The protein kinase C gamma (PKCγ) gene is mutated in spinocerebellar ataxia type 14 (SCA14). In this study, we investigated the effects of two SCA14 missense mutations, G118D and C150F, on PKCγ function. We found that these mutations increase the intrinsic activity of PKCγ. Direct visualization of labelled PKCγ in living cells demonstrates that the mutant protein translocates more rapidly to selected regions of the plasma membrane in response to Ca2+ influx. These results point to specific alterations in mutant PKCγ function that could lead to the selective neuronal degeneration of SCA14.
Spinocerebellar ataxia type 14 (SCA14) is an autosomal dominant neurodegenerative disorder characterized by slowly progressive cerebellar dysfunction including gait and limb ataxia, dysarthria and abnormal eye movements, with the age of onset ranging from 33 to 43 years (Chen et al., 2003a; van de Warrenburg et al., 2003; Yabe et al., 2003). Atrophy of the cerebellum is apparent by MRI (van de Warrenburg et al., 2003) and a reduction of cerebellar Purkinje cells has been observed at autopsy (Brkanac et al., 2002). Recently, PRKCG, the gene encoding the conventional protein kinase C (PKC) family member, PKCγ, was identified as the SCA14 disease gene (Chen et al., 2003a). In contrast to the repeat expansion mutations responsible for other forms of spinocerebellar ataxia, missense mutations cause SCA14. To date, the majority of mutations have been found in exons 4 and 5 of the PRKCG gene (Chen et al., 2003a; van de Warrenburg et al., 2003; Yabe et al., 2003; M.C. Fahey, M.A. Knight, J.H. Shaw, R.J.M. Gardner, D. duSart, M.R. Delatycki, P.C. Gates, E. Storey, unpublished data), with one reported in exon 18 (Stevanin et al., 2004). The effects of these mutations on the activity and function of PKCγ is not known. Furthermore, it is not clear how these mutations lead to loss of Purkinje cells and cerebellar dysfunction.
PKCs are known to play an important role in Purkinje cells. For example, cerebellar long-term depression (LTD), which probably regulates certain types of motor learning, requires PKC activity (Daniel et al., 1998). PKC appears to influence LTD by regulating the endocytosis of glutamate receptors on Purkinje cells (Chung et al., 2000, 2003; Man et al., 2000; Wang and Linden, 2000; Xia et al., 2000). Other synaptic functions are also regulated by PKC proteins. Polyinnervation of Purkinje cells by climbing fibres persists into adulthood in both PKCγ knockout mice and mice expressing a transgenic PKCγ inhibitor (Chen et al., 1995; De Zeeuw et al., 1998). This has been explained as a failure to eliminate redundant synapses during cerebellar development (Kano et al., 1995).
Recently, PKC has also been implicated in the regulation of the size and complexity of Purkinje cell dendritic arbours. In the presence of PKC inhibitors or in the absence of PKCγ, the dendritic trees are enlarged and the number of branch points is increased (Metzger and Kapfhammer, 2000; Schrenk et al., 2002). In contrast, activation of PKC by phorbol-12-myristate-13-acetate (TPA) has a strong inhibitory effect on dendritic arbour growth in slice culture assays (Metzger and Kapfhammer, 2000). Thus, both positive and negative regulation of PKCs has profound effects on the Purkinje cell morphology and behaviour.
The conventional PKCs are dependent on 1,2 diacylglyerol (DAG) or phorbol esters, phospholipids and Ca2+ for full activity (Newton, 1997). DAG and phorbol esters bind to an amino-terminal cysteine-rich domain designated C1, which is composed of two similar subdomains, Cys1 and Cys2 (Newton, 1997). Six of the SCA14 mutations reported to date are in the Cys2 subdomain (Chen et al., 2003a; van de Warrenburg et al., 2003; Yabe et al., 2003; M.C. Fahey, M.A. Knight, J.H. Shaw, R.J.M. Gardner, D. duSart, M.B. Delatycki, P.C. Gates, E. Storey, unpublished data). Calcium and phospholipid bind to C2 domain. Interestingly the function of this domain appears to depend upon the phosphorylation status of a motif in the C-terminal region that corresponds to T674 in PKCγ and is regulated by autophosphorylation (Edwards and Newton, 1997). Phosphorylation at this site leads to an estimated 10-fold increase in affinity for Ca2+ and phosphatidylserine (Edwards and Newton, 1997). C-terminal phosphorylation and the N-terminal domains are important in regulating the intrinsic activity of the conventional PKCs and their affinity for components of the plasmid membrane. In turn, the cycling of conventional PKCs between the cytosol and membrane compartments is likely to regulate the access of these kinases to key substrates.
In this study, we investigated the effect of two SCA14 mutations; G118D (van de Warrenburg et al., 2003) and C150F (M.C. Fahey, M.A. Knight, J.H. Shaw, R.J.M. Gardner, D. duSart, M.B. Delatycki, P.C. Gates, E. Storey, unpublished data) on PKCγ function. We found that these mutations in the Cys2 subdomain increase the intrinsic kinase activity of PKCγ. By directly visualizing the subcellular localization of a PKCγ-green fluorescent protein (GFP) fusion protein in living cells, we observed that mutant PKCγ has enhanced calcium-induced membrane translocation. The possible role of aberrant PKCγ-mediated signal transduction in the selective loss of Purkinje cells in SCA is discussed.
Material and methods
The PRKCG cDNA was amplified from IMAGE clone 5764695 IMAGE Consortium, http://image.llnl.gov/) by PCR using the following primers: forward 5′-GCC GAA TTC ACC ATG GCT GGT CTG GGC CCC-3′ and reverse 5′-CGG GGA TCC CGC ATG ACG GGC ACA GGC AC-3′, respectively. This introduced an EcoRI site at the 5′ end and a BamHI site at the 3′ end, which facilitated cloning into the respective sites of a pEGFP-N1 expression plasmid (Clontech, Palo Alto, CA, USA). The Gly118Asp and Cys150Phe mutations were introduced into PRKCG cDNA with the Quikchange II Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA) (Fig. 1). This technique involves amplifying a mutant version of the cDNA using the high fidelity polymerase, pfu. For the Gly118Asp mutant, the following primers were employed: forward 5′-GCG ACC ACT GTG ACG CCC TCC TCT ACG G-3′ and reverse 5′-CCG TAG AGG AGG GCG TCA CAG TGG TCG C-3′, respectively. For the Cys150Phe mutant, the primers were: forward 5′-CGT GCC CTC CCT GTT TGG TGT CGA CCA CAC-3′ and reverse 5′-GTG TGG TCG ACA CCA AAC AGG GAG GGC ACG-3′, respectively. In addition to the indicated mutations, translationally silent changes created restriction sites for HgaI and SalI in the Gly118Asp and Cys150Phe mutants respectively, to facilitate screening. All constructs were verified by sequencing.
Cell culture and transfection
COS-7 cells (ATCC) were cultured in Dulbecco's Modified Eagle medium (DME); high glucose medium supplemented with 10% fetal bovine serum, penicillin (100 U/ml), streptomycin (100 mg/ml), and glutamate (100 mg/ml) in humidified atmosphere with 5% CO2 at 37°C. One day before transfection, 0.8 × 106 cells were seeded onto 60 mm culture dishes and transiently transfected with 3 mg of each plasmid with Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA) for 5 h in normal culture medium without antibiotics. After transfection, the cells were cultured at 37°C at least for 24 h before imaging.
Time-lapse study of expressing cells
Transfected cells (0.1 × 106 cells) were plated onto DT glass-bottom culture dishes (Bioptechs, Butler, PA, USA) and cultured for an additional 16 h before use. Before stimulation, the culture media was replaced by Hank's Balanced Salt Solution (5.33 mM KCl, 0.441 mM KH2PO4, 4.17 mM NaHCO3, 137.93 mM NaCl, 1 mM CaCl2 and 0.338 mM Na2H2PO4 supplemented with 5 mM HEPES). The cells were treated with either TPA (5 mM; Sigma, Atlanta, GA, USA), or calcium ionophore A23187 (10 mM; Sigma) at 21°C. Time-lapse experiments were performed for 50 min; every 30 s an image was acquired digitally with a DeltaVision microscope (Applied Precision, Issaquah, WA, USA).
Twenty-four hours after transfection, the cells were plated on glass coverslips coated with poly-l-lysine in 12-well culture dishes and grown for at least 16 h before stimulation as described above. At times indicated, the cells were washed once with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde in PBS for 20 min at room temperature. After three washes with PBS, the cells were permeabilized with 0.3% Triton X-100 for 10 min and washed again three times with PBS. The cells were stained with Alexa fluor 594 phalloidin (Molecular Probes Eugene, OR, USA) in 3% bovine serum albumin in Tris-buffered saline for 60 min. The images were collected using a DeltaVision microscope and images were deconvolved using the Softworx program (Applied Precision).
Transfected cells were harvested and lysed in ice-cold RIPA buffer (0.15 M NaCl, 1% Nonidet P-40, 1 % sodium deoxycholate, 0.1% sodium dodecyl sulphate (SDS), 10 mM sodium phosphate pH 7.0, 2 mM EDTA, and 14 mM 2-mercaptoethanol) supplemented with complete mini protease inhibitor cocktail (Roche, Palo Alto, CA, USA) and phosphatase inhibitor cocktail I (Sigma). Cell lysates were incubated for 20 min on ice, sonicated for 10 s and centrifuged at 14 000 r.p.m. for 20 min at 4°C. Where indicated, transfected cells were stimulated with TPA (5 mM) for 50 min. The protein levels were quantified with the colorimetric BCA protein assay reagent kit (Pierce, Iselin, NJ, USA) before equal amounts were loaded on 12% Novex Tris-glycine mini-gels (Invitrogen). After electrophoresis, the proteins were transferred onto nitrocellulose filters (Invitrogen) and blocked with 5% milk in preparation for western blotting. The following antibodies were used in the western blots as indicated in the legends: anti-PKC (α, β and γ) (Upstate), anti-GFP Living Colors peptide (Clontech) antibodies, anti-PKC-phospho-T514, anti-phospho-T655 and phospho-T674 (Biosource) anti-phospho-Ser-PKCγ substrate, and anti-phospho-MARCKS (Cell Signaling, Beverly, MA, USA).
PKCγ kinase assay
PKCγ-GFP protein was immunoprecipitated from cell lysate (200 mg) of transfected cells with 4 mg anti-GFP Living Colors peptide antibody, which was immobilized on 50 ml protein A/G PLUS-agarose beads (Santa Cruz, Santa-Cruz, CA, USA). Immunoprecipitates were washed three times with ice-cold lysis buffer, followed by two washes with PAN buffer (100 mM NaCl, 10 mM Tris pH 7.0, with the mini protease inhibitor cocktail), and finally resuspended in kinase assay dilution buffer II (Upstate, Charlottesville, VA, USA). Immunoprecipitated proteins were assayed for PKCγ kinase activity by monitoring the incorporation of 32P into the substrate peptide (QKRPSQRSKYL) employing the PKC kinase system (Upstate) and measured with a Trilux scintillation counter (Wallac, Boston, MA, USA). The amounts of PKCγ proteins used in the assays were determined by direct comparison with known amounts of GFP by western blotting. These values were used to determine the specific kinase activity of the wild-type and mutant PKCγ proteins in units of 32P incorporation (arbitrary unit)/fmoles of kinase.
PKCγ autophosphorylation in the presence of SCA14 mutations
In order to study the effects of SCA14 mutations on the kinase function and subcellular distribution of the PKCγ protein, we generated expression constructs encoding an in-frame fusion of PKCγ and GFP. The GFP tag allowed us to distinguish the exogenously introduced kinases based on increased size relative to endogenous PKC isozymes, and by immunoreactivity with anti-GFP antibodies. The GFP tag also allowed us to analyse the subcellular distribution of the PKCγ kinases in living cells by fluorescence microscopy.
We chose to study the effects of two distinct SCA14 mutations found in the Cys2 subdomain (Fig. 1). The mutations G118D, identified in the Netherlands, and the C150F, identified in Australia, were introduced by site-directed mutagenesis into the PKCγ coding region. These fusion proteins were used to compare the activity and subcellular distribution of the wild-type and mutant PKCγ kinases.
Since phosphorylation is known to play a central role in the regulation of PKC isozymes, we examined the major phosphorylation sites of PKCγ. Using phosphospecific antibodies, we were able to monitor phosphorylation at T514, T655 and T674. COS-7 cells were transfected with the wild-type and two mutant PKCγ constructs at approximately the same efficiency based on the percentage of green fluorescent cells (data not shown). Typically, equal amounts of PKCγ (wild-type) and PKCγ (G118D) were recovered in cell lysates from transfected cells. However, lower amounts of PKCγ (C150F) were observed in total cell lysates, possibly reflecting reduced protein stability of this mutant protein (data not shown). After normalizing PKCγ protein levels, we observed approximately equal phosphorylation status at T514, T655 and T674 (Fig. 2). The mutations in the Cys2 Subdomain, therefore, do not appear to alter the phosphorylation of PKCγ by PDK1 or alter the autophosphorylation of the kinase.
PKCγ mutants are distinguished from wild-type kinase by their response to Ca2+
Previous characterization of PKCγ has shown that it translocates to the plasma membrane in response to treatment with a Ca2+ ionophore (Sakai et al., 1997). The calcium-induced translocation is reversible and may cycle a number of times during a one-hour period. We performed a time-lapse study with calcium-stimulated wild-type and mutant PKCγ-GFP transfected cells to determine the influence of SCA14 mutations on this event. Calcium-induced translocation of both the wild-type and mutant PKCγ-GFP proteins was observed in the first seconds of Ca2+ stimulation (Fig. 3). The PKCγ-GFP was released from the membrane ∼15 s later, and returned to a diffuse cytoplasmic, predominantly perinuclear location. The earliest translocation that we detected differed between wild-type and mutant PKCγ. A higher proportion of the GFP-tagged mutant kinase translocated to the membrane than the wild-type protein (Fig. 3, 10 s). The majority of the wild-type PKCγ-GFP remained in the cytosol. In addition, we observed a qualitative difference in the distribution of membrane binding between wild-type and mutant PKCγ-GFP. Although in the case of the wild-type protein, the distribution appeared to be uniform along the plasma membrane, the mutant kinases were clearly enriched in discreet membrane domains. These regions had the appearance of membrane ruffles, and co-staining with fluorescently tagged phalloidin indicated they were rich in filamentous actin (data not shown).
Since the C1 domain that harbours the SCA mutations has been shown to regulate membrane binding via DAG or phorbol esters, we studied the redistribution of PKCγ in transfected COS-7 cells in response to the phorbol ester, TPA. Wild-type and mutant PKCγ-GFP had similar patterns of cytosolic localization in unstimulated cells (Fig. 4A; left panels). TPA induced the redistribution of wild-type and mutant PKCγ-GFP from the cytosol to the membrane with similar kinetics and to a similar extent (Fig. 4A). Unlike the Ca2+ response, the TPA induced-membrane targeting was not reversed over the one-hour time course of the study. After TPA stimulation, we observed co-localization of actin and the PKCγ-GFP proteins (Fig. 4B). Interestingly, membrane ruffling was induced by TPA treatment. This effect was enhanced in the mutant PKCγ-GFP expressing cells compared with cells expressing wild-type kinase; the mutant kinases were targeted to actin-rich membrane ruffles (Fig. 4B, arrow). After treatment, 45% of the wild-type PKCγ-GFP expressing cells displayed membrane ruffling (n = 100) compared with 82% of the mutant PKCγ-GFP expressing cells (n = 100).
SCA14 mutations increase PKCγ kinase activity
Although the SCA14 mutations lie outside of the catalytic domain of PKCγ, it is possible that they lead to a change in kinase activity by altering the protein's conformation. The activity of the wild-type and mutant PKCγ-GFP proteins was therefore measured directly by in vitro kinase assays. The presence of the GFP tag provided a convenient handle to purify the recombinant proteins away from endogenous kinases. The activities of the kinases were assayed in extracts from unstimulated cells, and kinase activities were normalized for protein input by western blotting against known quantities of GFP. The specific kinase activity of the PKCγ with the G118D and C150F mutations increased in comparison to the wild-type kinase (Fig. 5). Interestingly, the C150F mutation, which led consistently to reduced protein recovery (data not shown), resulted in a greater increase in specific activity. Compared with the wild-type, the G118D mutation did not alter significantly the amount of protein recovered from the transfected cells.
The autosomal dominant spinocerebellar ataxias are generally caused by a repeat expansion mutation in otherwise unrelated disease genes. However, missense mutations in two genes encoding signal transduction proteins, fibroblast growth factor 14 (FGF14) (van Swieten et al., 2003) and PRKCG (Chen et al., 2003a; van de Warrenburg et al., 2003; Yabe et al., 2003; Stevanin et al., 2004), also lead to spinocerebellar ataxias. In this study, we found that two different SCA14 mutations lead to increased PKCγ kinase activity, although they reside outside the kinase domain. The mutant kinases showed a more pronounced redistribution to membrane domains than the wild-type kinase at early time points in response to Ca2+ treatment. We propose that SCA14 mutations in the regulatory C1 domain result in disinhibition of the PKCγ kinase, and this initiates a sequence of events that ultimately leads to loss of Purkinje cells.
The C1 domain of the PKC family members mediates interactions with DAG and phorbol esters (Newton, 1995). The translocation of PKCγ to the cellular membranes in response to TPA appears to be unaffected by these mutations. This is in line with a previous study that showed both the Cys1 and Cys2 subdomains of PKCγ bound to TPA and DAG with comparable affinities (Ananthanarayanan et al., 2003). Therefore, even if the SCA14 mutations prevented TPA binding to the Cys2 subdomain, binding to the Cys1 subdomain would be predicted to compensate. While it is formally possible that the mutations elevated the affinity of the Cys2 subdomain for TPA, this seems unlikely. Anomalies in PKCγ function, such as altered membrane targeting, were apparent in the absence of TPA. Therefore, the influence of the Cys2 subdomain mutations on TPA effects appears to be minimal. A similar rationale may extend to DAG stimulation of PKCγ; however, this was not tested directly in this study. It has been shown previously that the Cys1 subdomain of PKCγ is sufficient for translocation in response to TPA or a DAG analogue (Oancea et al., 1998). This suggests that, for some functions, the Cys1 and Cys2 subdomains of PKCγ are redundant.
A pseudosubstrate, which is found at the beginning of the C1 subdomain corresponding to residues 19–31 of PKCγ, is thought to suppress the activity of PKC kinases (Newton, 1997). Like PKC substrates, it is rich in basic amino acids; however, the pseudosubstrate contains an alanine in place of a serine or threonine. The pseudosubstrate is thought to block access to the active site of the catalytic domain, leading to autoinhibition of the kinase. Upon binding to DAG or phorbol esters, conformational changes in the C1 domain are predicted to release the pseudosubstrate from the active site and allow access to substrates (Srinivasan et al., 1996; Owens et al., 1998). Recently, physical interaction between the C1 and C2 domains of PKCα has been demonstrated (Slater et al., 2002). The SCA14 mutations and DAG or phorbol ester treatments may cause conformational changes in the N-terminal regulatory domain by disrupting a similar interaction in PKCγ. Data presented in this manuscript are consistent with a model whereby mutations in the C1 domain result in a conformational change that reduces the ability of the pseudosubstrate to suppress PKCγ kinase activity. Structural analysis of the mutant kinases will be required to confirm this hypothesis.
Understanding of the SCA14 disease mechanism may be advanced by a better understanding of which substrate proteins are more highly phosphorylated in Purkinje cells from affected individuals or animal models of the disease. It is interesting that, although the PKCγ isozyme is expressed broadly in the brain, the Purkinje cells appear to be particularly sensitive to the altered activity of mutant PKCγ. It has recently been shown that PKC inhibitors support Purkinje cell survival in cerebellar slice cultures (Ghoumari et al., 2002). The mechanism for this protection is not clear. However, it has also been shown that activation of PKC can prevent the retrograde transport of neurotrophins (Ozsarac et al., 2003), which is important for neuronal survival (Ginty and Segal, 2002). Therefore, increased PKCγ activity may reduce the viability of Purkinje cells by hindering the propagation of trophic signals.
Anomalous signal transduction appears to be at the core of SCA14 disease. This begs the question of whether other SCA diseases are also caused by aberrant signalling. Mutations have been identified in the FGF14 gene, which encodes a ligand for a tyrosine kinase receptor, in a form of spinocerebellar ataxia (van Swieten et al., 2003). These mutations likely reduce the function of the protein, since mice homozygous for disruption in the FGF14 gene develop ataxia and paroxysmal dyskinesia (Wang et al., 2002). In SCA12, a non-coding triplet repeat has been identified in the PP2R2B gene, which encodes a regulatory subunit of the serine threonine phosphatase, PP2A (Holmes et al., 1999, 2001). It appears that the triplet repeat expansion in this gene would change the expression of the regulatory subunit and may consequentially alter the phosphorylation of a subset of PP2A substrates. A serine phosphorylation site, S776, in ataxin-1 is required for full development of disease in an animal model for SCA1 (Emamian et al., 2003). This serine is a target for the AKT kinase and regulates binding to the scaffolding protein 14-3-3 (Chen et al., 2003b). It remains to be seen whether the expression of mutant forms of ataxin-1 itself leads to activation of AKT or other kinases, or the suppression of phosphatases that regulate the phosphorylation of this site. Using the Scansite consensus program (http://scansite.mit.edu/motifscan_seq.phtml) at low stringency, we found that PKCγ is predicted to phosphorylate S695 of ataxin-1, although the functional consequence of this is unknown. Interestingly, ataxin-1 levels have been reported to be lower in the Purkinje cells of an SCA14 patient, and PKCγ levels were found to be lower in a mouse model for SCA1 (Skinner et al., 2001; Chen et al., 2003a). This suggests a functional interaction of the SCA1 and SCA14 gene products and a possible convergence in the disease mechanisms. The identification of signalling pathways that are compromised by SCA diseases offers the hope of therapeutic interventions for these disorders.
We wish to thank Addis Taye for assistance with the western blots, Jim Nagle (NINDS sequencing facility) for the confirmatory sequencing and the members of the Neurogenetics Branch for many fruitful discussions. This work was supported by NINDS intramural funds.
1Neurogenetics Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, USA, and 2Department of Medical Genetics, University Medical Centre Utrecht, Utrecht, Netherlands