Epidemiological studies indicate that parkinsonism is the most prevalent movement disorder, manifesting multiple risk factors, and predominantly affecting the aged population. Pathogenesis is strongly associated with both genetic susceptibility and environmental factors. Current etiologic hypotheses concerning 'idiopathic' PD favor genetic susceptibility-by-environment (GxE) interactions [7, 10]. In the past decade, genetic studies have shown that relatively rare, inherited mutations cause familial forms of the disease; these studies have also provided important insights into the role of molecular networks in the development of hereditary as well as sporadic PD. Emerging evidence suggests that oxidative/nitrosative stress, possibly due to pesticide exposure , may serve as a primary event in PD pathogenesis for the more common sporadic or idiopathic form of PD. Recent studies have suggested that oxidative/nitrosative stress may interfere with normal function of the UPS in PD [5, 15, 33–35]. However, direct evidence for protein modification by free radicals resulting in DA neurodegeneration is limited . In the present study, our direct detection of sulfonated derivatives and subsequent aggregation of parkin in cell-based PD model, an increase in aggregated parkin in rats and primates exposed to mitochondrial complex I inhibitors and in PD brains, in conjunction with finding that parkin function is regulated by ROS, yields mechanistic insight into the chemical reactions of parkin under oxidative stress and their effect on UPS impairment. Conversely, UPS dysfunction has been suggested to cause oxidative/nitrosative stress . Thus, these two processes may develop into a vicious cycle that contributes to aberrant protein accumulation and neurodegeneration in PD.
Here we utilized an unbiased, "top-down" mass spectrometry approach combined with molecular and cell biology methods to elucidate the chemical nature of the posttranslational modifications (PTMs) of cysteine residues in parkin in response to environmental insults. With this approach, we provide evidence specifically linking these PTMs to parkin function. Initially, we discovered parkin sulfonation in vitro in response to exposure to pathological concentrations of H2O2, as well as in cells exposed to MPP+. We documented parkin sulfonation by immunoblot analysis using a newly developed antibody against sulfonated peptides, and found that sulfonated parkin was significantly increased in the "Insoluble" fraction of cells exposed to MPP+. Catalase significantly reduced sulfonated parkin to levels approximating normal conditions in control cells.
We documented that environmental insults, including the mitochondrial complex I inhibitors MPP+, rotenone, or other pesticides, increase levels of oxidative stress, at least in part, in the form of H2O2. We further found that pathological levels of H2O2 produce a vicious cycle of increased and then decreased E3 ligase activity. Under these conditions of oxidative stress, we also observed inclusion body-like aggregates of parkin by immunoreactivity in parkin-overexpressing SH-SY5Y cells. As in vivo confirmation of these findings, we observed an increase in the amount of aggregated parkin immunoreactivity in the striatum of rotenone-exposed rats and in the nigrostriatal system of MPTP-lesioned monkeys. hNSCs transplanted into MPTP-lesioned monkeys seem to reduce parkin aggregation in the host substantia nigra. This finding, in parallel with immunohistochemistry in the striatum of rotenone-exposed rats, suggests that hNSC transplants offer an opportunity to ameliorate parkin's ubiquitin ligase function by attenuating parkin aggregation, which would otherwise render the protein dysfunctional. Additionally, we found that parkin solubility was significantly decreased in human postmortem PD brains compared to control brains without CNS pathology. Moreover, we demonstrate that an increase in the level of sulfonated parkin correlates with the insolubility of parkin in human PD brains, suggesting pathophysiological relevance of parkin sulfonation in PD. However, there is a caveat in this finding in that we could not demonstrate direct evidence for parkin sulfonation by immunoprecipitation since none of the parkin antibodies available to us were suitable for immunprecipitation. There are at least three possible reasons preventing us from isolating parkin by immunoprecipitation: (1) We have only limited amounts of postmortem human PD brain to start with, especially from the dopamine fiber-rich regions of the corpus striatum (mainly caudate); (2) The amounts of "Soluble" parkin available for immunoprecipitation are relatively low compared to SDS or urea-dissolved "Insoluble" parkin; and (3) It is often observed that SDS or urea used on insoluble proteins may interfere with the ability of the antibodies to immunoprecipitate. Thus, this type of immunoprecipitation experiment is not feasible with current methods. Nonetheless, our findings on tissue lysates do show a significant correlation between increased sulfonation and insoluble parkin in human PD postmortem brains.
Using complementary MS strategies ─ high-resolution Q-TOF MS and high-throughput ion trap MS/MS ─ we mapped the PTMs of parkin under conditions of oxidative stress and encountered sulfination/sulfonation of specific cysteine residues. We found that the RING and IBR cysteine-rich domains manifested these oxidized modifications. These PTMs modulated parkin E3 ligase activity, affected ubiquitination-mediated protein degradation, and contributed to parkin aggregation. Prior mutagenesis experiments on parkin causing mutation-induced protein misfolding have demonstrated that Cys residues both within and outside of the RING-IBR-RING domain are important in maintaining protein solubility [27, 28, 38]. The two putative sites of S-nitrosylation on parkin, Cys241 and Cys260, which were first reported by our group  and others, match the predicted S-nitrosylation consensus motif, making them the most likely candidates for physiological modification. S-nitrosylation may also promote further oxidation reactions such as sulfonation, as observed in the present study.
With regard to related nitrosylation/oxidation reactions, S-nitrosylation of Prx interrupts the normal redox cycle of Prx in detoxifying ROS, and thus results in accumulation of cellular peroxides . Therefore, S-nitrosylation of Prx may also contribute to oxidative stress-induced neuronal cell death in PD. These previously published results from our group suggest that the mechanism of interplay between SNO-Prx and Prx-SO3H may be different from that of SNO-parkin and parkin-SO3H.
A variety of markers and indices in PD patients and animal models have suggested that derangements in mitochondria complex I activity and consequent oxidative/nitrosative stress are important contributors to sporadic PD . At least three mitochondrial complex I inhibitors, including MPTP, paraquat and rotenone, are capable of simulating many features of sporadic PD and provide valuable models for PD investigation [10, 40]. Impaired mitochondrial complex I leads to increased oxidative stress, free radical formation, and reduction in ATP formation, rendering neurons more vulnerable to glutamate-related excitotoxicity. ROS/RNS, such as endogenous H2O2 and NO, are implicated in the pathogenesis of PD. In fact, similar to ROS, NO can lead to secondary oxidative modification on parkin in vivo [21, 22], and pathological levels of RNS in combination with ROS can produce synergistic cytotoxic effects by irreversibly S-nitrosylating and then further oxidizing proteins as well as other cellular constituents [41, 42]. Increased oxidative stress contributes to a cascade leading to DA neuron degeneration predominantly in the pars compacta of the substantia nigra. The occurrence of oxidative stress in PD is supported by both postmortem analyses and studies demonstrating the capacity of oxidizing toxins to induce nigrostriatal degeneration [43, 44]. Previous reports have suggested that ROS/RNS can affect proteasomal function [19, 26, 33, 45, 46]. Indeed, recent studies by our laboratory and others have shown that nitrosative stress can modulate parkin E3 ubiquitin ligase activity and subsequently impair UPS function [21, 22]. Parkin may form a functional complex with PINK1 and DJ-1 , but dysfunctional ubiquitination in the face of oxidative/nitrosative stress results in the loss of the intrinsic neuroprotection mechanism, and thus mimics familial PD in the absence of mutation of one of the genes encoding these proteins [21, 22]. Moreover, recent work has shown that PINK1 recruits parkin to the outer membrane of impaired mitochondria, and parkin heralds mitophagy through its ubiquitination of outer mitochondrial membrane proteins . As ROS and RNS are closely related and interact with each other, our present results indicate that oxidative and nitrosative stress act similarly on parkin activity. Substantial evidence supports the notion that high levels of basal oxidative stress exist in the substantia nigra pars compacta in the normal brain and that these levels are increased in PD.
As we report in the present study, oxidative stress can lead to sulfonation of the cysteine residues of parkin, which can affect protein tertiary structure, decrease parkin solubility, and affect parkin E3 ligase activity. These changes may contribute to the etiology of sporadic PD. Parkin possesses 35 cysteine residues, corresponding to a cysteine content of 7.5% (human proteomic average is 2.3% ). In addition, at least six different cysteine mutants have been experimentally linked to parkin dysfunction (http://www.uniprot.org/uniprot/O60260). These findings suggest that cysteine residues in parkin are essential for its function, including protein folding and ubiquitination. Yet these cysteines also predispose this highly-expressed CNS protein to chemical modifications under nitrosative and oxidative stress, including S-nitrosylation and sulfonation. We speculate that such oxidative changes may result in structural changes in the protein similar to those produced by hereditary cysteine substitutions linked to parkinsonism, e.g., substitution of cysteine to Y(212), Y(253), stop(256), G(289), F(431), or R(441). In fact, we have found evidence for cysteine residues that are sulfonated in parkin are associated with aggregation of the protein. Hence, oxidation of structurally or functionally critical cysteine residues might represent a molecular point of convergence in the pathogenesis of PD, connecting hereditary mutations that affect parkin solubility and function with adverse environmental insults resulting in similarly detrimental oxidative modifications of parkin at a posttranslational level.