SEARCH

SEARCH BY CITATION

Keywords:

  • ALS;
  • nitrosative stress;
  • protein disulfide isomerase;
  • S-nitrosylation;
  • SOD1

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

A major hallmark of mutant superoxide dismutase (SOD1)-linked familial amyotrophic lateral sclerosis is SOD1-immunopositive inclusions found within motor neurons. The mechanism by which SOD1 becomes aggregated, however, remains unclear. In this study, we aimed to investigate the role of nitrosative stress and S-nitrosylation of protein disulfide isomerase (PDI) in the formation of SOD1 aggregates. Our data show that with disease progression inducible nitric oxide synthase (iNOS) was up-regulated, which generated high levels of nitric oxide (NO) and subsequently induced S-nitrosylation of PDI in the spinal cord of mutant SOD1 transgenic mice. This was further confirmed by in vitro observation that treating SH-SY5Y cells with NO donor S-nitrosocysteine triggered a dose-dependent formation of S-nitrosylated PDI. When mutant SOD1 was over-expressed in SH-SY5Y cells, the iNOS expression was up-regulated, and NO generation was consequently increased. Furthermore, both S-nitrosylation of PDI and the formation of mutant SOD1 aggregates were detected in the cells expressing mutant SOD1G93A. Blocking NO generation with the NOS inhibitor N-nitro-l-arginine attenuated the S-nitrosylation of PDI and inhibited the formation of mutant SOD1 aggregates. We conclude that NO-mediated S-nitrosylation of PDI is a contributing factor to the accumulation of mutant SOD1 aggregates in amyotrophic lateral sclerosis.

Abbreviations used
ALS

amyotrophic lateral sclerosis

ER

endoplasmic reticulum

FUS

fused in sarcoma

iNOS

inducible NOS

NNA

N-nitro-l-arginine

nNOS

NOS-neuronal

NOS

nitric oxide synthase

PDI

protein disulfide isomerase

SNOC

S-nitrosocysteine

SNO-PDI

S-nitrosylated PDI

SOD

superoxide dismutase

UPR

unfolded proteins response

Mutations in the Cu, Zn-superoxide dismutase (SOD1) gene have been identified as a possible cause of a subset of familial amyotrophic lateral sclerosis (Deng et al. 1993; Rosen 1993), an adult-onset neurodegenerative disease characterized by degeneration of motor neurons in the spinal cord, brainstem, and motor cortex. Aggregates of misfolded mutant SOD1 are commonly associated with amyotrophic lateral sclerosis (ALS), as seen at post-mortem examination. SOD1 is an intracellular homodimeric metalloprotein that forms a stable intra-subunit disulfide bond. Several factors are involved to drive SOD1 to acquire the propensity to misfold, and favor the disulfide-reduced SOD1 monomers to convert into oligomeric and aggregated species (Rakhit et al. 2004; Arnesano et al. 2004; Doucette et al. 2004). These factors include improper metallation of the protein, genetic mutations, loss of disulfide bound, and post-translational modifications. There are four cysteine residues in SOD1, located at amino acids 6, 57, 111, and 146. The formation of disulfide bonds is mediated by oxidation of the thiol groups of cysteine residues. Previous studies have established that detergent-insoluble mutant SOD1 aggregates accumulated in the spinal cords of mutant SOD1 transgenic mice are extensively cross-linked by disulfide bonds (Deng et al. 2006; Furukawa et al. 2006; Wang et al. 2006). In vitro cell culture models also demonstrated that disulfide bond formation between mutant SOD1 proteins could either trigger oligomerization or facilitate the bonding to stabilize aggregates' structures (Cozzolino et al. 2008; Niwa et al. 2007b). Furthermore, in both cell culture and mouse models, mutant SOD1 lacking the native intramolecular disulfide bond is a major component of the insoluble SOD1 aggregates (Karch et al. 2009). Indeed, the disulfide-reduced subunits of SOD1 are susceptible to disulfide-linked multimerization upon oxidative stress, indicating that the regulation of disulfide-bond formation can affect the formation of SOD1 aggregates.

Although wild-type SOD1 is found predominantly in the cytoplasm, the accumulation of mutant SOD1 aggregates is observed at the mitochondrial surface and in the intermembrane space of mitochondria. Studies demonstrate that this intermembrane space-targeted mutant SOD1 causes neurite mitochondrial fragmentation, impaired mitochondrial dynamics, and neuronal toxicity (Magrane et al. 2009). This aberrant deposition of mutant SOD1 in mitochondria contributes to ALS pathogenesis (Cozzolino et al. 2009; Sotelo-Silveira et al. 2009). Over-expression of glutaredoxin 2, a thiol-disulfide oxidoreductase located in mitochondrial, reduces the mitochondrial fragmentation and mutant SOD1 aggregation, and protects its metabolic activity (Ferri et al. 2010). Furthermore, mutant SOD1 forms monomers or insoluble high molecular weight multimers within the endoplasmic reticulum (ER) (Kikuchi et al. 2006). Protein disulfide isomerase (PDI) is an enzyme critical for proper protein folding in the ER. PDI can introduce disulfide bonds into proteins (oxidation), break disulfide bonds (reduction), and catalyze thiol/disulfide exchange (isomerization), thus facilitating disulfide bond formation, reaction rearrangements, and structural stability (Lyles and Gilbert 1991). In many neurodegenerative disorders and cerebral ischemia, the accumulation of protein aggregates results in ER dysfunction (Conn et al. 2004; Rao and Bredesen 2004), but up-regulation of PDI represents an adaptive response, which may offer neuroprotection by promoting protein refolding (Hetz et al. 2005; Ko et al. 2002). Uehara and colleagues (Uehara et al. 2006) demonstrate that in Parkinson's disease and related disorders, nitric oxide (NO)-mediated S-nitrosylation of PDI inhibits PDI function, leads to dysregulated protein folding within the ER, and consequently results in ER stress and neuronal cell death. The role of PDI in SOD1 inclusions remains unclear. There is only circumstantial evidence showing that pharmacological inhibition of PDI enzymatic activity increased the presence of mutant SOD1 inclusions (Atkin et al. 2006).

S-nitrosylation is an important biological reaction of NO that involves the covalent addition of NO to thiol groups of cysteine residues of proteins to form S-nitrosothiols (SNOs). This selective post-translational modification can affect many cellular processes and regulate protein function, stability, localization, and protein–protein interactions (Hess et al. 2005).

Seeing as nitrosative stress is linked with excessive glutamate receptor activation, excitotoxicity, and oxidative stress, NO is believed to play a key pathogenic role in neurodegenerative disorders (Nakamura and Lipton 2010; De Palma et al. 2008), including ALS (Uehara 2007). Increased levels of NO is linked to the toxicity of mutant SOD1 in neuroblastoma cells (Arciello et al. 2010). In ALS pathogenesis, inflammatory response characterized by the accumulation of activated astrocytes and microglia is another pathological hallmark (Beers et al. 2011). Inducible nitric oxide synthase (iNOS) expresses at high levels in inflammatory-activated glia induces various damages including S-nitrosylation of PDI (Bal-Price and Brown 2001; Brown and Bal-Price 2003). We report here that with disease progression, expression of iNOS and formation of S-nitrosylated PDI (SNO-PDI) were increased in the spinal cords of mutant SOD1 transgenic mice. Treatment with the NO donor S-nitrosocysteine (SNOC) in SH-SY5Y cells triggered a dose-dependent formation of SNO-PDI. Expression of SOD1G93A in SH-SY5Y cells promoted iNOS expression and consequently NO generation, resulting in increased S-nitrosylation of PDI. Blocking NO generation with the NOS inhibitor N-nitro-l-arginine (NNA) significantly attenuated formation of SNO-PDI and mutant SOD1 aggregates. Our data suggest that NO-mediated S-nitrosylation of PDI is a contributing factor to the accumulation of mutant SOD1 aggregates in ALS.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Transgenic mice

The mutant SOD1 transgenic mice (B6.Cg-Tg(SOD1-G93A)1Gur/J and B6.Cg-Tg(SOD1-G37R)42Dpr/J), and the wild-type SOD1 transgenic mice (B6.Cg-Tg(SOD1)2Gur/J) were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). Inbred male mice between 1 and 13 months of age were used for this experiment. Mice were genotyped by PCR with the following sense and antisense primers: 5′-CATCAGCCCTAATCCATC-3′, 5′-CGCGACTAACAATCAAAG-3′. Disease onset was determined as the time when mice reached their peak body weight before the denervation-induced muscle atrophy and weight loss as described previously (Lobsiger et al. 2009). End-stage was defined as the time at which a mouse cannot right itself within 30 s when placed on its side. This is an endpoint frequently used for mutant SOD1 transgenic mice and one that was consistent with the requirements of the Animal Care and Use Committee of the University of Manitoba. One-month-old G93A mutant transgenic mice were considered pre-symptomatic; disease onset occurred when these mice were 3 months old, and disease end-stage occurred when these mice were 6 months old. Five-month-old G37R mutant transgenic mice were pre-symptomatic; disease-onset occurred when these mice were 10 months old, and disease end-stage occurred when these mice were 13 months old. The use and maintenance of the mice described here were performed in accordance with the Guide of Care and Use of Experimental Animals of the Canadian Council on Animal Care. Mice tissues were collected and homogenized in 10 volumes of lysis buffer consisting of 1% (w/v) Triton X-100, 50 mM Tris-HCl, Ph 7.4, 300 mM NaCl, 5 mM EDTA, 0.02% (w/v) sodium azide, 10 mM iodoacetamide, 1 mM phenylmethylsulfonyl fluoride and 2 μg/mL leupeptin. Homogenates were centrifuged at 20 000 g for 30 min at 4°C. Supernatants are collected for analysis. Protein concentration was measured by the BCA method as described by the manufacturer (Pierce, Rockford, IL, USA).

Plasmids, cell culture, and transfection

G93A and WT SOD1 constructs tagged with EGFP were constructed in the same manner as those previously reported (Turner et al. 2005). Briefly, the SOD1WT template for PCR amplification uses the following primers: 5′-GCGCGCGTCGACAAGCATGGC-3′ (forward), 5′-GCGCGCGTCGACGCTTGGGCGATCCCAAT-3′ (reverse). Primers were designed to introduce a SalI site to allow subcloning into pEGFP-N1 (Clontech, Palo Alto, CA, USA) and to remove the SOD1 translation stop codon. Additional mutant SOD1G93A plasmid was generated via site-directed mutagenesis of the SOD1WT template using the Quik Change kit (Stratagene, La Jolla, CA, USA), with the use of the following primers: G93A: 5′-CTGCTGACAAAGATGCTGTGGCCGATGTGTC-3′ (forward) and 5′-GACACATCGGCCACAGCATCTTTGTCAGCAG-3′ (reverse). All the plasmid constructions were verified by automated sequencing. SH-SY5Y human neuroblastoma cells were maintained in Dulbecco's modified Eagle's/F-12 (1 : 1) medium, supplemented with 10% fetal calf serum, 2 mM l-glutamine, 100 IU/mL penicillin, and 100 μg/mL streptomycin, at 37°C in an atmosphere of 5% CO2 in air. Eighty percentage confluent cells were transfected with the indicated plasmids using Lipofectamine (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. Cells were analyzed 48 h after transfection.

Inclusion quantification and subcellular fractionation

SH-SY5Y cells were transiently transfected with the indicated plasmids carrying various SOD1s tagged with EGFP. Aggregate-positive cells were counted as a percentage of total EGFP-positive cells. If an EGFP-positive cell had one or many aggregates, the aggregates' score was one. Harvested cells were lysed in 50 mM Tris-HCl, pH 7.5, containing 150 mM NaCl, 0.1% sodium dodecyl sulfate, 1% triton X-100, and 1% protease inhibitors. Lysates were centrifuged at 1000 g for 10 min. The cleared lysates were centrifuged at 25 000 g for 50 min to obtain the supernatant and pellet fractions. Protein from the resulting supernatants (soluble fraction) and the pellets (insoluble fraction) after 15 s sonication in radio-immunoprecipitation assay buffer without sodium deoxycholate were analyzed by western blot.

Measurement of NO level

Spinal cord lysates (30 mL) were collected and incubated with an equal volume of Griess reagent for 20 min at 22°C (Lee et al. 2003). The concentration of nitrite (NO2), which is formed by the spontaneous oxidation of NO under physiological condition, was determined by measuring the absorbance at 540 nm. The value was normalized to the protein concentration of lysates. The NO formed by the cells was determined by the Griess reaction with a minor change. Briefly, 40 μL cell culture fluid, 10 μL NADPH and 40 μL basal solution (0.03 M phosphate-buffered saline, 1.25 mM glucose-6-phosphate, 400 U/L glucose-6-phosphate dehydrogenase, 200 U/L nitrate reductase) were incubated in a 96-well microtiter plate for 45 min at 22°C. Next, 50 μL Griess reagent was added, and the solutions were incubated for 20 min at 22°C. Finally, the absorbance of the samples was measured at 540 nm. NO2 concentrations were calculated from a standard curve of Sodium nitrite (NaNO2).

Biotin-switch assay for detection of S-nitrosylated PDI

Briefly, mouse spinal cord tissue extracts and cell lysates were prepared in HENC buffers (250 mM Hepes pH 7.5, 1 mM EDTA, 0.1 mM neocuproine, 0.4% CHAPS). Typically, 1 mg of cell lysates and up to 2 mg of tissue extracts was used. The blocking buffer (2.5% sodium dodecyl sulfate, 20 mM methyl methane thiosulphonate [MMTS] in HEN buffer) was mixed with the samples and incubated for 30 min at 50°C to block free thiol groups. After removing excess MMTS by acetone precipitation, nitrosothiols were reduced to thiols with 1 mM ascorbate. The newly formed thiols were then linked with the sulfhydryl-specific biotinylating reagent N-[6-(biotinamido)hexyl]-3′-(2′-pyridyldithio) propionamide (Biotin-HPDP). The biotinylated proteins were pulled down with Streptavidin–Agarose beads. Western blot analysis was performed to detect the amount of PDI remaining in the samples (Walker et al. 2010).

Preparation of S-Nitrosocysteine and N-nitro-l-arginine

S-nitrosocysteine is a NO donor that can spontaneously decompose to generate NO. SNOC decays quickly within a half-life in the range of 2–3 min. A 100 mM stock of SNOC was produced immediately before each use from a mixture of 100 mM l-cysteine and 100 mM NaNO2 by acidification with 5% (v/v) with 10 N HCl. The solution turned from clear to rose-colored after completion of the reaction. It was applied within minutes of its synthesis. After synthesis, stocks of SNOC were left at room temperature for several days to degrade fully to ‘old SNOC’ for use as a control. NNA is an irreversible inhibitor of constitutive NOS-neuronal (nNOS) and a reversible inhibitor of inducible nitric oxide synthase (iNOS). A 5 mM stock NNA was prepared by dissolving 1.096 g NNA in 1 L warmed saline and infused through a 0.22-μm syringe tip filter. NNA (100 μM) was added to plates 1 h before the transfection.

Immunoblotting

Protein samples (20 μg) were loaded on 12% sodium dodecyl sulfate–polyacrylamide gel for electrophoresis and then transferred to the PVDF membrane. Membranes were blocked with 5% (w/v) milk in Tris-buffered saline -T buffer (10 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween-20) for 1 h and incubated with primary antibodies for 16 h at 4°C. The primary antibodies are as follows: SOD1 (1 : 2000; Santa Cruz Biotechnology, Santa Cruz, CA, USA), iNOS (1 : 1000; Santa Cruz Biotechnology), PDI (1 : 1000, Cedarlane). β-actin was used as a loading control (1 : 2000; Santa Cruz Biotechnology). Blots were washed three times in Tris-buffered saline-T buffer then probed with horseradish peroxidase -conjugated goat anti-rabbit or goat anti-mouse antibodies at 1 : 2500 for 1 h at 22°C, and then developed using chemiluminescence reagents (PerkinElmer, Waltham, MA, USA). Quantification of band intensities was performed by densitometric analysis using quantity one (Bio-Rad Laboratories, Hercules, CA, USA).

Statistical analysis

All data were tested using one-way analysis of variance (anova) with Tukey's post hoc test. A p-value of 0.05 or less was judged to be significant and results were expressed as mean ± SEM.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

iNOS protein levels were up-regulated in the spinal cords of mutant SOD1 transgenic mice and increased with disease progression

Spinal cord extracts from non-transgenic mice, wild-type SOD1 transgenic mice, and mutant SOD1 transgenic mice (G93A and G37R) were analyzed by western blotting and probed with iNOS antibodies. A band of immunoreactivity at 130 kDa was detected and it was consistent with the molecular weight of full-length iNOS protein (Heneka and Feinstein 2001; Mungrue et al. 2003). Both G93A and G37R mutant SOD1 transgenic mice had greater levels of iNOS expression in the spinal cords compared with the non-transgenic mice and wild-type SOD1 transgenic mice (Fig. 1a). In non-transgenic and wild-type SOD1 transgenic mice, iNOS levels varied to some extent in different individual mice, probably because of individual variability between the mice. However, statistic analysis of data of three independent experiments using six mice in each group showed no significant difference in iNOS expression in the spinal cords of both non-transgenic mice and wild-type SOD1 transgenic mice (Fig. 1a). Furthermore, the iNOS levels in mutant transgenic mice were increased during disease progression. The levels of iNOS expression in G93A and G37R mutant SOD1 transgenic mice at disease end-stage reached an optical density of almost 3- and 2-folds, respectively, compared with the average density from mice at the pre-symptomatic stage (Fig. 1c and d). However, there were no significant changes in iNOS expression levels in both non-transgenic mice and wild-type SOD1 transgenic mice in the course of development (Fig. 1b).

image

Figure 1. Inducible NOS (iNOS) protein levels in the spinal cords of non-transgenic and superoxide dismutase (SOD1) transgenic mice. (a) Full-length iNOS protein is more enriched in the spinal cords of mutant SOD1 transgenic mice (both G93A and G37R) than in the age-matched non-transgenic mice and wild-type SOD1 transgenic mice (6 m). (b) iNOS expression does not change during the lifetime of non-transgenic mice and wild-type SOD1 transgenic mice. (c and d) iNOS expression in the spinal cords of G93A and G37R mutant transgenic mice is increased with the disease progression and reaches a peak at disease end-stage. Values represent the mean ± SE of three independent experiments; *< 0.05; **< 0.01 by one-way anova with Tukey's post hoc test.

Download figure to PowerPoint

NO concentrations were increased in the spinal cords of mutant SOD1 transgenic mice and the NO production was up-regulated during disease progression

To measure the tissue concentrations of NO (measured nitrite concentration by Griess reagent), spinal cord extracts were prepared as described in the materials and methods section. The concentration of NO was significantly increased in mutant SOD1 transgenic mice (both G93A and G37R), as compared with the wild-type SOD1 transgenic mice and non-transgenic mice (Fig. 2a). Both non-transgenic mice and wild-type SOD1 transgenic mice did not produce significant alterations in the levels of NO in the course of development (Fig. 2b). Furthermore, with the progression of disease, the NO production was increased in the spinal cord tissues of mutant SOD1 transgenic mice, and reached a peak at disease end-stage (Fig. 2c and d). This up-regulation of NO levels as the disease progresses was significant. This change correlated well with the increase in iNOS expression during disease progression. Compared with the wild-type SOD1 transgenic mice and non-transgenic mice, the mutant SOD1 transgenic mice had a higher net amount of NO production, which was probably because of the enhanced inflammation and gliosis with a concomitant increasing in iNOS expression and iNOS-derived NO generation in the absence of nNOS because of motor neuron death.

image

Figure 2. Nitric oxide (NO) levels in the spinal cords of non-transgenic and superoxide dismutase (SOD1) transgenic mice. (a) NO increases significantly in the spinal cords of mutant SOD1 transgenic mice (both G93A and G37R) compared with that of the age-matched non-transgenic mice and wild-type SOD1 transgenic mice (6 m). (b) NO production remains unchanged during the lifetime of non-transgenic mice and wild-type SOD1 transgenic mice. (c and d) NO generation in the spinal cords of G93A and G37R mutant SOD1 transgenic mice increases significantly at disease end-stage, compared with the pre-symptomatic and onset stages. Given are the mean ± SE from five independent measurements; *< 0.05; **< 0.01; ***< 0.001 by one-way anova with Tukey's post hoc test.

Download figure to PowerPoint

PDI was increased and S-nitrosylated in the spinal cords of mutant transgenic mice, and up-regulation and S-nitrosylation of PDI developed as disease progresses

Quantitative western blot analysis was employed to detect the PDI expression in the spinal cords of the mice model of ALS. PDI expression was nearly equivalent in the spinal cords of non-transgenic mice and wild-type SOD1 transgenic mice. However, in both G93A and G37R mutant SOD1 transgenic mice, PDI levels were significantly up-regulated, compared with the non-transgenic and wild-type SOD1 transgenic mice (Fig. 3b). PDI expression increased gradually with disease progression, and peaked at disease end-stage in mutant SOD1 transgenic mice, but the non-transgenic and wild-type SOD1 transgenic mice did not have variable expression of PDI in the course of development (Fig. 3c). The PDI expression levels in the spinal cords of G93A mutant SOD1 transgenic mice was nearly 3-fold greater at the end-stage of ALS than that at the pre-symptomatic stage (Fig. 3d). Similarly, PDI in the spinal cords of G37R mutant SOD1 transgenic mice at end-stage revealed a significant increase compared with that at the pre-symptomatic stage (Fig. 3e).

image

Figure 3. Total protein disulfide isomerase (PDI) and S-nitrosylated PDI (SNO-PDI) levels in the spinal cords of non-transgenic and superoxide dismutase (SOD1) transgenic mice. (a) Without ascorbate or biotin-HPDP treatment, there is no SNO-PDI. (b) PDI is up-regulated and S-nitrosylated in the spinal cords of mutant SOD1 transgenic mice (both G93A and G37R at 6 m). (c) There are no changes in PDI expression and S-nitrosylation of PDI during the lifetime of non-transgenic mice and wild-type SOD1 transgenic mice. (d and e) PDI is increased and S-nitrosylated with disease progression in both G93A and G37R SOD1 mutant transgenic mice. SNO-PDI level is much higher at disease end-stage than pre-symptomatic and onset stages, and this difference is significant in densitometric quantitation. Data are presented as mean ± SE; *< 0.05; **< 0.01; ***< 0.001 by one-way anova with Tukey's post hoc test.

Download figure to PowerPoint

We investigated whether aberrant generation of NO was through activation of iNOS-mediated S-nitrosylation of PDI in mutant SOD1 transgenic mice. Using a biotin-switch assay, we detected that S-nitrosylation of PDI occurred in ALS (Fig. 3b). The specificity of the biotinylation reaction was confirmed by almost no detection of S-nitrosylated PDI in the samples without treatment of ascorbate. Ascorbate is required to enhance the chemical decomposition of nitrosothiol groups required for reaction with the biotinylating reagent biotin-HPDP (Jaffrey et al. 2001). In addition, there was no detection of S-nitrosylated PDI in the absence of biotin-HPDP (Fig. 3a). Despite up-regulation of total PDI in mutant SOD1 transgenic mice, S-nitrosylated PDI levels were abundant in spinal cord tissue from end-stage G93A (Fig. 3d) and G37R (Fig. 3e) mutant SOD1 transgenic mice. However, SNO-PDI was virtually undetectable in the pre-symptomatic stage. This trend of SNO-PDI level consisted with the change of iNOS expression and NO level during the disease progression. In addition, barely any SNO-PDI was found in non-transgenic mice and wild-type SOD1 transgenic mice. To rule out the possibility that the detectable SNO-PDI in mutant SOD1 transgenic mice was because of the up-regulation of total PDI expression, we deliberately increased total protein loading to enhance total PDI level in wild-type SOD1 mice group (Fig. 3a). However, we could not detect the presence of SNO-PDI in this group. Furthermore, in the mutant SOD1 transgenic mice group, with less manipulated total PDI levels because of less total protein loading, had detectable SNO-PDI. These data demonstrate that NO-mediated S-nitrosylation of PDI is a common feature of the mutant SOD1-linked transgenic mice (Turner and Talbot 2008), which is the most widely accepted mouse model of ALS.

Mutant SOD1 induced up-regulation of iNOS expression and NO generation

To investigate whether mutant SOD1 expression would affect iNOS expression and NO generation, the human neuroblastoma SH-SY5Y cells were transiently transfected with vectors encoding either SOD1wt or mutant SOD1G93A tagged with EGFP. Control cells were not transfected with plasmids (Fig. 4a). Forty-eight hours after transfection, cells were harvested to determine the iNOS expression and NO production. Immunoblotting of cell lysates indicated that iNOS expression was unchanged in cells transfected with SOD1wt, as compared with the control cells. In contrast, mutant SOD1G93A expression in cells resulted in a significant increase in iNOS expression compared with both control cells and cells expressing SOD1wt (Fig. 4b). Furthermore, cells expressing mutant SOD1G93A had much higher NO generation than control cells and cells with SOD1wt expression (Fig. 4c). This result indicates that mutant SOD1 expression in cells triggers nitrosative stress, which results in increased iNOS expression and iNOS-derived NO generation.

image

Figure 4. The inducible NOS (iNOS) expression and nitric oxide generation in SH-SY5Y cells expressing superoxide dismutase (SOD1)wt or mutant SOD1G93A. (a) Cells transfected with SOD1wt or SOD1G93A have equivalent hSOD1-EGFP expression, compared with control cells without transfection. (b) SH-SY5Y cells expressing SOD1G93A have increased iNOS expression compared with control cells and cells expressing SOD1wt. (c) Nitric oxide production from cells with SOD1G93A expression is much higher than that from control cells or cells expressing SOD1wt. Data are presented as mean ± SE; *< 0.05; **< 0.01; ***< 0.001 by one-way anova with Tukey's post hoc test.

Download figure to PowerPoint

Exposure SH-SY5Y cells to SNOC triggered the formation of SNO-PDI, mutant SOD1G93A expressing in SH-SY5Y cells induced SNO-PDI formation, which was blocked by the NOS inhibitor NNA

To investigate whether exogenously generated NO can induce S-nitrosylation of PDI, NO donor SNOC was used as a reagent to transfer NO+ to cysteine thiols (Cys-SH) of the PDI's redox modulator sites. SH-SY5Y cells were exposed to freshly prepared SNOC in various concentrations and decayed (old) SNOC and then subjected to the biotin-switch assay. Exposing the cells to SNOC resulted in SNO-PDI formation. Further increase in NO release from SNOC led to a marked increase in abundance of SNO-PDI. SNOC behaved in a dose-dependent manner (Fig. 5a). However, the control cells subjected to old SNOC did not exhibit S-nitrosylation of PDI. Using the same conditions, we found that mutant SOD1G93A expression induced SNO-PDI formation. Formation of S-nitrosylated PDI was observed only in cells expressing mutant SOD1G93A, not in the control cells or cells with SOD1wt expression; this reaction was suppressed by the NOS inhibitor NNA, which dramatically reduced the SNO-PDI formation. SNOC reinforced the SNO-PDI formation, as SNOC exposure promoted the SNO-PDI levels in mutant SOD1G93Aexpressing cells (Fig. 5b). This result suggests that mutant SOD1G93A induces S-nitrosylation of PDI probably results from up-regulation of iNOS expression and NO generation, which can be manipulated by introducing NO donors or NOS inhibitors.

image

Figure 5. Exogenous NO donor and mutant superoxide dismutase (SOD1)G93A expression induce S-nitrosylation of protein disulfide isomerase (PDI) in SH-SY5Y cells. (a) Old S-nitrosocysteine (SNOC) has no effect on the S-nitrosylated PDI (SNO-PDI) formation; freshly prepared SNOC induces S-nitrosylation of PDI. (b) SH-SY5Y cells expression SOD1G93A have SNO-PDI formation, not in cells expressing SOD1wt. Exposing the cells with SOD1G93A expression to SNOC (200 μM) promotes the SNO-PDI formation; N-nitro-l-arginine (NNA) (100 μM) suppresses the SNO-PDI formation. Data are presented as mean ± SE; *< 0.05; **< 0.01; ***< 0.001 by one-way anova with Tukey's post hoc test.

Download figure to PowerPoint

SNOC promoted and NNA suppressed the mutant SOD1 aggregates formation in transfected SH-SY5Y cells

To assess the effects of SNOC on mutant SOD1 aggregates, transfected SH-SY5Y cells were exposed to SNOC. The cells expressing SOD1wt and the control cells showed a widespread fluorescence in the cytoplasm, whereas in the cells transfected with mutant SOD1G93A, large and prominent cytoplasmic protein inclusions were observed (Fig. 6a). The proportion of cells containing inclusions was quantified. In the presence of SNOC, the mutant SOD1G93A formed inclusions with increased frequency (Fig. 6b). The effect of SNOC was also monitored using a different assay for protein aggregation: detergent insolubility that is usually used for aggregated proteins including mutant SOD1G93A. Mutant SOD1G93A was consistently found to be enriched in the pellet fraction, whereas SOD1wt was found in the soluble fraction (Fig. 6a). When the transfected cells were exposed to SNOC, there was a statistically significant increase in the amount of insoluble SOD1G93A in the pellet (Fig. 6c). These results indicate that SNOC induces further formation of inclusion or aggregates of SOD1G93A in the cells, which probably operate through the mechanism of S-nitrosylation of PDI. To further elucidate the importance of iNOS activation in SNO-PDI-related SOD1 aggregation, NNA was added to the medium on the day of transfection. NNA treatment inhibited the formation of SOD1G93A inclusions in transfected cells (Fig. 6b). Furthermore, the level of insoluble SOD1G93A in the pellet was reduced after NNA treatment, which inhibited NO generation (Fig. 6c).

image

Figure 6. The effect of S-nitrosocysteine (SNOC) and N-nitro-l-arginine (NNA) on the SOD1 aggregates formation in transfected SH-SY5Y cells. (a) SH-SY5Y cells expressing SOD1G93A have inclusions in the cytoplasm. EGFP-hSOD1 is clearly detectable in the pellet fraction of SH-SY5Y cells expressing SOD1G93A. (b) Exposing SOD1G93A transfected cells to SNOC leads to a statistically significant increase in inclusion formation compared with the transfected cells without SNOC treatment; NNA treatment suppress inclusion formation in SOD1G93A transfected cells. (c) SNOC exposure increases EGFP-hSOD1 expression in pellet fractions of SOD1G93A transfected cells; NNA treatment decreases EGFP-hSOD1 level in the insoluble pellet fraction of transfected cells. Data are presented as mean ± SE; *< 0.05; **< 0.01;***< 0.001 by one-way anova with Tukey's post hoc test.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

There are three subtypes of NOS in the nervous system. The two constitutive forms of NOS-neuronal and endothelial NOS are mainly sources of NO production. The third subtype-iNOS as its name indicates is induced by acute inflammatory stimuli. iNOS is induced mainly by activated astrocytes and microglia cells in various neurodegenerative diseases. Once iNOS is expressed, it produces neurotoxic amounts of NO chronically, without any requirement for further activation (Saha and Pahan 2006). In ALS, increased tissue levels of NO in the lumbar spinal cord of mutant ALS transgenic mice are mainly produced by iNOS located in astrocytes, seeing as nNOS-positive motor neurons are depleted, while iNOS-positive activated glial cells are increased in ALS mice (Lee et al. 2009). Furthermore, pharmacological inhibition of iNOS shows the significant effects of delaying disease-onset and extending survival in ALS mice (Chen et al. 2010). The findings of our study demonstrate that iNOS protein expression is up-regulated in the spinal cords of both G93A and G37R SOD1 mutant transgenic mice compared with non-transgenic mice and wild-type SOD1 transgenic mice. In addition, the level of iNOS increased during the progression of disease, and it reached a peak at the end-stage of ALS. These observations demonstrate that iNOS is involved in the causal mechanisms of motor neuron degeneration in ALS mice. The role of iNOS in the pathogenesis of ALS is less clear (than nNOS), probably because study on its effect is very controversial. For example, iNOS-knockout mice are protected against neuronal injury in several models of neurodegenerative diseases, including Alzheimer's disease (Nathan et al. 2005) and Parkinson's disease (Dehmer et al. 2000). In ALS, the G93A-high-mutant SOD1 mouse, iNOS gene deletion has a significantly prolonged survival (Martin et al. 2007). However, a recent study of three different strains of iNOS-knockout showed no effect of iNOS on the infarct size after transient focal brain ischemia (Pruss et al. 2008). This may suggest that iNOS is more likely to be toxic to chronic neurodegeneration, rather than to acute neuronal injury (Brown 2010). We focused on iNOS because earlier studies on ALS mice (Almer et al. 1999; Sasaki et al. 2001; Martin et al. 2007) and human patients (Phul et al. 2000) have indicated that this isoform of NOS could be vital in ALS, and its role in the neuropathology of ALS has been under appreciated. The importance of iNOS comes from its properties that are different from endothelial NOS and nNOS. Homodimeric iNOS is catalytically active when expressed and it is active for extended periods with a maximum of 10-fold greater than other NOS isoforms, generating a very high output of NO (Lowenstein and Padalko 2004).

We also studied the concentrations of NO in the spinal cord extracts in mutant SOD1 transgenic mice (both G93A and G37R), wild-type SOD1 transgenic mice, and non-transgenic mice. As predicted, compared with the wild-type SOD1 transgenic mice and non-transgenic mice, the mutant SOD1 transgenic mice had a higher net amount of NO production, which was probably because of the increase in inflammation and gliosis with a concomitant up-regulation of iNOS expression. Furthermore, NO concentration was significantly increased in the spinal cord extracts of mutant SOD1 transgenic mice at end-stage of ALS. The up-regulation of NO level during disease progression presumably resulted from activation of astrocytes and microglia, which induced iNOS up-regulation in the spinal cords of mutant SOD1 transgenic mice. The unaltered NO levels in both non-transgenic mice and wild-type SOD1 transgenic mice indicated that the constitutive level of NO was probably because of nNOS residued in motor neurons. Our results imply an interesting connection between robust iNOS expression and excessive NO generation in mutant transgenic mice at end-stage of ALS. Up-regulation of iNOS may account for the large amount of NO generation at disease end-stage.

Nitric oxide mediates cellular signaling pathways that regulate broad aspects of physiological processes. NO has been implicated in neurotransmission, synaptic plasticity, and neuromodulation in CNS (Garthwaite and Boulton 1995). Excessive generation of NO and its derivatives has also been implicated in the pathogenesis of neurodegenerative disorders (Calabrese et al. 2000). For example, high levels of NO induce neuronal death by causing inhibition of mitochondrial cytochrome oxidase in neurons (Brown and Cooper 1994). The inhibition of neuronal respiration leads to depolarization and glutamate release, followed by excitotoxicity via the NMDA receptor (Jekabsone et al. 2007; Golde et al. 2002). Normally, NO mediates the physiological and pathophysiological effects via stimulation of guanylate cyclase to form cyclic guanosine – 3′,5′-monophosphate (cGMP) or through S-nitrosylation of regulatory protein thiol groups (Isaacs et al. 2006). S-nitrosylation involves the covalent addition of a NO group to a critical cysteine thiol/sulfhydryl to form an S-nitrosothiol derivative. This S-nitrosylated modification can influence the function of a broad spectrum of proteins as well as the protein–protein interaction (Hara et al. 2005). Our studies have found that PDI is S-nitrosylated in spinal cord tissues of end-stage mutant SOD1 transgenic mice (both G93A and G37R). However, the S-nitrosylated PDI was virtually undetectable in our pre-symptomatic stage mice. In addition, in non-transgenic mice and wild-type SOD1 transgenic mice, there was no S-nitrosylation of PDI. This finding suggests that S-nitrosylation of PDI probably inactivates the normal properties of PDI and it may contribute to the pathogenesis of ALS. Consistent with this result, we also found that PDI expression levels at the end-stage of disease were significantly up-regulated in spinal cords of both G93A and G37R SOD1 mutant transgenic mice as compared with the non-transgenic and wild-type transgenic mice. Similarly, a previous study confirmed significant up-regulation of PDI expression at disease end-stage in both G93A mutant mice and rats as compared with that of the non-transgenic controls (Atkin et al. 2006). However, there are some varieties of PDI expression at different disease stages in mutant transgenic rats relative to the corresponding non-transgenic rat. Ferri and colleagues found that PDI expression was up-regulated at the pre-symptomatic stage (8 weeks), but at 16 weeks and end stage, PDI levels were not significantly higher in the spinal cords of G93A mutant transgenic rats as compared with non-transgenic rats (Ahtoniemi et al. 2008). Another study showed that PDI levels in CSF of G93A mutant transgenic rats were most prominently elevated at the disease onset stage when compared with non-transgenic rats (Atkin et al. 2008). All these studies set the age-matched non-transgenic mice or rats as control. To examine the change of PDI expression during disease progression, absolute PDI levels were quantified as a ratio relative to the corresponding β-actin anounts at each stage. We found that the PDI level increased gradually with the progression of disease and peaked at disease end-stage in both G93A and G37R mutant transgenic mice. Even though the up-regulation of PDI represents an adaptive response to provide potential neuroprotection (Conn et al. 2004; Hetz et al. 2005; Ko et al. 2002; Tanaka et al. 2000), NO-mediated S-nitrosylation of PDI probably affects its normal function and promotes aggregate formation. The formation of SOD1 aggregates in ALS was probably through destabilizing mutant SOD1 with reactive reduced cysteine residues (Ahtoniemi et al. 2008). Cysteine residues are crucial for SOD1 stability, and a non-physiological intermolecular disulfide bond between cysteine 6 and 111 in mutant SOD1 was found to be associated with mutant SOD1 aggregates' formation (Niwa et al. 2007a). PDI may get involved in rearranging the cysteine residues of SOD1, and S-nitrosylation of PDI toward the end-stage of ALS may facilitate the formation of non-physiological disulfide bonds and promote SOD1 aggregates formation. This could be a possible explanation to the finding that cysteine-reduced SOD1 levels increased with disease progression and reach the peak at end-stage in G93A mutant transgenic rats (Ahtoniemi et al. 2008). Furthermore, an association of PDI and mutant SOD1 aggregates was identified, seeing as high levels of PDI that recruit to abnormal inclusions were observed in both G93A transgenic mice and ALS patients (Atkin et al. 2006, 2008). In addition, PDI was found to be co-localized with SOD1 in neuronal cytoplasmic inclusions (Honjo et al. 2011). Given the chaperone activity of PDI, it is possible that PDI may interact with insoluble mutant SOD1 to form abnormal inclusions; this may partially explain the findings of other studies that PDI levels were not most prominently up-regulated at the disease end-stage in mutant transgenic rat when compared with non-transgenic mice (Atkin et al. 2008; Ahtoniemi et al. 2008). A greater portion of PDI may be associated with mutant SOD1 in abnormal inclusion at end-stage of ALS, and then it cannot be detected in the soluble fraction of tissue lysates.

Protein disulfide isomerase is a ubiquitous, highly conserved redox enzyme from the thioredoxin superfamily, and it is mainly located in the ER (Noiva 1999). During protein folding in the ER, PDI facilitates proper protein folding and helps to maintain the structural stability of the mature protein (Lyles and Gilbert 1991). As a consequence, PDI is considered a molecular chaperone capable of stabilizing the correct folding of substrate proteins. It also facilitates ER-associated degradation of misfolded proteins (Lee et al. 2010). Through interacting with the ER transmembrane protein Derlin-1, PDI is involved in retro-translocation of misfolded cholera toxin from the ER to the cytoplasm. (Moore et al. 2010). As we know, aberrant protein folding and further protein aggregates are associated with various neurodegenerative diseases, including ALS. The accumulation of misfolded protein in the ER results in ER stress that triggers the protective unfolded proteins response (UPR). The UPR entails the induction of chaperone molecules, the degradation of misfolded proteins, and inhibition of protein translation (Zhang and Kaufman 2006). Prolonged ER stress can nonetheless lead to activation of apoptosis (Xu et al. 2004). Studies involving pancreatic β cells, macrophages (Gotoh and Mori 2006), and cerebellar granule cells (He et al. 2004) have demonstrated that NO can induce ER stress. However, the molecular basis remains unclear. Furthermore, although the involvement of NO in neurodegeneration has been widely accepted, the chemical relationship between nitrosative stress and formation of protein aggregates has remained obscure. Our findings indicate that S-nitrosylation of PDI may hold some of the answers to these questions. Studies have shown that excitotoxic activation of nNOS leads to excessive NO generation, which causes S-nitrosylation of the active-site thiols of PDI, inhibiting its isomerase and chaperone activities (Uehara et al. 2006). In this regard, NO blocks the protein's protective effect via S-nitrosylation of PDI, which leads to accumulation of misfolded and polyubiquitinated proteins, resulting in prolonged UPR activation, and thus persistent ER stress, which induces apoptosis. Our study showed that, when SH-SY5Y cells were expressing mutant SOD1G93A, iNOS expression was significantly increased in compared with cells expressing SOD1wt. This finding was relatively inconsistent with the previous study that had showed expressing SOD1wt or mutant SOD1G93A in NSC-34 cells did not alter iNOS expression, which remained at a constitutive level. This variation may be because of the controlled SOD1wt or mutant SOD1G93A expression in NSC-34 cells. In that study, the cell clones expressing low and similar amounts of various SOD1 proteins upon induction. The induced human SOD1 expression was lower than murine endogenous SOD1 expression; this low expression of human expression may not be extensive enough to trigger the possible correlation between mutant SOD1 expression and iNOS expression. Furthermore, the increased iNOS level in mitochondria has been observed in G93A mutant transgenic mice and the deletion of iNOS gene significantly extended the survival of mice. However, the activity of iNOS is more crucial to its biological role in ALS pathogenesis. The effect of mutant SOD1 in alteration of iNOS activity was further investigated. We found that cells expressing mutant SOD1G93A had much higher NO generation, and this increased NO generation could be reversed by the use of NNA (non-selective inhibitor of NOS) (data not shown). Consequently, ALS-linked SOD1 is highly associated with up-regulated iNOS expression and increased iNOS-derived NO generation. However, another enzyme responsible for NO production, nNOS, was found to be down-regulated in SH-SY5Y cells with mutant SOD1G93A expression in a study (Aquilano et al. 2003). The down-regulation of nNOS-derived NO production was also evidenced. It has been suggested that NO released from nNOS activity keeps iNOS inhibited under normal conditions. However, under pathophysiological conditions, down-regulation of nNOS is a necessary condition to promote the iNOS expression and the release of large amounts of NO (Qu et al. 2001). For example, in a rat model of inflammatory bowel disease, nNOS down-regulation could induce iNOS over-expression (Porras et al. 2006). Our results could be in agreement with the hypothesis that down-regulation of nNOS is associated with up-regulation of iNOS in SH-SY5Y cells expressing mutant SOD1G93A. However, the nNOS expression and the mechanism for the association between iNOS and nNOS will certainly deserve further investigation. We introduced the physiological NO donor, SNOC to provoke S-nitrosylation of PDI. The dose-dependent manner of SNOC-induced S-nitrosylation of PDI confirmed that PDI was S-nitrosylated by NO-related species. The role of PDI in protecting against mutant protein aggregation in ALS is supported by various studies. For example, when mutant SOD1 expressing NSC-34 cells were treated with bactracin, an inhibitor of PDI, the formation of SOD1 inclusions was increased (Atkin et al. 2006). Another example was siRNA-mediated knockdown of PDI resulted in increased formation of mutant SOD1 inclusion in neuroblastoma cells (Walker et al. 2010). However, the PDI up-regulation in ALS was not enough to protect against mutant SOD1 aggregate formation, as S-nitrosylation of PDI may affect its enzymatic activity and promote the aggregates formation. We found that NO-mediated S-nitrosylation of PDI was probably involved in the formation of mutant SOD1 aggregates, seeing as NO donor SNOC-induced S-nitrosylation of PDI in a dose-dependent manner. Furthermore, the inclusions in mutant SOD1G93A expressing cells were enhanced by SNOC. Our studies also showed that, with the use of NOS inhibitor NNA to suppress S-nitrosylation of PDI, the SOD1 inclusions were inhibited, and the level of insoluble mutant SOD1 in the pellet fraction was decreased as well. This finding indicates that in the cell model of ALS, manipulating the NO levels by using a NO donor or NOS inhibitor, affects the formation of SNO-PDI. S-nitrosylation of PDI may highly correlate with mutant SOD1 aggregates formation. As S-nitrosylation of PDI could inhibit its chaperone activity in rearrangement of protein folding, allow the misfolded proteins to accumulate, and finally contribute to neuronal cell death. In addition, the subcellular redistribution of PDI has recently been implicated in the pathogenesis of ALS. Studies have demonstrated that the reticulon family of proteins can modulate PDI distribution. They found that reticulon over-expression causes a change of PDI localization, from a normal ER distribution to a less homogenous punctuate pattern (Yang et al. 2009). In ALS mice, knocking down the expression of the reticulon-4A, B proteins accelerate disease processes, possibly resulting from the prevention of reticulon-mediated PDI redistribution (Yang et al. 2009). Furthermore, co-localized inclusions of PDI with mutant SOD1, and TAR DNA-binding protein 43 kDa (TDP-43) have been found in ALS patients (Honjo et al. 2011). Under cellular stress, PDI may leave the ER and then accumulate with SOD1 or TDP-43 in the cytosol. PDI also accumulates in the swollen neuritis, the disturbance of axon transport was probably because of the loss of PDI function (Honjo et al. 2011). Another ALS-related protein, fused in sarcoma (FUS), is found to be associated with PDI. Mutant FUS inclusions in human ALS lumbar spinal cords are co-localized with PDI. As FUS contains cysteine residues, similar to mutant SOD1 and mutant TDP-43, it may physically interact with PDI, and the chaperone function of PDI may have a protective role in refolding misfolded FUS protein. Overall, these findings indicate that post-translational modifications and subcellular redistribution of PDI are involved in regulation of PDI's function in ALS, with potential implications for disease pathogenesis.

In summary, this study examined the propensity of S-nitrosylated PDI contributing to the accumulation of mutant SOD1 aggregates in ALS. As excessive production of NO derived from iNOS up-regulation is thought to be a contributing factor in ALS, our elucidation of an NO-mediated pathway to dysfunction of PDI by S-nitrosylation provides a mechanistic link between the production of free radicals and aberrant protein accumulation in ALS. The outcome of this study shall open up new therapeutic approaches to prevent aberrant protein misfolding by targeted-prevention of nitrosylation of specific proteins such as PDI in the future. Moreover, enhancing the action of PDI may represent a novel strategy for the treatment of ALS.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This work was supported by CIHR, the Muscular Dystrophy Association (MDA) USA and the National Natural Science Foundation of China (U0632007). The authors have no conflict of interest to declare.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References