Parkin-mediated ubiquitination regulates phospholipase C-gamma1.

Mutations in parkin cause autosomal recessive forms of Parkinson's disease (PD), with an early age of onset and similar pathological phenotype to the idiopathic disease. Parkin has been identified as an E3 ubiquitin ligase that mediates different types of ubiquitination, which has made the search for substrates an intriguing possibility to identify pathological mechanisms linked to PD. In this study, we present PLCgamma1 as a novel substrate for parkin. This association was found in non-transfected human neuroblastoma SH-SY5Y cells as well as in stable cell lines expressing parkin WT and familial mutants R42P and G328E. Analysis of cortical, striatal and nigral human brain homogenates revealed that the interaction between parkin and PLCgamma1 is consistent throughout these regions, suggesting that the interaction is likely to have a physiological relevance for humans. Unlike many of the previously identified substrates, we could also show that the steady-state levels of PLCgamma1 is significantly higher in parkin KO mice and lower in parkin WT human neuroblastoma cells, suggesting that parkin ubiquitination of PLCgamma1 is required for proteasomal degradation. In line with this idea, we show that the ability to ubiquitinate PLCgamma1 in vitro differs significantly between WT and familial mutant parkin. In this study, we demonstrate that parkin interacts with PLCgamma1, affecting PLCgamma1 steady state protein levels in human and murine models with manipulated parkin function and expression levels. This finding could be of relevance for finding novel pathogenic mechanisms leading to PD.

Introduction

Parkinson’s disease (PD) is a neurodegenerative movement disorder characterized by neuronal loss mainly affecting dopaminergic neurons in substantia nigra pars compacta and the appearance of intracellular inclusions termed Lewy bodies [1, 2]. The recent identification and understanding of mutations causing PD has given important clues as to the pathological processes underlying this highly complex neurodegenerative disorder [3]. Various deletion and point mutations in the parkin gene have been shown to cause autosomal recessive juvenile Parkinsonism (AR-JP) an inherited form of PD with similar phenotype [4].

Parkin is a 52 kD E3 ubiquitin ligase [5] expressed in multiple tissues. E3 ubiquitin ligases function in the ubiquitin/proteasome pathway by binding to protein substrates and targeting them to the proteasome for degradation. It has also been reported that parkin can mediate degradation-independent ubiquitination [6]. Parkin has a ubiquitin-like sequence (UBL) at the amino terminal region and two RING (really interesting new gene) finger motifs at the carboxy-terminal region separated by an in-between-RING (IBR) domain [7]. The UBL domain plays a key role in bringing parkin to the proteasome [8]. The RING-IBR-RING motifs are important for interaction with the substrates and the E2 conjugating enzymes UbcH7 and UbcH8 [9].

Several substrates for parkin have been identified, including the septins CDC-rel 1 and 2, cyclin E, aminoacyl-tRNA synthetase cofactor, p38, O-glycosylated α-synuclein (α-Sp22), Pael-R, synaptotagmin XI, synphilin-1, Eps15, far upstream sequence element [FUSE]-binding protein 1 (FBP1) and protein interacting with C-kinase 1 (PICK1) [10-18]. It has been suggested that parkin loss of function (i.e. mutation in both alleles) leads to an accumulation of its substrates, which then leads to cell death. To date, only two parkin substrates, aminoacyl-tRNA synthetase cofactor, p38 and FBP1, have been reported to accumulate in parkin knockout mice as well as in sporadic PD subjects [16, 19], whereas accumulation of α-Sp22 and CDC-rel 2 has been shown in parkin ARJP subjects [12, 20].

Parkin ubiquitination of the substrate Eps15 lead to regulation of the epidermal growth factor receptor (EGFR). Lack of parkin results in enhanced endocytosis and degradation of EGFR and also reduced downstream phosphoinositide 3-kinase (PI(3)K)-Akt signalling. Phospholipase C γ1 (PLCγ1) is part of this pathway, being phosphorylated both by EGFR and Akt [21]. Considering these data, we investigated whether parkin is involved in regulation of PLCγ1, an enzyme known to be degraded in the proteasome following ubiquitination by the E3 ubiquitin ligase, c-Cbl [22].

We identified PLCγ1 as a novel substrate for parkin and report that PLCγ1 interacts with parkin both in dopaminergic neuroblastoma SH-SY5Y cells and in different regions of human brain. Ubiquitination of PLCγ1 is enhanced in parkin WT cells compared to cells expressing the parkin ARJP R42P and G328E mutants. PLCγ1 accumulates in the brains of parkin knockout mice, whereas overexpression of parkin WT in SH-SY5Y cells leads to reduced protein levels of PLCγ1 in comparison to non-transfected (NT) and parkin ARJP mutant cells.

Experimental procedures

Human brain and parkin KO mice brain samples

Post-mortem human brain material was obtained from the Huddinge Brain Bank (Karolinska University Hospital, Sweden) with approval by the Human Ethics Committee of Karolinska University Hospital. Three brains from individuals with non-neurological disorders (two males, 66 and 83 years old, one 85-year-old female) were used. All brains had a post-mortem delay between 24 and 48 hrs. Whole brains from four parkin knockout mice [23] and four wild-type littermates were also used. Homogenization of brain samples was performed on ice in 50 mM Tris-HCl (pH 7.4) plus 150 mM NaCl, 1% (v/v) Triton-X100 and complete protease inhibitors (Hoffmann-LaRoche Inc., Nutley, NJ, USA). Samples were centrifuged at 1000 ×g and protein concentration of the supernatants containing soluble proteins were frozen until analysis by immunoblot.

Treatments, DNA-constructs, transfections and cell culture

Human dopaminergic SH-SY5Y neuroblastoma cells were treated with 10 μM of proteasome inhibitor Lactacystin (Calbiochem, Darmstadt, Germany) for 6 hrs. The constructs used in this study were described previously [24]. Transfections of human dopaminergic SH-SY5Y neuroblastoma cells with various expression vectors were performed using Lipofectamine™2000 (Invitrogen, Stockholm, Sweden) according to the manufacturer’s instructions. Stable transfectants were selected with 200 μg/ml geneticin (Invitrogen, Stockholm, Sweden). Cells were cultured at 37°C, 5% CO2, in Eagle’s Minimal Essential Medium with Glutamax containing 10% foetal calf serum (FCS). Transfected cells were supplemented with 200 μg/ml geneticin. All cell culture supplies were purchased from (Invitrogen, Stockholm, Sweden). The expression of WT and mutant parkins were confirmed by sequencing following reverse transcription.

Immunoblot analysis

Cells were lysed in lysis buffer (20 mM Tris-HCl, 137 mM Na Cl, 2 mM EDTA, 2% Nonidet P-40, 2% Triton-X100 and protease inhibitor cocktail (Sigma-Aldrich, Stockholm, Sweden). Brains and cell extract protein amounts were quantified using the BSA protein assay kit (Pierce, Rockford, IL, USA). Equivalent amounts of protein were separated using 10% acrylamide gels. Proteins were transferred to a nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany). Western immunoblotting was performed using either anti-parkin (Cell signaling, Danvers, MA, USA), anti-PLCγ1, anti-PLCβ, anti-c-Cbl, anti-ubiquitin, anti-His antibodies or β-actin (BD Biosciences, San Jose, CA) with overnight incubations at a 1:1000 dilution. The secondary antibodies were anti-rabbit or anti-mouse horseradish peroxidase-linked, and were used at 1:2000 dilution for 1 hr in room temperature. Detection was made by the ECL method and exposure to Hyper film MP (Amersham Biosciences, Piscataway, NJ, USA).

Co-immunoprecipitation

The lysates containing 200 μg protein in 200 μl lysis buffer were pre-cleared first with Protein A/G-Sepharose beads (Amersham Biosciences, Piscataway, NJ, USA). The lysates were immunoprecipitated overnight with 1 μg of anti-parkin or anti-PLCγ antibody. Thereafter, new beads were added and lysates incubated for 2 hrs at 4°C with rocking. The immune complexes were washed 3 times with lysis buffer and analysed by Western blot.

Immunocytochemistry

SH-SHY5 cells were grown on glass coverslips and fixed in ice-cold methanol for 10 mins on ice. Methanol was subsequently discarded and cells were wahsed with PBS. Following 1 h blocking in 5% BSA, 0.2% Triton X-100, 2% goat serum (Sigma Aldrich, Darmstadt, Germany) in PBS, cells were incubated with primary parkin anti rabbit IgG (Cell Signaling, 1:100) and PLC l anti mouse IgG (Jackson Immuno Research 1:500) together with DAPI (Sigma Aldrich, Darmstadt, Germany) and incubated for 1 h at RT. Finally, cells were rinsed several times in PBS with 0.2% Triton X-100 and mounted with Fluorescent Mounting Medium (Dako, CA, USA).

In vitro ubiquitin conjugation assay

PLCγ1 and parkin immunoprecipitates were washed 3 times with washing buffer containing 20 mM HEPES, 10 mM KCl, 1.5 mM MgCl2, 25 μM N-Carbobenzoxy-Leu-Leu-Leucinal (MG132), 10 nM dithiothreitol, (Sigma-Aldrich, Stockholm, Sweden), 10 mM N-ethylmaleimide, protease inhibitor cocktail (Sigma-Aldrich, Stockholm, Sweden), 50 mM NaF, 1 mM Na3VO4. The immune complexes were analysed by Western blot and normalized to contain similar levels of PLCγ1 and parkin. Thereafter, they were mixed and incubated together for 10 min. at 30°C with agitation in a reaction buffer containing 10% (vol/vol) reticulocyte lysate (Promega, Nacka, Sweden), 1 μM dithiothreitol, 20 mM HEPES, 10 mM MgCl2 (Sigma-Aldrich, Stockholm, Sweden), 1 mM ATP, 1 mg/ml hexokinase, 20 mM deoxyglucose, 25 μM MG132 and His-tagged ubiquitin (5 μg/ml) (Sigma-Aldrich, Stockholm, Sweden). The reaction was terminated by washing twice with washing buffer. The immune complexes were analysed by Western blot.

Statistical analyses

Analyses of differences were carried out by ANOVA followed by Fisher’s PLSD post hoc test. A value of P < 0.05 was considered statistically significant.

Results

PLCγ1 levels are dependent on the expression level of functional parkin

Western blot analysis of human dopaminergic SH-SY5Y stable cell lines expressing WT, R42P or G328E ARJP mutant versions of parkin revealed that PLCγ1 levels are lower in cells overexpressing WT parkin (Fig. 1A and B). Cells overexpressing ARJP mutant versions of parkin showed similar levels of PLCγ1 as in NT cells. Protein levels of parkin were approximately 10–12 times higher in transfectants compared to NT cells.

Figure 1

PLCl levels are decreased in cells overexpressing WT Parkin, and increased in Parkin knockout mice. (A) Immunoblotting of parkin, PLCγ1 and PLCβ levels in cell lysates from SH-SY5Y cells transfected with parkin (WT) and the ARJP parkin mutations R42P and G328E. (B) Quantification of the PLCγ1 levels from the cells in A. Data are expressed as percentage of values for NT cells and represent the mean ± S.D. of four independent experiments (*, P < 0.05; anova followed by Fisher’s PLSD post hoc test). (C) Immunoblots of brain homogenates from wild-type (WT) and parkin knockout mice (KO) using antibodies against PLCγ1, parkin and β-actin. (D) Quantification of PLCγ1 levels in WT and parkin KO mice. Data were normalized as ratio of PLCγ1 to actin levels, and expression as percentage of values for WT animals. Bars represent means ± S.D. of four animals per group, (*, P < 0.05; anova followed by Fisher’s PLSD post hoc test).

In order to further elucidate whether the impact of functional parkin levels on PLCγ1 was species specific, we analysed brain material from parkin KO and WT mice. Immunoblotting of whole brain homogenates shows that PLCγ1 levels are higher when parkin is completely abolished, as compared to WT mice (Fig. 1C). Quantification of the PLCγ1 band intensity shows that this difference between mouse strains was significant (Fig. 1D).

Parkin and PLCγ1 are interacting in human dopaminergic neuroblastoma

Since PLCγ1 has been shown previously to be degraded in the proteasome through ubiquitination by the E3 ubiquitin ligase c-Cbl which is a member of the RING domain family, and parkin is also an E3 ubiquitin ligase in the same family, we analysed if there was a ligase-substrate interaction between parkin and PLCγ1 using two different approaches. Protein extracts from NT and parkin transfected cells immunoprecipitated with parkin and PLCγ1 antibodies revealed that there was an interaction between parkin and PLCγ1 in all cell lines (Fig. 2A). The input for this immunoblot is presented as Figure 1A and the interaction with c-Cbl is shown as a positive control for the immunoprecipitation. In order to visualize the interaction between parkin and PLCγ1, we performed confocal microscopy of parkin transfectants. We noted that in parkin WT overexpressing cells PLCγ1 and parkin co-localize mostly in plasma membrane, whereas in the parkin ARJP mutants the interaction was more scattered. Consistent with the immunoprecipitation results, the detected fluorescence representing parkin (red) and PLCγ1 (green) overlapped in all the parkin overexpressing cell lines. Since endogenous parkin was impossible to be detected with this method, we were not able to confirm the interaction in NT cells. As negative controls, immunoprecipitation and co-staining experiments were also performed with the IgG antibody PKCɛ together with parkin, where neither overlapping signal in confocal microscopy nor IgG interaction was found (Supporting Fig. S1).

Figure 2

PLCγ1 is associated with parkin. (A) Lysates from non-transfected (NT) and in stably transfected human SH-SY5Y neuroblastoma cells with wild-type (WT) parkin and the ARJP parkin mutations R42P and G32E were immunoprecipitated with either anti-parkin (upper panel) or anti-PLCγ1 antibodies (middle and lower panels) and detected by Western blotting with anti-PLCγ1, anti-parkin or anti-c-Cbl antibodies. (B) Confocal micrographs of parkin WT, R42P and G328E cells were double stained with anti-parkin (red) and anti-PLCγ1 (green) antibodies. Superimposed images showed co-localization of parkin and PLCγ1 in parkin WT, R42P and G328E cells (yellow).

Parkin and PLCγ1 are associated in humancortical, striatal and nigral areas

The observed interaction between parkin and PLCγ1 in human neuroblastoma cells led us to further investigate if the same observations could be seen in human brain material. Since one of the pathological hallmarks of PD is nigral degeneration affecting dopaminergic input to the striatum, we decided to analyse substantia nigra, striatum and cortex. Immunoprecipitation of human brain homogenates showed that the interaction between parkin and PLCγ1 is consistent in all the three examined regions from three different individuals (Fig. 3A). Western blot analysis of the input for the precipitation uncovered a regional difference in protein levels of PLCγ1 and c-Cbl, which were lower in the nigral fractions but present in striatum and cortex (Fig. 3B). Though c-Cbl could be recognized as a co-precipitate to PLCγ1 in the substantia nigra, but persistently at very low levels. This finding suggests that in the substantia nigra parkin may play a stronger regulatory role in regulating PLCγ1 levels than does c-Cbl. A lower molecular weight band was also detected with the PLCγ1 antibody in all the fractions, but more abundant in the substantia nigra.

Figure 3

The association of PLCγ1 in different brain regions. (A) Immunoblotting of PLCγ1 and parkin in homogenates Cortex (Cx), Striatum (St) and Substantia Nigra (SN) for three human brains. (B) Homogenates for the same samples seen in figure A were immunoprecipitated with either anti-parkin (upper panel) or anti-PLCγ1 (lower panels) antibodies and detected by Western blotting with anti-PLCγ1, anti-parkin or anti-c-Cbl antibodies.

Parkin WT successfully ubiquitylates PLCγ1 and this function is impaired in parkin ARJP mutants

As previously reported by others, we found that proteasome inhibition by lactacystin results in an accumulation of PLCγ1 (Fig. 4A). One suggested mechanistic impairment caused by parkin ARJP mutations is a deficient E3 ubiquitin ligase activity. To test this function on PLCγ1, we performed an in vitro ubiquitin conjugation assay. Two major bands were detected, corresponding to full length (150 kD) and a highly ubiquitylated PLCγ1 (approximately 200 kD), respectively. Levels of PLCγ1 associated ubiquitin were identified with immunoblotting with a His-tag antibody (Fig. 4B). Quantification of ubiquitin revealed that parkin ARJP mutants had significantly lower levels of PLCγ1 associated ubiquitin (Fig. 4C).

Figure 4

ARJP parkin mutations cause a defective PLCγ1 ubiquitination. (A) Immunoblotting of PLCγ1 in human SH-SY5Y neuroblastoma cells untreated or treated with proteasome inhibitor Lactacystin (10 μM, 6 hrs). (B) In vitro ubiquitin conjugation assay of PLCγ1 in cells transfected with parkin WT or the ARJP parkin mutations R42P and G328E. Parkin and PLCγ1 immunoprecipitates were normalized to contain similar levels of these proteins and incubated in the presence of reticulocyte extracts and His-tagged-ubiquitin. Samples were analysed by Western blotting with antibodies against PLCγ1 (upper panel), parkin (middle panel), anti-his (lower panel).

Discussion

Parkin mutant ARJP patients have a similar neuropathological profile as the idiopathic disease, making studies on the functions of parkin important for distinguishing PD pathology. How parkin dysfunction participates in ARJP pathogenesis is still under debate, but it is likely that deficient protein ubiquitination by parkin would lead to impaired degradation, aberrant localization or altered activity of substrates and that this consequence takes part in the pathological mechanisms behind PD. This idea has stimulated the search for parkin substrates of which defective ubiquitination could be linked to neurodegeneration. Previous studies have identified a range of important substrates. However, few are accumulated and others are not targets for degradation depending on the type of ubiquitination by parkin.

Our major finding in this study is the identification of PLCγ1 as a novel substrate for the E3 ubiquitin ligase parkin. This interaction is consistent in human SH-SY5Y cells as well as in cortical, striatal and nigral regions of the human brain. We also show increased ubiquitination of PLCγ1 by parkin WT compared to R42P and G328E mutants, resulting in a lower steady state PLCγ1 protein level in parkin WT overexpressing SH-SY5Y neuroblastoma cells. In line with previous studies, the R42P and G328E mutants appear to be loss-of-function mutations, thus cells with ectopic expression of these mutants do not differ in PLCγ1 levels compared to NT cells due to the preserved endogenous expression of parkin between all cell lines. The R42P mutation that is situated in the UBL domain and the G328E mutation that is found in the IBR-domain can both interact successfully with PLCγ1. Thus, the apparent difference in ubiquitination and PLCγ1 degradation could be resulting from deficient catalytic activity. Though we did not detect a typical poly-ubiquitination pattern, but possibly multiple mono-ubiquitinations, directing us to think that the difference in PLCγ1 degradation could result from the activity of other ubiquitin ligases, where parkin mono-ubiquitination is a pre-requisite. There is also a possibility that parkin is initiating PLCγ1 translocation, which would lead to impaired degradation as a downstream consequence. We detected that PLCγ1 co-localize differently in parkin ARJP mutants, supporting this hypothesis but further experiments have to be done to confirm this observation.

When we expanded this finding to a murine model, we found that the level of PLCγ1 was moderate but significantly increased in parkin KO animals, suggesting that parkin E3 ubiquitin ligase activity is important for maintaining balanced levels of PLCγ1. Among the many previously reported parkin substrates, only aminoacyl-tRNA synthetase cofactor p38 and FBP1 have been shown to accumulate in tissue from parkin null mice. These proteins are believed to be authentic parkin substrates targeted for proteasomal degradation by poly-ubiquitination, since they have also been shown to accumulate in MPTP-treated mice and in the brain of sporadic PD subjects.

Supporting the finding of Wang et al. (2006) that parkin is a regulator of EGFR endocytosis, by ubiquitination of Eps15 and PI3K-Akt signalling, we show that parkin also regulates PLCγ1 that is part of the same signalling pathway. It is therefore likely that parkin regulates tyrosin kinase receptor downstream signalling in several critical steps of the cascade. Additionally, PLCγ1 plays an important role in cellular proliferation, differentiation and motility. It is part of the PLC superfamily that is divided into six isotypes, namely, β, γ, δ, ɛ, γ, η[25-29]. Tyrosine kinases activate PLCγ1, which catalyses the hydrolysis of phosphoinositides (PI) to inositol triphosphate (IP3) and diacylglycerol, two second messengers that regulate the mobilization of the intracellular calcium and protein kinase C (PKC) activation, respectively [30]. PKC signalling results in various cellular responses including differentiation, gene expression and apoptosis [31]. Also calcium is central for different cellular processes from regulation of enzyme activity and gene expression to apoptosis [32, 33]. Dopaminergic neurons in SN are more exposed to oxidative stress and higher intracellular calcium concentrations due to their pace-making feature, making them more vulnerable compared to other neurons in the human brain [34]. It has been suggested that cell death in Parkinson’s disease may be associated with increased calcium concentrations [35].

Interestingly, PLCγ1 levels are significantly lower in substantia nigra compared to striatum and cortex. One reason for this regional discrepancy could be a higher sensitivity to PLCγ1 downstream signalling cascades within the nigral regions, which then would require high proteasomal regulation of PLCγ1 to maintain accurate protein level. Changes in PLCγ1 protein levels would probably have a substantial effect. Since c-Cbl is the only other known E3 ubiquitin ligase responsible to maintain this balance, the relative lower levels of c-Cbl in the substantia nigra fractions may suggest that the E3 ubiquitin ligase of parkin is particularly important in this region.

Taken together, we have demonstrated that parkin is associated and participates in the ubiquitination of PLCγ1. PLCγ1 is one of the few parkin substrates that accumulate in parkin knockout mouse. Moreover, ARJP parkin mutants have a defective PLCγ1 ubiquitination. Parkin has most likely a physiological important role in the regulation of PLCγ1 in substantia nigra, since we found a low c-CBl/PLCl interaction in this area. Our studies provide new insights for understanding the function of parkin by the identification of the new substrate PLCγ1.