons, and in human postmortem brain (Fiesel et al. 2015a). As expected, p-Ser65-Ub signals partially colocalize with total Ub as well as mitochondrial and lysosomal markers in small cytoplasmic granules that increase with mitochondrial stress, disease, and age.
Homozygous or compound heterozygous loss of PARKIN function is unequivocally linked to familial, early-onset forms of PD. In addition, a role for heterozygous mutations and PARKIN inactivation in idiopathic, late-onset PD has been suggested. Loss of PARKIN results in sensitization whereas overexpression of PARKIN has been shown to provide neuroprotection against various environmental and genetic insults and even to extend lifespan of Drosophila. Furthermore, many somatic and few germline mutations are found in cancer tissues and cell lines and a tumor suppressive role for PARKIN is emerging. Loss of PARKIN increases susceptibility to tumorigenesis, promotes proliferation and tumor formation, whereas overexpression reduces growth of cancer cells in vitro and in vivo. While the exact pathways and mechanisms resulting in PD and/or cancer remain to be elucidated, PARKIN has been shown to act as a broadly cytoprotective E3 Ub ligase with widespread cellular functions. The best characterized enzymatic and biological functions of PARKIN have been linked to mitophagy. Together with PINK1, PARKIN orchestrates a complex quality control to regulate numerous aspects of mitochondrial biology including morphology and transport, function and metabolism, as well as biogenesis and degradation. Despite the selective degeneration of DA neurons in PD, PARKIN is widely expressed and dysfunctions in PARKIN and mitochondrial quality control have been linked to several human diseases affecting not only brain but also other tissues with high-energy demand including muscle, liver, and kidney. Though much remains to be learned, recent detailed understanding of PARKIN’s structure as well as enzymatic and biological functions has spurred efforts to develop novel biomarkers of mitochondrial dysfunction and small molecules that activate and/or enhance its E3 Ub ligase functions. Given that PARKIN (and PINK1)-mediated mitochondrial quality controls appears to be a fundamental, cell biological pathway, this stress response may provide new avenues to treat several age-related human maladies.
- Caulfield TR, Fiesel FC, Moussaud-Lamodiere EL, Dourado DF, Flores SC, Springer W. Phosphorylation by PINK1 releases the UBL domain and initializes the conformational opening of the E3 ubiquitin ligase Parkin. PLoS Comput Biol. 2014;10:e1003935. doi: 10.1371/journal.pcbi.1003935.PubMedPubMedCentralCrossRefGoogle Scholar