Mutations in leucine-rich repeat kinase 2 (LRRK2) cause autosomal-dominant Parkinsonism with pleomorphic pathology including deposits of aggregated protein and neuronal degeneration. The pathogenesis of LRRK2-linked Parkinson's disease (PD) is not fully understood. Here, using co-immunoprecipitation, we found that LRRK2 interacted with synphilin-1 (SP1), a cytoplasmic protein that interacts with α-synuclein and has implications in PD pathogenesis. LRRK2 interacted with the N-terminus of SP1 whereas SP1 predominantly interacted with the C-terminus of LRRK2, including kinase domain. Co-expression of SP1 with LRRK2 increased LRRK2-induced cytoplasmic aggregation in cultured cells. Moreover, SP1 also attenuated mutant LRRK2-induced toxicity and reduced LRRK2 kinase activity in cultured cells. Knockdown of SP1 by siRNA enhanced LRRK2 neuronal toxicity. In vivo Drosophila studies, co-expression of SP1 and mutant G2019S-LRRK2 in double transgenic Drosophila increased survival and improved locomotor activity. Expression of SP1 protects against G2019S-LRRK2-induced dopamine neuron loss and reduced LRRK2 phosphorylation in double transgenic fly brains. Our findings demonstrate that SP1 attenuates mutant LRRK2-induced PD-like phenotypes and plays a neural protective role.
Parkinson's disease (PD) is a common neurodegenerative disorder with two pathological hallmarks: the selective loss of dopaminergic neurons and the presence of Lewy bodies. The pathogenesis of PD is incompletely understood, and includes both genetic and environmental contributions. Genetic causes of PD and parkinsonism-related phenotypes include leucine-rich repeat kinase 2 (LRRK2), alpha-synuclein, glucocerebrosidase, Parkin, PINK1 and others (1,2). Mutations in the LRRK2 cause familial PD and contribute to some sporadic PD as well (3–6), making mutant LRRK2 one of the most common known causes of PD thus far. LRRK2 PD cases have pleomorphic pathology very similar to idiopathic PD, including Lewy bodies, nigral degeneration and/or neurofibrillary tau-positive tangles (7,8). The LRRK2 protein (2527 aa) is widely expressed at relatively low levels. The LRRK2 protein contains several functional domains including kinase, GTPase and protein interaction domains, suggesting that LRRK2 has multiple functions (9).
LRRK2 has kinase and GTPase activities (10,11). The physiological substrate(s) of LRRK2 kinase activity are still incompletely understood (10,11). LRRK2 can be autophosphorylated (12). Previous studies have reported that LRRK2 can phosphorylate myelin basic protein (a generic substrate), as well as 4E-BP, s15, radixin, ezrin and moesin (13–16). Expression of PD-linked mutant LRRK2 (e.g. G2019S) causes neuronal degeneration in mammalian cells in culture and in drosophila (12,14,17–20). Reduction of LRRK2 kinase and guanosine-5′-triphosphate binding activities can protect against mutant LRRK2-induced toxicity (21–23). Studies of LRRK2 kinase inhibitors further support a role for LRRK2 kinase activity in PD pathogenesis (24–28).
LRRK2 also contains protein–protein interaction domains (6). LRRK2 has been shown to interact with several proteins related to multiple cellular functions (11,29). The physiological function of LRRK2 is unclear. Studies suggest that LRRK2 is involved in cytoskeleton arrangement, chaperone machinery, synaptic vesicle endocytosis, protein translational machinery, the MAPK signaling cascades, cell death and ubiquitin/autophagy protein degradation pathways (11,29). Given that synphilin-1 (SP1) interacts with α-synuclein and parkin (30,31), we hypothesized that SP1 may also be associated with LRRK2, and could modify LRRK2 pathogenesis.
SP1 is a cytoplasmic protein which was first identified to interact with α-synuclein (30). SP1 co-localizes with α-synuclein in Lewy bodies in brains of PD patients (32). An anti-SP1 antibody labels the central core of Lewy bodies while α-synuclein is generally present more peripherally. Previously, we and others report that overexpression of SP1 in cultured cells protects against cell death caused by toxin exposure (33,34). Co-expression of α-synuclein and SP1 in cultured cells promotes the formation of Lewy-body-like inclusions (30,31,35). SP1 attenuates α-synucleinopathy and enhances the formation of aggresomes in vitro and in transgenic mice (36–39). In this study, we investigate the relationship between SP1 and LRRK2 using in vitro and in vivo model systems with biochemistry, cell biology and genetic approaches. Our studies suggest that SP1 interacts with LRRK2 and can modulate LRRK2 cellular pathogenesis.
LRRK2 interacts with SP1
To investigate the relationship between LRRK2 and SP1, we co-transfected FLAG-LRRK2 and hyaluronic acid (HA)-SP1 into HEK 293T cells. The resulting cell lysates were subjected to co-immunoprecipitation (IP) assays. LRRK2 showed a specific interaction with SP1 (Fig. 1A). Figure 1A shows IP with FLAG-LRRK2 and detection of HA-SP1, demonstrating that SP1 co-immunoprecipitated with LRRK2. Conversely, SP1 was immunoprecipitated using anti-HA antibodies, followed by anti-FLAG LRRK2 immunoblotting (Fig. 1B), showing that LRRK2 bound with HA-SP1. Moreover, PD-linked mutant LRRK2 variants, R1441C and G2019S, were immunoprecipitated with SP1 and displayed increases in binding (Fig. 1C).
Predominant interaction regions of SP1 and LRRK2
To determine which regions of SP1 and LRRK2 were responsible for the interaction, a series of constructs containing different regions of SP1 tagged with HA were co-transfected with full-length LRRK2. As shown in Figure 2, SP1 N-terminal region (1–349 aa) co-immunoprecipitated with LRRK2. Conversely, when different regions of FLAG-LRRK2 were co-transfected with full-length HA-SP1, the C-terminal fragments of LRRK2 including C-terminal of Ras in complex protein domain (COR), kinase and WD 40 domains were predominantly associated with SP1 (Fig. 3).
SP1 promoted LRRK2-induced cytoplasmic inclusions
Previously, we and others showed that overexpression of human LRRK2 can cause cytoplasmic inclusions (17). Since SP1 promotes α-synuclein inclusion formation (30,31,35), we tested whether SP1 alters LRRK2-linked inclusions. When LRRK2 was co-transfected with SP1, there was about 3-fold increase in the percentage of cells with aggregates (Fig. 4). Co-expression of PD-linked mutant G2019S-LRRK2 with SP1 resulted in up to 38% cells with inclusions. About 80% of these inclusions contained LRRK2 and SP1. Additionally, 90% of these inclusions were positively stained by anti-ubiquitin antibodies, which were similar to Lewy bodies.
SP1 attenuated LRRK2-induced neuronal degeneration in vitro
Expression of mutant LRRK2 causes neuronal degeneration (17,18). To determine whether the interaction of SP1 and LRRK2 could alter cell viability, SH-SY5Y cells were co-transfected with GFP and various LRRK2 constructs along with SP1 constructs using Lipofectamine. Viable cells were defined as having at least one smooth extension with twice the length of the cell body, and they were counted by an investigator who was kept unaware of the experimental condition. Consistent with previous findings (17,18), PD-linked mutants, R1441C and G2019S, caused significant cell toxicity compared with WT LRRK2 and vector control in human neuroblastoma SH-SY5Y cells (Fig. 5A and B). Co-expression of SP1 significantly protects against mutant LRRK2-induced neuronal degeneration.
To further investigate the protective role of SP1, siRNA targeting mouse SP1 was co-transfected into mouse primary cortical neurons. There was about an 88% reduction of SP1 mRNA by siRNA knockdown compared with the control RNA group, which indicated that the majority of neurons were expressed SP1 siRNA. Three days post-transfection of SP1 siRNA, neurons were co-transfected with GFP and various LRRK2 constructs at 1:15 ratio for 48 h. Knockdown of SP1 in GFP/vector expressing neurons decreased neuronal survival ∼20% (Fig. 5D), while knockdown of SP1 in GFP/wild-type LRRK2 expressing neurons reduced neuronal survival ∼50% (Fig. 5D). Moreover, knockdown of SP1 in GFP/mutant LRRK2 (G2019S or R1441C) reduced neuronal survival ∼70% (Fig. 5D). These results indicated that SP1 played a protective role against LRRK2 toxicity.
SP1 reduced LRRK2 kinase activity
PD-linked mutant LRRK2 increases kinase activity and reduction of kinase activity attenuates mutant LRRK2 toxicity (18,23). To assess whether SP1 alters LRRK2 kinase activity, we performed an in vitro LRRK2 autophosphorylation assay. Various FLAG-tagged LRRK2 constructs with or without SP1 were transfected into HEK 293T cells. The resulting cell lysates were immunoprecipitated using anti-FLAG antibodies. The immunoprecipitates were subjected to LRRK2 autophosphorylation assays. Expression of SP1 significantly reduced mutant G2019S-LRRK2 kinase activity (Fig. 6).
SP1 suppressed G2019S-LRRK2-induced PD-like phenotypes in drosophila
Previously, we generated a G2019S-LRRK2 transgenic Drosophila model, in which expression of G2019S-LRRK2 in neurons results in dopamine (DA) neuron degeneration, locomotor impairment and early death in flies (20). To further study the role of SP1 interactions with LRRK2 in vivo, we generated a double transgenic fly line which expresses both human SP1 and mutant G2019S-LRRK2 under a pan-neuron promoter, Elav-Gal4 driver or a DA neuron promoter, ddc-Gal4 driver. Expression of SP1 alone in DA neurons did not cause any PD-like neurodegeneration phenotypes. However, co-expression of SP1 in DA neurons significantly increased survival (Fig. 7A) and improved locomotor impairment (Fig. 7B) caused by mutant G2019S-LRRK2 expression in double transgenic flies. Moreover, co-expression of SP1 also protected against mutant G2019S-LRRK2-induced DA neuron degeneration (Fig. 8A) and reduced the G2019S-LRRK2 phosphorylation (Fig. 8B).
In this study, we found that LRRK2 interacted with SP1. In co-IP experiments, LRRK2 interacted preferentially with the N-terminal region of SP1, and SP1 predominantly interacted with the C-terminal region of LRRK2, including COR, kinase and WD domains. In cell transfection studies, co-expression of SP1 and LRRK2 increased cytoplasmic protein aggregates, reduced mutant LRRK2 protein kinase activity and attenuated LRRK2 toxicity. With in vivo Drosophila studies, expression of SP1 in neurons suppressed G2019S-LRRK2-induced PD-like phenotypes by increasing survival, improving locomotor impairment, reducing DA degeneration and reducing LRRK2 kinase activity in double transgenic flies. These findings suggest that SP1 is a protective interaction partner of LRRK2.
We originally identified that SP1 interacts with α-synuclein at 30–543 aa, and, when co-expressed, promotes α-synuclein aggregates (30). We further found that SP1 interacts with parkin at 214–556 aa, and is a substrate of parkin, which promotes the ubiquitinated aggregate formation that containing α-synuclein and parkin (31,35). SP1 also interacts with other proteins (40). SP1 is a ubiquitously expressed cytoplasmic protein enriched in neurons. SP1 co-localizes with α-synuclein in Lewy bodies in brains of PD patients (32). These findings suggested that SP1 may play a role in PD pathogenesis. We now demonstrate that SP1 can interact with LRRK2. Expression of SP1 promotes LRRK2 aggregate formation in cultured cells. The N-terminal 1–349 aa of SP1 predominantly interacts with LRRK2, which partially overlaps with the sites that interact with α-synuclein and parkin (30,31). Thus, SP1 may be working as a bridge protein to sequester the toxic proteins forming inclusions. Recent studies show that SP1 contains signal sequence for protein aggregation (38), suggesting that SP1 promotes the formation of inclusions. In addition, an in vitro study shows that inclusions formed by SP1 and α-synuclein appear to be more prevalent in non-apoptotic cells (38), supporting an idea that inclusions represent a protective cellular response to toxic protein (41).
SP1 predominantly interacts with LRRK2 C-terminal regions including COR, kinase and WD domains. This interaction may cover up the GTPase and kinase regions of LRRK2 to reduce LRRK2 activity. Alternatively, this interaction may alter the LRRK2 protein conformation thereby reducing kinase activity although this remains further crystal structure studies. Moreover, the increased interactions between SP1 and mutant LRRK2 variants compared with wild-type LRRK2 suggest a protective role for SP1 against mutant LRRK2 toxicity, which could be through increasing the binding with enzymatic regions and reducing elevated kinase activity. In fact, our results showed that co-expression of SP1 did reduce mutant LRRK2 kinase activity in vitro. In SP1 and LRRK2 double transgenic flies, there was less LRRK2 phosphorylation (activation) compared with those of LRRK2 transgenic flies. Most importantly, expression of SP1 attenuates mutant LRRK2 toxicity in cultured cells and reduces DA neuron degeneration in double transgenic flies. Knockdown of SP1 by targeting siRNA enhanced LRRK2 toxicity. Thus, reduction of LRRK2 kinase activity may be the major contributing factor for SP1 protecting against LRRK2 toxicity. SP1 may also act protection through other pathways. Previous reports have shown that SP1 displays a protective function against staurosporine and 6-hydroxydopamine toxicity by reducing the hydrolysis of procaspase-3, decreasing poly (adenosine diphosphate-ribose) polymerase cleavage and reducing p53 transcriptional activity and expression (34). We recently showed that SP1 has a neurotrophic effect in vitro (33), and that SP1 attenuates mutant α-synuclein toxicity in a mouse model by reducing caspase-3 activation and promoting aggresome formation, which is linked to autophagic clearance of inclusions and toxic proteins (39).
In conclusion, we demonstrated that SP1 interacts with LRRK2, promotes the formation of cytoplasmic aggregates containing LRRK2/SP1, reduces LRRK2 kinase activity and protects against mutant LRRK2 toxicity. These results suggest that SP1 can play a protective role against LRRK2-linked toxicity. These studies provide novel understanding of SP1 and LRRK2.
Materials and Methods
Anti-FLAG antibodies were obtained from Sigma. Anti-LRRK2 and anti-phospho-LRRK2 (S935) were from Michael J. Fox Foundation. Anti-HA and anti-myc antibodies were obtained from Santa Cruz Biotechnology and Roche Molecular Biochemicals. Fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit immunoglobulin G (IgG) and CyTM3-conjugated goat anti-mouse IgG were obtained from Jackson ImmunoResearch. Hoechst 33342 was obtained from Molecular Probes. Cell culture media, N2 and N27 supplements were from Invitrogen.
Plasmids and transfection
HA tagged full length and truncated human SP1 were described previously (30,31,35). Flag tagged full length (wild type and G2019S mutant) and truncated human LRRK2 constructs were described previously (17,42). Plasmids were transiently transfected into HEK 293T and SH-SY5Y cells with Lipofectamine Plus or Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol.
Cell cultures, IP and western blot analysis
Human HEK 293T and SH-SY5Y cells were grown in the media as described (17). IP experiments from transfected cell lysates were performed with anti-FLAG, or anti-HA antibodies and protein G Plus/protein A-agarose (Amersham Pharmacia Biotech). The resulting immunoprecipitates or cell lysates were separated using 4–12% NuPAGE Bis-Tris gels and transferred onto polyvinylidene difluoride (PVDF) membranes (Invitrogen). The membranes were blocked in TBST (10 mm Tris·HCl, pH 7.4/150 mm NaCl/0.1% Tween 20) containing 5% non-fat milk and then probed with different antibodies. Proteins were detected by using enhanced chemiluminescence reagents (NEN).
Cell viability assays
SH-SY5Y cell viability assays were conducted as described previously (17). Cells were co-transfected pcDNA3.1-GFP with various FLAG-LRRK2 constructs and/or HA-SP1 (at 1:15) for 24 h in 10% fetal bovine serum (FBS) OPTI-I media and then changed to DMEM with N2 supplement for 24 h. Viable GFP-positive cells (neurons) with at least one smooth extension (neurite) with twice the length of the cell body were counted.
Mouse primary cortical neuronal cultures and electroporation transfection
Mouse primary cortical neuronal cultures were derived from CD-1 outbred mice (The Jackson Laboratory) at embryonic day 15 or 16 as described previously (17). Neurons were grown on laminin- and poly-D-lysine-coated plates (BD Biosource, San Diego) using neurobasal medium containing Glutamax, B-27 supplement and penicillin/streptomycin. SP1 siRNA and the corresponding scrambled control RNA oligonucleotides were ordered from Dharmacon (Chicago, IL, USA), and transfected by using an electroporation-based method with Nucleofector (Amaxa Biosystem, Cologne, Germany) before placing the neurons in the plates as described previously (33). Three days post-transfection of siRNA, the knockdown effects of SP1 siRNA were confirmed by reverse transcription polymerase chain reaction (RT-PCR) using the primers for mouse SP1: 5′-CTCAAGACCATCCCAGCACT-3′ and 5′-TCAGTGGAGAAACTCGCTTCA-3′. The separated set of neurons with siRNA transfection was then further co-transfected with GFP and various LRRK2 constructs at 1:15 ratio for 48 h using lipofectAMINE 2000 (Invitrogen) according to the manufacturer's instructions. The neuron viability assays were performed by quantifying the viable GFP-positive neurons with at least one smooth extension (neurite) with twice the length of the cell body as described previously (17).
Cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100 and processed as described (17). Cell samples were probed using anti-FLAG or anti-myc as primary antibodies, then incubated with secondary antibodies: Cy3-conjugated anti-mouse or FITC-conjugated anti-rabbit. The nuclei were stained with Hoechst 33342, and the images were captured by fluorescence and confocal microscopy (LSM 510 and Axiovert 100, Zeiss). The number of cells with cytoplasmic aggregates was counted in each experimental group by an investigator who was unknown the experimental condition.
In vitro kinase assays
LRRK2 proteins were immunoprecipated from transfected cell lysates and washed twice with kinase assay buffer (Cell Signaling) containing Phosphatase Inhibitor Cocktails 1 and 2 (Sigma). Kinase assays were performed in 55 μl kinase assay buffer containing 10 μCi of [γ-32P]ATP (3000 Ci/mmol), 50 mm MgCl2 and 500 μm adenosine triphosphate (ATP) for 90 min as described previously (23). Kinase reaction samples were separated on 4–12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and blotted onto PVDF membranes. LRRK2 autophosphorylation was quantified with a PhosphorImager (Bio-Rad Molecular Imager FX).
The elav-Gal4 and ddc-Gal4 fly lines were obtained from the Bloomington Drosophila Stock Center. USA-SP1, UAS- LRRK2 and UAS-LRRK2 G2019S were generated previously and maintained in our laboratory (20,43). All Drosophila were maintained on regular fly food at 25°C and humidity at 55%. Fresh food media were changed every 3–4 days.
Survival and locomotor activity assays
Cohorts of 60 flies from each genotype were monitored for survival. Mortality was scored weekly and analyzed by using Kaplan–Meier survival curves. Climbing assay was performed to test locomotor activity as described previously (20). Briefly, 60 flies from each group were tapped to the bottom of a graduated cylinder (diameter 1.5 cm; length 25 cm). Flies that could climb to or above the median line of the cylinder within 10 s were counted.
Drosophila whole-mount brain immunostaining and DA neuron quantification
Fly heads at 7 weeks of age were used to perform whole-mount immunostaining with anti-tyrosine hydroxylase (TH) antibodies as described previously (20). Eight fly brains in each group were incubated with mouse monoclonal anti-TH (Immunostar) antibodies as the primary antibodies, and followed by addition of Alexa Fluor 568 or Alexa Fluor 488 goat anti-mouse IgG (Invitrogen) as secondary antibodies. The numbers of TH-positive neurons in all DA clusters (except paired anterolateral medial clusters) were counted under fluorescent microscopy (Zeiss LSM 250). The paired anterolateral medial clusters contain many DA neurons with very high fluorescent density to prevent precise counting.
Quantitative data are expressed as arithmetic mean ± SEM based on at least three separate experiments. Sigmastart 3.1 statistical software (Aspire Software International, VA, USA) was used to analyze the difference between groups by ANOVA. A P-value <0.05 was considered significant.
This work is support by NIH grants: R21NS055684 to W.W.S. and RO1NS055252 to C.A.R.
We thank Yi Yu for assistance and technical support.
Conflict of Interest statement. None declared.