Four-and-a-half LIM domains proteins are novel regulators of the protein kinase D pathway in cardiac myocytes.

1.3 MB·English

by  Stathopoulou, Konstantina

Biochem.J.(2014)457,451–461(PrintedinGreatBritain) doi:10.1042/BJ20131026 451
Four-and-a-half LIM domains proteins are novel regulators of the protein
kinase D pathway in cardiac myocytes
KonstantinaSTATHOPOULOU*†,FriederikeCUELLO*†,AlexandraJ.CANDASAMY*,ElizabethM.KEMP*,ElisabethEHLER*, g
RobertS.HAWORTH*andMetinAVKIRAN*1
r
*CardiovascularDivision,King’sCollegeLondonBritishHeartFoundationCentre,LondonSE17EH,U.K. o
†InstituteofExperimentalPharmacologyandToxicology,UniversityMedicalCenterHamburg-Eppendorf,Martinistr.52,Hamburg20146,Germany
.
j
m
e
PKD (protein kinase D) is a serine/threonine kinase implicated response to endothelin 1, but not to the α -adrenoceptor
1 h
in multiple cardiac roles, including the phosphorylation of the agonistphenylephrine.Incontrast,selectiveknockdownofFHL2
classIIHDAC5(histonedeacetylaseisoform5)andtherebyde- expression caused a significant reduction in PKD activation c
repression of MEF2 (myocyte enhancer factor 2) transcription and HDAC5 phosphorylation in response to both stimuli.
o
factoractivity.InthepresentstudyweidentifyFHL1(four-and- Interestingly, neither intervention affected MEF2 activation by
a-halfLIMdomainsprotein1)andFHL2asnovelbindingpartners endothelin 1 or phenylephrine. We conclude that FHL1 and i
forPKDincardiacmyocytes.Thiswasconfirmedbypull-down FHL2 are novel cardiac PKD partners, which differentially b
assaysusingrecombinantGST-fusedproteinsandheterologously facilitatePKDactivationandHDAC5phosphorylationbydistinct
.
orendogenouslyexpressedPKDinadultratventricularmyocytes neurohormonalstimuli,butareunlikelytoregulateMEF2-driven w
orNRVMs(neonatalratventricularmyocytes)respectively,and transcriptionalreprogramming.
byco-immunoprecipitationofFHL1andFHL2withGFP–PKD1 w
fusion protein expressed in NRVMs. In vitro kinase assays
showed that neither FHL1 nor FHL2 is a PKD1 substrate. Keywords:cardiacmyocyte,four-and-a-halfLIM(FHL),histone w
SelectiveknockdownofFHL1expressioninNRVMssignificantly deacetylase, neurohormonal stimulation, protein kinase, signal
inhibited PKD activation and HDAC5 phosphorylation in transduction.

l
a
INTRODUCTION The functional significance of PKD1in cardiac myocyte n
(patho)physiology has recently started to be unveiled by both
The PKD (protein kinase D) family of serine/threonine kinases in vitro and in vivo studies. We have shown previously that r
consistsofthreemembers,PKD1,PKD2andPKD3,andbelongs PKD1mayregulatecardiacmyofilamentfunctionandtheCa2+ u
to the CaMK (Ca2+/calmodulin-dependent protein kinase) sensitivity of contraction by phosphorylating cTnI (inhibitory o
superfamily. These PKD isoforms share the common structural subunit of cardiac troponin) at Ser22/Ser23 [7,8] and cMyBP-C
features of a C-terminal catalytic domain and an N-terminal (cardiac myosin-binding protein C) at Ser302 [9]. Furthermore, J
regulatory domain. Components of the regulatory domain PKD1 has been proposed to facilitate cardiac hypertrophy
autoinhibit the activity of the catalytic domain in unstimulated through the phosphorylation of HDAC5 (histone deacetylase l
a
cells and promote PKD association with the plasma and isoform 5) at Ser259 and Ser498 [10]. Nuclear HDAC5 associates
intracellularmembranesafterstimulationwithhormones,growth with and represses the activity of MEF2 (myocyte enhancer c
factors,neurotransmitters,chemokinesandbioactivelipids[1,2]. factor 2) transcription factors, which drive the transcriptional
i
In cardiac myocytes, the most abundantly expressed PKD reprogrammingthatprecipitatespathologicalcardiachypertrophy m
family member is PKD1, which is activated after stimulation andremodelling.Inresponsetopro-hypertrophicneurohormonal
of diverse GPCRs (G-protein-coupled receptors) that signal via stimuli, activated PKD1 phosphorylates HDAC5 at Ser259 and e
Gα ,includingα -adrenergic,ET1(endothelin1)andangiotensin Ser498, thus inducing the binding of 14-3-3 proteins to these
q 1 h
II receptors [3–5]. The principal PKD activation mechanism sitesandrevealingaNES(nuclearexportsequence)thattriggers
involves recruitment of the kinase to plasma or intracellular HDAC5 extrusion from the nucleus to the cytosol, through c
membranesbyDAG(diacylglycerol)andtransphosphorylationof a mechanism that is mediated by the CRM1 (chromosome
o
itsactivationloopataminoacidresiduesSer744andSer748(amino region maintenance 1) protein [10,11]. HDAC5 nuclear export
acidnumberingreferstomurinePKD1)byactivatednovelPKC de-represses MEF2 transcriptional activity, which then drives i
(protein kinase C) isoforms. The resulting PKD activation then pro-hypertrophic gene expression [12–14]. Studies in mice B
leads to both autophosphorylation at residue Ser916 and trans- with cardiac-specific deletion [15] or overexpression [16] of
phosphorylation of PKD substrates, which include transcription PKD1 corroborate a key role for PKD1in pathological cardiac
factors,proteinsinvolvedincellmotilityandvesiclefissionfrom remodelling, and PKD1 expression and activation have been
theGolgiapparatus,otherkinasesandsarcomericproteins[1,2,6]. showntobeincreasedinfailinghumanmyocardium[17].
Abbreviations: ARVM, adult rat ventricular myocyte; BPKDi, bipyridyl PKD inhibitor; CaMK, Ca2+/calmodulin-dependent protein kinase; caPKD,
constitutivelyactivecatalyticdomainofPKD;cMyBP-C,cardiacmyosin-bindingproteinC;CRM1,chromosomeregionmaintenance1;cTnI,inhibitory
subunitofcardiactroponin;ERK,extracellular-signal-regulatedkinase;ET1,endothelin1;FHL,four-and-a-halfLIMdomains;HDAC,histonedeacetylase;
IVK,invitrokinase;MEF2,myocyteenhancerfactor2;MOI,multiplicityofinfection;MuRF,muscleRINGfinger;NRVM,neonatalratventricularmyocyte;
PE,phenylephrine;pfu,plaque-formingunit;PKC,proteinkinaseC;PKD,proteinkinaseD;TAC,transverseaorticconstriction.
1 Towhomcorrespondenceshouldbeaddressed([email protected]).
© 2014 The Author(s) (cid:2)c TheAuthorsJournalcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
452 K.Stathopoulouandothers
The key roles proposed for PKD activity in cardiac with an adenoviral vector expressing murine PKD1 at a MOI
(patho)physiologymakeimprovedunderstandingofthemolecu- (multiplicityofinfection)of100pfu(plaque-formingunit)/cell.
larmechanismsunderlyingboththeupstreamregulationandthe ARVMs were maintained in culture for 24h before use in
downstream actions of this kinase in the heart an imperative. experiments.
Towards this objective, in a previous study [7], we performed
a yeast two-hybrid screen of a human cardiac cDNA library,
whichidentifiedFHL2(four-and-a-halfLIMdomainsprotein2) IsolationandcultureofNRVMs(neonatalratventricularmyocytes)
asanovelbindingpartnerforthePKD1catalyticdomain.Inthe
NRVMs were isolated and maintained in culture as described
presentstudy,wehaveconfirmedandcharacterizedtheinteraction
previously[8],andtransferredtomaintenancemedium[DMEM
offull-lengthPKD1withFHL2aswellasthehighlyhomologous
(Dulbecco’s modified Eagle’s medium)/M199 4:1 (v/v), 100
FHLisoformFHL1(bothofwhichareabundantlyexpressedin
i.u./ml penicillin and streptomycin] 24h after isolation. For
the heart [18]) in cardiac myocytes and explored the potential
RNAiexperiments,themaintenancemediumuseddidnotcontain
functionalsignificanceoftheseFHLisoformsinregulatingPKD
antibiotics.
activityanddownstreamactionsinthatcelltype.
ExpressionofrecombinantFHLproteins
EXPERIMENTAL
PlasmidsencodingGST–FHL1andGST–FHL2fusionproteins
Materials were transformed into Escherichia coli BL21 (DE3) pLysS
cells (Invitrogen). Cells were grown at 37◦C in LB medium
Rabbit polyclonal antibodies against phosphorylated supplemented with 100μg/μl ampicillin and 100μM ZnCl .
(pSer744/Ser748 and pSer916) PKD were from Cell Signaling 2
When D reached 0.6–0.8, cells were induced to express the
Technology.RabbitpolyclonalantibodiesagainsttotalPKDwere 600
GST-fused proteins for 3h by adding IPTG to 1mM. Cells
fromCellSignalingTechnologyandSantaCruzBiotechnology. were then centrifuged at 5000g for 5min (4◦C) in a Sorvall
A rabbit polyclonal antibody against phosphorylated (pSer498)
RC5Bcentrifuge(GSArotor),pelletswereresuspendedinPBS
HDAC5 and a mouse monoclonal antibody against FHL1 were
containing 1% (v/v) Triton X-100 and CompleteTM protease
fromAbcam.AmousemonoclonalantibodyagainstFHL2was
inhibitors (Roche) and suspensions were then sonicated on ice
from Medical and Biological Laboratories. Mouse monoclonal
(threetimesfor15s,SonicsMaterialsUltrasonicProcessor).For
antibodiesagainstGFPandGSTwerefromRocheandSantaCruz
thepurificationofsolublerecombinantGSTandGST–FHL1,the
Biotechnologyrespectively.Expressionvectorsforexpressionof lysatewasclearedbycentrifugationat12000gfor30min(4◦C)
recombinantmouseFHL1andhumanFHL2proteinsweregifts
(rotorSS34)andtheproteinwaspurifiedbypassingthecleared
from Stephan Lange (Department of Medicine, UC San Diego,
lysate through a glutathione–Sepharose 4B prepacked column
CA,U.S.A.)[19].RecombinantPKDcatalyticdomainexpressed
(GE Healthcare) according to the manufacturer’s instructions.
in Sf21 insect cells was a gift from Harold Jeffries and Peter J.
For the purification of insoluble recombinant GST–FHL2, the
Parker (London Research Institute, Cancer Research UK) and lysatewascentrifugedat20000gfor30min(4◦C)inaBeckman
recombinant human TnI was a gift from Douglas G. Ward and
L60 ultracentrifuge (rotor SW 41 Ti). The pellet was then
Ian Trayer (School of Biosciences, University of Birmingham,
resuspended in PBS containing 1% (v/v) Triton X-100 and
Birmingham, U.K.) [7]. BPKDi (bipyridyl PKD inhibitor) was
protease inhibitors and centrifuged again, as described above.
custom-synthesized by the MRC Protein Phosphorylation Unit,
The pellet was then dissolved in PBS containing 1% (v/v)
UniversityofDundee,Dundee,U.K.[20].Theadenoviralvectors
Triton X-100, protease inhibitors and 8M urea and centrifuged
used to express mouse PKD1 (AdV:wtPKD1) and enhanced at20000gfor15min(4◦C).Thesupernatant,whichcontained
GFP (AdV:GFP) were prepared as described previously [8].
thedenaturedrecombinantGST–FHL2,wasdialysedinrefolding
The adenoviral vector encoding human PKD1 tagged at the
buffercontaining50mMTris/HCl,pH8.0,50mMNaCl,1mM
N-terminus with GFP (AdV:GFP–PKD1) was a gift from Jody
DTT and 0.1mM ZnCl . GST–FHL2 was purified using a
Martin (Department of Cell and Molecular Physiology, Loyola 2
glutathione–Sepharose4Bprepackedcolumn,asnotedabove.
University Chicago, IL, U.S.A.) [17]. The MEF2-luciferase
reporter construct was a gift from Eric Olson (Department of
Molecular Biology, UT Southwestern, TX, U.S.A.) [14] and
GSTpull-downassay
the promoter from this (comprising three tandem repeats of the
MEF2-bindingsitefromthedesminenhancer)wassubclonedinto For the GST pull-down assay, lysates from ARVMs (treated
apGL4.24vectorencodingthesyntheticluciferasereportergene with vehicle or 100nM PMA for 10min) or NRVMs (treated
luc2P(Promega).ThevectorencodingRenillaluciferasereporter with vehicle or 10nM ET1 for 10min) prepared in lysis
(pRL-null)wasfromPromega.ET1wasfromCalbiochem.PMA buffer [containing 50mM Tris/HCl, pH7.4, 0.05mM ZnCl2,
and PE (phenylephrine) were from Sigma–Aldrich. Secondary 2mM DTT and 1% (v/v) Triton X-100] were clarified by
antibodies and ECL kits were from GE Healthcare. All other centrifugation (25000g, 20min, 4◦C) in an Eppendorf 5417R
chemicals were from Sigma–Aldrich, Life Technologies or centrifuge.Eachsupernatantwaspre-clearedbyincubationwith
FisherScientific,unlessotherwisestated.Theinvestigationwas glutathione–Sepharose 4B beads (GE Healthcare) for 1h under
performed in accordance with the Home Office Guidance on rotation(4◦C),thensplitinequalvolumesandincubatedfor2h
the Operation of the Animals (Scientific Procedures) Act 1986 under rotation (4◦C) with 5μg of recombinant proteins (GST–
(U.K.). FHL1,GST–FHL2andGST)pre-boundfor1htoglutathione–
Sepharose 4B beads (4◦C). The complexes were pelleted by
centrifugation(1000g,1min,4◦C),washedthreetimesinlysis
buffer, resuspended in 2× SDS/PAGE sample buffer [125mM
IsolationandcultureofARVMs(adultratventricularmyocytes)
Tris/HCl,pH6.8,4%(w/v)SDS,20%(v/v)glycerol,6%(v/v)
ARVMs were isolated and maintained in culture as described 2-mercaptoethanol and 0.02% Bromophenol Blue], boiled and
previously [8]. Attached cardiac myocytes were transduced storedat −20◦Cuntiluse.
©(cid:2)c 20T1h4e TAhueth oArusthJoour(rsn)alcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
FHL1andFHL2arenovelregulatorsofPKD 453
Immunoprecipitation Luciferaseassays
Onedayafterisolation,NRVMsweretransducedwithadenoviral NRVMsweretransfectedinitiallywithactiveorscrambledsiRNA
vectorsexpressingeitherGFP–PKD1(atMOIof50pfu/cell)or sequences,asdescribedabove,and24hlaterwithacombination
GFP (at MOI of 10 pfu/cell). After 24h, NRVMs were treated of0.6μgofMEF2-luc2Pluciferasereporterconstructand0.4μg
for 20min with either ET1 (10nM) or vehicle and GFP–PKD1 of Renilla luciferase reporter construct, the latter using Escort
or GFP was immunoprecipitated by using the GFP-Trap®_A III transfection reagent (Sigma). The medium was changed
kit (Chromotek), according to the manufacturer’s instructions. to transfection medium without serum 5h later and the cells
Briefly, NRVMs were lysed in a buffer containing 10mM incubatedforafurther3h,beforetheyweretreatedwithvehicle,
Tris, pH7.5, 150mM NaCl and 0.5mM EDTA, supplemented ET1 (10nM) or PE (3μM, in the presence of 1μM atenolol)
with 0.5% Nonidet P40 and protease inhibitors. Lysates were for 18–24h. Luciferase assays were performed using the Dual-
incubated on ice for 30min and subsequently centrifuged for Glo®LuciferaseassaysystemandaGloMax®20/20luminometer
10minat20000g(4◦C).Theclearedlysateswerethendiluted (Promega), according to the manufacturer’s instructions. In
to a final volume of 1ml with dilution buffer (containing experiments where BPKDi was used, NRVMs were processed
10mM Tris/HCl, pH7.5, 150mM NaCl, 0.5mM EDTA and in a similar manner, but without the siRNA transfection step;
supplemented with protease inhibitors) and incubated under BPKDi(3μM)orvehiclewasadded30minbeforetreatmentwith
rotation for 2h (4◦C) with GFP-Trap®_A beads. Beads were vehicle, ET1 or PE. Firefly luciferase activity was normalized
thenwashedthreetimesindilutionbuffer,resuspendedin100μl against Renilla luciferase activity to account for transfection
of 2× SDS/PAGE sample buffer, boiled and stored at −20◦C efficiency.DatawereexpressedrelativetothescrambledsiRNA-
untiluse. control group (in experiments where FHL protein levels were
knocked down by RNAi) or to the vehicle-control group (in
experimentswherePKDactivitywaspharmacologicallyinhibited
Invitrophosphorylationassay withBPKDi).
Theinvitrophosphorylationassaywascarriedoutasdescribed
previously[7].RecombinantGST,GST–FHL1,GST–FHL2and
Immunoblotanalysis
cTnI,aloneorincombination(200pmoleach)asindicated,were
incubatedfor30minat37◦C,intheabsenceorpresenceofactive Immunoblotanalysiswasperformedasdescribedpreviously[7],
PKD(35ng)togetherwith100μMATP,supplementedwith[γ- using specific antibodies for total or phosphorylated proteins as
32P]ATPinkinaseassaybuffer(30mMTris/HCl,pH7.4,15mM indicated. Specific protein bands were detected by ECL (GE
MgCl ,1mMDTTand0.05mMZnCl ).Proteinswereresolved Healthcare) and phosphorylation status was quantified using
2 2
on 12% acrylamide gels, which were subsequently dried and a calibrated densitometer (GS-800TM, Bio-Rad Laboratories).
subjectedtoautoradiography. Data were normalized to total protein loading (measured by
densitometryintherespectiveimmunoblotsafterCoomassieBlue
staining) and expressed relative to an internal standard sample
RNAi includedinaseparatelane,topermitconsolidationofexperiments
analysedondifferentgels.
Synthetic siRNA sequences (denoted as ‘active’ siRNAs) and
theirrespectivescrambledcounterparts(denotedas‘scrambled’
siRNAs) with 3(cid:3) dTdT overhangs were purchased from Life Statisticalanalysis
Technologies.Thesequencesusedwere:5(cid:3)-GCCUGAAGUGCU-
UUGACAA-3(cid:3) (active siRNA sequence for FHL1), 5(cid:3)-GCCA- Statistical comparisons were performed by one-way ANOVA
AUGCGUUGUAUGCAA-3(cid:3) (scrambled siRNA sequence for followed by the Newman–Keuls post-hoc test. Quantitative
FHL1), 5(cid:3)-GCAAGGACUUGUCCUACAA-3(cid:3) (active siRNA data are given as means+−S.E.M. and P<0.05 was considered
sequence for FHL2) and 5(cid:3)-GCAUCAGCUGUAUCGACAA-3(cid:3) significant.
(scrambledsiRNAsequenceforFHL2).NRVMsweretransfected
with 40 nmol of active or scrambled siRNA sequences using
RESULTS
Lipofectamine RNAimax (Life Technologies), 24h after they
were transferred to maintenance medium. The cells were then PKDinteractswithbothFHL1andFHL2
incubatedfor48hintransfectionmediumcontaining22%(v/v)
The potential interactions of PKD with FHL1 and FHL2 were
M199,4%(v/v)horseserumand74%(v/v)modifiedDulbecco’s
initiallyinvestigatedthroughpull-downassaysusingrecombinant
balanced salt solution buffer (116mM NaCl, 1mM NaH PO ,
2 4 GST–FHL1andGST–FHL2proteinsandcelllysatesfromAR-
0.8mMMgSO ,5.5mMglucose,32.1mMNaHCO and1.8mM
4 3 VMsthatheterologouslyexpressedmousePKD1followingaden-
CaCl ,pH7.2),afterwhichtimetheywereserum-starvedfor3h
2 oviralgenetransfer.ARVMsweretreatedwithvehicleorPMA(to
before treatment for 20min with vehicle, ET1 (10nM) or PE
activatePKD)toascertainthepotentialimpactofPKDactivation
(3μM,inthepresenceof1μMatenololadded10minbeforePE).
Atenololwasusedtoinhibitanyβ-adrenergicreceptoractivation, statusonanyinteractionwithFHL1orFHL2.PKD1proteinwas
pulleddownbybothGST–FHL1andGST–FHL2,butnotbyGST
whichisknowntocounteractPKDactivation[20,21].
alone(Figure1A,firstpanel),indicatingaspecificinteractionwith
bothFHL1andFHL2.Asexpected,PMAmarkedlyincreasedthe
phosphorylationstatusofPKD1atSer744/Ser748andSer916,reflect-
PharmacologicalinhibitionofPKDactivity
ingactivationofthekinase.PhosphorylatedPKD1associatedwith
IntheexperimentswheretheselectivePKDinhibitorBPKDiwas bothFHL1andFHL2inthispull-downassay(Figure1A,second
used,NRVMs were processedas describedabovefor theRNAi andthirdpanels),indicatingthatthePKD1–FHL1/2association
experiments,butwithoutthetransfectionstep.BPKDi(3μM)or occursindependentlyofthePKD1activationstatus.Baitprotein
vehiclewasadded30minbeforetreatmentwithvehicle,ET1or (GST–FHL1, GST–FHL2 or GST) content was comparable in
PE. controlandstimulatedgroups(Figure1A,fourthandfifthpanels).
© 2014 The Author(s) (cid:2)c TheAuthorsJournalcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
454 K.Stathopoulouandothers
Figure 2 Co-immunoprecipitation of endogenous FHL1 and FHL2 with
heterologouslyexpressedPKD1
NRVMs transduced with AdV:GFP–PKD1 or AdV:GFP were treated with vehicle (C) or
ET1 (10nM) for 20min and protein complexes associated with GFP–PKD1 or GFP were
immunoprecipitatedusingGFP-Trapmethodology.AssociationofGFP–PKD1orGFPwith
endogenouslyexpressedFHL1andFHL2wasdetectedbyimmunoblot(IB)analysisofthe
immunoprecipitates (IP) using specific antibodies against FHL1 or FHL2. Expression and
immunoprecipitationofGFP–PKD1andGFPwereverifiedbyimmunoblotanalysisusingaGFP
antibody,whereasactivationstatusofthePKD1componentofGFP–PKD1wasdeterminedby
immunoblotanalysisusingaphospho-specificpSer916PKDantibody.Samplesofthelysates
beforetheIP(input,10%)werealsoincludedintheimmunoblotanalyses.Dataarerepresentative
ofthreeindependentexperiments.
Figure1 Pull-downofheterologouslyexpressedorendogenousPKD1with immunoblot assessment of PKD autophosphorylation status at
recombinantFHL1andFHL2 Ser916 (Figure2,thirdpanel).Immunoblotanalysiswithananti-
GFP antibody confirmed comparable expression and equally
(A)ARVMstransducedwithAdV:wtPKD1weretreatedwithvehicle(C)orPMA(100nM)for
efficient immunoprecipitation of GFP–PKD1 and GFP in the
10min.LysateswereincubatedwithpurifiedGST–FHL1,GST–FHL2orGSTbaitproteinsbound
toglutathione–SepharosebeadsandthePKDpull-downwasassessedbyimmunoblot(IB) differentstudygroups(Figure2,fourthandfifthpanels).Taken
analysisusingtotalorphospho-specific(pSer916orpSer744/Ser748)PKDantibodies.Samples together, the above results provide complementary biochemical
ofthelysatesbeforethepull-downassay(input,10%)werealsoincludedintheimmunoblot evidence that FHL1 and FHL2 are novel interaction partners
analysesandareshowninseparatepanelsreflectingdifferentexposuretimes.Baitprotein for PKD1 and that this interaction occurs independently of the
contentofthepull-downswasdeterminedbyimmunoblotanalysisusingaGSTantibody.(B)
activationstatusofthekinase.
Asin(A),butusinglysatesfromNRVMstreatedwithvehicle(C)orET1(10nM)for10min.
Immunoblotsarerepresentativeofthreeindependentexperiments.
FHL1andFHL2arenotphosphorylatedbyPKD
ToexploretheinteractionofendogenousPKDwithFHL1and Ourpreviousstudiesthatbuiltonthediscoveryofnovelcardiac
FHL2, we additionally performed GST pull-down assays using PKD1interactionpartnersthroughayeasttwo-hybridscreen[7]
cell lysates from NRVMs, in which PKD1 is more abundantly have led to the identification of cTnI and cMyBP-C as PKD
expressed than in ARVMs [3]. PKD1 was pulled down by substratesandthemappingandfunctionalanalysisofthepertinent
GST–FHL1 and GST–FHL2, but not by GST alone, reflecting substratephosphorylationsites[7–9].Inananalogousapproach,
an interaction between the endogenous kinase and both FHL1 weexploredwhetherFHL1andFHL2arealsoPKDsubstrates,
andFHL2(Figure1B,firstpanel).Theseinteractionswerealso byperformingIVK(invitrokinase)assaysinwhichthecaPKD
retainedfollowingtheactivationofendogenousPKD(Figure1B, (constitutively active catalytic domain of PKD) was incubated
second and third panels), this time achieved by exposure to a with recombinant GST, GST–FHL1 or GST–FHL2 proteins in
physiologicalstimulus,ET1.Asabove,baitprotein(GST–FHL1, the presence of [γ-32P]ATP. Autoradiographic analysis revealed
GST–FHL2 or GST) content was comparable in control and nodetectablephosphateincorporationintoGST,GST–FHL1or
stimulatedgroups(Figure1B,fourthandfifthpanels). GST–FHL-2, even though robust phosphorylation of cTnI was
We next investigated whether the PKD1–FHL1/2 interaction readily detectable (Figure 3A). To explore whether FHL1 or
occurs in the cellular environment of cardiac myocytes. To FHL2 might affect directly the catalytic activity of PKD, we
address this, we expressed GFP–PKD1 or GFP in NRVMs by also performed IVK reactions in which caPKD was incubated
adenoviralgenetransferandemployedGFP-Trapmethodologyto with cTnI in the absence or presence of each of the GST-fused
efficientlyimmunoprecipitateproteincomplexesassociatedwith proteins. These experiments revealed that cTnI phosphorylation
these proteins. Immunoblot analysis revealed that endogenous was not affected by the presence of GST, GST–FHL1 or GST–
FHL1andFHL2bothco-immunoprecipitatedwithGFP–PKD1, FHL2(Figure3B).SuchfindingsindicatethatFHL1andFHL2
but not with GFP, under basal conditions and after stimulation are not PKD substrates, which is consistent with the fact that
with ET1 (Figure 2, first and second panels). Activation of theaminoacidsequencesoftheseproteinsdonotcontainPKD
the PKD1 component of the heterologously expressed GFP– substrate consensus phosphorylation motifs, as revealed by in
PKD1 fusion protein in response to ET1 was verified by silicoanalysis(http://phospho.elm.eu.org).Theyalsosuggestthat
©(cid:2)c 20T1h4e TAhueth oArusthJoour(rsn)alcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
FHL1andFHL2arenovelregulatorsofPKD 455
Figure3 FHL1andFHL2arenotPKDsubstrates
(A)AutoradiogramsofinvitrophosphorylationassaysusingrecombinantcaPKDandrecombinantGST,GST–FHL1,GST–FHL2orcTnIproteinassubstrate.(B)Assayswerealsoconductedusing
cTnIassubstrate,intheabsenceorpresenceofadditionalrecombinantGST,GST–FHL1orGST–FHL2protein.(C)Histogramshowingquantitativedatafromexperimentsasillustratedin(B).
Reactionswereperformedinthepresenceof[γ-32P]ATPandproteinswereresolvedbySDS/PAGE(12%gel).ProteinloadingwasconfirmedbyCoomassieBluestaining.Dataarerepresentativeof
threeindependentexperiments,andthequantitativedatain(C)aremeans+−S.E.M.(n=3).
interaction with FHL1 or FHL2 does not inhibit the catalytic As illustrated in Figure 4, both stimuli significantly increased
activityofcaPKDtowardsanestablishedsubstrate,cTnI. PKD autophosphorylation at Ser916 and transphosphorylation
at Ser744/Ser748 in control cells transfected with scrambled
siRNA.KnockingdownFHL1expressionsignificantlyattenuated
FHL1andFHL2facilitatePKDphosphorylationincardiacmyocytes
the increase in PKD phosphorylation in response to ET1
To explore the potential roles of FHL1 and FHL2in regulating (Figure 4A), but had no significant effect on the increase in
the activation and functions of endogenous PKD in cardiac PKDphosphorylationinresponsetoPE(Figure4B).Incontrast,
myocytes, we applied a loss-of-function approach and knocked knockdown of FHL2 expression significantly attenuated the
down FHL1 or FHL2 protein expression by RNAi through increases in PKD phosphorylation in response to both ET1
transfection of NRVMs with synthetic siRNA duplexes specific (Figure 4C) and PE (Figure 4D), with a greater inhibitory
foreachFHLmRNAsequence.TransfectionwithFHL1siRNA effect on the latter. These findings suggest that FHL1 and
caused an approximately 75% decrease in FHL1 protein FHL2 facilitate PKD activation by multiple neurohormonal
expressionrelativetocontrolcellstransfectedwithascrambled stimuli,potentiallyinastimulus-dependentmanner,bypromoting
siRNA sequence, with no apparent effect on FHL2 expression transphosphorylation of the PKD activation loop. Interestingly,
(Supplementary Figures S1A and S1B at http://www.biochemj. theeffectsofsimultaneousknockdownofbothFHL1andFHL2
org/bj/457/bj4570451add.htm). Similarly, transfection with onPKDphosphorylationweresimilartotheeffectsofselective
FHL2 siRNA caused a comparable decrease in FHL2 protein FHL2 knockdown, with no indication of an additive effect
expression relative to cells transfected with a scrambled siRNA (Supplementary Figure S2 at http://www.biochemj.org/bj/457/
sequence, with no compensatory change in FHL1 expression bj4570451add.htm).
(Supplementary Figures S1C and S1D). We then examined the FHL2hasbeenreportedtopromotethestabilityofmyocardin
consequences of the selective knockdown of each FHL isoform andmyocardin-relatedtranscriptionfactor-Abyprotectingthese
on PKD activation in response to stimulation with ET1 or PE. proteinsfromproteasome-mediateddegradation[22],raisingthe
© 2014 The Author(s) (cid:2)c TheAuthorsJournalcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
456 K.Stathopoulouandothers
Figure4 EffectofFHL1orFHL2knockdownonET1-andPE-inducedphosphorylationofendogenousPKD
NRVMsweretransfectedwitheitherscrambledsiRNAoractivesiRNAduplexestargetedatFHL1(AandB)orFHL2(CandD)transcripts.After48h,cellsweretreatedwithvehicle(C)orET1
(10nM)(AandC)orvehicle(C)orPE(3μM)(BandD)for20min.PhosphorylationstatusofendogenousPKDatSer916andSer744/Ser748wasdeterminedbyimmunoblot(IB)analysisusing
appropriatephospho-specificantibodies.Proteinloadingwasconfirmedusingananti-PKDantibodyandbyCoomassieBluestaining.Individualimmunoblotsillustraterepresentativeexperiments,
andhistogramsshowquantitativedataasmeans+−S.E.M.(n=7–8).*P<0.05.
©(cid:2)c 20T1h4e TAhueth oArusthJoour(rsn)alcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
FHL1andFHL2arenovelregulatorsofPKD 457
possibilitythat,ifFHL1orFHL2similarlyprotectedPKDprotein OnepossibleexplanationforthislackofeffectofFHLisoform
from degradation, the apparent effects of FHL knockdown on knockdownonMEF2activation,despiteasignificantattenuation
PKD phosphorylation described above might actually reflect a of PKD activation, is that under our experimental conditions
decrease in total PKD protein levels. We therefore analysed PKDactivityisnotanimportantmediatorofMEF2activationin
total PKD protein levels in samples after FHL1 or FHL2 responsetoET1orPE.TohelpconfirmtheroleofPKDactivity
knockdown.Despiteatendency,therewasnosignificantreduction in MEF2 activation in NRVMs, therefore, we also investigated
in PKD protein expression with knockdown of either FHL1 or the effects of the selective PKD inhibitor BPKDi [20,23] under
FHL2(SupplementaryFigureS3athttp://www.biochemj.org/bj/ thesameexperimentalconditions.Consistentwithpreviouswork
457/bj4570451add.htm). Thus the negative effects of reduced [23],pre-treatmentofNRVMswithBPKDicompletelyabolished
FHL isoform expression by RNAi on ET1- and PE- ET1-andPE-inducedincreasesinbothPKDautophosphorylation
induced increases in PKD autophosphorylation at Ser916 and and HDAC5 phosphorylation (Figure 7A). Furthermore, pre-
transphosphorylation at Ser744/Ser748 (Figure 4) most likely treatment with BPKDi significantly attenuated ET1- and PE-
indicate an attenuation of PKD activation, rather than simply induced increases in MEF2 activity (Figures 7B and 7C). Thus
reflectingreducedPKDproteinexpression. it appears that FHL1 and FHL2 facilitate PKD activation and
HDAC5phosphorylation,butnotMEF2activationfollowingET1
andPEstimulation,despitethefactthatPKDactivityisindeedan
FHL1andFHL2facilitateHDAC5phosphorylationincardiac importantmediatorofbothHDAC5phosphorylationandMEF2
myocytes activationundertheseexperimentalconditions.
ToexplorewhethertheattenuationofPKDactivationisparalleled
by reduced phosphorylation of a functionally important cellular
PKD substrate, we also studied the effects of FHL1 and DISCUSSION
FHL2 knockdown on the phosphorylation status of HDAC5
The present study provides novel information concerning the
at Ser498, an established PKD substrate in cardiac myocytes
regulation of PKD activity and functions in cardiac myocytes,
[10,11].SelectiveFHLisoformknockdowninNRVMsattenuated
throughthefollowing principalfindings:(i) PKDinteracts with
HDAC5 phosphorylation in response to stimulation with ET1
two members of the FHL protein family, FHL1 and FHL2; (ii)
or PE (Figure 5), in a manner that mirrored the effects of
FHL1andFHL2arenotPKDsubstratesanddonotinhibitPKD
these loss-of-function interventions on PKD phosphorylation
catalytic activity; (iii) FHL1 and FHL2 differentially regulate
(Figure 4). Thus selective knockdown of FHL1 expression
PKDactivationafterneurohormonalstimulation;and(iv)FHL1
significantlyattenuatedtheincreaseinHDAC5phosphorylation
and FHL2 do not regulate MEF2 activity after neurohormonal
in response to ET1 compared with control cells transfected
stimulation.
with a scrambled siRNA sequence (Figure 5A), but had a non-
FHL1andFHL2arenon-enzymaticproteinsthatareexpressed
significant effect on the increase in HDAC5 phosphorylation in
primarilyinstriatedmuscles,includingtheheart[18,24,25].Their
response to PE (Figure 5B). In contrast, selective knockdown
importancefornormalmusclefunctionisevidencedbythefact
of FHL2 expression significantly attenuated the increases in
thatmutationsinthegenesencodingtheseproteinsfhl1andfhl2,
HDAC5 phosphorylation in response to both ET1 (Figure 5C)
areassociatedwithvariousskeletalandcardiacmyopathies[25].
and PE (Figure 5D). Therefore it appears that, in response
The multiple LIM domains comprising FHL1 and FHL2 form
to neurohormonal stimulation in NRVMs, FHL1 and FHL2
a tandem zinc-finger structure that provides a modular protein-
facilitate not only PKD activation, but also phosphorylation
binding interface, through which FHL1 and FHL2 function as
of the downstream PKD substrate HDAC5. Once again, the
adaptors or scaffolds to support the assembly of multimeric
effectsofsimultaneousknockdownofbothFHL1andFHL2on
protein complexes and regulate the localization and activity of
HDAC5phosphorylationweresimilartotheeffectsofselective
their partners [24,25]. Examples include specific interactions
FHL2 knockdown, with no indication of an additive effect
betweenFHL1andFHL2withanarrayoftranscriptionfactors,
(SupplementaryFigureS2).
such as AP-1 (activator protein 1) [26] and HIF-1 (hypoxia-
inducible factor 1) [27]. FHL proteins may co-activate or co-
repress these transcription factors, depending on the identity
FHL1andFHL2donotregulateMEF2activationincardiac
of the interaction partner, the stimulus and the cell type [24–
myocytes
27], thereby exerting regulatory effects on gene expression.
PKD-mediated phosphorylation of HDAC5 at Ser498 is believed In addition, FHL1 and FHL2 have been reported to form
to trigger HDAC5 nuclear export, with the consequent de- biologically significant interactions with the α5β1 and α7β1
repressionofMEF2activity[13,14].Wethereforehypothesized integrins [28,29], the muscle-specific RING finger proteins
thatknockingdownFHL1orFHL2expressionmightalsohave MuRF1 and MuRF2 [30], the metabolic enzymes creatine
an impact on ET1- and PE-induced MEF2 activation, as a kinase,adenylatekinaseandphosphofructokinase[19]andother
consequenceofattenuatedPKDactivationandreducedHDAC5 proteinkinaseslikeERK2(extracellular-signal-regulatedkinase
phosphorylation. To test this hypothesis, MEF2 transcriptional 2)[31,32].SuchinteractionsdefinethefunctionalrolesofFHL1
activity was monitored in NRVMs that were transfected andFHL2inregulatingbiologicalprocessessuchascelldivision,
with a MEF2-luciferase reporter. As expected, stimulation of differentiation, migration and metabolism [24,25]. The present
control NRVMs transfected with scrambled siRNA with either study adds PKD to the growing list of interaction partners for
ET1 or PE significantly increased MEF2 activity (Figure 6). FHL1 and FHL2 and provides evidence that these interactions
Surprisingly, however, transfection with active siRNA targeted areoffunctionalimportanceinregulatingPKDactivityincardiac
at FHL1 (Figures 6A and 6B) or FHL2 (Figures 6C and 6D) myocytesinresponsetoneurohormonalstimulation.
transcripts had no significant effect on ET1- and PE-induced PreviousstudieshaveexploredtherolesthatFHL1andFHL2
increases in MEF2 activity, despite again achieving marked may play in cardiac myocyte (patho)physiology through the
reductionsinFHLproteinexpression(SupplementaryFigureS4 targeted disruption of fhl1 and fhl2 in mice. Such studies have
athttp://www.biochemj.org/bj/457/bj4570451add.htm). revealedthatthegloballossofFHL1orFHL2proteindoesnot
© 2014 The Author(s) (cid:2)c TheAuthorsJournalcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
458 K.Stathopoulouandothers
Figure5 EffectofFHL1orFHL2knockdownonET1-andPE-inducedphosphorylationofendogenousHDAC5atSer498
NRVMsweretransfectedwitheitherscrambledsiRNAoractivesiRNAduplexestargetedatFHL1(AandB)orFHL2(CandD)transcripts.After48h,cellsweretreatedwithvehicle(C)orET1(10nM)
(AandC)orvehicle(C)orPE(3μM)(BandD)for20min.ThephosphorylationstatusofendogenousHDAC5wasdeterminedbyimmunoblot(IB)analysisusingaphospho-specificpSer498
HDAC5antibody.ProteinloadingwasconfirmedbyCoomassieBluestaining.Individualimmunoblotsillustraterepresentativeexperiments,andhistogramsshowquantitativedataasmeans+−S.E.M.
(n=7–8).*P<0.05.
impair cardiovascular development, or cause any spontaneous be isoform-specific and dependent on the nature of the stress
cardiac phenotype [32–34]. They have also shown that the stimulusthattriggersthehypertrophicresponse.Interestingly,at
targeted deletion of either fhl1 or fhl2 does not induce a leastinsomesettings,FHL1andFHL2appeartohavedifferential
compensatory change in the cardiac expression of the other effects on cardiac myocyte ERK activity [31,32], raising the
isoform [32,34], which is reminiscent of our observations possibilitythattheERKpathwaymaybecriticalinmediatingthe
with individual knockdown of FHL1 or FHL2 expression in regulationofcardiachypertrophybyFHLproteins.Furthermore,
NRVMs (Supplementary Figure S1). Nevertheless, when the recent evidence suggests a role for FHL2in the regulation of
heartsofgene-targetedmiceweresubjectedtodifferenttypesof calcineurin/nuclear factor of activated T-cell signalling, another
stress stimuli, some overt differences in the manifestation pro-hypertrophicpathway[35].FHL2repressescalcineurin,with
of pathological phenotypes became apparent. For example, whichitco-localizesatthesarcomere[35].LossofFHL2from
in comparison with their wild-type littermates, fhl1−/− mice thesarcomerehasbeenobservedinhumanheartfailuresamples,
have been reported to develop diminished cardiac hypertrophy suggesting that changes in the subcellular distribution of the
followingTAC(transverseaorticconstriction)orcardiac-specific proteinmaybeassociatedwithdiseasedevelopment[36].
Gα overexpression,suggestingapro-hypertrophicroleforFHL1 InourinvitrostudiesinNRVMs,selectiveknockdownofeither
q
[32].Onthecontrary,fhl2−/− micepresentedwithacomparable FHL1 or FHL2 inhibited PKD activation by pro-hypertrophic
magnitude of cardiac hypertrophy relative to wild-type controls neurohormonal stimuli and attenuated the downstream phos-
after TAC [34,35], but mounted an exaggerated hypertrophic phorylationofHDAC5,albeitwithsomedifferencesintheirim-
response after chronic β-adrenergic receptor stimulation by pactontheresponsestoET1againstPE(discussedlater).Never-
isoproterenolinfusion[33],suggestingananti-hypertrophicrole theless,MEF2activationbythesestimuliwasunaffectedbyFHL1
forFHL2inthelattersetting.ThustherolesthatFHL1andFHL2 or FHL2 knockdown (Figure 6). These data suggest that FHL
playinregulatingcardiachypertrophyintheinvivosettingmay protein-mediated regulation of PKD activity is unlikely to play
©(cid:2)c 20T1h4e TAhueth oArusthJoour(rsn)alcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
FHL1andFHL2arenovelregulatorsofPKD 459
Figure6 EffectofFHL1orFHL2knockdownonET1-andPE-inducedactivationofendogenousMEF2
NRVMsweretransfectedwitheitherscrambledsiRNAoractivesiRNAduplexestargetedatFHL1(AandB)orFHL2(CandD)transcripts.After24h,cellswereco-transfectedwitha3×MEF2-firefly
luciferasereportervectorandaRenillaluciferasecontrolvectorandtreatedwithvehicle(C)orET1(10nM)(AandC)orvehicle(C)orPE(3μM)(BandD)forafurther18–24h.Luciferaseactivity
incelllysateswasassessedbyinvitroluminescenceassays,withluciferaseactivitynormalizedforRenillaluciferaseactivityineachsampletocorrectfortransfectionefficiency.Histogramsshow
quantitativedataasmeans+−S.E.M.(n=4).*P<0.05.
Figure7 EffectofpharmacologicalPKDinhibitiononET1-andPE-inducedHDAC5phosphorylationandMEF2activation
(A)NRVMsweretreatedwithvehicle(C),ET1(10nM)orPE(3μM)for20min,followinga30-minpre-treatmentwithvehicleortheBPKDi(3μM).PKDautophosphorylationatSer916,totalPKD
expressionandHDAC5phosphorylationatSer498wereassessedbyimmunoblot(IB)analysisusingappropriateantibodies,asindicated.MembraneswerestainedwithCoomassieBluetoconfirm
equalproteinloading.Dataarerepresentativeofthreeindependentexperiments.(BandC)NRVMswereco-transfectedwitha3×MEF2-fireflyluciferasereportervectorandaRenillaluciferase
controlvectorandtreatedwithvehicle(C),ET1(10nM)orPE(3μM)for18–24h,followinga30-minpre-treatmentwithvehicleortheBPKDi(3μM).Luciferaseactivityincelllysateswas
assessedbyinvitroluminescenceassays,withluciferaseactivitynormalizedforRenillaluciferaseactivityineachsampletocorrectfortransfectionefficiency.Histogramsshowquantitativedataas
means+−S.E.M.(n=4).*P<0.05.
© 2014 The Author(s) (cid:2)c TheAuthorsJournalcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.
460 K.Stathopoulouandothers
acriticalroleincontrollingMEF2-drivencardiacmyocytetran- MEF2-driven transcriptional reprogramming towards cardiac
scriptional reprogramming towards hypertrophy, despite its im- hypertrophy.
pactonHDAC5phosphorylation.Inthisregard,althoughpharma-
cologicalPKDinhibitionabolishedtheneurohormonalinduction
ofPKDautophosphorylationandHDAC5transphosphorylation, AUTHORCONTRIBUTION
itincompletely(albeitsignificantly)inhibitedMEF2activationin
Metin Avkiran, Friederike Cuello, Elisabeth Ehler and Robert Haworth conceived and
ourstudies(Figure7).Itispossible,therefore,thatacombination instigatedtheproject.KonstantinaStathopoulou,FriederikeCuelloandMetinAvkiran
ofPKD-independentmechanisms,suchasthephosphorylationof planned the experiments, the majority of which were performed by Konstantina
classIIHDACsbyCaMKII[37,38]orGRK5(G-protein-coupled StathopoulouwithcontributionsfromFriederikeCuello,AlexandraCandasamy,Elizabeth
receptor kinase 5) [39], or their redox-mediated nuclear export KempandRobertHaworth.KonstantinaStathopoulouandMetinAvkirananalysedthedata
andwrotethepaper.
[20,40], or indeed HDAC-independent mechanisms mediated
by p38-MAPK (mitogen-activated protein kinase) [41], ERK5
[42,43]ortheacetyltransferasep300[44],maybesufficienttoin-
ACKNOWLEDGEMENTS
ducesignificantMEF2activationandtherebyMEF2-driventran-
scriptional reprogramming towards cardiac hypertrophy. In this WethankShineyRejiandUpamaliPerera(CardiovascularDivision,King’sCollegeLondon,
context,itisalsoimportanttonotethatpharmacologicalinhibition London,U.K.)fortechnicalassistanceandStephanLange(DepartmentofMedicine,UC
SanDiego,CA,U.S.A.)forkindlyprovidingthebacterialexpressionvectorsforGST–FHL1
ofPKDactivityhasbeenshownnottoattenuatecardiachyper-
andGST–FHL2.
trophyindifferentmodelsintheratinvivo[45],althoughcardiac-
specificdeletionofPKD1inthemousesignificantlyinhibitedcar-
diachypertrophyinresponsetoTACorchronicisoproterenolinfu- FUNDING
sion[15].AplausibleinterpretationofsuchdataisthatPKD1reg-
ulatescardiachypertrophythroughamechanismthatisindepend- ThisworkwassupportedbytheMedicalResearchCouncil[grantnumberG0800206].
entofitskinaseactivity,inwhichcaseFHL1-andFHL2-mediated
regulationofneurohormonalPKDactivationwouldbemorelikely
REFERENCES
toregulateotherPKD-mediatedfunctions.Inthecardiaccontext,
PKDactivityhasbeenimplicatedintheregulationofmyofilament 1 Avkiran,M.,Rowland,A.J.,Cuello,F.andHaworth,R.S.(2008)ProteinkinaseDinthe
calcium sensitivity and dynamics [7–9], voltage-gated calcium cardiovascularsystem:emergingrolesinhealthanddisease.Circ.Res.102,157–163
channel activity [46], protection against ischaemia/reperfusion 2 Steinberg,S.F.(2012)RegulationofproteinkinaseD1activity.Mol.Pharmacol.81,
injury [47] and insulin resistance [48], and the potential roles 284–291
3 Haworth,R.S.,Goss,M.W.,Rozengurt,E.andAvkiran,M.(2000)Expressionandactivity
of FHL proteins in regulating such processes warrant further
ofproteinkinaseD/proteinkinaseCμinmyocardium:evidenceforα1-adrenergic
investigation.
receptor-andproteinkinaseC-mediatedregulation.J.Mol.Cell.Cardiol.32,1013–1023
The possible mechanism(s) through which FHL1 and FHL2 4 Bossuyt,J.,Chang,C.W.,Helmstadter,K.,Kunkel,M.T.,Newton,A.C.,Campbell,K.S.,
mayregulatePKDactivationalsorequireconsideration.Previous Martin,J.L.,Bossuyt,S.,Robia,S.L.andBers,D.M.(2011)Spatiotemporallydistinct
studiesinratcardiacmyocyteshaveshownthatneurohormonal proteinkinaseDactivationinadultcardiomyocytesinresponsetophenylephrineand
PKD activation is achieved through the phosphorylation of endothelin.J.Biol.Chem.286,33390–33400
its activation loop, principally by the novel PKC isoform 5 Iwata,M.,Maturana,A.,Hoshijima,M.,Tatematsu,K.,Okajima,T.,Vandenheede,J.R.,
PKCε [5,21]. Interestingly, proteomic analysis of the cardiac VanLint,J.,Tanizawa,K.andKuroda,S.(2005)PKCε–PKD1signalingcomplexat
PKCε signalling complex has identified FHL2 as a component Z-discsplaysapivotalroleinthecardiachypertrophyinducedbyG-proteincoupling
receptoragonists.Biochem.Biophys.Res.Commun.327,1105–1113
protein [49]. Moreover, PKCε association with another LIM-
6 Fu,Y.andRubin,C.S.(2011)ProteinkinaseD:couplingextracellularstimulitothe
domain protein, enigma homologue, has been proposed to regulationofcellphysiology.EMBORep.12,785–796
play a key role in its substrate targeting [50]. Given that 7 Haworth,R.S.,Cuello,F.,Herron,T.J.,Franzen,G.,Kentish,J.C.,Gautel,M.and
ET1- and PE-induced phosphorylation of PKD at Ser744/Ser748 Avkiran,M.(2004)ProteinkinaseDisanovelmediatorofcardiactroponinI
was diminished by FHL1 or FHL2 knockdown in the present phosphorylationandregulatesmyofilamentfunction.Circ.Res.95,1091–1099
study (Figure 4), it is plausible that FHL proteins may act as 8 Cuello,F.,Bardswell,S.C.,Haworth,R.S.,Yin,X.,Lutz,S.,Wieland,T.,Mayr,M.,
scaffolds that facilitate the co-localization of PKCε and PKD Kentish,J.C.andAvkiran,M.(2007)ProteinkinaseDselectivelytargetscardiactroponin
IandregulatesmyofilamentCa2+sensitivityinventricularmyocytes.Circ.Res.100,
and thus the phosphorylation and activation of the latter, in a
864–873
signal-responsive manner. Additionally, the differential effects
9 Bardswell,S.C.,Cuello,F.,Rowland,A.J.,Sadayappan,S.,Robbins,J.,Gautel,M.,
of FHL1 and FHL2 knockdown on ET1- compared with PE- Walker,J.W.,Kentish,J.C.andAvkiran,M.(2010)Distinctsarcomericsubstratesare
inducedPKDphosphorylationatSer744/Ser748andSer916(Figure4) responsibleforproteinkinaseD-mediatedregulationofcardiacmyofilamentCa2+
suggest that the two FHL proteins might associate with distinct sensitivityandcross-bridgecycling.J.Biol.Chem.285,5674–5682
subcellular pools of PKD downstream of the ET1 and α - 10 Vega,R.B.,Harrison,B.C.,Meadows,E.,Roberts,C.R.,Papst,P.J.,Olson,E.N.and
1
adrenergic receptors. In this context, it is interesting to note McKinsey,T.A.(2004)ProteinkinasesCandDmediateagonist-dependentcardiac
hypertrophythroughnuclearexportofhistonedeacetylase5.Mol.Cell.Biol.24,
recent evidence which suggests that ET1 and PE may induce
8374–8385
spatiotemporally distinct patterns of PKD activation in cardiac
11 Harrison,B.C.,Roberts,C.R.,Hood,D.B.,Sweeney,M.,Gould,J.M.,Bush,E.W.and
myocytes [4]. The potential contribution of FHL isoforms
McKinsey,T.A.(2004)TheCRM1nuclearexportreceptorcontrolspathologicalcardiac
to such compartmentalized regulation also requires further geneexpression.Mol.Cell.Biol.24,10636–10649
investigation. 12 McKinsey,T.A.,Zhang,C.L.andOlson,E.N.(2000)Activationofthemyocyteenhancer
In conclusion, in the present study, we identify FHL1 factor-2transcriptionfactorbycalcium/calmodulin-dependentproteinkinase-stimulated
and FHL2 as novel interaction partners for PKD in cardiac bindingof14-3-3tohistonedeacetylase5.Proc.Natl.Acad.Sci.U.S.A.97,
myocytes and we show that FHL1 and FHL2 regulate PKD 14400–14405
13 McKinsey,T.A.,Zhang,C.L.,Lu,J.andOlson,E.N.(2000)Signal-dependentnuclear
activation in response to neurohormonal stimulation. Although
exportofahistonedeacetylaseregulatesmuscledifferentiation.Nature408,
such regulation affects HDAC5 phosphorylation, it does not
106–111
control neurohormonal MEF2 activation, suggesting that PKD
14 Lu,J.,McKinsey,T.A.,Nicol,R.L.andOlson,E.N.(2000)Signal-dependentactivation
regulationbyFHLproteinsislikelytobeoffunctionalimportance oftheMEF2transcriptionfactorbydissociationfromhistonedeacetylases.Proc.Natl.
in PKD-mediated (patho)physiological processes other than Acad.Sci.U.S.A.97,4070–4075
©(cid:2)c 20T1h4e TAhueth oArusthJoour(rsn)alcompilation(cid:2)c 2014BiochemicalSociety
The author(s) has paid for this article to be freely available under the terms of the Creative Commons Attribution Licence (CC-BY) (http://creativecommons.org/licenses/by/3.0/)
which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited.