Background
Parkinson’s disease (PD) is characterized by a progressive loss of dopaminergic neurons in the
substantia nigra pars compacta and intracellular protein inclusions termed Lewy bodies whose main component is α-synuclein [
1,
2]. Several point mutations have been reported in α-synuclein: A53T, A30P, E46K and H50Q, all of which result in familial forms of PD [
3‐
7]. Therefore α-synuclein plays a key role in the pathogenesis of PD. The function of α-synuclein is complex involving the regulation of neurotransmitter release, exocytosis and trafficking of synaptic vesicles and co-chaperone activity [
8‐
12]. α-Synuclein physiologically exists as an unfolded or membrane-bound monomer but is capable of forming oligomers, fibrils and finally inclusion bodies under pathological conditions [
2,
13]. Notably, increasing evidence points to α-synuclein oligomers rather than fibrils as the toxic species leading to neurodegeneration [
14‐
17].
ALS is another neurodegenerative disease that is characterized by a progressive loss of the upper and lower motor neurons resulting in spasticity and paresis. SOD1 is directly associated with a familial form of ALS with more than 100 different mutations in the SOD1 linked to ALS [
18,
19]. SOD1 physiologically dimerizes and forms a non-covalently bound homodimer catalyzing the oxidation of O
2˙
− to H
2O
2 or O
2 [
20]. Like α-synuclein, SOD1 pathologically aggregates forming soluble oligomers, insoluble fibrils and inclusion bodies [
21‐
23].
The co-occurrence of ALS and PD has been reported on Guam and in the Kii Peninsula of Japan and several cases have also been described apart from the Western Pacific [
24‐
30]. Additionally, extrapyramidal symptoms due to nigrostriatal dysfunction have been reported in ALS patients [
31,
32] indicating that PD related pathological features may play a role in ALS. Indeed, several studies suggest an involvement of α-synuclein in ALS, for instance prominent phosphorylated α-synuclein inclusions were found in ALS-PD complex in Kii Peninsula [
33]. In addition to the Kii Peninsula, other ALS cases with α-synuclein inclusions have been reported [
25,
34,
35]. Furthermore, increased α-synuclein expression was identified in glial cells and in spheroids of the spinal cord of ALS patients and in the anterior horn cells of the spinal cord, in the hippocampus and cerebellum of mice expressing mutated SOD1 (G93A) [
36,
37].
Although the literature, mentioned above, suggest an involvement of α-synuclein in ALS and a few studies have found that α-synuclein and SOD1 co-localize in the same protein aggregates [
34,
38], hardly anything is known about a possible molecular α-synuclein-SOD1 interaction.
Therefore we asked whether SOD1 and α-synuclein directly interact and influence their respective oligomerization processes. To address these questions, we used human material, a mouse model for PD, and a PD cell culture model to determine the molecular interaction of α-synuclein and SOD1 and the impact of disease associated mutants of either α-synuclein or SOD1 on the interaction. Furthermore, using an α-synuclein or SOD1 protein complementation assay, we explored whether α-synuclein or SOD1 directly influence their oligomerization characteristics and exert cross-seeding activities.
Discussion
This study provides evidence for a novel interaction between α-synuclein and SOD1. Using a
Gaussia luciferase protein fragment complementation assay and immunoprecipitation approaches, we found α-synuclein-SOD1 interaction in α-synuclein and SOD1 transfected cells and in conditioned medium, but also in mouse brain tissue and human erythrocytes. Our results from the
Gaussia luciferase protein-fragment complementation assay show that there is a high degree of α-synuclein-SOD1 interaction which is comparable to α-synuclein oligomerization (S1 + S2) indicating that α-synuclein-SOD1 binding might have a comparable binding affinity and kinetic to α-synuclein oligomerization. Since endogenous SOD1 might compete with hGluc tagged SOD1 for the binding to hGluc tagged α-synuclein, the SOD1 and α-synuclein interaction might be even more pronounced in vivo. Furthermore, our microscopy data support the finding of α-synuclein-SOD1 interaction by detecting a co-localization of SOD1 and α-synuclein in wt mouse and human brain tissue. The interaction of α-synuclein and SOD1 could also be proven with co-IPs of co-transfected cells, mouse brains of wt and Thy1-α-synuclein tg mice and human erythrocytes. These findings suggest that the α-synuclein-SOD1 interaction might take place in all organs co-expressing α-synuclein and SOD1. Since SOD1 is ubiquitously expressed and α-synuclein occurs in all organs tested so far except the liver [
47,
48], the α-synuclein-SOD1 binding might take place in other organs, too.
As a first step towards understanding the molecular consequence of the α-synuclein and SOD1 interaction and its relevance to human diseases, we investigated if PD related mutations in α-synuclein and ALS related mutations in SOD1 modify the α-synuclein-SOD1 interaction. The results of the luciferase assay show that familiar A30P and A53T α-synuclein mutations have reduced interaction capabilities with SOD1 compared to wt α-synuclein in living cells and conditioned medium. The extracellular complementation assay where the same molarity of mutated and wt hGluc tagged α-synuclein is used confirms the reduced interaction of familiar A30P and A53T α-synuclein mutations to SOD1. In contrast to familial α-synuclein mutation, our co-IP studies show that G85R and G93A SOD1 known to cause familial ALS have a higher tendency to bind α-synuclein compared to wt SOD1. As expression levels of G85R, G93A and wt SOD1 are not equal, we considered co-IP as an appropriate method to examine the influence of disease related mutations in SOD1. We observed that G85R and G93A SOD1 increasingly bind to α-synuclein compared to wt SOD1. Taken together, our results show that pathological mutations in α-synuclein and SOD1 alter the interaction of α-synuclein and SOD1 suggesting a relevance of the interaction of α-synuclein and SOD1 in human diseases.
In addition to our findings on a molecular interaction of α-synuclein and SOD1, we also investigated whether α-synuclein and SOD1 influence their oligomerization capabilities. Recent studies highlight cross-seeding activities where one protein accelerates the aggregation of other proteins involved in neurodegeneration. Interestingly, we found that α-synuclein accelerates SOD1 dimerization. Consequently, reduced SOD1 dimerization was detected in α-synuclein knockdown cell lines. SOD1 homodimers are enzymatic active whereas aggregates of SOD1 seem to have a reduced enzyme activity [
49]. Since we did not find an influence of α-synuclein on the enzymatic activity of SOD1, we conclude that α-synuclein increases the SOD1 oligomerization and higher molecular weight species rather than solely SOD1 dimerization.
Seeding effects of α-synuclein to other aggregation prone proteins by a prion-like principle is in accordance with other studies [
43,
44]. Notably, a cross-seeding activity for α-synuclein with other aggregating proteins has already been described: α-synuclein accelerates the aggregation of tau and amyloid-β, proteins implicated in Alzheimer’s disease (AD), and huntingtin, a protein involved in Huntington’s disease [
45,
46,
50‐
53]. The seeding activity of α-synuclein to tau has even been shown in vivo [
51]. Future in vivo studies will be needed to investigate if the α-synuclein is also able to promote SOD1 oligomerization in vivo.
Moreover, our data indicate that SOD1 oligomerization is even more enhanced in presence of A30P or A53T α-synuclein compared to wt α-synuclein. A plausible explanation would be that more unbound SOD1 is available because A30P and A53T α-synuclein bind less to SOD1 than wt α-synuclein.
Notably, SOD1 with G93A mutation, but not G85R and wt SOD1, was able to promote the α-synuclein oligomerization. Our findings could be a possible explanation why increased α-synuclein was found in spinal cord, hippocampus and cerebellum of mice expressing G93A SOD1 compared to mice expressing wt SOD1 [
37]. Interestingly, spinal cord homogenate of G93A tg mice have been shown to accelerate SOD1 aggregation formation in vitro and in vivo and to mediate the transmission of motor symptoms [
54,
55]. Thus, our study supports previous studies demonstrating seeding properties of G93A SOD1.
Although we show that α-synuclein co-localizes with SOD1 in protein aggregates and promotes the SOD1 aggregation, there are only few studies showing α-synuclein and SOD1 in the same protein inclusions [
34,
38]. It might be possible that the interaction of SOD1 and α-synuclein is disrupted in diseases or α-synuclein seeds SOD1 without being internalized into protein inclusions. Thus, future studies in a larger number of patient post-mortem material are needed to determine if α-synuclein and SOD1 co-occur in the same protein aggregates.
Methods
Plasmid generation
Fusion constructs α-synuclein-hGluc1 (S1) and α-synuclein-hGluc2 (S2) are based on pcDNA3.1-Zipper-hGluc(1) and pcDNA3.1-Zipper-hGluc(2) and have been described previously [
40,
56]. α-synuclein was replaced by SOD1 to generate SOD1-hGluc1 (SOD1-1) and SOD1-hGluc2 (SOD1-2). Fusion constructs SOD1-3 and SOD1-4 were created by cloning hGluc1 or hGluc2, respectively, with a linker to the n-terminus of SOD1 into pcDNA3.
Cell culture and transfection
H4 (human neuroglioma) were cultivated at 37 °C in 5 % CO2 in OptiMEM® or DMEM (both from Life technologies, Carlsbad, USA) supplemented with 10 % FBS (Sigma, St. Louis, USA; or Life technologies, Carlsbad, USA). Cells were transfected 24 h after plating with Superfect® (Quiagen, Chatsworth, USA) or Fugene6® (Promega, Fitchburg, USA) transfection reagent according to the manufacturers’ instructions. Cells were further cultivated at 37 °C in 5 % CO2 for 24 h.
Cell lysis
Cells were washed with PBS and scraped from dishes on ice with cold TritonX lysis buffer (150 mM NaCl, 50 mM TrisHCl pH 7,4, 0,1 % TritonX) containing protease inhibitor (Roche Diagnostics). Different lysis buffers were used for co-IP (see below). Lysate was centrifuged at 10000 g for 10 min at 4 °C and total protein concentrations of the supernatants were determined with the BCA assay (Pierce Biotechnology, Rockford, USA) according to the manufacturer’s instructions.
Gaussia luciferase protein-fragment complementation assay
The Gaussia luciferase protein-fragment complementation assay was performed either with transfected living cells in a 96-well plate, lysates of transfected cells or with conditioned medium. After transferring 100 μl of cultured medium to a new 96-well plate, living cells in 96-well plate were washed with PBS and 100 μl of a serum- and phenolred- free medium was added per well. Cell lysates were prepared as described above, adjusted to the same protein concentration and 100 μl of the adjusted cell lysates were added to each well of a 96-well plate. The luciferase activity was measured by a luminometer (Multilabel plate reader Victor3, Perkin Elmer) that automatically applied coelenterazine (Nanolight, Pinetop, USA), a cell permeable substrate of the luciferase, to a final concentration of 20 μM.
Treatment of transfected H4 cells with recombinant α-synuclein
24 h post transfection, H4 cells in 96-well plates or 6 cm dishes were washed with PBS and treated with serum- and phenolred- free medium containing 7 μM of recombinant α-synuclein (Roche, Rotkreutz, Swiss) or the same volume of the corresponding solvent (PBS). After 24 h, Gaussia luciferase protein-fragment complementation assay with living cells in 96-well plates or western blot with cells in 6 cm dishes was performed.
To detect protein-protein binding ex vivo, we utilized extracellular complementation assay with lysates from H4 cells overexpressing either α-synuclein- or SOD1-hGluc alone. The concentrations of α-synuclein and SOD1 were quantified by ELISA (Invitrogen, Carlsbad, CA, USA; Enzo life science, New York, USA, respectively) and adjusted to the same molarity. α-synuclein/SOD1 fused to either luciferase half were mixed and incubated in a spinning wheel at 4 °C for 12 h before measuring luciferase activity.
Western blotting
Proteins were separated by standard SDS-PAGE using the NuPAGE® system (Invitrogen, Carlsbad, CA, USA), followed by a transfer to PVDF membranes (Millipore, Billerica, USA). To improve the detection of endogenous α-synuclein of H4 cells, the membranes were fixed with 0,4 % PFA prior blocking as described previously [
57]. Following antibodies were used: rabbit-anti-β-actin, (1:2000, Sigma, St. Louis, USA or 1:1000, abcam, Cambridge, UK), mouse-anti-α-synuclein (4B12, 1:3000, Covance, Princeton, USA), mouse-anti-α-synuclein (LB-509, 1:1000, Covance, Princeton, USA), rabbit-anti-SOD1 (ADI-SOD-100, 1:2000, Enzo life science, New York, USA), sheep-anti-SOD1 (1:1000, Calbiochem, Farmingdale, USA), HRP coupled secondary antibodies (1:1000, Life Technologies, Carlsbad, USA, or SouthernBioTech, Birmingham, USA).
SOD1 activity gels
H4 cells were lysed in dH
2O by 3 freeze-thaw cycles and centrifuged at 15000 g for 10 min at 4 °C. 20 μg total protein per lane, combined with 5x loading buffer (62,5 mM TrisHCl pH 6,8, 30 % Glycerol, 0,05 % bromphenol blue), were separated in 12 % non-reducing Tris-Glycin gel. Gel was stained, as described previously [
42], in 2,45 mM nitro blue tetrazolium (NBT) solution for 20 min and then in developer solution (28 mM TEMED, 28 μM riboflavin, 36 mM KH
2PO
4 pH 7,4) for 15 min in darkness. Gel was illuminated until sufficient contrast between light line (SOD1 activity) and background was achieved.
Co-immunoprecipitation from H4 cells
The co-immunoprecipitation (co-IP) was performed using a adapted version of the ReCLIP method (reversible cross-link immunoprecipitation) [
58,
59]. In brief, H4 cells were washed 2 times with PBS 24 h post transfection. Then, lysine residues of the interacting proteins were crosslinked by incubation with 1 mM of the cell permeable dithiobissuccinimidylpropionate (DSP) at RT for 30 min. After incubation with quenching solution (20 mM TrisHCl pH 7,4; 5 mM L-cysteine in PBS) for 10 min and washing step with PBS, cells were lysed with NP-40 lysis buffer (150 mM NaCl, 10 mM TrisHCl pH 7,4, 1 mM EDTA, 1 mM EGTA, 0,5 % (v/v) NP-40, 5 % (v/v), complete mini protease inhibitor). Lysates were centrifuged at 10000 g for 10 min at 4 °C and adjusted to the same protein concentration with the respective lysis buffer. To remove unspecific protein binding to the beads, 20 μl Protein A and G mag sepharose Xtra magnetic beads (GE Healthcare, Chalfont St Giles, UK) were incubated with lysates for 1 h at 4 °C with end-over rotation and then separated from the lysates. Pre-cleared lysates were incubated with primary antibodies (mouse-anti-c-myc, Roche, Mannheim, Germany; mouse-anti-α-synuclein (4B12), Covance, Princeton, USA; rabbit-anti-SOD1 (ADI-SOD-100), Enzo life science, New York, USA; rabbit-anti-SOD1, gift from Dennis Dickson, Mayo Clinic Jacksonville [
60]) or, as control, with IgGs from the same species as the primary antibody’s species (ChromPure Mouse IgG and ChromPure rabbit IgG, both Jackson Immuno, West Grove, USA) with end-over rotation overnight, followed by incubation with Protein G and A mag sepharose Xtra magnetic beads for 2 h on the next day. After washing the beads 3 times with lysis buffer, bound proteins were eluted and de-crosslinked by incubating with 40 μl of NuPAGE® LDL sample buffer (Invitrogen, Carlsbad, USA) and dithiothreitol (DTT) at a final concentration of 80 mM at RT for 15 min and then boiled at 95 °C for 5 min before loading on a gel.
Co-IP from brain and erythrocytes
For co-IP with whole mouse brain homogenate, BDF1 wt mouse and Thy1-α-synuclein-tg mouse [
61] (both 7 months old, female) were anaesthetized with ketamine and a 10 % (w/v) brain homogenate was made in PBS by mechanical homogenization. Human erythrocytes were separated from mononuclear blood cells by density gradient centrifugation of EDTA-whole blood on Histopaque (Sigma Aldrich, St. Louis, USA) at 500 g for 30 min at RT. Erythrocytes and brain homogenates were incubated with 10 mM DSP for 30 min, followed by lysis with same volume of TritonX lysis buffer (100 mM NaCl, 50 mM TrisHCl pH 7,4, 5 mM EDTA, 0,3 % v/v TritonX, 5 % (v/v), complete mini protease inhibitor) for 30 min on ice. For co-IP without crosslinking, lysis buffer was added directly to homogenized mouse brain. Afterwards, brain homogenate lysates were centrifuged at 9000 g for 10 min at 4 °C and co-IP was performed as described above. In case of erythrocyte lysates, antibodies were first coupled to the magnetic beads by incubating antibodies and beads in lysis buffer with 2 % BSA overnight at 4 °C with end-over rotation. Then, the antibody coupled beads were incubated with pre-cleared lysates at 4 °C for 2 h, followed by the standard co-IP protocol.
Generation of stabile α-synuclein knockdown cell lines
Endogenous α-synuclein was knocked down by shRNA to establish two stable α-synuclein knockdown cell lines derived from H4 cells. The oligonucleotides coding for α-synuclein shRNA, TRCNOOO320 and TRCN3736 (Sigma-Aldrich, St. Louis, USA), and scrambled shRNA were cloned into pLK0 harboring puromycine resistance. Lentiviruses were produced in HEK cells after transfection of shRNA plasmid, HIV-gag/pol (psPAX2, gift from Didier Trono, Addgene plasmid # 12260) and VSVG glycoprotein (pMD2, gift from Didier Trono, Addgene plasmid # 12259) using CalPhosMammalian transfection kit (Clontech, Mountain View, USA) according to the manufacturer’s instructions. 24 h after transfection, cells were washed with PBS and fresh DMEM with 10 % FCS was added. 48 h and 72 h post transfection, medium containing lentiviruses were collected and centrifuged at 300 g for 5 min to remove floating cells. H4 cells were infected using 80 % medium containing lentiviruses, 20 % DMEM with 10 % FCS and 6 μg/mL polybrene. After 24 h of incubation, cells were washed with PBS and stably transfected cells were selected with puromycine at a final concentration of 6 μg/ml.
Immunhistochemistry of mouse brain tissue
Wt C57Bl/6 mice (1–3 months old, female) were anesthetized with ketamin/xylazin and perfused with PBS. After incubation of brains with 30 % sucrose for 24 h at 4 °C, brain tissue was embedded in TissueTek®O.C.T (Sakura, Alphen a.d.R., Nehterland), frozen at −80 °C and cut in 12 μm sections using a cryostat. Sections were fixed with 2 % PFA for 30 min, permeabilized with 0,5 % saponine for 10 min and treated with 3 % H2O2 for 10 min, followed by blocking with Roti®-ImmunoBlock (Roth, Karlsruhe, Germany) for 1 h. Samples were co-immunostained with primary antibodies mouse-anti-α-synuclein (1:200, syn-1,BD,New Jersey, USA) and rabbit-anti-SOD1 (1:100, ADI-SOD100, Enzo life science, New York, USA) in Roti®-ImmunoBlock for 1 h 45 min at RT. After washing with PBS, sections were blocked with 5 % goat-serum in PBS for 15 min and incubated with fluorophore conjugated secondary antibodies (1:750, goat-anti-rabbit-Alexa546 and goat-anti-mouse-Alexa-488, both Life technology, Carlsbad, USA) in 5 % goat serum for 1 h. Then sections were washed with PBS, incubated with xylol for 2 min and 100 % ethanol for 3 min and coverslipped using DAPI Fluoromont®G (SouthernBioTech, Birmingham, USA). To avoid unspecific binding of secondary antibodies, sections were also single stained with either α-synuclein or SOD1 primary antibody and both secondary antibodies.
Immunhistochemistry of human brain sections
Immunohistochemistry of midbrain sections of a patient with DLB was performed as described previously [
62]. In brief, 5 μm thick paraffin-embed human midbrain sections were de-paraffinized by xylol and a descending series of alcohols and subjected to antigen retrieval by steaming in dH
2O for 30 min. Sections were blocked with 5 % goat serum in PBS-T for 1 h and incubated with primary antibodies mouse-anti- α-synuclein( 1:100, 4B12, Covance, Princeton, USA) and rabbit-anti-SOD1 (1:70, [
60]) in 2,5 % goat serum at 4 °C overnight. After washing with PBS, sections were incubated with secondary antibodies (1:500, goat-anti-rabbit-Alexa568 and goat-anti-mouse-Alexa-488, both Life technology, Carlsbad, USA) in 1 % goat-serum for 1 h at RT. Subsequently sections were washed with PBS, blocked with 1 % Sudan Black (Sigma, St.Louis, USA) for 2 min and coverslipped with Fluoromont®G (SouthernBioTech, Birmingham, USA). As control, sections were stained without primary antibodies.
Confocal and fluorescent microscopy
Confocal microscopy of mouse brain sections was performed using a Carl Zeiss LSM 710 laser scanning microscope (LSM 710 NLO, Carl Zeiss, Oberkochen, Germany) and a LD C-Apochromat 63x/1,15 W Korr objective. Confocal images were analyzed with the ZEN2010 software (Carl Zeiss Microimaging GmbH, Jena, Germany).
Immunostained mouse brain sections were also evaluated using a Carl Zeiss Axio Observer.A1 microscope and digital camera (AxioCamMRm, Zeiss, Oberkochen, Germany). Apart from same brightness and contrast correction for all images, no additional image processing was performed.
Ethics, consent and permissions
Human blood sample collection was performed in accordance with the declaration of Helsinki and approved by the Ethic Committee of Ulm University. All volunteers gave informed written consent to participate in the study. Animal studies were performed in compliance with the National Institute of Health guidelines for the use of experimental animals.
Quantification of western blots
Western blots were quantified with ImageJ (Version 1,48). Expression levels of proteins of interest were standardized to protein expression level of a loading control (e.g. β-actin). At least three western blots of three biological replicates/samples (n = 3) were quantified and used for statistical analysis.
Statistical analysis
Graph Pad Prism (Version 5,04 and 6,05) was used to calculate the mean, standard error of mean (SEM) and to perform statistical analysis. The values are shown as mean ± SEM. Unless indicated otherwise, figures show pooled data from several independent replicates. To pool the data from independent experiments, the data of each experiment were normalized to the mean of the control and then used for statistical analysis. One way ANOVA/Tukey’s multiple comparison test and two tailed, unpaired student’s t-test were used to calculate the p-values. P values less than 0,05 were considered as significant.
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
AMH and WPR performed the experiments. VG performed preliminary experiments. AF cloned the SOD1 plasmid constructs. MSF generated the L1 and L2 plasmid constructs. AMH, JHW, KMD and PJM analyzed the results and designed the study. ACL gave conceptual input. AMH and KMD wrote the manuscript. All authors read and approved the final manuscript.