2339Research Article

IntroductionThe protein kinase C (PKC) family encompasses a group ofserine/threonine kinases that have important roles in many cellularprocesses. This family consists of ten isoforms that are divided intothree subclasses: conventional (PKCα, PKCβI, PKCβII and PKCγ),novel (PKCδ, PKCε, PKCη and PKCθ) and atypical (PKCζ andPKCι). This classification is based on their sensitivity for the secondmessengers calcium and diacylglycerol. The conventional isoformsrespond to both activators, whereas novel PKCs only react todiacylglycerol, and atypical PKCs are not activated by either calciumor diacylglycerol.

Recently, PKCγ (also known as PRKCG) was identified as thedisease-causing gene for the neurodegenerative disorderspinocerebellar ataxia type 14 (SCA14) (Chen et al., 2003). SCA14is characterized by a slowly progressive cerebellar ataxiaaccompanied by slurred speech and abnormal eye movements. Theataxic syndrome is due to loss of Purkinje cells leading tocerebellar neurodegeneration. PKCγ is the brain-specific memberof the PKC family and is highly expressed in Purkinje cell somaand dendritic processes, and regulates expression of long-termdepression at the parallel fiber of the Purkinje cell synapse,controlling synaptic plasticity (Daniel et al., 1998; Schrenk et al.,2002). Furthermore, PKCγ activity seems to control theelimination of climbing fibers during cerebellar developmentbecause Purkinje cells from PKCγ mutant mice persist withmultiple climbing fibers into adulthood (Chen et al., 1995; Kanoet al., 1995). These studies show the importance of PKCγ activityin Purkinje cell development and neuronal connectivity; however,

the molecular mechanisms by which PKCγ controls theseprocesses are still poorly understood.

Whereas many types of SCA are caused by coding polyglutaminerepeat expansion mutations in the encoding proteins, missensemutations in PKCγ were identified to cause SCA14 in families withdifferent genetic backgrounds (Alonso et al., 2005; Chen et al., 2003;Dalski et al., 2006; Fahey et al., 2005; Klebe et al., 2005; Stevaninet al., 2004; van de Warrenburg et al., 2003; Vlak et al., 2006; Yabeet al., 2003). Up to now, 23 different mutations have been identifiedthroughout the coding region of PKCγ, although about 75% of allSCA14 mutations are located in the highly conserved C1Bsubdomain of the regulatory domain of PKCγ. The impact of SCA14missense mutations in this particular subdomain remains unknown,but we hypothesized that the mutated amino acids are possiblycrucial for the structure of the C1B subdomain and could thereforeinduce conformational changes leading to altered regulation ofPKCγ activity.

PKC remains inactive in a ‘closed’ conformation viaintramolecular interactions. This involves binding of the pseudo-inhibitory substrate that is localized at the N-terminus of the variableprotein region (V1 domain) to the substrate-binding core in thecatalytic domain (Pears and Parker, 1991). In addition, threephosphorylation events are required for PKC to become fullymaturate, but still inactive. The first phosphorylation step iscatalyzed by phosphoinositide-dependent protein kinase 1 (PDPK1),which is followed by two autophosphorylation events (Dutil et al.,1994; Dutil et al., 1998). Upon activation, PKC rapidly translocatesfrom the cytoplasm to the plasma membrane and interacts with

Spinocerebellar ataxia type 14 (SCA14) is a neurodegenerativedisorder caused by mutations in the neuronal-specific proteinkinase C gamma (PKCγ) gene. Since most mutations causingSCA14 are located in the PKCγ C1B regulatory subdomain, weinvestigated the impact of three C1B mutations on theintracellular kinetics, protein conformation and kinase activityof PKCγ in living cells. SCA14 mutant PKCγ proteins showedenhanced phorbol-ester-induced kinetics when compared withwild-type PKCγ. The mutations led to a decrease inintramolecular FRET of PKCγ, suggesting that they ‘open’PKCγ protein conformation leading to unmasking of the

phorbol ester binding site in the C1 domain. Surprisingly,SCA14 mutant PKCγ showed reduced kinase activity asmeasured by phosphorylation of PKC reporter MyrPalm-CKAR, as well as downstream components of the MAPKsignaling pathway. Together, these results show that SCA14mutations located in the C1B subdomain ‘open’ PKCγ proteinconformation leading to increased C1 domain accessibility, butinefficient activation of downstream signaling pathways.

Key words: Spinocerebellar ataxia, Protein kinase C gamma, GFP,FLIM, FRET

Summary

PKCγ mutations in spinocerebellar ataxia type 14affect C1 domain accessibility and kinase activityleading to aberrant MAPK signalingDineke S. Verbeek1,*, Joachim Goedhart2, Laurie Bruinsma1, Richard J. Sinke3 and Eric A. Reits1

1Department of Cell Biology and Histology, Academic Medical Center, University of Amsterdam, Meibergdreef 15, 1105 AZ Amsterdam,The Netherlands2Section Molecular Cytology, Swammerdam Institute for Life Sciences, Centre for Advanced Microscopy, University of Amsterdam,The Netherlands3Department of Biomedical Genetics, University Medical Center Utrecht, University of Utrecht, The Netherlands*Author for correspondence (e-mail: D.S.Verbeek@medgen.umcg.nl)

Accepted 22 April 2008Journal of Cell Science 121, 2339-2349 Published by The Company of Biologists 2008doi:10.1242/jcs.027698

Jour

nal o

f Cel

l Sci

ence

JCS ePress online publication date 24 June 2008

2340

phospholipids, leading to the release of the pseudo-inhibitorysubstrate from the catalytic domain resulting in an ‘open’ and activeconformation (Newton, 2001). This translocation process is initiatedby calcium binding to the C2 domain and the subsequently bindingof diacylglycerol to the two C1 subdomains C1A and C1B (Newton,2004; Oancea and Meyer, 1998; Rodriguez-Alfaro et al., 2004; Sakaiet al., 1997). In the case of PKCγ, C1A and C1B bind its ligandwith comparable affinity in vitro (Quest et al., 1994; Quest andBell, 1994). The V1-domain that contains the pseudo-inhibitorysubstrate controls the accessibility of the C1 domain for phorbolester, which guarantees that PKC is sequentially activated bycalcium and diacylglycerol signals (Oancea and Meyer, 1998). Inaddition, masking of the phorbol ester binding sites in the C1 domainis also regulated via intramolecular interactions between the C2domain and the C-terminal V5 domain, because mutations in thesedomains make PKCα respond to DAG and increase the phorbolsensitivity of PKCβ (Stensman and Larsson, 2007).

Here we show that SCA14 mutations located in the C1Bsubdomain enhance PKCγ translocation from cytoplasm to plasmamembrane upon phorbol ester stimulation when compared withwild-type PKCγ. These mutations apparently increase theaccessibility of the C1 domain for phorbol ester by unmasking thephorbol ester binding sites of the C1 domain. Fluorescence lifetimeimaging microscopy (FLIM) analysis using YFP-PKCγ-CFP fusionproteins showed significantly lower intramolecular FRET signalswhen SCA14 mutations are present, demonstrating that the C1B

subdomain mutations result in an ‘open’ protein conformation.Surprisingly, the increased kinetics resulted in a decrease of PKCγkinase function upon activation. By using the PKC phosphorylationreporter MyrPalm-CKAR, we show that the mutations causingSCA14 reduce PKCγ kinase activity in living cells. In addition,reduced extracellular signal-regulated kinase (ERK) nucleartranslocation, and ERK phosphorylation was also observed in cellsexpressing mutant PKCγ. Together, these findings show that reducedPKCγ activity and aberrant mitogen-activated protein kinase(MAPK) signaling underlies SCA14.

ResultsPKCγ protein expression and subcellular localizationIn order to investigate the effects of SCA14 mutations on theintracellular distribution and kinetics of PKCγ in living cells, wegenerated both green and red fluorescently tagged (GFP/RFP) fusionproteins of both wild-type and SCA14 mutant PKCγ. In addition,we generated similarly tagged fusion proteins for the isolated C1Bsubdomains (Fig. 1B). The SCA14 mutations that were used in thisstudy include G118D, V138E and C142S (van de Warrenburg etal., 2003; Vlak et al., 2006) (Fig. 1A), which are all located in theC1B subdomain. Both full-length PKCγ and C1B were transientlyexpressed in HEK293T cells, followed by the analysis of cell lysatesby western blotting, which showed that all constructs wereefficiently expressed (Fig. 1C). Visualization of the intracellulardistribution of the proteins in living cells showed that PKCγ had a

cytoplasmic distribution, whereas themuch smaller C1B subdomain wasdistributed throughout the cytoplasmand nucleus (Fig. 1D). The threeSCA14 mutations did not alter thecellular distribution of either full-length PKCγ or the C1B subdomain.No massive aggregation was observedin cells expressing wild-type ormutant PKCγ, as was shown by Sekiand colleagues (Seki et al., 2005).Only a low percentage of cellsshowed accumulation of full-lengthPKCγ-GFP protein in the perinuclearregion of the cells, probably becauseof high levels of PKCγ overexpression(data not shown).

Journal of Cell Science 121 (14)

Fig. 1. Scheme and protein expression levels of wild-type and SCA14 mutantPKCγ. (A) Schematic representation of the primary structure of PKCγ,including the V1-domain containing the pseudoinhibitory substrate, theregulatory domain comprising the C1 and C2 domains and the catalyticdomain comprising the C3 and C4 domains. SCA14 mutations used in thisstudy are indicated in black. (B) Schematic representation of the constructs.Both full-length PKCγ and the C1B subdomain are fused to GFP and RFP. (C)Western blot for PKCγ, demonstrating equal expression of the different PKCγalleles (left). Similar expression levels of the C1B fusion proteins were alsodetected using anti-GFP (right). Actin was used as a loading control. (D)Confocal images of transiently transfected HeLa cells with either full-lengthPKCγ-GFP or C1B-GFP fusion plasmid. SCA14 mutations do not alter thecellular localization and distribution of full-length PKCγ and C1B proteins inHeLa cells. Scale bars: 10 μm.

Jour

nal o

f Cel

l Sci

ence

2341Mutant PKCγ affects MAPK signaling

SCA14 mutations located in the C1B subdomain enhancePKCγ kinetics upon phorbol ester activationModelling and mutational studies have shown that multiple residuesin the C1B subdomain of PKCδ are critical for Zn2+ ion coordinationand phorbol ester binding, including the residues Val138 and Cys142(Kazanietz et al., 1995; Quest et al., 1994; Xu et al., 1997).Val138Gly and Cys142Gly mutations completely abolish phorbolester binding to the C1B subdomain. Given the homology betweenthe different isoforms, we aimed at investigating whether the SCA14mutations G118D, V138E and C142S would affect PKCγ activationupon phorbol ester stimulation.

We started by measuring the kinetics of PMA-inducedtranslocation of both wild-type C1B and mutant C1B subdomainsin living cells. Initially, all experiments were performed using singleplasmid transfections; however, to exclude cell-cell variation andto get a more accurate comparison between wild-type and mutant,co-transfection experiments were carried out with wild-type GFPand PKCδ mutant RFP plasmids. Similar results were obtained forboth experiments (data not shown). HeLa cells were cotransfectedwith both wild-type C1B-GFP and mutant C1B-RFP tosimultaneously image both translocation events using confocalmicroscopy. The translocation kinetics of G118D mutant C1Bsubdomain by PMA was similar to wild-type C1B (Fig. 2A,B). Bycontrast, no translocation of the V138E and C142S mutant C1Bsubdomains was observed upon PMA stimulation. Similar resultswere obtained when the GFP and RFP fluorophores were exchangedbetween the wild-type and mutant domains, excluding the possibilitythat this effect is fluorophore dependent (data not shown). Takentogether, the results demonstrate that SCA14 mutations, which areall located in the C1B subdomain, apparently work via a differentmechanisms, because some, but not all, of the disease-relatedmutations abolish PMA-induced translocation.

Next, we examined whether the SCA14 mutations also affectPMA kinetics of full-length PKCγ, because it has been shown that

the C1A subdomain of PKCγ can also bind PMA with comparableaffinity as C1B (Ananthanarayanan et al., 2003). Therefore, werepeated the PMA-induced translocation experiment using full-length wild-type PKCγ-GFP and mutant PKCγ-RFP. Surprisingly,none of the SCA14 mutations impaired PMA-induced translocation,but resulted in enhanced translocation kinetics when compared withwild-type PKCγ (Fig. 3A,B). This effect was also not dependenton the fused fluorophores, because exchange of the GFP and RFPfluorophores showed identical results (data not shown). IncreasedPMA-induced translocation has also been observed for a PKCγwitha deleted V1 domain, which contains the pseudo-inhibitory substrate(Oancea and Meyer, 1998). Their results suggest that theaccessibility of the C1 domain for phorbol ester is restricted by a‘clamp’ between the V1 domain and the catalytic domain. Removalof the V1 domain results in an ‘open’ protein conformation, whichleads to exposure of the phorbol binding sites in the C1 domain.To test whether SCA14 mutations in PKCγ lead to similartranslocation kinetics, we compared the kinetics of SCA14-mutantPKCγ-GFP and deleted V1 mutant, deltaV1-PKCγ-GFP upon PMAstimulation. Both mutants showed similar rates of translocation uponPMA stimulation, suggesting that SCA14 mutations may affect theregulation of the catalytic domain by the V1-clamp (Fig. 3C). Inaddition, we also tested whether the constitutively activepseudosubstrate mutant A24E-PKCγ showed increased translocationkinetics upon PMA stimulation. The A24E mutation leads toconstant exposure of the pseudosubstrate, leading to completeabolishment of the V1-clamp to the catalytic domain and inducestherefore an ‘open’ protein conformation (Pears et al., 1990). Indeed,the translocation kinetics of A24E-PKCγ to the plasma membraneupon PMA stimulation was similar to deltaV1-PKCγ and SCA14-mutant PKCγ (data not shown).

These data suggest that the enhanced translocation kinetics ofSCA14 mutant PKCγ upon PMA stimulation might result fromincreased accessibility of the overall C1 domain for phorbol ester,

Fig. 2. SCA14 mutations located inC1B affect C1B translocationkinetics upon phorbol esterstimulation. (A) Confocal images ofHeLa cells cotransfected with C1B-GFP and SCA14 mutant C1B-RFP.PMA activation induces irreversibletranslocation of C1B-GFP and C1BG118D-RFP to the plasmamembrane. By contrast, the V138Eand C142S C1B-RFP did notrespond to PMA activation. Imagesshown were recorded prior to (0)and 200 seconds after addition of400 nM PMA. Scale bars: 10 μm.(B) Translocation analysis ofcotransfected C1B-GFP and mutantG118D, V138E or C142S C1B-RFPshowed that the V138E and C142Smutations impair phorbol ester (400nM PMA)-induced membranetranslocation. Each of the traces in Band C represent mean ± s.e.m. of 4-6cells from at least three independentexperiments.

Jour

nal o

f Cel

l Sci

ence

2342

as a result of conformation changes under the influence of disease-related mutations in the PKCγ protein.

SCA14 mutant PKCγ shows increased membrane targetingwithout affected C2 domain functioningAs PKC is sequentially activated by calcium (C2 domain) andphorbol ester (C1 domain), we wanted to determine the effect of aphysiological stimulus on the translocation kinetics of PKCγ.Therefore, we tested whether SCA14 mutations would affecthistamine stimulation of PKCγ, because activation of the histaminereceptor induces the rapid release of endoplasmic-reticulum-storedcalcium and production of diacylglycerol. HeLa cells weretransiently cotransfected with wild-type (red) and SCA14-mutant(green) PKCγ and imaged in real time using histamine stimulationand subsequent PMA and ionomycin treatment to obtain full PKCγtranslocation. Although no difference was detected in thetranslocation pattern between two wild-type PKCγ protein fused toGFP and RFP (Fig. 4A), the amplitude of translocation of the SCA14

mutant PKCγ was 1.5±0.03 (G118D; mean ± s.e.m.), 1.9±0.1(V138E) and 1.4±0.05 (C142S) times significantly increased whencompared with wild-type PKCγ (G118D; V138E; C142S unpairedt-test, n=3-5; P=0.0002, P=0.0026 and P=0.0007, respectively) (Fig.4B-E). The graphs are representatives of at least three independentexperiments. These results indicate that the SCA14 mutationsenhance the amount of PKCγ protein at the plasma membrane uponstimulation of cells under physiological conditions.

To resolve whether the enhanced membrane targeting of SCA14-mutant PKCγ was solely caused by the increased accessibility ofthe C1 domain for phorbol ester or also by changes in Ca2+ bindingto the C2 domain, we measured PKCγ translocation to the plasmamembrane upon Ca2+ ionophore stimulation. This ionophoreactivation will result in opening of Ca2+ pores in the plasmamembrane leading to a rapid increase in intracellular Ca2+. Nodifference in translocation kinetics was observed between wild-typeand SCA14 mutant PKCγ (Fig. 4F), indicating that stimulation viathe C2 domain was not affected.

Overall, these results show that SCA14 mutations located in theC1B subdomain principally affect the phorbol ester sensitivity ofthe C1 domain, which leads to increased membrane targeting ofSCA14 mutant PKCγ.

Unmasking of the C1 domain via SCA14 mutations isassociated with altered PKCγ conformationOur kinetic studies suggest that the SCA14 mutations located inthe C1B subdomain unmask the C1 domain, leading to exposureand increased accessibility of the phorbol ester binding sites. Thiscould be the result of destabilization of the ‘closed’ conformation,

Journal of Cell Science 121 (14)

Fig. 3. SCA14 mutations located inC1B enhanced PKCγ translocationkinetics upon phorbol esterstimulation. (A) Confocal imagesof HeLa cells cotransfected withwild-type PKCγ-GFP and SCA14mutant (G118D, V138E andC142S) PKCγ-RFP. PMAactivation induces irreversibletranslocation of PKCγ to theplasma membrane. Images shownwere recorded 140 and 440seconds after addition of 400 nMPMA. Scale bars: 10 μm.(B) Translocation analysis showedthat PMA (400 nM) stimulationresulted in an increase intranslocation kinetics of all threeSCA14 mutant PKCγ proteinscompared with wild-type PKCγ tothe plasma membrane.(C) Translocation analysis of wild-type PKCγ-GFP, SCA14 mutantPKCγ-GFP (G118D, V138E andC142S) and ΔVPKCγ-GFP uponPMA (400 nM) stimulation. Eachof the traces in B and C representmean ± s.e.m. of 4-6 cells from atleast three independentexperiments.

Jour

nal o

f Cel

l Sci

ence

2343Mutant PKCγ affects MAPK signaling

which will consequently provoke unmasking of the C1domain. To examine the differences in PKCγ proteinconformation caused by C1B SCA14 mutations, we tookadvantage of fluorescence resonance energy transfer (FRET)analysis. It has been shown previously that intramolecularFRET can be used to probe protein conformation in singleliving cells by tagging the protein of interest with both a donorand an acceptor fluorophore (Giepmans et al., 2006;Miyawaki, 2003). The FRET signal is dependent on thedistance and orientation of the two fluorophores and alteredprotein conformations of PKCγ may lead to a changeddistance between the N- and C-terminus and subsequentchange in the FRET signal. PKCγ FRET constructs weregenerated by tagging the N-terminus with EYFP and the C-terminus with ECFP.

To examine whether FRET can be used to distinguishbetween different conformations, we measured FRETefficiencies of the wild-type PKCγ, which is in a ‘closed’conformation, and the A24E mutant which cannot bind thepseudosubstrate (N-terminus) to the substrate-binding groveand consequently remains in an ‘open’ conformation.Although FRET was observed in these constructs, nodifferences in FRET efficiency were measured (data notshown). Since it is known that the ECFP and EYFP have thetendency to dimerize, the ECFP was replaced by amonomerized, brighter variant (Kremers et al., 2006), toabolish intramolecular dimerization. FLIM was used tomeasure FRET efficiencies (van Munster and Gadella, 2005)using SCFP3A as a control. Both phase and modulationlifetimes of SCFP3A were strongly reduced in the wild-typeYFP-PKCγ-SCFP3A protein, yielding a FRET efficiency of29% (Fig. 5A-C). By contrast, the SCFP3A lifetime of theA24E-PKCγ mutant was close to values of the unquenchedSCFP3A with a relatively low FRET efficiency of 11% basedon the phase lifetime (Fig. 5B,C) and 10% based on themodulation lifetime. These results fit with the model in whichthe N- and C-terminus are close together in the closedconformation of inactive wild-type PKCγ, whereas the terminiare further apart into an open conformation of theconstitutively active A24E mutant. These results showed thatintramolecular FRET can be used to distinguish between theclosed and open PKCγ conformation in unstimulated cells.When we monitored the changes in protein conformationupon activation of PKC by PMA, we consistently observeda FRET increase for all constructs (data not shown), whereasa decrease was expected, as this would correspond to the openactive conformation. Similar observations have been reportedfor the PKCδ isoform (Braun et al., 2005). This can beexplained by accumulation of PKC at the plasma membraneinduced by PMA. Owing to the relative high surfaceconcentration of PKC molecules at the plasma membrane,intermolecular aspecific FRET takes place. Therefore, thisapproach did not allow monitoring changes in conformationupon activation of PKC.

Subsequently, we incorporated the SCA14 mutations, G118D,V138E and C142S into the wild-type PKCγ FRET construct. TheSCA14 mutant PKCγ gave rise to significantly lower FRETsignals than wild-type PKCγ (G118D; V138E; C142S, allP<0.0001, unpaired t-test, n=5) (Fig. 5A-C). The three SCA14mutant PKCγ displayed approximately similar FRET efficienciesranging from 14% to 16% (Fig. 5C), based on the phase lifetime

(and 11-12% based on the modulation lifetime). Interestingly, theseFRET signals were similar to the value measured for A24E-PKCγ(Fig. 5A,B). Although the FRET efficiency measured in theSCA14 mutants and A24E mutant is significantly reduced relativeto wild-type PKCγ, it cannot be concluded from these experimentsthat the conformations are identical. All proteins were properlyexpressed, excluding the possibility that the lower FRET signal

Fig. 4. SCA14 mutant PKCγ shows increased membrane targeting but does notaffect C2 domain functioning. (A) Translocation analysis of HeLa cellscotransfected with wild-type PKCγ-RFP and wild-type PKCγ-GFP. Both proteinsinduced rapid reversible translocation to the plasma membrane upon histaminestimulation. No difference was observed in translocation pattern between the twowild-type PKCγ proteins fused to different fluorophores. Furthermore, wild-typePKCγ-RFP was cotransfected with SCA14 mutant G118D (B), V138E (C) andC142S (D) PKCγ-GFP. No difference was observed in membrane dissociationbetween wild-type and SCA14-mutant PKCγ. Increased translocation amplitudeswere observed for SCA14 mutant PKCγ compared with wild-type PKCγ. Graphsshow the translocation curve of single cells, but represent three independentexperiments. (E) Quantification of the amount of PKCγ translocated to the plasmamembrane upon histamine stimulation. Bars represent the ratio between themaximal amplitude of wild-type PKCγ and the maximal amplitude of SCA14-mutant PKCγ. Data are means ± s.e.m. of three experiments each on three cells(unpaired t-test, **P<0.01; ***P<0.001). (F) No differences were observed in C2domain-induced translocation kinetics of full-length PKCγ-GFP and mutant G118D,V138E or C142S PKCγ-GFP in response to 5 μM A23187. The curves are mean ±s.e.m. of 3-5 cells and the data are representative of three experiments.

Jour

nal o

f Cel

l Sci

ence

2344

was due to degradation of the SCA14-mutant PKCγ proteins (datanot shown).

In conclusion, these results demonstrate that SCA14 mutationslocated in the C1B subdomain change PKCγ from a closedconformation into a different, most likely, open conformation inwhich the C- and N-termini are further apart.

SCA14 mutant PKCγ shows reduced kinase activation in livingcellsSince SCA14 mutant PKCγ display enhanced PMA kinetics andexhibit a different protein conformation, we hypothesized that C1BSCA14 mutations alter the phosphorylation capacity and thusactivity of PKCγ. In order to measure PKCγ activity in living cells,we used the PKC phosphorylation reporter, MyrPalm-CKAR,which is a membrane-bound intramolecular FRET reporter (Violinet al., 2003). Immediate phosphorylation of the PKC substratesequence upon PKC activation causes a conformational change thatreduces the amount of FRET between the CFP and YFPfluorophores. Unfortunately, the PKC-substrate peptide sequenceis not selectively accessible for PKCγ but also permitsphosphorylation of PKCα and PKCβ. Therefore, we first testedwhether activation of expressed wild-type PKCγwould lead to lowerFRET ratios than activation of endogenous PKC (α and β) levelsin HeLa cells. HeLa cells were transiently transfected withMyrPalm-CKAR or cotransfected with wild-type PKCγ-RFP. First,we stimulated the cells with histamine, which only induced partialphosphorylation of MyrPalm-CKAR, but did not show muchdifference in FRET ratio and magnitude of activation between cellsoverexpressing wild-type and endogenous PKC in response tohistamine (Fig. 6A). Since cells react rather heterogeneously tostimulation by histamine, we decided to subsequently add PMA toobtain the maximal amplitude of the reporter corresponding tomaximal phosphorylation. Cells overexpressing PKCγ showedsignificant lower FRET ratios (24.2±0.4; mean ± s.e.m.) (P<0.0001,unpaired t-test, n=5) compared with cells that did not express PKCγ(15.3±0.2), demonstrating that we could detect PKCγ-specificCKAR phosphorylation (Fig. 6A). We then investigated whether

changes in PKCγ protein conformation due to unmasking of the C1domain are reflected in a different MyrPalm-CKAR FRET ratio.Surprisingly, significantly lower FRET efficiencies (G118D; V138E;C142S: 20.6±0.4; 18.4±0.5; 18.0±0.6; mean ± s.e.m.; P=0.0002;P=0.0001; P<0.0001, unpaired t-test, n=3-5) were observed in cellsexpressing SCA14 mutant PKCγ than cells containing wild-typePKCγ upon PMA stimulation, indicating less MyrPalm-CKARphosphorylation (Fig. 6B). These results suggest that the SCA14mutant PKCγ has less capability to phosphorylate cellular substratesand thus subsequently exhibit reduced kinase activity in living cells.This finding was confirmed in cell lysate of cells overexpressingwild-type or SCA14 mutant PKCγ using an anti-phosphoserine-PKC-substrate antibody. We observed an overall reduction inkinase activity in cell lysates of SCA14 mutant PKCγ whencompared with wild-type PKCγ, because a number of unidentifiedPKC targets showed a reduction in phosphorylation upon PMAstimulation (data not shown). To examine whether reduced SCA14mutant PKC activity upon activation is the result of increased basalmembrane activity due to the open conformation, we used the PKCinhibitor Gö6976 to block basal PKC activity. Upon addition ofGö6976 to unstimulated cells expressing either wild-type PKCγ orSCA14 mutant PKCγ, we observed similar increased YFP/CFPFRET ratios in both cases (data not shown). This result suggeststhat the MyrPalm-CKAR reporter is constitutively phosphorylatedto a similar extent by either wild-type PKCγ or SCA14 mutantPKCγ.

Since these experiments only report on the phosphorylation ofan artificial substrate or unknown PKC targets, we also wanted toexamine whether relevant cellular target proteins were also affectedby reduced phosphorylation. Therefore, we studied thephosphorylation status of the known intracellular PKC targetmyristoylated alanine-rich C-kinase substrate (MARCKS).MARCKS is a multifunctional protein implicated in cell motility,membrane trafficking and mitogenesis, because its activity regulatesgrowth cone adhesion and pathfinding (Gatlin et al., 2006; Myatet al., 1997). We determined the phosphorylation level ofendogenous MARCKS in lysates of cells transfected with either

Journal of Cell Science 121 (14)

Fig. 5. SCA14 mutations located in C1Bopen PKCγ protein conformation.(A) Fluorescence lifetime imaginganalysis of the indicated constructs,showing in each column, from left toright, the fluorescence intensity, false-color map of the fluorescence lifetimecalculated from the phase shift (tau-phi)and histogram of the lifetime distributionwith the same false-color scale as thelifetime map. Scale bar: 10 μm. (B) Thetable summarizes the average phase andmodulation lifetimes (± s.d.). The numberof cells measured is indicated by n.(C) FRET efficiency (mean ± s.e.m.;unpaired t-test, ***P<0.001) andquantification are calculated based on thephase lifetimes, using the SCFP3Alifetime as the control.

Jour

nal o

f Cel

l Sci

ence

2345Mutant PKCγ affects MAPK signaling

wild-type or SCA14 mutant PKCγ upon PMA activation, using ananti-phospho-MARCKS antibody. Immunoblotting showed no basalphosphorylation of MARCKS, but when PMA was added, reducedMARCKS phosphorylation in cells expressing SCA14 mutant PKCγwas observed compared with that in cells containing wild-type PKCγ(Fig. 6C,D).

These experiments demonstrate that SCA14 mutations locatedin the C1B subdomain affect PKCγ kinase activity, resulting inreduced phosphorylation of downstream targets.

SCA14 mutant PKCγ affects downstream MAPK signalingThe MAPK pathways that activate ERK play an important role incell mobility, cell migration, dendrite spine plasticity and long-termpotentiation in learning and memory functions (De Zeeuw et al.,1998; Pullikuth and Catling, 2007; Robinson et al., 1998).Misregulation of cell migration has already been associated withneurodegeneration (Hafezparast et al., 2003; Vallee et al., 2001)and as result, we hypothesized that aberrant MAPK and ERKsignaling may be affected in SCA14. The movement of ERK2 intothe nucleus has been suggested to be important for the long-term

consequences of ERK activation (Horgan and Stork, 2003).Therefore, we determined ERK2 cytoplasmic-nuclear shuttlingusing confocal microscopy. GFP-ERK2 was transiently transfectedin HeLa cells either with wild-type PKCγ-RFP or SCA14 mutantPKCγ-RFP and stimulated them with PMA. PKC activation induceda transient GFP-ERK2 nuclear translocation within 10 minutes ofaddition of PMA (Fig. 7A,B). Translocation analysis showed asignificant reduction in nuclear accumulation (G118D; V138E;C142S: P=0.0024; P=0.0009; P=0.0011, unpaired t-test, n=3) ofGFP-ERK2 in the nucleus of cells coexpressing SCA14 mutantPKCγ (G118D; V138E; C142S: 2.0±0.04; 1.7±0.03; 1.7±0.06; mean± s.e.m.) compared with cells coexpressing wild-type PKCγ(3.0±0.1) or cells only expressing GFP-ERK2 (2.9±0.1) (Fig.7B,C). To exclude the idea that reduced nuclear ERK translocationwas caused by basally higher ERK levels in the nucleus of cellscoexpressing SCA14 mutant PKCγ, we determined nuclear ERK2-GFP expression levels of cells coexpressing either wild-type PKCγor SCA14 mutant PKCγ using western blotting. No increasednuclear ERK2-GFP expression was observed in cells expressingSCA14 mutant PKCγ compared with that cells expressing wild-

Fig. 6. SCA14 mutations in C1B reduce PKCγ kinase activity in living cells. (A) HeLa cells were transiently transfected with either MyrPalm-CKAR alone orcotransfected with wild-type PKCγ-RFP or the different SCA14 mutant PKCγ-RFP. Intramolecular FRET was measured by determining the ratio of the YFP/CFPfluorophore intensities of the MyrPalm-CKAR molecule. Upon phosphorylation, less FRET from CFP to YFP will be measured because conformational changesincrease the distance between the N- and C-terminus. MyrPalm-CKAR shows reduced phosphorylation, as was shown by less loss of FRET in the presence ofSCA14 mutant PKCγ (G118D, V138E and C142S) when compared with wild-type PKCγ upon stimulation with 500nM PMA. The data traces were corrected bythe background images acquired prior to adding ligand. The corrected traces were normalized to 1 by dividing the traces by the average baseline FRET ratio.(B) Bars represent the FRET efficiencies of the MyrPalm-CKAR reporter in the absence and presence of wild-type PKCγ and SCA14 mutant PKCγ (unpairedt-test, ***P<0.001). (C) Western blot analysis shows reduced MARCKS phosphorylation in whole cell lysates of SCA14-mutant PKCγ upon PMA stimulation(400 nM) compared with cells containing wild-type PKCγ using anti-phospho-MARCKS antibody. Top panel shows equal PKCγ protein levels in all cell lysates.(D) Quantification of band intensity. Bar graph represents the phosphorylation of MARCKS normalized to the amount of MARCKS and PKCγ, and related to thevalue obtained for cells expressing wild-type PKCγ. Data in B and D represent mean ± s.e.m.

Jour

nal o

f Cel

l Sci

ence

2346

type PKCγ (data not shown). As nuclear accumulation is associatedwith ERK phosphorylation (Costa et al., 2006), we also determinedthe phosphorylation profile of endogenous ERK in cells upon PKCactivation. Notably, no significant basal phosphorylation of ERKwas observed in unstimulated cells overexpressing wild-type PKCγor any of the SCA14 mutants. In accordance with the translocationassay, reduced levels of ERK phosphorylation were observed incells expressing SCA14 mutant PKCγ when compared with cellsexpressing wild-type PKCγ (Fig. 7D,E).

Together these data show that SCA14 mutations located in theC1B subdomain affect PKCγ kinase activity, lead to lower ERKphosphorylation levels and accordingly reduce nuclear ERKtranslocation. Both events will reduce MAPK signaling, and mighttherefore lead to reduced expression of target genes affecting neuriteoutgrowth and survival.

DiscussionIn resting cells, the accessibility of the C1 domain for phorbol esteris well regulated by masking of the C1 domain, which prevents

exposure of the phorbol binding sites. This ‘closed’ proteinconformation is necessary to keep PKC inactive. Although twoSCA14 mutations, V138E and C142S, completely abolishtranslocation of the isolated C1B subdomain to the plasmamembrane upon phorbol ester stimulation (Fig. 2A,B), the G118Dmutation does not seem to affect C1B PMA-induced kinetics.However, when present in the full-length PKCγ protein, all threeSCA14 mutations enhanced PKCγ translocation to the plasmamembrane upon phorbol ester stimulation (Fig. 3A,B). Thesefinding suggest that SCA14 mutations not only affect the regulatoryfunction of C1B subdomain, but also affect the C1A domain. Thispossibility would reconcile our findings with previous studiesshowing that the C1A and C1B subdomains of PKCγ have similaraffinities for phorbol ester in vitro (Ananthanarayanan et al., 2003).They thus have redundant roles in PKCγ phorbol ester activation,because even the SCA14 mutant PKCγ lacking a functional C1Bsubdomain showed enhanced PMA-induced translocation.Furthermore, the residues V138 and C142 have already been shownto be crucial for phorbol ester binding (Kazanietz et al., 1995), and

modelling of the crystal structure of the C1Bsubdomain also demonstrated that the mutated residuesindeed led to a dysfunctional phorbol ester binding site(our unpublished data). Interestingly, despite properC1B subdomain activation upon phorbol esterstimulation, the G118D mutation also enhanced thetranslocation kinetics of the full-length PKCγ protein,similarly to the other mutated proteins. Apparently, the

Journal of Cell Science 121 (14)

Fig. 7. SCA14 mutant PKCγ reduces ERK2nuclear accumulation and phosphorylation.(A) Confocal images of HeLa cellstransiently transfected with GFP-ERK2 andcotransfected with either wild-type PKCγ-RFP or PKCγ G118D-RFP (V138E andC142S mutants confocal data not shown).Activation of PKCγ with PMA inducedGFP-ERK2 nuclear accumulation in time.Images shown are recorded 250 and 500seconds after addition of 400 nM PMA.Scale bar: 10 μm. (B) Translocationanalysis shows that GFP-ERK2 nucleartranslocation was reduced in cellscoexpressing SCA14 mutant PKCγ whencompared with cells that express only GFP-ERK2 or coexpress wild-type PKCγ-RFP.(C) Bars represent the ratio ofnuclear/cytosolic ERK fluorescence in theabsence and presence of wild-type PKCγand SCA14 mutant PKCγ (unpaired t-test,**P<0.01; ***P<0.001). (D) Western blotanalysis showed that endogenous ERK1and ERK2 are phosphorylated after PKCγactivation with 400 nM PMA for 10minutes. Reduced ERK2 phosphorylationwas observed in the presence of SCA14mutant PKCγ (G118D, V138E and C142S)compared with cells expressing wild-typePKCγ. Furthermore, equal ERK and PKCγlevels were observed using anti-ERK andanti-PKC antibodies. (E) Quantification ofthe band intensity. Bar graph represents thephosphorylation of ERK normalized to theamount of ERK and PKCγ and related tothe value obtained for cells expressing wildtype PKCγ. Data in C and E represent mean± s.e.m.

Jour

nal o

f Cel

l Sci

ence

2347Mutant PKCγ affects MAPK signaling

SCA14 mutations affect PKCγ activation via a different mechanismbut with a similar outcome.

Enhanced PKCγ translocation upon phorbol ester stimulation wasalso observed when the V1-domain containing the pseudoinhibitorysubstrate was deleted in PKCγ. This study showed that removal ofthe V1-domain increased the accessibility of the C1 domain forphorbol ester (Oancea and Meyer, 1998). SCA14 mutations locatedin the C1B subdomain apparently affect PKCγ kinetics in a similarway, by altering the conformation of the protein, leading tounmasking of the C1 domain as was shown in Fig. 3C. The changedconformation of the SCA14 mutants also became obvious from ourFRET studies (Fig. 5). We showed that all three SCA14 mutationsled to an open PKCγ protein conformation, which may be similarto that of PKCγ with mutations in the V1 domain (ΔV and A24E),because all three variants showed enhanced PMA-induced kinetics.

Although the SCA14 mutations in the C1B subdomain showedsimilar kinetics and conformational changes as modification ordeletion of the upstream V1-domain, they could also affectfunctioning of the other regulatory domain, the C2 domain (Oanceaand Meyer, 1998; Slater et al., 2002; Stahelin et al., 2005; Stensmanet al., 2004). However, we did not observe any effect of the SCA14mutations on C2 functioning as measured by PKCγ calciumsensitivity (Fig. 4E). Furthermore, the SCA14 mutations located inthe C1B subdomain did not impair PKCγ membrane dissociationas shown by the reversible histamine translocation. However,SCA14 mutant PKC did show increased membrane binding uponhistamine stimulation, demonstrating that SCA14 mutant PKCγ hasincreased affinity for the plasma membrane under physiologicalconditions (Fig. 4A-D).

As enhanced PMA kinetics and open PKC protein conformationsare always associated with active PKC, we hypothesized that SCA14mutant PKCγ would increase signaling pathways upon activation.To our surprise, the SCA14 mutations located in the C1B subdomainreduced PKCγ kinase activity in living cells as demonstrated usingthe fluorescent membrane-bound reporter MyrPalm-CKAR (Fig.6). Moreover, both MARCKS and a more downstream target ERK2showed reduced PKCγ-mediated phosphorylation, which led toreduced MARCKS and MAPK activation in living cells (Fig. 6Cand Fig. 7). Importantly, we did not observe enhanced basalphosphorylation of CKAR, MARCKS or ERK as a result ofoverexpression of SCA14 mutants. These results suggest that if theSCA14 mutants have increased basal activity, this activity iseffectively counteracted in cells by phosphate activity.

PKC has been shown to regulate the structure and function ofthe cell by phosphorylating various cytoskeletal or functionalproteins, such as MARCKS or annexins (Keenan and Kelleher,1998). MARCKS has been implicated in regulating cell attachment,membrane ruffling and spreading (Myat et al., 1997). Furthermore,its activity is important for the reorganization of cytoskeletalstructures, which are crucial events in growth cone formation anddendrite outgrowth (Geddis and Rehder, 2003; Matsuoka et al.,1998). Although there may be several possibilities that may explainthe SCA14 disease pathology, defects in MARCKS-dependent celladhesion or spreading may explain the neurodegenerativephenotype. The role of ERK signaling in neurodegenerativedisorders is more complex, because both protective and deleteriousroles for ERK activation in neuronal cells have been described(Apostol et al., 2006; Colucci-D’Amato et al., 2003; Lievens et al.,2005). The diversity of ERK activation is not surprising, becausethe pathway integrates various signals depending on the cellularcontext, including cell proliferation, differentiation, migration and

survival (Bhalla and Iyengar, 1999; Chuderland and Seger, 2005;Hetman and Gozdz, 2004). This diversity is also mediated by severalfactors, such as cell or tissue type, the duration of activation, aswell as the subcellular localization (Ebisuya et al., 2005). However,our results suggest that reduced ERK signaling underlies SCA14disease.

To address the question of how increased PKCγ kinetics resultingfrom an open conformation corresponds to reduced activity, we arecurrently studying the maturation process of SCA14 mutant PKCγand performing membrane-binding studies to determine maximalPKC activity upon membrane interactions. We suggest that an openPKCγ protein conformation at resting state affects its maturationprocess via PDPK1. The maturation of PKCs after synthesisinvolves the phosphorylation of the enzyme at multipleserine/threonine sites. Phosphorylation by PDPK1 is the firstphosphorylation event in this maturation process and acts as an on-off switch for PKC activity. As PDPK1 preferably interacts andcovalently modifies substrates when they are in open conformations(Sonnenburg et al., 2001), we hypothesize that SCA14 mutationsaffect the proper maturation process of PKCγ leading to reducedPKC kinase activity.

Interestingly, the neurological phenotype of SCA14 patients(heterozygous mutations) is more severe than that of the PKCγ-knockout mice (Chen et al., 1995). Therefore, the increase in SCA14mutant PKCγ kinetics but decrease in activity may explain thedominant-negative effect of the mutated PKCγ protein, because itmight compete with wild-type PKCγ for its ligands at the plasmamembrane. A similar competition between PKC isoforms forphorbol ester binding has been observed previously (Lenz et al.,2002), where agonist stimulation induced selective membranerecruitment of classical and/or novel PKC isoforms. Thisphenomenon may also explain why even less ERK phosphorylationwas observed in cell lines overexpressing SCA14 mutant PKCγcompared with cells that did not overexpress PKCγ (Fig. 7D).

Although SCA14 is caused by mutations in the PKCγ gene, otherSCA types that show comparable disease phenotypes are causedby mutations in genes encoding other signaling proteins. Theseproteins include tau-tubulin-associated kinase 2 (SCA11), α1subunit of the P/Q-type calcium channel (SCA6), regulatory subunitof the protein phosphatase PP2A (SCA12) and fibroblast growthfactor 14 (SCA27) (Holmes et al., 2001; Houlden et al., 2007; vande Leemput et al., 2007; van Swieten et al., 2003; Zhuchenko etal., 1997). Future research is required to determine whether thesedifferent proteins are components of shared signaling routes or evenare linked within one common signaling cascade that underlies SCAdisease in general. Furthermore, FGF signaling, Ca2+ influx andphosphatase activity also regulate PKC functioning, and place PKCγat an important position in neuronal signaling pathways. Therefore,it is not surprising that PKC is involved in various diseases, suchas cancer, cardiac and lung diseases, diabetes and neurodegenerativedisorders, such as SCA14. To select PKC as a target for therapeuticapproaches, we need to know more about PKC isoform differencesto generate selective inhibitors or activators. Neverthelessinvestigating ‘naturally occurring’ PKC mutations may help us tounderstand the mechanisms of PKC activation and regulation andthe subsequent effects on downstream signaling.

In conclusion, we show that SCA14 mutations located in the C1Bsubdomain of PKCγ enhance PMA-induced kinetics and an openPKCγs protein conformation. This results in increased membranetargeting upon physiological stimuli but leads to decreased kinaseactivity and downstream MAPK signaling. We propose that SCA14

Jour

nal o

f Cel

l Sci

ence

2348

mutations unmask the C1 domain, expose the phorbol binding sitesby opening of the PKCγ protein conformation, which subsequentlyresults in increased accessibility of the whole C1 domain for phorbolester. The increased accessibility subsequently leads to increasedPMA translocation kinetics. Unexpectedly, these eventscorresponded with decreased kinase activity, as shown by thereduced phosphorylation capacity of PKCγ in living cells. As aconsequence, SCA14 mutant PKCγ reduced ERK activation, whichhighlights the MAPK signaling pathway for possible therapeuticintervention in SCA14 disease.

Materials and MethodsDNA constructsThe wild-type and SCA14 mutant Gly118Asp PKCγ-GFP constructs were describedpreviously (Verbeek et al., 2005). The Val138Glu and the Cys142Ser mutations wereintroduced into PRKCG cDNA with the Quikchange II Site-Directed Mutagenesiskit (Stratagene, La Jolla, CA). For the Val138Glu and the Cys142Ser mutations thefollowing primers were used: Val138Glu forward, 5�-GTC GCG AGA TGA ACGAGC ACC GGC GCT GTG TG-3�; Val138Glu reverse, 5�-CAC ACA GCG CCGGTG CTC GTT CAT CTC GCA GC-3�; Cys142Ser forward, 5�-CGT GCA CCGGCG CTC TGT GCG TAG CGT GC-3�; and Cys142Ser reverse, 5�-GCA CGC TACGCA CAG AGC GCC GGT GCA CG-3�. The CIB cDNAs were amplified from thecorresponding PKCγ-GFP constructs by PCR using the following primers: forward,5�-GCC GAA TTC ACC ATG CAC AAG TTC CGC CTG CAT AG-3� and reverse,5�-CGG GGA TCC CGG CAC AGG GAG GGC ACG CT-3� generating an EcoRIsite at the 5� end and a BamHI site at the 3� end. This facilitated cloning into theEGFP-N1 plasmid. In addition, the EGFP-N1 plasmid (Clontech) was replaced by aRFP-N1 expression vector (Campbell et al., 2002). The MyrPalm-CKAR constructand YFP-PKCδ-CFP were kindly provided by A. Newton (University of Californiaat San Diego, La Jolla, CA) and by B. Blumberg (National Cancer Institute NIH,Bethesda, MD), respectively. The EYFP-PKCδ-ECFP construct was used as templateto generate the EYFP-PKCγ-ECFP and the EYFP-PKCγ-SCFP3A constructs. PKCδwas removed from the EYFP-ECFP vector using XhoI and BamHI. Wild-type andSCA14-mutant PKCγ inserts flanked with XhoI and BamHI site were amplified fromthe corresponding PKCγ-GFP plasmids using the following primer: forward, 5�-GCCCTC GAG ACC ATG GCT GGT CTG GGC CCC-3� and reverse, 5�-CGG GGATCC CGC ATG ACG GGC ACA GGC AC-3� creating XhoI and BamHI sites onthe 3� end and 5� end, respectively. The ECFP was subsequently replaced withSCFP3A, a previously described brighter, monomeric variant of CFP (Kremers etal., 2006), using the flanking BamHI and NotI restriction sites. The GFP-MARCKSconstruct was kindly provided by H. Mogami (Hamamatsu University School ofMedicine, Hamamatsu, Japan) and the GFP-ERK2 was cloned by P. Stork (OregonHealth & Science University, Portland, OR). All constructs were verified bysequencing.

Cell culture and transfectionHeLa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM); highglucose supplemented with 10% fetal bovine serum and 5% penicillin (100 U/ml),streptomycin (100 mg/ml) and glutamate (100 mg/ml). The cultures were maintainedat 37°C in an atmosphere of 10% CO2. The cells were dissociated using trypsinizationand plated (0.3�10–6 cells) onto glass coverslips (24 mm; Fisher Scientific,Braunsweg, Germany) in a six-well plate and transfected after 1 day. Each well ofcells was treated with a mixture of 1 μg DNA and 2 μl Fugene 6.0 transfection reagentin serum-free medium (100 μl transfection per well; Roche Diagnostics, Indianapolis,IN). After transfection, the cells were cultured at 37°C for at least 24 hours. HEK293T cells were cultured in Iscove’s modified Dulbecco’s medium (IMDM); highglucose supplemented with 10% fetal bovine serum and 5% penicillin (100 U/ml),streptomycin (100 mg/ml) and glutamate (100 mg/ml) in an atmosphere of 5% CO2.One day before transfection, 0.1�10–6 cells were seeded into 12-well plates andtransiently transfected with 0.5 μg DNA using Fugene reagent. After transfection,the cells were cultured at 37°C for 48 hours.

Live cell microscopyBefore imaging, the culture medium was replaced by Hank’s balanced salt solutioncontaining CaCl2 supplemented with 5 mM HEPES. Translocation analysis wasperformed by confocal microscopy (Leica Sp2) using a 63� objective at roomtemperature (21-25°C). Fluorescent images of GFP and RFP constructs were obtainedin sequential mode. The time-lapse experiments were controlled by the manufacturer’sacquisition software and the images were acquired every 12 seconds. To measurekinase activity in single living cells, the FRET reporter MyrPalm-CKAR was imagedon a Zeiss 200M inverted fluorescence microscope equipped with a 40� Plan-Neofluar(1.3 NA) oil-immersion lens (Zeiss, Oberkochen, Germany). The set-up wascontrolled by MetaMorph software. Excitation light from a Cairn Xenon Arc lampwas selected by a monochromator (Cairn Research, Kent, UK) (420 nm, bandwidth30 nm) and reflected by a 455DCLP dichroic mirror. Emission was passed through

470/30 nm (CFP) and 535/30 nm (YFP) band-pass filters. Images were acquired witha Photometrics CoolSnap HQ CCD camera with 4�4 binning and an exposure timeof 200 ms per image. The time interval was 4 seconds and 121 images were acquired.The average fluorescence intensity of individual cells was measured using ImageJ(http://rsb.info.nih.gov/ij/), background subtracted and normalized to the initialfluorescence. Subsequently the YFP/CFP ratio was calculated as a measure of kinaseactivity. FLIM experiments and subsequent image analysis were performed asdescribed (Goedhart et al., 2007). A 442 nm helium-cadmium laser (125 mW, Melles-Griot, Carlsbad, CA) was intensity modulated at a frequency of 75.1 MHz. The lightwas reflected by a 455DCLP dichroic mirror and the CFP emission was passed througha D480/40 band-pass emission filter (Chroma Technology). A 63� objective (PlanApochromat NA 1.4 oil) was used for all measurements. The FRET efficiency E wascalculated according to: E=(1–(τDA/τD))�100% in which τDA is the fluorescencelifetime of the donor in presence of the acceptor (i.e. the PKC FRET constructs) andτD is the fluorescence lifetime of the donor (i.e. unfused SCFP3A) in absence of theacceptor. Since frequency domain FLIM yields a phase lifetime and a modulationlifetime, the FRET efficiency can be calculated based on τπ and τM.

ImmunoblottingThe cells were harvested 48 hours after transfection and lysed with ice-cold 2� Trisbuffer (100 mM Tris-HCl pH 7.5, 300 mM NaCl, 10 mM EDTA pH 8.0, 1% Triton-X100) supplemented with complete mini protease inhibitor cocktail (Roche, PaloAlto, CA) and phosphatase inhibitor cocktail (Sigma). Cell lysates were incubatedfor 30 minutes on ice and centrifuged at 14,000 r.p.m. for 20 minutes at 4°C. Proteinconcentrations in the lysates were determined with the Bradford protein assay andequal protein levels were loaded on 10% SDS-PAGE gels. After electrophoresis, theproteins were transferred onto a nitrocellulose membrane filter (Schleicher & Schuell,Dassel, Germany) and blocked in 5% milk to prepare for western blotting. The westernblots were analyzed with anti-PKC (α,β and γ; Upstate), anti-GFP (Santa Cruz), anti-phospho-(Ser)-PKC-substrate (Cell Signaling), anti-phosho-p44/42-MAP kinase(Thr202/Tyr204) (Cell Signaling), anti-p44/42 MAP kinase (Cell Signaling), anti-MARCKS phosphoSer152/156 (Chemicon), and anti-MARKCS (Chemicon)antibodies.

We would like to thank Anastassis Perrakis (NKI, Amsterdam) forhis technical assistance with the structure modelling of the C1B domain,and we thank Dorus Gadella (UvA, Amsterdam) for valuable advice.

ReferencesAlonso, I., Costa, C., Gomes, A., Ferro, A., Seixas, A. I., Silva, S., Cruz, V. T., Coutinho,

P., Sequeiros, J. and Silveira, I. (2005). A novel H101Q mutation causes PKCgammaloss in spinocerebellar ataxia type 14. J. Hum. Genet. 50, 523-529.

Ananthanarayanan, B., Stahelin, R. V., Digman, M. A. and Cho, W. (2003). Activationmechanisms of conventional protein kinase C isoforms are determined by the ligandaffinity and conformational flexibility of their C1 domains. J. Biol. Chem. 278, 46886-46894.

Apostol, B. L., Illes, K., Pallos, J., Bodai, L., Wu, J., Strand, A., Schweitzer, E. S.,Olson, J. M., Kazantsev, A., Marsh, J. L. et al. (2006). Mutant huntingtin alters MAPKsignaling pathways in PC12 and striatal cells: ERK1/2 protects against mutant huntingtin-associated toxicity. Hum. Mol. Genet. 15, 273-285.

Bhalla, U. S. and Iyengar, R. (1999). Emergent properties of networks of biologicalsignaling pathways. Science 283, 381-387.

Braun, D. C., Garfield, S. H. and Blumberg, P. M. (2005). Analysis by fluorescenceresonance energy transfer of the interaction between ligands and protein kinase Cdeltain the intact cell. J. Biol. Chem. 280, 8164-8171.

Campbell, R. E., Tour, O., Palmer, A. E., Steinbach, P. A., Baird, G. S., Zacharias, D.A. and Tsien, R. Y. (2002). A monomeric red fluorescent protein. Proc. Natl. Acad.Sci. USA 99, 7877-7882.

Chen, C., Kano, M., Abeliovich, A., Chen, L., Bao, S., Kim, J. J., Hashimoto, K.,Thompson, R. F. and Tonegawa, S. (1995). Impaired motor coordination correlateswith persistent multiple climbing fiber innervation in PKC gamma mutant mice. Cell83, 1233-1242.

Chen, D. H., Brkanac, Z., Verlinde, C. L., Tan, X. J., Bylenok, L., Nochlin, D.,Matsushita, M., Lipe, H., Wolff, J., Fernandez, M. et al. (2003). Missense mutationsin the regulatory domain of PKC gamma: a new mechanism for dominant nonepisodiccerebellar ataxia. Am. J. Hum. Genet. 72, 839-849.

Chuderland, D. and Seger, R. (2005). Protein-protein interactions in the regulation of theextracellular signal-regulated kinase. Mol. Biotechnol. 29, 57-74.

Colucci-D’Amato, L., Perrone-Capano, C. and di Porzio, U. (2003). Chronic activationof ERK and neurodegenerative diseases. BioEssays 25, 1085-1095.

Costa, M., Marchi, M., Cardarelli, F., Roy, A., Beltram, F., Maffei, L. and Ratto, G.M. (2006). Dynamic regulation of ERK2 nuclear translocation and mobility in livingcells. J. Cell Sci. 119, 4952-4963.

Dalski, A., Mitulla, B., Burk, K., Schattenfroh, C., Schwinger, E. and Zuhlke, C.(2006). Mutation of the highly conserved cysteine residue 131 of the SCA14 associatedPRKCG gene in a family with slow progressive cerebellar ataxia. J. Neurol. 253, 1111-1112.

Daniel, H., Levenes, C. and Crepel, F. (1998). Cellular mechanisms of cerebellar LTD.Trends Neurosci. 21, 401-407.

Journal of Cell Science 121 (14)

Jour

nal o

f Cel

l Sci

ence

2349Mutant PKCγ affects MAPK signaling

De Zeeuw, C. I., Hansel, C., Bian, F., Koekkoek, S. K., van Alphen, A. M., Linden, D.J. and Oberdick, J. (1998). Expression of a protein kinase C inhibitor in Purkinje cellsblocks cerebellar LTD and adaptation of the vestibulo-ocular reflex. Neuron 20, 495-508.

Dutil, E. M., Keranen, L. M., DePaoli-Roach, A. A. and Newton, A. C. (1994). In vivoregulation of protein kinase C by trans-phosphorylation followed by autophosphorylation.J. Biol. Chem. 269, 29359-29362.

Dutil, E. M., Toker, A. and Newton, A. C. (1998). Regulation of conventional proteinkinase C isozymes by phosphoinositide-dependent kinase 1 (PDK-1). Curr. Biol. 8, 1366-1375.

Ebisuya, M., Kondoh, K. and Nishida, E. (2005). The duration, magnitude andcompartmentalization of ERK MAP kinase activity: mechanisms for providing signalingspecificity. J. Cell Sci. 118, 2997-3002.

Fahey, M. C., Knight, M. A., Shaw, J. H., Gardner, R. J., du Sart, D., Lockhart, P. J.,Delatycki, M. B., Gates, P. C. and Storey, E. (2005). Spinocerebellar ataxia type 14,study of a family with an exon 5 mutation in the PRKCG gene. J. Neurol. Neurosurg.Psychiatr. 76, 1720-1722.

Gatlin, J. C., Estrada-Bernal, A., Sanford, S. D. and Pfenninger, K. H. (2006).Myristoylated, alanine-rich C-kinase substrate phosphorylation regulates growth coneadhesion and pathfinding. Mol. Biol. Cell 17, 5115-5130.

Geddis, M. S. and Rehder, V. (2003). The phosphorylation state of neuronal processesdetermines growth cone formation after neuronal injury. J. Neurosci. Res. 74, 210-220.

Giepmans, B. N., Adams, S. R., Ellisman, M. H. and Tsien, R. Y. (2006). The fluorescenttoolbox for assessing protein location and function. Science 312, 217-224.

Goedhart, J., Vermeer, J. E., Adjobo-Hermans, M. J., van Weeren, L. and Gadella,T. W. (2007). Sensitive detection of p65 homodimers using red-shifted and fluorescentprotein-based FRET couples. Plos One 2, e1011.

Hafezparast, M., Klocke, R., Ruhrberg, C., Marquardt, A., Ahmad-Annuar, A., Bowen,S., Lalli, G., Witherden, A. S., Hummerich, H., Nicholson, S. et al. (2003). Mutationsin dynein link motor neuron degeneration to defects in retrograde transport. Science 300,808-812.

Hetman, M. and Gozdz, A. (2004). Role of extracellular signal regulated kinases 1 and2 in neuronal survival. Eur. J. Biochem. 271, 2050-2055.

Holmes, S. E., Hearn, E. O., Ross, C. A. and Margolis, R. L. (2001). SCA12: an unusualmutation leads to an unusual spinocerebellar ataxia. Brain Res. Bull. 56, 397-403.

Horgan, A. M. and Stork, P. J. (2003). Examining the mechanism of Erk nucleartranslocation using green fluorescent protein. Exp. Cell Res. 285, 208-220.

Houlden, H., Johnson, J., Gardner-Thorpe, C., Lashley, T., Hernandez, D., Worth, P.,Singleton, A. B., Hilton, D. A., Holton, J., Revesz, T. et al. (2007). Mutations in TTBK2,encoding a kinase implicated in tau phosphorylation, segregate with spinocerebellar ataxiatype 11. Nat. Genet. 39, 1434-1436.

Kano, M., Hashimoto, K., Chen, C., Abeliovich, A., Aiba, A., Kurihara, H., Watanabe,M., Inoue, Y. and Tonegawa, S. (1995). Impaired synapse elimination during cerebellardevelopment in PKC gamma mutant mice. Cell 83, 1223-1231.

Kazanietz, M. G., Wang, S., Milne, G. W., Lewin, N. E., Liu, H. L. and Blumberg, P.M. (1995). Residues in the second cysteine-rich region of protein kinase C delta relevantto phorbol ester binding as revealed by site-directed mutagenesis. J. Biol. Chem. 270,21852-21859.

Keenan, C. and Kelleher, D. (1998). Protein kinase C and the cytoskeleton. Cell. Signal.10, 225-232.

Klebe, S., Durr, A., Rentschler, A., Hahn-Barma, V., Abele, M., Bouslam, N., Schols,L., Jedynak, P., Forlani, S., Denis, E. et al. (2005). New mutations in protein kinaseCgamma associated with spinocerebellar ataxia type 14. Ann. Neurol. 58, 720-729.

Kremers, G. J., Goedhart, J., van Munster, E. B. and Gadella, T. W., Jr (2006). Cyanand yellow super fluorescent proteins with improved brightness, protein folding, andFRET Forster radius. Biochemistry 45, 6570-6580.

Lenz, J. C., Reusch, H. P., Albrecht, N., Schultz, G. and Schaefer, M. (2002). Ca2+-controlled competitive diacylglycerol binding of protein kinase C isoenzymes in livingcells. J. Cell Biol. 159, 291-302.

Lievens, J. C., Rival, T., Iche, M., Chneiweiss, H. and Birman, S. (2005). Expandedpolyglutamine peptides disrupt EGF receptor signaling and glutamate transporterexpression in Drosophila. Hum. Mol. Genet. 14, 713-724.

Matsuoka, Y., Li, X. and Bennett, V. (1998). Adducin is an in vivo substrate for proteinkinase C: phosphorylation in the MARCKS-related domain inhibits activity in promotingspectrin-actin complexes and occurs in many cells, including dendritic spines ofneurons. J. Cell Biol. 142, 485-497.

Miyawaki, A. (2003). Visualization of the spatial and temporal dynamics of intracellularsignaling. Dev. Cell 4, 295-305.

Myat, M. M., Anderson, S., Allen, L. A. and Aderem, A. (1997). MARCKS regulatesmembrane ruffling and cell spreading. Curr. Biol. 7, 611-614.

Newton, A. C. (2001). Protein kinase C: structural and spatial regulation by phosphorylation,cofactors, and macromolecular interactions. Chem. Rev. 101, 2353-2364.

Newton, A. C. (2004). Diacylglycerol’s affair with protein kinase C turns 25. TrendsPharmacol. Sci. 25, 175-177.

Oancea, E. and Meyer, T. (1998). Protein kinase C as a molecular machine for decodingcalcium and diacylglycerol signals. Cell 95, 307-318.

Pears, C. J. and Parker, P. J. (1991). Domain interactions in protein kinase C. J. CellSci. 100, 683-686.

Pears, C. J., Kour, G., House, C., Kemp, B. E. and Parker, P. J. (1990). Mutagenesisof the pseudosubstrate site of protein kinase C leads to activation. Eur. J. Biochem. 194,89-94.

Pullikuth, A. K. and Catling, A. D. (2007). Scaffold mediated regulation of MAPKsignaling and cytoskeletal dynamics: a perspective. Cell. Signal. 19, 1621-1632.

Quest, A. F. and Bell, R. M. (1994). The regulatory region of protein kinase C gamma.Studies of phorbol ester binding to individual and combined functional segmentsexpressed as glutathione S-transferase fusion proteins indicate a complex mechanismof regulation by phospholipids, phorbol esters, and divalent cations. J. Biol. Chem. 269,20000-20012.

Quest, A. F., Bardes, E. S. and Bell, R. M. (1994). A phorbol ester binding domain ofprotein kinase C gamma. Deletion analysis of the Cys2 domain defines a minimal 43-amino acid peptide. J. Biol. Chem. 269, 2961-2970.

Robinson, M. J., Stippec, S. A., Goldsmith, E., White, M. A. and Cobb, M. H. (1998).A constitutively active and nuclear form of the MAP kinase ERK2 is sufficient for neuriteoutgrowth and cell transformation. Curr. Biol. 8, 1141-1150.

Rodriguez-Alfaro, J. A., Gomez-Fernandez, J. C. and Corbalan-Garcia, S. (2004). Roleof the lysine-rich cluster of the C2 domain in the phosphatidylserine-dependent activationof PKCalpha. J. Mol. Biol. 335, 1117-1129.

Sakai, N., Sasaki, K., Ikegaki, N., Shirai, Y., Ono, Y. and Saito, N. (1997). Directvisualization of the translocation of the gamma-subspecies of protein kinase C in livingcells using fusion proteins with green fluorescent protein. J. Cell Biol. 139, 1465-1476.

Schrenk, K., Kapfhammer, J. P. and Metzger, F. (2002). Altered dendritic developmentof cerebellar Purkinje cells in slice cultures from protein kinase Cgamma-deficient mice.Neuroscience 110, 675-689.

Seki, T., Adachi, N., Ono, Y., Mochizuki, H., Hiramoto, K., Amano, T., Matsubayashi,H., Matsumoto, M., Kawakami, H., Saito, N. et al. (2005). Mutant protein kinaseCgamma found in spinocerebellar ataxia type 14 is susceptible to aggregation and causescell death. J. Biol. Chem. 280, 29096-29106.

Slater, S. J., Seiz, J. L., Cook, A. C., Buzas, C. J., Malinowski, S. A., Kershner, J. L.,Stagliano, B. A. and Stubbs, C. D. (2002). Regulation of PKC alpha activity by C1-C2 domain interactions. J. Biol. Chem. 277, 15277-15285.

Sonnenburg, E. D., Gao, T. and Newton, A. C. (2001). The phosphoinositide-dependentkinase, PDK-1, phosphorylates conventional protein kinase C isozymes by amechanism that is independent of phosphoinositide 3-kinase. J. Biol. Chem. 276,45289-45297.

Stahelin, R. V., Wang, J., Blatner, N. R., Rafter, J. D., Murray, D. and Cho, W. (2005).The origin of C1A-C2 interdomain interactions in protein kinase Calpha. J. Biol. Chem.280, 36452-36463.

Stensman, H. and Larsson, C. (2007). Identification of acidic amino acid residues in theprotein kinase C alpha V5 domain that contribute to its insensitivity to diacylglycerol.J. Biol. Chem. 282, 28627-28638.

Stensman, H., Raghunath, A. and Larsson, C. (2004). Autophosphorylation suppresseswhereas kinase inhibition augments the translocation of protein kinase Calpha in responseto diacylglycerol. J. Biol. Chem. 279, 40576-40583.

Stevanin, G., Hahn, V., Lohmann, E., Bouslam, N., Gouttard, M., Soumphonphakdy,C., Welter, M. L., Ollagnon-Roman, E., Lemainque, A., Ruberg, M. et al. (2004).Mutation in the catalytic domain of protein kinase C gamma and extension of thephenotype associated with spinocerebellar ataxia type 14. Arch. Neurol. 61, 1242-1248.

Vallee, R. B., Tai, C. and Faulkner, N. E. (2001). LIS1: cellular function of a disease-causing gene. Trends Cell Biol. 11, 155-160.

van de Leemput, J., Chandran, J., Knight, M. A., Holtzclaw, L. A., Scholz, S., Cookson,M. R., Houlden, H., Gwinn-Hardy, K., Fung, H. C., Lin, X. et al. (2007). Deletionat ITPR1 underlies ataxia in mice and spinocerebellar ataxia 15 in humans. PLoS Genet.3, e108.

van de Warrenburg, B. P., Verbeek, D. S., Piersma, S. J., Hennekam, F. A., Pearson,P. L., Knoers, N. V., Kremer, H. P. and Sinke, R. J. (2003). Identification of a novelSCA14 mutation in a Dutch autosomal dominant cerebellar ataxia family. Neurology61, 1760-1765.

van Munster, E. B. and Gadella, T. W. (2005). Fluorescence lifetime imaging microscopy(FLIM). Adv. Biochem. Eng. Biotechnol. 95, 143-175.

van Swieten, J. C., Brusse, E., de Graaf, B. M., Krieger, E., van de Graaf, R., deKoning, I., Maat-Kievit, A., Leegwater, P., Dooijes, D., Oostra, B. A. et al. (2003).A mutation in the fibroblast growth factor 14 gene is associated with autosomal dominantcerebellar ataxia [corrected]. Am. J. Hum. Genet. 72, 191-199.

Verbeek, D. S., Knight, M. A., Harmison, G. G., Fischbeck, K. H. and Howell, B. W.(2005). Protein kinase C gamma mutations in spinocerebellar ataxia 14 increase kinaseactivity and alter membrane targeting. Brain 128, 436-442.

Violin, J. D., Zhang, J., Tsien, R. Y. and Newton, A. C. (2003). A genetically encodedfluorescent reporter reveals oscillatory phosphorylation by protein kinase C. J. Cell Biol.161, 899-909.

Vlak, M. H., Sinke, R. J., Rabelink, G. M., Kremer, B. P. and van de Warrenburg, B.P. (2006). Novel PRKCG/SCA14 mutation in a Dutch spinocerebellar ataxia family:expanding the phenotype. Mov. Disord. 21, 1025-1028.

Xu, R. X., Pawelczyk, T., Xia, T. H. and Brown, S. C. (1997). NMR structure of a proteinkinase C-gamma phorbol-binding domain and study of protein-lipid micelle interactions.Biochemistry 36, 10709-10717.

Yabe, I., Sasaki, H., Chen, D. H., Raskind, W. H., Bird, T. D., Yamashita, I., Tsuji, S.,Kikuchi, S. and Tashiro, K. (2003). Spinocerebellar ataxia type 14 caused by a mutationin protein kinase C gamma. Arch. Neurol. 60, 1749-1751.

Zhuchenko, O., Bailey, J., Bonnen, P., Ashizawa, T., Stockton, D. W., Amos, C., Dobyns,W. B., Subramony, S. H., Zoghbi, H. Y. and Lee, C. C. (1997). Autosomal dominantcerebellar ataxia (SCA6) associated with small polyglutamine expansions in the alpha1A-voltage-dependent calcium channel. Nat. Genet. 15, 62-69.

Jour

nal o

f Cel

l Sci

ence