The S-nitrosylation of parkin attenuated the ubiquitination of divalent metal transporter 1 in MPP+-treated SH-SY5Y cells

Abnormal iron accumulation caused by elevated levels of divalent metal transporter 1 (DMT1) contributes to progressive neurodegeneration in Parkinson's disease (PD). Parkin is a E3 ubiquitin ligase for the ubiquitination of DMT1. S-nitrosylated parkin (SNO-parkin) is commonly observed in PD. However, the effects of S-nitrosylation on the E3 ubiquitin ligase activity of parkin for the ubiquitination of DMT1 in PD are largely unknown. To elucidate the role of S-nitrosylated parkin and DMT1 in PD, SH-SY5Y cells were transfected with parkin, being treated with S-nitrosoglutathione (GSNO) and 1-methyl-4-phenylpyridinium (MPP+). The results showed increased levels of oxidized nitric oxide (NO) and S-nitrosylated parkin after the treatment of GSNO and MPP+ in parkin-transfected cells. Consistently, increased levels of DMT1, iron uptake and cell viability were observed. Interestingly, inhibition of S-nitrosylated parkin reduced the level of DMT1. Further, S-nitrosylation of parkin significantly inhibited the ubiquitination of DMT1. When HEK293T cells were transfected with plasmid of parkin with single site mutation (Cys241A, Cys260A, Cys323A), ubiquitination of DMT1 was also inhibited. However, the cells cotransfected with plasmids containing all three mutations, GSNO treatment did not affect the ubiquitination of DMT1. The expression of SNO-parkin and DMT1 protein in substantia nigra increased significantly gradually after 2 h, 4 h and 24 h with MPTP injection. These results indicate that the S-nitrosylation of parkin inhibits its E3 ubiquitin ligase activity for the ubiquitination of DMT1, which contributes to iron accumulation and degenerative process in PD. Targeted S-nitrosylation could provide a potential therapeutic strategy against PD.


S-nitrosylation of parkin resulted in an increase of DMT1 level in parkin-transfected SH-SY5Y cells.
Parkin-transfected SH-SY5Y cells were treated with 200 μM GSNO for 24 h. Under such conditions, parkin was S-nitrosylated in Biotin-Switch Assay (Fig. 2a,b, P = 0.0002). Meanwhile, a 2.77-fold significantly increase in DMT1 levels was observed, compared to those in cells with GSH treatment (Fig. 2a,c, P = 0.0003). Interestingly, levels of DMT1 elevated in Parkin-GSNO group, compared with Empty vector-GSNO group (Fig. 2a,c , P = 0.025). The results indicated that S-nitrosylation of parkin inhibited the expression of DMT1. Confirmly, we still found that transfected parkin downregulated the protein level of DMT1 (Fig. 2a,c, P = 0.013).

GSNO and MPP + treatment increased the levels of oxidized NO in parkin-transfected SH-SY5Y cells.
Excessive levels of NO cause a variety of protein thiol nitrosylations. In order to determine the nitrosylation environment in cells with the treatment of GSNO and MPP + , we measured the amount of oxidized NO. Due to the rapid conversion of NO to nitrite (NO 2 ) and nitrate (NO 3 ), the total amount NO 2 and NO 3 was used to represent NO production. The results showed that the oxidized NO (NO 2 + NO 3 ) levels were increased significantly by 46.5% (Fig. 3, P = 0.004, parkin-GSNO vs. parkin-GSH) and 60.2% (Fig. 3, P = 0.0007, parkin-MPP + vs. parkin-NS) in the GSNO and MPP + treatment groups, respectively. When NOS inhibitor L-NNA was added, the levels of oxidized NO were decreased by 35.1% (Fig. 3, P = 0.0009, parkin-MPP + + L-NNA vs. parkin-MPP + ). These results indicated that both GSNO and MPP + could induce increased levels of oxidized NO, leading to protein thiol nitrosylations.

S-nitrosylation impaired the E3 ubiquitin ligase activity of parkin and inhibited the ubiquitination of DMT1.
To further investigate the influence of nitrosative stress on the E3 ubiquitin ligase activity of parkin, we performed intracellular ubiquitination experiments. As shown in Fig. 6a, When Flag-parkin joined, immunoprecipitated DMT1 elevated the levels of HA-Ub, which indicated parkin mediated ubiquitination of DMT1. However, the ubiquitination of DMT1was significantly was decreased with GSNO treatment. The results indicate that the S-nitrosylation of parkin inhibits its E3 ubiquitin ligase activity for the ubiquitination of DMT1.
To examine whether the mutation of S-nitrosylated parkin had an effect on its ubiquitination function, HEK293T cells were transfected with plasmid of parkin with the single site mutations (Cys241A, Cys260A, Cys323A, respectively). The results indicated that S-nitrosylation of parkin significantly inhibited the ubiquitination of DMT1 in single site mutant parkin-transfected HEK293T cells (Fig. 6b). However, when cells were cotransfected with all three mutant parkin plasmids, GSNO treatment did not affect the ubiquitination of DMT1 (Fig. 6c). These results prompted problably that residues Cys241, Cys260 and Cys323 worked together to regulate the E3 ubiquitin ligase activity of SNO-parkin.

Discussion
In this study, we observed that nitrosative stress induced by GSNO or MPP + promoted S-nitrosylation of parkin, resulting in the elevated expression of DMT1 and subsequent iron influx related cell death. The Cys241, Cys260 and Cys323 residues in parkin cooperate in the E3 ubiquitin ligase activity by S-nitrosylation. These results demonstrated that S-nitrosylation inhibits the E3 ubiquitin ligase activity of parkin, leading to impaired ubiquitination of DMT1and iron accumulation during the degenerative process in PD. Abnormal iron accumulation is involved in several neurological disorders. Iron delivery systems provide a requisite supply of iron for mitochondrial respiratory chain complexes and iron-containing proteins [23][24][25][26][27] . Over the last decade, a number of proteins regulated by transcriptional and posttranscriptional processes have been identified that are involved in both iron absorption and homeostatic iron trafficking 28,29 . DMT1 plays an important role in the function of transferrin, which is the major iron transporter that allows iron to enter mammalian cells and to be excreted from endosomes 30 . As reported, the expression of DMT1 is partially regulated by the proteasomal pathway 31,32 . Two different E3 ubiquitin ligases participate in the ubiquitination of DMT1: parkin and WW domain-containing protein 2 (WWP2) 15,33,34 . As an important post-translational modification, ubiquitination regulates cell biological processes in many ways 35 . The attachment of ubiquitin or the polyubiquitin chain to the protein regulates turnover, localization and activity of substrate.
The parkin protein consists of a ubiquitin-like domain in its N-terminus, a unique parkin-specific domain, an IBR domain, RING1 and RING2 domains 36 . Parkin ubiquitinates a variety of cytosolic and outer mitochondrial membrane substrates during mitochondrial depolarization 37 . As reported, parkin is an E3 ubiquitin ligase that is responsible for the ubiquitination of DMT1. Overexpression of parkin in SH-SY5Y cells resulted in a reduction of the 1B-DMT1 isoform and manganese transport 15 . In addition, a few of substrates were found to be accumulated in parkin deficient mice brain or in disease phenotypes. P62, a critical regulator for protein quality control, inclusion body formation, selective autophagy and diverse signaling pathways, is a new substrate of parkin 38 . Parkin interacts with the substrate pyruvate kinase M2 (PKM2) to regulates the glycolysis pathway and affects the cell metabolism 39 . CHOP was identified as a parkin substrate. Parkin blunts excessive CHOP to prevent ER stress-induced cell death and adverse cardiac ventricular remodeling 40 . Parkin inactivation leads to accumulation of the parkin interacting substrate (PARIS, ZNF746) that plays an important role in dopamine cell loss 41 . Substrates of parkin also include CDCrel-1, synphilin, and cyclin E. Parkin catalyzes the binding of ubiquitin to these substrates 42,43 . Ubiquitination, phosphorylation, and acetylation are reversible post-translational modifications that can trigger or inhibit the activity of parkin 14 . The phosphorylation of Ser131, Ser101, Ser127 and Ser378 led to the abnormal aggregation of parkin without specifically affecting its E3 ubiquitin ligase activity 44 . However, the phosphorylation of Tyr143 inhibited parkin E3 ubiquitin ligase activity during PD in vivo and in vitro 45,46 . www.nature.com/scientificreports/ Additionally, parkin regulates its own ubiquitination by the Lys48 proteasome-dependent ubiquitin chain 12,14 .
The precise location and function of parkin during ubiquitination remain to be determined. The excessive production of NO in PD causes nitrative stress damage to nigral dopaminergic neurons. Nitrification-induced protein thiol nitrosylation was shown to be involved in the pathogenesis of PD 16 . S-nitrosylation inhibited the E3 ubiquitin ligase activity of parkin, which led to the decrease of the ubiquitination and degradation of substrates, abnormal protein clearance, and, finally, the death of dopaminergic neurons 17,22 . SNO-parkin was detected in brains from both animal PD models and human PD patient 22 . Parkin dysfunction caused by nitrosative stress suggests a probable link between free radical production and abnormal protein accumulation in PD. In this study, GSNO or MPP + treatment elevated the level of oxidized NO, resulting in the S-nitrosylation of parkin. Meanwhile, the expression of DMT1, iron uptake and cell death were significantly increased also.By cell viability detection, we can infer that DMT1 is degraded by parkin and could not transfer excessive iron ions into cells, which will change the cell viability. However, GSNO plays an impired role in several ways, it could not reflect protective effect of parkin in cells with GSNO treatment. The effect of parkin mutant, which is related to the function of S-nitrosylation, on cell viability could not be reflected, too.
The production of SNO-containing proteins led to changes in protein conformation, enzyme activity, protein-protein interactions and cell localization 47,48 , which then affected protein function 16 . Five thiolation sites were found at cysteine residues (amino acids 59, 95, 182, 212 and 377) in parkin protein. But it was unclear whether mutations at these sites had an effect on the regulation of ubiquitination 49 . Cys241, Cys260 22 in the RING1 domainand and Cys323 21 in the IBR domain are S-nitrosylated sites in parkin. Based on this, individual parkin mutants (Cys241A, Cys260A, and Cys323A) by site-directed mutagenesis were generated. However, whether there is any other S-nitrosylation site in parkin is still unclear.
In conclusion, the current study showed that the S-nitrosylation of parkin inhibited its E3 ubiquitin ligase activity, preventing the ubiquitination and degradation of DMT1. Increased DMT1 levels resulted in enhanced iron uptake, which aggravated iron accumulation in PD. Our study suggested that S-nitrosylation of parkin was essential for cell survival, which provided a new target for the prevention and treatment of PD.
Cell transfection and treatment. SH-SY5Y cells were seeded at a density of 1 × 10 5 /cm 2 in 6-well plates.
When the cells grew to 70%-80% confluence, a pCMV6 vector that contained cDNA encoding parkin was transfected into cells1 (1 μg/well). The plasmid was mixed with the Lipofectamine 2000 (Thermo Fisher Scientific, Waltham, MA, USA) for 30 min. Then serum-free medium was added for 4 h, followed by serum-containing medium. The SH-SY5Y cell line transiently overexpressing parkin (Parkin-SY5Y cells) has been prepared. As a negative control, an empty vector was transfected into cells.
SH-SY5Y cells were exposed for 24 h with 200 μM of GSNO (Santa Cruz Biotechnology, CA, USA), or glutathione (GSH; Santa Cruz Biotechnology, CA, USA), as the negative control. Cells were also exposed to 200 μM of MPP + (Sigma, Ronkonkoma, NY, USA), or normal saline (NS, 0.9% saline), as the blank control. In some cases, 1 mM of the broad spectrum NOS inhibitor NG-nitro-L-Arginine (L-NNA; Enzo Biochem, NY, USA) was added. In other cases, 100 μM of Fe 2+ was added.
In vivo experiment. The experimental animals are provided by Qingdao Institute for drug control. The operations on animals are in accordance with the rules of experimental animals. Male C57BL/6 mice aged 8-10 weeks were randomly divided into six groups: control group, 2 h group, 4 h group, 24 h group, 48 h group and72 h group, with an average of 6 mice in each group. 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) was dissolved in normal saline to prepare a concentration of 6.25 mg / ml. Mice were intraperitoneally injected with MPTP (20 mg/kg) once every two hours, while the control group was intraperitoneally injected with the same volume of normal saline, a total of 4 times. After the last injection, at 2 h, 4 h, 24 h, 48 h, 72 h, the substantia nigra and striatum were collected for tissue protein extraction.
Biotin-switch assay for detection of S-nitrosylated protein. S-nitrosylated protein was detected using a biotin switch assay according to the instructions of a S-nitrosylated Protein Detection Assay Kit (Cayman Chemical, Ann Arbor, MI, USA) 17 . Firstly, the free sulfhydryl group in the sample was blocked with methyl methanethiol sulfonate (MMTS). Secondly, the excess blocking agent was removed by acetone precipitation, and the thiol nitroso group (-SNO) was selectively reduced to free thiol by ascorbate. Finally, biotin-HPDP was added to label the newly synthesized thiol group. The biotin-labeled protein was adsorbed by streptavidinagarose, subjected to denaturing polyacrylamide gel electrophoresis (SDS-PAGE). www.nature.com/scientificreports/ for 5-10 min. The protein samples were subjected to gel electrophoresis and then transferred to PVDF membranes. After blocking with 10% skimmed milk powder for 2 h at room temperature, the membranes were incubated with anti-parkin antibody ( 50 . The assay buffer was added to the collected cells and centrifuged to remove insoluble impurities. The enzyme cofactor and nitrate reductase were added to the supernatant of each well to convert the nitrate to nitrite. Then an enhancer and a DAN probe were added to the sample to react the nitrite with the fluorescent probe. The addition of NaOH enhanced the fluorescence yield, which was proportional to the total nitric oxide production. The fluorescence was measured at wavelengths of 360 nm and 450 nm using a microplate reader. Extrapolate sample readings from the standard curve plotted using the following equation:

Western blotting analysis.
Concentration of nitrate and nitrite in the test samples is calculated as: Sa = Amount of nitrate + nitrite in the sample well (conc). Sv = Sample volume added into the reaction well (µL). D = Sample dilution factor.

Calcein loading of cells and ferrous iron influx assay.
Calcein is a stable fluorescent molecule, which could be quenched rapidly once combining with ferrous iron 11,51,52 . After different treatment, cells were seeded on coverslips in the 24-well plate and washed with HBS three times. Next, cells were incubated with 1 µM calcein-AM for 30 min (Yeasen, Shanghai, China) at 37 °C. Then the slides were removed and fixed in a heated chamber connected to a perfusion apparatus. Calcein fluorescence was recorded at 488 nm excitation and 525 nm emission wavelengths. And 10 scans were performed every 3 min for about 7 s, and 1 mm ferrous iron was infused at the same time. Additionally, ascorbate acted as a chelator to maintain the iron in solution. The X-Y-T scan mode was used to obtain the data.
Cell viability assay. Cell viability was detected using CellTiter-Glo Luminescent Cell Viability Assay Kit (G7570, Promega, Madison, WI, USA) according to the manufacturer's protocol. Cells were seeded in a 96-well plate. 24 h later, cells are treated in various ways according to "Cell transfection and treatment" as mentioned before. The reagent was added directly to cells in culture medium with a volume equivalent to the amount of medium. Mix for 2 min, and after 10-min incubation, detect the emitted luminescence using a microplate reader (Molecular Device, M5, Sunnyvale, CA, USA).
Ubiquitination assays. In ubiquitination assays, HEK293T cells were used. Cells were transfected with Flag-WT-parkin, Myc-DMT1, and HA-ubiquitin. Plasmid of parkin with three mutations, Flag-C241A-parkin, Flag-C260A-parkin and Flag-C323A-parkin, were transfected into HEK293T cells respectively or together. 24 h later, cells were treated with GSNO for additional 24 h. The details are the same as described in "Cell transfection". 10 μM of proteasome inhibitor MG132 (ApexBIO, Houston, USA,) was added 8 h prior to cell collection to block the rapid degradation of polyubiquitinated proteins. The cell lysates were immunoprecipitated using anti-Myc antibody. The ubiquitination level of DMT1 was detected by anti-HA antibody.
Statistical analysis. Prism 8.0 software (GraphPad Software Inc., La Jolla, CA, USA) was applied to analyze the data. When comparing two sets of data, P value is greater than 0.05, which obeying normal distribution. Independent Samples Test is selected. When multiple groups of data are compared, with the test of homogeneity of variance, P value is greater than 0.05. One-way ANOVA is selected. Besides, the ferrous iron influx and pulse chase assay were analyzed by Two-way Repeated Measures ANOVA. Data are normally distributed by Shapiro Wilk test (P value is greater than 0.05). According to the hypothesis of Mauchly's sphere, the variance is homogeneous (P value is greater than 0.05). The effect of interaction was not statistically significant. Data are shown as the mean ± SEM. P < 0.05 was considered statistically significant.

Data availability
The data used to support the findings of this study are available from the corresponding author upon request.