Protein sulfonation recognized in vitroby an antibody raised against per-sulfonyl-BSA

To investigate protein posttranslational modifications by oxidation, we generated a broad-spectrum polyclonal antibody (pAb) that could identify protein sulfonation by recognizing epitopes containing cysteine sulfonic acid [Cys(SO₃^-)]-modified residues (see Methods). Due to the extremely small size of the specificity-determining side chain of sulfonated cysteine (i.e., methyl sulfonate), a large amount of antigen containing multiple potential epitopes was deemed necessary to boost immunoreactivity for antibody generation. Potential epitopes of interest included the Cys(SO₃^-) side chains together with the adjacent peptide bonds, or pairs of adjacent sulfonated cysteines. Our strategy was to oxidize the 35 cysteine residues in BSA to generate per-sulfonyl-BSA, containing up to 35 peptide Cys(SO₃^-) groups in various protein sequence environments. The synthesis of per-sulfonyl-BSA was performed following a protocol described previously [23]. Using the antibody thus generated, we initially examined putative sulfonation of a number of cysteine-containing proteins after oxidative stress engendered by H₂O₂ in vitro. Dot blots with various amounts of BSA were used to examine the dose-dependence of oxidized proteins recognized by the sulfonation pAb (Figure 1A). Next, immunoblot analysis with the sulfonation pAb revealed that an increase in BSA sulfonation correlated to the concentration of H₂O₂ that BSA was exposed to (Figure 1B).

Validation of the sulfonation antibody by in vitrocompetition assays

To further assess the specificity of the sulfonation antibody in vitro, we examined another known cysteine-containing protein, protein-tyrosine phosphatase 1B (PTP1B), which has cysteine residues that are known to be oxidized [24]. We used H₂O₂ to induce sulfonation of PTP1B and then conducted a competition assay to test for the ability of structurally related amino acids to compete with the cysteic acid epitopes on PTP1B for binding to the anti-sulfonation pAb. We found that there was an approximately 50% increase in PTP sulfonation (PTP-SO₃H) after exposure to 200 μM H₂O₂ in vitro (Figure 1C and 1D). Free l-cysteic acid (Cys-SO₃H) competed virtually completely with PTP-SO₃H for antibody binding. In contrast, a structurally-related amino acid with negative charges, l-cysteine sulfinic acid (Cys-SO₂H), competed less well for anti-sulfonation pAb binding. Taken together, these results suggest that the sulfonation antibody was able to recognize oxidant-induced protein sulfonation in a variety of peptide environments, potentially representing a "pan-sulfonation antibody".

ROS-induced parkin sulfonation observed in vitro and in vivo

To investigate parkin oxidation, we used the sulfonation pAb that we had generated to initially examine parkin sulfonation in vitro after oxidative stress engendered by H₂O₂. To further ensure the specificity of the antibody in identifying sulfonated proteins, we employed another well-studied antibody, which recognizes solely sulfonated peroxiredoxin (Prx-SO₃H) [14], as a positive control (Figure 1E). Immunoblot analysis revealed that H₂O₂ increased parkin sulfonation in vitro in a dose-dependent manner (Figure 1F). Taken together, these results suggest that our "pan-sulfonation antibody" may recognize various sulfonated proteins because it was able to react with proteins containing multiple amino-acid sequences surrounding the sulfonated cysteine residue. To test for the presence and effect of sulfonated parkin in cell-based models of PD, we used either parkin-overexpressing SH-SY5Y cells exposed to toxic levels of the mitochondrial complex I inhibitor--MPP ⁺ , or primary cultures of striatal neurons exposed to the pesticide rotenone. We detected ROS generation in these models with two fluorogenic probes, the cell-permeable fluorogenic dye chloromethyl-2',7' dicholor-dihydrofluorescein diacetate (CM-H₂DCFDA, DCF) or hydroethidine (HEt; Invitrogen, San Diego, CA). Both cell-based models displayed significant increases in ROS, while pretreatment with cell-permeable catalase, a specific reductase of H₂O₂, reduced ROS production and protected dopaminergic neurons from death (Figure 2A-2C and Additional file 1, Figure S1A). We also quantified the increase in DCF fluorescence intensity following MPP⁺ exposure using a Synergy 4 microplate reader (BioTek, Winooski, VT) and obtained results (Additional file 1, Figure S1B) similar to those previously observed under fluorescence microscopy in a report linking ROS generation to MPP⁺ exposure [25]. Next, parkin-overexpressing SH-SY5Y cells exposed to MPP ⁺ were lysed into "Soluble" and "Insoluble" fractions. We enriched for recombinant myc-parkin with anti-myc antibody, and found evidence that exposure to MPP⁺ resulted in a significant increase in parkin sulfonation in the "Insoluble" fraction. Pretreatment 1 hour prior to MPP⁺ exposure with cell-permeable catalase attenuated sulfonation and aggregation of parkin into the insoluble fraction (Figure 2D and 2E). These findings suggest that oxidative stress induced by MPP⁺ results in parkin sulfonation and aggregation.

MS analysis of parkin oxidation

It is most important to link structure with function, that is, to identify the critical residues in parkin whose oxidative state regulates the E3 ligase activity. Previous studies of parkin function utilized a predominantly genetic approach and focused on various point mutations of parkin found in familial PD patients [17, 26–28]. Such a "bottom-up" approach limits the study to known mutations, and mutations of some of the cysteine-containing enzymes were more arbitrary and lacked genetic support. Therefore, we took an unbiased "top-down" approach by using liquid chromatography MS (LC/MS)-based comprehensive analysis to determine post-translational modifications of cysteine residues in parkin exposed to ROS.

Full-length recombinant human parkin was exposed to various concentrations of H₂O₂ in vitro, followed by in-solution and in-gel trypsin digestion. The resulting tryptically-digested peptides were subjected to two complementary LC/MS platforms - high resolution nanoLC quadrupole time-of-flight (Q-TOF) MS and high throughput nanoLC ion trap tandem MS (see MS workflow in Additional file 1, Figure S2), as we have described previously [19, 21, 29]. LC/MS data were then converted into a DeCyder™ MS-compatible format [30]. DeCyder™ MS allows for quantitative proteomics analysis to compare the retention time and m/z value of each identified peak observed in two different LC/MS runs. The MS data generated from either Q-TOF or ion trap were compared to confirm the similarity of their elution profiles. Differentially sulfinated/sulfonated peptides were detected by comparing sample intensity maps and peaks of interest (m/z values), selected as precursor ions for subsequent peptide identification and quantification.

We used DeCyder™ MS for multi-spectrum quantitative MS analysis and examined the stoichiometry of the cysteine modifications of parkin based on their chromatography retention time and relative intensity of the identified masses. Q-TOF MS data were displayed in DeCyder™ "2D gel" format with insets showing higher resolution of the parkin RING-I domain peptide, "SPVLVFQCNSR" (Figure 3A-3D). This tryptically-digested parkin peptide was further confirmed by ion trap tandem MS to contain a double-charged ion with a carbamidomethyl (CAM)-modified cysteine (representing alkylation of the free cysteine thiol group by iodioacetamide (IAM) prior to trypsin digestion) at m/z 654.7 (Figure 3G), or sulfonated (SO₃H) cysteine at m/z 649.4 (Figure 3H). The sulfinated/sulfonated cysteine-modified peptides in H₂O₂-exposed parkin were compared to the vehicle-treated control and calculated by the DeCyder™ MS PepMatch module. Such quantitative DeCyder™ MS analysis revealed that the sulfonated "SPVLVFQCNSR" peptide (Figure 3F) was not found in the spectra of the vehicle-treated control (Figure 3E). Overall, persistent sulfinated/sulfonated cysteine-containing peptides identified in the H₂O₂-exposed parkin by Q-TOF MS and ion trap MS/MS are listed in Table 1 and Table 2, respectively. There 63.6% of the sulfinated/sulfonated cysteine-containing peptides were only found in the H₂O₂-exposed parkin. Next, we compared the identified peptide oxidations (sulfination or sulfonation) of parkin with the sites of parkin mutations found in rare familial cases of PD (Figure 4). Almost three quarters (73%) of the sulfinated/sulfonated cysteines in parkin were identified in the RING and IBR (In-Between-RING) domains, indicating that these regions were highly reactive with ROS (Figure 4A). Among these identified cysteines in parkin (shown in Figure 4A), six cysteine residues (C²¹², C²⁵³, C²⁶⁸, C²⁸⁹, C⁴³¹, and C⁴⁴¹) were previously identified from the UniProt archive (http://www.uniprot.org/uniprot/O60260) as bearing point mutations related to rare familial PD (Figure 4B). Mutations of these cysteines alter parkin solubility, intracellular localization, and sensitivity to stress [27, 28]. These results are consistent with the notion that the RING and IBR domains contain regions undergoing cysteine modification during oxidative stress that modulate parkin E3 ligase activity and affect ubiquitination-mediated protein degradation, thus contributing to parkin aggregation. Moreover, these findings suggest that sulfonation of these cysteine residues in the more common sporadic cases of PD may mimic the effect of rare mutations observed in familial PD.

Oxidative stress induces parkin auto-ubiquitination and aggregation in cell-based and in vivoanimal models of PD

We next examined whether the ubiquitin E3 ligase activity of parkin was affected by oxidative stress/ROS. We found that exposure of parkin-overexpressing SH-SY5Y cells to H₂O₂ resulted in increased auto-ubiquitination within 2 hours, followed by a decrease in activity 6 hours later, reflecting a biphasic effect (Figure 5A). Previously, we reported a similar biphasic pattern in parkin activity after exposure to NO [21]. One possible explanation for this biphasic action is that the free radical modification of parkin initially activates its E3 ubiquitin ligase activity, and the resulting auto-ubiquitination subsequently leads to its inactivation. Excessive ubiquitinated proteins can potentially overwhelm the proteasome and result in aberrant protein accumulation, as has been observed with parkin in LBs of PD patients [4, 15].

Emerging genetic studies indicate that a large number of familial PD-linked point mutations of parkin alter its solubility in cells [11, 17, 26–28]. We thus investigated if the decrease in parkin solubility that we observed in sporadic cases of PD resulted in part from oxidative stress-induced perturbations in the redox state of parkin, producing sulfonation. We studied whether such oxidative stress could affect the cellular localization of parkin in both cell-based and animal PD model systems. We found that parkin-overexpressing SH-SY5Y cells exposed to MPP⁺ exhibited a dramatic increase in parkin aggregation (Figure 5B and 5C). Administration of catalase prior to MPP⁺ exposure prevented the aggregation of parkin, consistent with the notion that H₂O₂-mediated sulfonation of parkin affects its aggregation state. To confirm that the MPP⁺-induced parkin sulfonation (as shown in Figure 2D and 2E) resulted in parkin aggregation via oxidative stress, we exposed parkin-overexpressing SH-SY5Y cells to various concentrations of H₂O₂. Proteins were extracted and separated into "Soluble" and "Insoluble" fractions. These fractions were then subjected to SDS-PAGE and immunoblot analysis to detect parkin. We found a significant increase in parkin in the "Insoluble" fraction and a decrease in the "Soluble" fraction in H₂O₂-treated SH-SY5Y cells compared to controls; the increase in insoluble parkin occurred in a dose-dependent manner with regard to hydrogen peroxide exposure (Additional file 1, Figure S3).

To recapitulate the effects of exposure of neurotoxins in vivo, we injected 1 mg/kg/day of rotenone intraperitoneally (ip) into rats for 5 days. We found an increase in inclusion body-like parkin immunoreactivity in the striatum of rotenone-exposed rats compared to vehicle-treated controls (Figure 5D). We observed a similar pattern of parkin immunoreactivity in aggregates from monkeys exposed to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (Figure 5E), consistent with our prior findings for α-synuclein aggregates in the nigrostriatal system [31]. Moreover, human neural stem cells (hNSCs) transplanted into MPTP-lesioned monkeys appeared to diminish parkin aggregation in the host substantia nigra, similar to that of nonlesioned monkeys with or without hNSCs (Figure 5E).

Selective increase in insoluble parkin levels and sulfonation in idiopathic PD brains

Previously, we and others demonstrated that nitrosative stress induced by RNS affects parkin E3 ubiquitin ligase activity [21, 22]. This alteration in E3 ligase activity may contribute to LB formation with the accumulation of insoluble, aberrant proteins including parkin and its substrates [4, 15, 32]. In the present study, since we found that oxidative stress induced by ROS also affects parkin E3 ligase activity, we asked if such insult might also affect aberrant parkin accumulation in idiopathic PD. We initially examined brain tissues from human patients with LB dementia compared to control brains from patients dying of non-CNS causes (Additional file 1, Table S1). Brain tissue was extracted and separated into "Soluble" and "Insoluble" fractions. These fractions were then subjected to SDS-PAGE and immunoblot analysis to detect parkin. We found a significant increase in parkin in the insoluble compared to the soluble fraction of brains obtained from patients with PD and diffuse LBs (Figure 6A and 6B). Note that in PD compared to control brains, the absolute level of parkin increased in the insoluble fraction representing an increase in aggregated parkin.

Next, we examined human postmortem brain tissue for immunodetection of parkin sulfonation in order to demonstrate pathophysiological relevance of increased sulfonation to parkin insolubility in PD. To obtain sufficient tissue for the analysis, we pooled tissue lysates from two cases in each lane on SDS-PAGE for analysis of parkin sulfonation via immunoblot (Additional file 1, Table S2). We found that parkin sulfonation in the "Insoluble" fractions was significantly increased in the PD brains compared to the controls (Figure 6C and 6D). These results demonstrate a significant correlation between increased sulfonation and insoluble parkin in PD brains.

Expression of recombinant parkin

Recombinant parkin was expressed and purified as previously described [21]. Briefly, full-length parkin cDNA was cloned into a pGEX-4T-2 vector, and the constructs transduced into BL21 cells. Expression of recombinant GST-parkin was induced by isopropyl β-D-1-thiogalactopyranoside (IPTG) according to the manufacturer's instructions (GE Healthcare, Uppsala, Sweden). After cell lysis, GST-parkin was purified on a column of glutathione Sepharose beads and further treated with thrombin to remove GST. Parkin (1 μg) was exposed to 0, 10, or 200 μM H₂O₂ at room temperature (RT) for 30 minutes (min) and then immunoblotted to detect parkin sulfonation.

Generation and evaluation of an anti-sulfonation polyclonal antibody

A polyclonal antibody (pAb) was developed in order to recognize diverse epitopes containing cysteine sulfonic acid [Cys(SO₃^-)] derivatives of intact proteins. Unlike other commercial sulfonation antibodies that react with only one oxidized protein (for instance of Prx-SO₃H), we attempted to design a polyclonal antibody that would react with multiple sulfonated peptides. For this purpose, we reasoned that BSA, which has multiple cysteine residues, could serve as a scaffold for the generation of several sulfonation antigens--per-sulfonyl-BSA. In brief, BSA (200 mg) was suspended in 8.88 ml of highly purified DMSO. Then 1 ml of 10 mg/ml iodine in DMSO was added, followed by 0.12 ml of 36% HCl. The mixture was then incubated at 45°C for 24 hours while shaking gently. After cooling to RT, the product was precipitated with 20 ml of cold acetone at 4°C for 1 hour. The precipitate was washed with 10 ml of acetone twice as above, and subsequently washed twice with cold methanol at 1-hour intervals at 4°C. The last precipitate was dried under a stream of nitrogen.

The final product was analyzed by mass spectrometry to verify the oxidative BSA modifications. The absence of reducible thiols in the product thus obtained was confirmed by Ellman's reagent (DTNB). Sulfonated BSA was used as the antigen and submitted to a commercial vendor, Abgent (San Diego, CA), to generate anti-sulfonation polyclonal anti-serum from rabbits with an ELISA titer higher than 1:10,000.

The anti-sulfonated BSA pAb was purified using protein A/G-agarose. Briefly, rabbit anti-serum containing the anti-sulfonated BSA pAb was mixed with an equal volume of 100 mM phosphate buffer (pH 7.4) and incubated with protein A/G-agarose at 4°C for 2 hours. After washing the beads with 100 mM phosphate buffer, the anti-sulfonated BSA pAb was eluted five times with 100 μl volumes of 100 mM glycine solution (pH 2.8), followed by neutralization of each eluate with 5 μl of 1 M Tris solution (pH 8.3).

Next, we sought to minimize any rabbit anti-serum containing antibodies that recognized the reduced form of BSA, i.e., BSA that had been reduced and converted to its CAM derivative by alkylation with IAM. To deplete any anti-reduced BSA pAb, the eluted anti-serum was incubated with reduced BSA-agarose, prepared as follows: 1 ml NHS (N-hydroxysuccinimide)-activated agarose was incubated with a 2 ml volume of BSA (10 mg/ml) in 100 mM phosphate buffer (pH 7.4) for 1 hour at RT; residual BSA was then removed by centrifugation. Following this step, 2 M ethanolamine (pH 7.4) was added to the beads and incubated at RT for 15 min. After washing the beads with 50 mM Tris (pH 7.4), 50 mM dithiothreitol (DTT) was added for 1 hour at RT, followed by washing with 1 mM DTT. Then 50 mM IAM was added for 1 hour at RT. Finally, the resulting anti-sulfonated BSA pAb was incubated with an equal volume of reduced BSA-agarose overnight at 4°C for further purification. After centrifugation, the supernatant was saved as the anti-sulfonation pAb. Dot blotting was performed to determine the dose-response of the pAb to antigen. To further evaluate the specificity of the anti-sulfonation pAb, competition assays were performed using chemicals with structures similar to the envisaged epitopes containing Cys(SO₃^-). Briefly, 10 μM BSA or another cysteine-containing protein, such as recombinant PTP1B in 50 mM Tris (pH 7.4), was exposed to 20 or 200 μM H₂O₂ for 30 min at RT. The resulting sulfonated BSA or PTP1B was then gel electrophoresed and transferred to nitrocellulose membrane for immunoblotting. The membrane was incubated with anti-sulfonation pAb at 4°C overnight. For the competition assays, the anti-sulfonation pAb was incubated with 1 mM l-cysteic acid or l-cysteine sulfinic acid at RT for 2 hours prior to immunoblotting.

For cell-based assays to assess the anti-sulfonation pAb, SH-SY5Y cells were exposed to 200 μM H₂O₂, and the resulting sulfonated proteins in cell lysates were subjected to electrophoresis for immunoblotting with the pAb. Peroxiredoxin (Prx) and sulfonated Prx (Prx-SO₃H; Lab Frontier, Seoul, Korea) served as controls [14].

Cell-based and animal models of PD

(i) PD cell-based models: Myc-tagged, parkin-overexpressing SH-SY5Y cells [21] were grown in Dulbecco's modification of Eagle's Medium (DMEM) containing 10% fetal bovine serum (FBS) and 500 μg/ml geneticin in a 5% CO₂ atmosphere. At 70% confluence, the medium was removed and cells were incubated with the mitochondrial complex I inhibitor, MPP ⁺ (Sigma, St. Louis, MO), in FBS-free DMEM for 18 hours. Primary cultures of the enriched striatal neurons were prepared and maintained for 14 days in vitro (DIV), as previously described [50] with minor modifications. Briefly, to prepare primary cultures, striatal tissues were dissected out from E16-18 time-pregnant Sprague-Dawley rats and dissociated in 0.05% trypsin for 35 min at 37°C. Striatal neurons were seeded in DMEM supplemented with heat-inactivated FBS and Ham's F12 (ratio 8:1:1), 24 U/ml penicillin, 24 μg/ml streptomycin, and 24 mM HEPES, pH 7.4 at a density of 0.67 × 10⁶ cells per 35-mm dish containing poly-l-lysine (PLL, Sigma, St. Louis, MO) coated glass cover slips. The cultures were replaced the following day with DMEM supplemented with 1% B27 neurobasal medium and 0.5 mM glutamate, and incubated in a 5% CO₂, 95% air humidified atmosphere at 37°C. Striatal neurons were exposed to 100 nM rotenone for 4 hours. To prevent H₂O₂ accumulation, 100 U/ml catalase (Sigma, St. Louis, MO) was added to the medium 1 hour prior to MPP ⁺ or rotenone exposure.

(ii) PD animal models: All animal experiments were performed as described previously [21] according to guidelines set by the relevant institutional Animal Care and Use Committees of the collaborating institutions. Briefly, Sprague-Dawley rats weighing 200-250 g were divided into two groups (n = 10 per group) to receive ip injections of freshly prepared rotenone (1 mg/kg/day in 50% DMSO and 50% polyethylene glycol) or vehicle only (as the control group) daily for 5 days. MPTP-lesioned monkey brain slices were generously provided by Dr. Evan Y. Snyder of the Sanford-Burnham Medical Research Institute with details of the preparation previously described [31].

Immunoblotting parkin from fractionated cells and human brains

Autopsied human brain samples were analyzed following the guidelines of the relevant Institutional Review Boards. Cell lysates and PD brain tissue homogenates were sequentially fractionated as previously described [28, 51] with minor modifications. Myc-tagged, parkin-overexpressing stable SH-SY5Y cells were lysed with 1% Triton X-100 in phosphate buffered saline (PBS) containing a cocktail of protease inhibitors (Sigma, St. Louis, MO) and sedimented at 22,000 g at 4°C for 15 min. Supernatant from the initial fractionation was harvested as the "Soluble" fraction, and the remaining pellet was washed once with 1% Triton X-100 in PBS before re-extraction with 1% SDS in PBS. After sedimentation at 22,000 g for 15 min at RT, the supernatant was collected as the "Insoluble" fraction. Equal amounts of protein from the "Soluble" and "Insoluble" fractions were used for immunoprecipitation (IP) with anti-myc mAb-Agarose (Santa Cruz, CA) in order to detect parkin. For IP, 1 mg of protein from the "Soluble" fraction, or 200 μg of protein from the "Insoluble" fraction diluted 10-fold with 1% Triton X-100 in PBS, was incubated with 40 μl of a 25% slurry of anti-myc mAb-agarose overnight at 4°C. After washing with 1% Triton X-100 in PBS on ice, the beads were re-suspended in an aliquot of SDS sample buffer and the eluant collected for detection of parkin sulfonation by electrophoresis under reducing conditions with subsequent immunoblot analysis.

Similarly, brain tissue from PD patients or control subjects without CNS disease was fractionated with 1% Triton X-100 (designated the "Soluble" fraction) followed by 8 M urea (the "Insoluble" fraction) in 0.5 M NaCl and 50 mM Tris at pH 7.4. Equal amounts of protein from the different fractions were loaded for immunoblotting with anti-parkin monoclonal (m)Ab (PRK8, Cell Signaling Tech. Danvers, MA). For all experiments, protein concentration was determined with a BCA protein assay kit.

ROS detection and immunocytochemistry

Cellular localization of parkin was examined in cultures by immunocytochemistry and confocal microscopy, as previously described [28]. Parkin-overexpressing SH-SY5Y cells were plated onto cover slips in 24-well culture plates. After two days, cells were incubated with 200 μM MPP ⁺ in FBS-free DMEM for 18 hours or with 100 nM rotenone for 4 hours, with or without the administration of 100 U/ml catalase 1 hour prior to exposure to neurotoxins. For ROS detection, during the final 30 min of exposure to either MPP⁺ or rotenone, 5 μM cell-permeable fluorogenic dye DCF or 1 μM HEt was applied to the cells, as suggested by the manufacturer. Cells were fixed in 4% paraformaldehyde (PFA), permeabilized with 0.1% Triton X-100, and incubated with anti-myc mAb for parkin, NeuN and MAP-2 mAb to identify neurons, or anti-tyrosine hydroxylase (TH) antibody for labeling dopaminergic neurons in primary striatal cultures, followed by second antibody conjugated to various fluorescent dyes. Cells were then counterstained for nuclear DNA with Hoechst dye 33342 (Invitrogen, San Diego, CA).

Immunohistology on tissue sections was performed with standard methods using a primary pAb for parkin (2132, Cell Signaling Tech, Danvers, MA) at a dilution of 1:500 plus a secondary Ab at a dilution of 1:200 with the Vector ABC kit (Burlingame, CA). Monkey brain sections were counterstained with Hoechst 33342. Fluorescent images were acquired on an Olympus X81 spinning-disc confocal microscope equipped with a Hamamatsu deep-cooled monochrome EM camera (C9100-13) and 3I-Slidebook™ software analysis package (Denver, CO). Immunostaining that omitted primary antibody and SH-SY5Y cells transfected with empty vector pcDNA3.1 were included as negative controls. To determine the number of cells containing parkin aggregates, ~200 myc-parkin positive cells were counted from three independent cover slips per experiment.

Polyubiquitination assays

For cell-based ubiquitination assays, parkin-overexpressing SH-SY5Y cells were transfected with hemagglutinin (HA)-tagged ubiquitin, exposed to various concentrations of H₂O₂, and then examined at different time points. The proteasome inhibitor MG132 (10 μM) was added to block degradation of poly-ubiquitinated proteins during the assay. Cell lysates were subjected to immunoprecipitation with anti-myc mAb for parkin, followed by Western blotting against HA for ubiquitin.

Mapping PTMs of cysteine residues in parkin

(i) Protein digestion: Recombinant human parkin was exposed to pathological doses of H₂O₂ in vitro followed by in-solution or in-gel trypsin digestion, as described previously [19, 21]. Protein digestion was carried out by in-solution digestion or by DigestPro in-gel digestion using a robotic workstation (Intavis, Koeln, Germany).

(ii) High-resolution mass spectrometry: LC-MS analyses were carried out using a high-resolution (5 ppm) Q-TOF API-US mass spectrometer (Waters Micromass, Manchester, UK) with an on-line nanoLC system (LC Packings, Sunnyvale, CA), as previously described [21]. We used a 75 μm I.D. × 15 cm PepMap C18 analytical column (LC Packings, Sunnyvale, CA) with 3-μm particle size. Mobile phase A was 2% acetonitrile (ACN)/0.1% formic acid (FA), while mobile phase B was 80% ACN/0.1% FA. The gradient was 2% ~ 40% mobile phase B in 75 min, followed by 100% mobile phase B for 5 min. The flow rate was set at 0.2 μl/min. Mass spectra of untreated parkin and H₂O₂-exposed parkin were analyzed for PTMs, including cysteine alkylation or sulfination (R-SO₂H)/sulfonation (R-SO₃H) and methionine oxidation using an in-house Mascot 2.0 server (Matrix Science Inc., London, UK). The modifications of tryptically-digested peptides of parkin were further confirmed by tandem MS spectra analysis as described below.

(iii) High-throughput ion trap tandem MS: We used packed Picofit capillary column with an integral spray tip (Picofit, 15 μm tip, New Objective, Woburn, MA) for reversed-phase nanoLC MS/MS. The column had an internal diameter of 75 μm, was 15 cm long, and packed with 5 μm C18 AQUASIL. A linear gradient elution was used, from buffer A (0.1% FA in water) to 50% buffer A + 50% buffer B (0.1% FA in ACN) in 100 min, and flow rate was set at 0.2 μl/min. Two ion-trap mass spectrometers were used for maximizing peptide detection: an LTQ linear ion trap mass spectrometer (Thermo Electron, San Jose, CA) and an HCTultra PTM discovery system (Brüker Daltonics, Bremen, Germany). The instruments were operated in a data-dependent mode, in which the first and second strongest ions were sequenced in each cycle with dynamic exclusion enabled and the collision energy set at 35%. For the nanoLC-MS/MS analysis by using an LTQ linear ion trap LC/MS, recombinant proteins were digested with 0.04 μg/μl trypsin for 15 hours. The resulting tryptic digests were then on-line separated on a 5-cm packed PicoFrit capillary column (New Objective, Woburn, MA) with an Eksigent nanoLC system (Dublin, CA). Initially, it was eluted with H₂O/ACN/FA (98:2:0.1, v/v/v) for 5 min, then a linear gradient increased ACN to 55% by 60 min, and finally decreased ACN to 2% by 65 min. MS/MS spectra were acquired in the data-dependent scanning mode with one full scan followed by four MS/MS scans on the most intense precursor ions with dynamic exclusion enabled. MS/MS data were then further analyzed with SEQUEST Sorcerer software (Sage-N Research, Milpitas, CA).

(iv) DeCyder™ MS analysis: LC/MS data generated by either Q-TOF MS or ion trap MS/MS were converted into a DeCyder™ MS v2.0 (GE Bioscience, Uppsala, Sweden) compatible format, as previously described [30], and the MS profiles were aligned based on the m/z and retention time values to ensure the reproducibility and consistency of the LC spectral profiles. Differentially sulfinated/sulfonated peptide candidates were detected and quantified by comparing the signal intensity maps generated by DeCyder™ MS software, and peaks of interest (m/z values and charge states) were selected as precursor ions for subsequent peptide identification.