The Role of Parkin in Familial and Sporadic Parkinson’s Disease

Abstract

Mutations in Parkin are the second most common known cause of Parkinson’s disease (PD). Parkin is an ubiquitin E3 ligase that monoubiquitinates and polyubiquitinates proteins to regulate a variety of cellular processes. Loss of parkin’s E3 ligase activity is thought to play a pathogenic role in both inherited and sporadic PD. Here we review parkin biology and pathobiology and its role in the pathogenesis of PD.

Introduction

Familial Parkinson’s disease (PD) with specific genetic defects may account for fewer then 10 percent of all cases of PD¹, however, the identification of these rare genes and their functions has provided tremendous insight into the pathogenesis of PD and opened up new areas of investigation^2–7 (Table 1). Five genes have been clearly linked to PD, and a number of other genes or genetic linkages have been identified that may cause PD. The first “PD-gene” (PARK1) was the gene encoding the presynaptic protein, alpha-synuclein^{8, 9}. The second “PD-gene” (PARK2) is caused by mutations in the gene for parkin¹⁰, and it leads to autosomal recessive PD (AR-PD) and is the subject of this review. The third “PD-Gene” (PARK7) results from mutations in DJ-1¹¹. The fourth “PD-Gene” (PARK6) results from mutations in PTEN Kinase 1 (PINK1)¹². The fifth “PD-Gene” (PARK8) is due to mutations in LRRK2^{13, 14}. Mutations in alpha-synuclein, parkin, DJ-1, PINK1, LRRK2 definitely cause PD. The identification of the genes for PARK1 (α-synuclein), PARK2 (parkin), PARK7 (DJ-1), PARK6 (PINK1) and PARK8 (LRRK2) has led to new insights and direction in PD research and pathogenesis.

Table 1.

Loci and genes linked to familial PD

Locus	Chromosomal Location	Gene	Mode of Inheritance
PARK1 / PARK4	4q21.3	α-synuclein	Autosomal Dominant
PARK2	6q25.2-27	Parkin	Autosomal Recessive
PARK3	2p13	Unknown	Autosomal Dominant
PARK5	4p14	UchL1	Autosomal Dominant
PARK 6	1p35-p36	PINK1	Autosomal Recessive
PARK7	1p36	DJ-1	Autosomal Recessive
PARK8	12p11q13.1	LRRK2/Dardarin	Autosomal Dominant
PARK9	1p36	ATP13A2	Autosomal Recessive (Kufer-Rakeb Syndrome)
PARK10	1p32	Unknown	Late-Onset Susceptibility Gene
PARK11	2q36-37	Unknown	Late-Onset Susceptibility Gene
PARK12	Xq21-q25	Unknown	X-Linked
PARK13	2p13.1	Omi/HtrA2	Autosomal Dominant

Parkin

Parkin belongs to a family of proteins with conserved ubiquitin-like domain (UBL) and really interesting new gene (RING) finger motifs¹⁵. Mutations in parkin cause autosomal recessive Parkinson’s disease (AR-PD)¹⁰. Mutations in parkin are the most common cause of AR-PD^{16, 17}. In addition, mutations in parkin may play a role in sporadic PD^16–18. In the limited neuropathologic studies of patients with parkin mutations, there is a selective loss of dopaminergic neurons of the substantia nigra and loss of noradrenergic neurons in the locus coeruleus with accompanying gliosis¹⁹. There a few cases with α-synuclein positive inclusions or Lewy pathology, the hallmark feature of PD^20–22, and others that lack of α-synuclein positive inclusions¹⁹. A few cases also show Tau-positive neurofibrillary tangles^{23, 24}. The relationship of parkin to α-synuclein pathology requires additional study as it is not clear from the human post-mortem studies published to date whether parkin and α-synuclein participate in the same pathogenic pathway. However, clinical studies would suggest that PD due to parkin and α-synuclein mutations are distinct clinical diseases, albeit with PD symptomatology¹⁷.

Parkin and Ubiquitination

Parkin functions as an ubiquitin E3 protein-ligase (Figure 1)^25–27. Parkin contains two RING finger domains separated by an in-between RING domain IBR. Ubiquitin is a 76-amino-acid protein produced from a number of ubiquitin precursor proteins encoded in the human genome. It is covalently attached to the lysine residue of substrate proteins in a process called ubiquitination. The process of ubiquitination occurs through the transfer of an ubiquitin molecule from an activated E1-enzyme to the conjugating E2 enzyme, where an E3-ligase catalyzes the transfer of the ubiquitin molecule from the E2 enzyme to a target substrate²⁸. The E3-enzyme usually confers substrate specificity and acts as a scaffold to facilitate the stoichiometric requirements of the covalent attachment of ubiquitin. The ubiquitin chain elongation factors, otherwise known as E4s, can catalyze the multi-ubiquitination of proteins bound to the E2–E3 complexes. The ubquitination reaction may end with the attachment of a single ubiquitin molecule, a process called mono-ubiquitination, or the attachment of an ubiquitin molecule to several lysine residues in the target protein (multiple mono-ubiquitination). E3-ligase proteins can also promote the attachment of ubiquitin molecules to lysine residues in the ubiquitin molecule already attached to a target substrate including on lysine residues 48 or 63 in the ubiquitin molecule to form a chain of ubiquitin molecules. A chain of at least four lysine-48 linkages act as a signal for proteasomal degradation. Monoubiquitination and lysine-63 chains tend to function in non-degradative signaling roles. Parkin appears to perform monoubiquitination and polyubiquitination with either lysine-48 or lysine-63 linkages.

Figure 1.

Parkin is multifunctional ubiquitin E3 ligase. Parkin functions in the ubiquitin proteasome system as an E3 ligase. Ubiquitination requires the E1 activating enzyme and the E2 conjugating enzyme. Parkin utilizes a variety of E2s including UBCH7, UBCH8 and UbcH13/Uev1. Parkin utilizes a variety of linkages including monoubiquitination and polyubiquitanation via lysine-48 and lysine-63 chains.

Monoubiquitination by parkin may under certain circumstances be involved with receptor turnover²⁹. Parkin mediated lysine-48 linkages are involved with protein degradation and parkin mediated lysine-63 linkages with protein inclusions^{30, 31}. Parkin appears to use both UbcH7 and UbcH8 as its E2^{25, 26}. Parkin also utilizes the endoplasmic reticulum (ER)-associated E2’s Ubc6 and Ubc7³². Additionally, parkin interacts with the E2 complex UbcH13/Uev1 in mediating lysine-63-linked polymerization of ubiquitin³³. The function and type of ubiquitin modification that parkin mediates is probably largely defined on the cellular context and the ubiquitin machinery that parkin utilizes. Purified, in vitro parkin ubiquitination reactions suggest that parkin mediates primarily mono-ubiquitination reactions^{34, 35}. However, addition of the chaperone-dependent ubiquitin ligase CHIP (COOH terminus of heat shock protein 70-interacting protein), allows parkin to poly-ubiquitinate^{34, 35} and it is likely that other E4-like factors cooperate with parkin in vivo. Thus, parkin is a multifunctional E3-ligase, which has the capable of performing a variety of ubiquitin linkages and cellular functions.

Parkin and PD

Disease causing mutations in parkin range from single base pair substitutions to small deletions and splice site mutations, to deletions that span hundreds of thousands of nucleotides^{36, 37}. The general view is that parkin-related PD arises from similar mechanisms. Along these lines, the simplest explanation is that parkin mutations serve to lead to a loss of parkin function. Parkin-linked PD where deletions span several exons is certainly consistent with a loss of parkin function. Nonsense-mediated decay would serve to destabilize any truncated transcripts that might be expressed leading to the absence of protein expression. Indeed, there is little evidence that truncated parkin protein is expressed in patients with exon deletions (Reviewed in West, Dawson and Dawson ³⁸).

Many missense mutations also appear to lead to a loss of parkin function through decreased catalytic activity, aberrant ubiquitination and impairment of proteasomal degradation and/or destabilization of parkin leading to insolubility or rapid proteasomal degradation of mutant parkin^{34, 35, 39, 40}. Thus, the general view is that disease-causing mutations in parkin lead to a loss of parkin function, albeit through different mechanisms.

Parkin may play a role in sporadic PD through common and frequent mutations^{16, 18}. In addition, it is inactivated due to nitrosative stress ^{41, 42}, dopaminergic stress⁴³ and oxidative stress^{44, 45}, which are key pathogenic processes in sporadic PD. Thus, loss of parkin E3-ligase activity may not only play a role in AR-PD, but sporadic PD as well (Figure 2).

Figure 2.

Parkin inactivation plays a role in both sporadic Parkinson’s disease (PD)and in patients with parkin mutations. Dopaminergic (DA), nitrosative (nitric oxide – NO), oxidative (reactive oxygen species – ROS) and MPTP intoxication can inactive parkin abrogating its ubiquitination and cytoprotective properties. In sporadic PD, parkin is inactivated through nitrosative and dopaminergic stress and autosomal recessive PD it is inactivated through a variety of mutations. The loss of parkin E3 ligase activity leads to the accumulation of AIMP2 and FBP1, which causes neurodegeneration through mechanisms that require further clarification.

Parkin Substrates

A number of parkin substrates have been identified and were recently reviewed by West, Dawson and Dawson³⁸. “CDCrel-1 was the first parkin substrate identified. It belongs to a family of GTPases called septins and is robustly expressed in the nervous system where it associates with synaptic vesicles²⁶. Adeno-associated viral mediated transduction of CDCrel-1 induces neurodegeneration⁴⁶. However, there is limited evidence that CDCrel-1 accumulates in the absence of parkin and that parkin modulates CDCrel-1 levels in vivo⁴⁷.

Parkin-associated endothelin receptor-like receptor (Pael-R) in another putative parkin substrate. It is a G-protein-coupled transmembrane protein with homology to the endothelin receptor type B³². Pael-R is primarily expressed in oligodendrocytes, but it is present dopaminergic neurons. Pael-R overexpression induces the unfolded stress response in cultured cells and becomes insoluble. Parkin attenuates the formation of insoluble Pael-R and its accompanying toxicity presumably through an ubiquitination dependent mechanism. Pan-neuronal Human Pael-R overexpression in Drosophila causes age-dependent selective degeneration of dopaminergic neurons lending some credibility to neurodegenerative specificity ⁴⁸. However, limited evidence suggests that parkin is indeed a native physiological factor responsible for regulating levels of Pael-R.

The alpha-synuclein interacting protein, synphilin-1, interacts with and is ubiquitinated by parkin leading to the formation of protein aggregates when over-expressed with alpha-synuclein in cell culture⁴⁹. Parkin preferentially mediates the formation of lysine-63 linked polyubiquitin chains onto synphilin-1³⁰. Recent studies suggest that lysine-63 mediated ubiquitination may participate in the degradation of inclusions by serving as signal for autophagic cargo when the ubiquitin proteasome system is dysfunctional ^{50, 51}. Parkin mediated lysine-63 ubiquitination may be play an important role in this process⁵². Thus, parkin may play a specialized role in inclusion formation and targeting proteins for autophagic clearance when the ubiquitin-proteasome system is dysfunctional.

Aminoacyl-tRNA synthetase interacting multifunctional protein type 2 (AIMP2), also named p38/Jtv1 was originally identified as a parkin substrate through a yeast two-hybrid screen⁵³. AIMP2 is present in Lewy bodies. Parkin can promote the degradation of over-expressed AIMP2 presumably via polyubiquitination and proteasome degradation. Viral overexpression of AIMP2 leads to selective degeneration of dopaminergic neurons and it accumulates in parkin-null mice and in patients with parkin mutations⁵⁴. Moreover, consistent with the notion that parkin is inactivated in sporadic PD, is the observation that AIMP2 also accumulates in the brains of sporadic PD. In a similar manner, the far up stream element binding protein 1 (FBP-1) accumulates in parkin knockout mice, patients with AR-PD due to parkin mutations, sporadic PD as well as the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) mouse model of Parkinson’s disease⁵⁵. Parkin is inactivated in the MPTP model by S-nitrosylation^{41, 42}. Thus, parkin substrates, such as AIMP2 and FBP-1 that are polyubiquitinated via lysine-48 chains should not only accumulate in parkin knockout mice and patients with parkin mutations, but also under conditions where parkin is inactivated such as MPTP-intoxication or sporadic PD. We propose that parkin substrates should fulfill at least these 4 criteria to be designated a true parkin substrates that are regulated by parkin-mediated ubquitination and the ubiquitin proteasomal system.

Yeast-two hybrid experiments followed by confirmation with co-immunoprecipitation and in vitro ubiquitination experiments identified a variety of other parkin substrates. Parkin interacts with α/β tubulin heterodimers and microtubules and acts to stabilize microtubule formation, potentially in an ubiquitin dependent manner⁵⁶. In vitro steady-state levels of synaptotagmin XI are decreased in the presence of parkin, and protein aggregates in PD were found immunoreactive for synaptotagmin XI⁵⁷. SEPT5_v2/CDCrel-2, another member of the septin family and a close homolog to CDCrel-1, has been reported as a parkin substrate and may accumulate in disease brains⁵⁸. In another study, parkin was found to interact with cyclin E in the context of a protein complex including hSel-10 and Cullin-1, and found to prevent the accumulation of cyclin E in kainate acid treated neurons⁵⁹. Parkin also binds to the RanBP2 protein in over-expression cell culture models and apparently influences the downstream ability of exogenous RanBP2 to sumoylate the HDAC4 protein due to ubiquitination of RanBP2 via parkin⁶⁰. None of the latter substrates fulfill the criteria for a “true” parkin substrate.

A number of other functions have been attributed to parkin. Parkin monoubiquitinates HSP70, but the physiologic importance of this modification is not known ⁶¹. Parkin polyubiquitinates misfolded DJ-1 via lysine-63 chains in overexpression studies and targets misfolded DJ-1 to agggresomes via binding to HDAC6 ⁵². Parkin-mediated monoubiquitination of the PDZ protein PICK1 regulates the activity of acid-sensing ion channels ⁶². Parkin reduces the cofilin-phosphorylation of LIM kinase-1 through ubiquitination ⁶³. Both wild type and mutant ataxin-2 seems to be a substrate for parkin and ataxin-2 toxicity is attenuated by parkin overexpression ⁶⁴. Parkin may also play a role in EGF receptor trafficking and PI(3) kinase signaling through interactions with the UIM protein, Eps15 ²⁹. Other parkin interactors and putative substrates have been identified ³⁸ and their role in parkin-mediated PD is not clear.

Parkin and Neuroprotection

Parkin acts as a multipurpose protective agent when over expressed in a variety of stressful paradigms (reviewed by West, Dawson and Dawson ³⁸). Parkin overexpression prevents mitochondrial swelling in PC-12 cells treated with ceramide or subjected to serum withdrawal ⁶⁵. Kainic acid excitotoxicity is attenuated by parkin over-expression in neurons ⁵⁹. Manganese-induced cell death is reduced by parkin overexpression ⁶⁶ and parkin protects against dopaminergic toxicity ⁶⁷. The exact mechanisms of how parkin overexpression protects against a variety of toxic insults is not known, but it seems to be dependent on its E3 ligase activity. Dopaminergic cell death was comparable between wild type and parkin null mice following MPTP or 6-OHDA intoxication ^{68, 69}, thereby suggesting that cell line models of parkin overexpression may not recapitulate in vivo experiments.” Moreover, expression of parkin may provide a non-physiologic protection to a variety of stressors, but endogenous levels of parkin do not participate in neuronal survival to these various stressors.

α-Synuclein toxicity in rat, drosophila, and in cellular models is reduced by parkin overexpression ^{48, 70, 71}. Parkin and α-synuclein fail to interact and cannot bind one another in most assays ^{49, 72}. There is one report that the interaction of parkin with alpha-synuclein requires post-translation modification, but this has not been replicated ²⁵. Thus, how parkin overexpression prevents α-synuclein toxicity is unclear. Indeed similar to exogenous stressors, parkin may be protecting against α-synuclein toxicity non-specifically. The toxicity and phenotype associated with mutant α-synuclein was not affected by the loss of parkin in a genetic cross between parkin-knockout mice and α-synuclein overexpressing mice ⁷³. Little if any biochemical evidence suggests that loss of parkin expression influences overexpressed α-synuclein toxicity as might be assumed from studies employing the reverse context, namely overexpressing both parkin and α-synuclein. The implication is that parkin may acquire novel (perhaps non-specific) attributes when expressed at non-physiological concentrations (reviewed by West, Dawson and Dawson ³⁸). Additional studies concerning endogenous parkin are required to understand its largely undefined role in protection against a variety of stressors including α-synuclein. More importantly is the focus on understanding how endogenous parkin functions and how it regulates the survival of dopamine neurons.

The diverse array of parkin substrates and its broad neuroprotective properties have hindered the generation of a consensus in the field on parkin’s physiologic function and pathologic role in PD. The majority of substrates are understood only by a limited number of experiments that, in general, fail to determine the effects of ubiquitination on the function of the host protein and whether the interaction has physiological relevance. The mouse parkin knockout models have not demonstrated a robust up-regulation of any protein as evident by several proteomic screens ^{47, 74}. It is difficult to reconcile a common biochemical pathway among the interacting substrates and there is no clear pre-existing genetic or biochemical data that might elevate a particular substrate to a more important status with the possible exception of AIMP2 and FBP-1 as they accumulate in several in vivo models of parkin dysfunction (reviewed by West, Dawson and Dawson ³⁸).

Parkin and Mitochondrial Function

Clues to the key determinant of parkin-mediated pathology may come from recent studies in drosophila. The absence of parkin in drosophila leads to mitochondrial pathology and apoptotic muscle degeneration and raises the possibility that similar mitochondrial impairment triggers the selective cell loss observed in AR-PD ^{75, 76}. Despite having only having mild deficits, parkin knockout mice have features of mitochondrial dysfunction and oxidative damage ⁷⁴ and parkin-deficient patients have decreased lymphocyte mitochondrial complex I activity ⁷⁷ providing further support to the notion that loss of parkin function leads to mitochondrial deficits. How parkin might regulate mitochondrial function is not known. A small fraction of parkin may reside at or near the mitochondria ⁷⁸, suggesting that parkin might regulate a mitochondrial protein that is important for mitochondrial function. However, the suitability for parkin antibodies to detect endogenous parkin raises questions about the mitochondrial localization of parkin ⁷⁹. Parkin also seems to enhance mitochondrial biogenesis through as yet unconfirmed mechanisms ⁷⁸. Further clues come from additional studies in drosophila. Loss of drosophila PINK1 also leads to defects in mitochondrial function resulting in male sterility, apoptotic muscle degeneration and minor loss of DA neurons mirroring the loss of drosophila parkin phenotype ^{80, 81}. The loss-of-function PINK1 phenotype is rescued by overexpression of parkin, but the loss-of-function parkin phenotype is not rescued by PINK1 suggesting that PINK1 and parkin, at least in part, function in the same pathway and that PINK1 functions upstream ⁸². PINK1 deficiency in human cells also results in mitochondrial abnormalities, which is ameliorated by enhanced expression of parkin ⁸³. The mechanism by which both parkin and PINK1 regulate mitochondrial function and integrity is not known, but the enlarged and swollen mitochondria in parkin and PINK1 deficient drosophila suggests that they regulate mitochondrial morphology ⁸⁴. It is conceivable that parkin regulates the stead-state level of a protein critical for maintaining mitochondrial function. We posit that this putative parkin substrate should accumulate in models of parkin inactivation such as Parkin knockouts and the MPTP intoxication model as in AR-PD due to parkin mutations and in sporadic PD. Moreover, PINK1 should regulate its interaction or ubiquitination by parkin. Future studies are required to identify this missing link.

Conclusions

Parkin is an ubiquitin E3 ligase that plays an important role in the pathogenesis of PD. Not only does parkin play a role in AR-PD, it seems to play important roles in the pathogenesis of sporadic PD as it is inactivated in sporadic PD due to nitrosative, oxidative and dopaminergic stress. Parkin is multi-functional E3 ligase that is capable of different ubiquitin modifications including monoubiquitination and polyubiquitination via lysine-48 or lysine-63 chains. True parkin substrates that are regulated by the ubiquitin proteasome system should accumulate in both AR-PD and sporadic PD as well as animal models of parkin inactivation such as parkin knockouts and the MPTP-intoxication model. AIMP2 and FBP-1 fulfill the criteria for true parkin substrates suggesting that they may play important roles in parkin-mediated PD. Parkin appears to be a multi-functional neuroprotective protein when overexpressed, but whether it plays such a broad protective role when expressed at endogenous levels seems unclear. Finally, recent studies suggest that parkin may play important roles in mitochondrial function in a common genetic pathway that is shared by PINK1. Understanding the relationship of parkin to mitochondrial function, it relationship to PINK1 and the role of true parkin substrates in these processes will lead to a greater understanding of the normal physiologic role of parkin and its role in the pathogenesis of PD.